Report No. 32674-PE Peru Rethinking Private Sector Participation in Infrastructure Towards Effective Public Private Partnerships/ Concessions in the Provision of Infrastructure Services June 30, 2006 Finance, Private Sector and Infrastructure Sector Management Unit Country Management Unit 6 Latin America and the Caribbean Region Document of the World Bank IBRD Vice President Pamela Cox Chief Economist Guillermo Perry Country Director Marcelo Giugale Sector Leader Franz Drees-Gross Task Manager J. Luis Guasch RethinkingPrivate Sector Participationin Infrastructure:Towards Effective Public Private PartnershipsKoncessionsinthe Provisionof Infrastructure Services in Peru Table of Contents* Acronyms ................................................................................................................................................................. ii Executive Summary ........................................................................................................................................... x 1. Private Participation Progress and Highlights ......................................................................... 1 1,l Introduction .......................................................................................................................... 1 1.2 New institutional frameworks ......................................................................................................... 3 1.3 Private participation progress: all sectors .. 1.4 Telecommunications................................................... 1 5 . Electricity .............................................................................. ...................1I 1.6 Water and Sanitation .......................................................................................... 1.7 Roads...................................................................................................................................................... 18 1.7.1 Arequipa - Matarani Road.......................................................................................................... 19 1.7.2 Ancon-Huacho-Pativilca Road .............................................. 21 1,8 Ports.......................................................................... .......................22 1.9 Railroads ................................................................................................................... 25 1.10 Airports.................................................................................................................................................. 28 2. Characteristics of the Privatization and Concession Processes .................................. 30 2.1 A warding process ................................................................................................. 30 2.1.1 Principles for transparency 2.1.2 Competitive bidding processes 2.2 Contract design ............................... 37 2.2.2 Unbundling a 2.2.5 Certainty an 2.2.6 Renegotiations ................................................ 60 2.2.7 Dispute res ............................. 66 2.3 Summary ................................................................... ..................69 3. Impacts of Private Participation .................................................................................................... 70 * This report was produced by J. Luis Guasch (Task Manager, LCSFP), Isabel Sanchez and Thomas Haven (Consultants). Background papers were provided by Maximo Torero, Lorena Alcazar, APOYO Consulting. Commentsby Alberto Pasco-Fontare gratefully acknowledged. iv 3 , l Impacts on Firm Performance 70 3.1.1 Telecommunications and Electricity..........,....... .................................. 72 3.1.2 Transportation................................................................................................................................ 74 3.2 . Impacts on Consumers.................................................................................................................... 76 3.2.1 Telecommunications and Electricity 3.2.2 Transportation 3.3 Impacts on Employment................................................................................................................. 87 3.3.1 Telecommunications and Electricity....................................................................................... 87 3.3.2 Transportation................................ 3.4 Impacts on government revenue 3.5 Economy-wide impacts: telecommunications and electricity .......... 3.6 International Comparisons................,........... ...,........................... , , , , ........, ..... , , , , , , , ,.......... .. .........94 , , , 3,7 Summary.1,.,.,......1.1,....1.......1...1..11(.,..,,,,,,.....,,,,,,,,,,..,....,,,,,,,,,,..,.,.,,..,,.,,,.,,,.,,..,,,.,,,..,.,.,,....,.,....,,,,,,,.., 98 4. Public perceptions of private participation ............................................................................. 99 4.1 Social context...................................................................................................................................... 99 4.1.1 Evaluation of Problems in Daily Life 99 4.1.2 Access to public services 00 4.2 Perceptions of the actual privatization process,...1..........111........I.).....).............,.........,,.,.,....., 101 4.2.1 Positive Aspects of Privatization............................................................................................ 103 4.2.2 Negative Aspects of Privatization .................................... 4.2.3 Quality of service 4.3 Lessons for the future.................................................................................................................... 110 5. Moving forward: A Road Map for the Peru Concession Program .............................. 114 5.1 Addressing social issues................................................................................................................ 117 5.1.1 Overall transparency ................................................................................................................. 120 5.1.2 Communications program 5.1.3 Community involvement 5.1.4 Accounting for the poor a 5.2 Improving concession contract design.................................................................................... 129 5.2.1 Concession awarding process ................................................................................................ 130 5.2.2 "Financial equilibrium" clauses for the operation of the concession 5.2.3 Renegotiation clauses and triggers for renegotiation 5.2.4 Penalties and incentives ........................................................ 5.2.5 Concession length and financing.......................................................................................... 136 5.2.6 Investment commitments ...................................................................................................... 137 5.2.7 Better allocation of concession risks ................................................................................... 137 5.2.8 Addressing Termination of the Concession 140 5.2.9 Arbitration rules and dispute resolution 140 5,3 Enhancing regulatory design....................................................................................................... 141 5.3.1 Determining future tariffs ........................................................................................................ 141 5.3.2 Regulatory structure: rate of return vs. price caps....................................................... 142 5.3.3 Cost of capital and how it should be determined ........................................................... 145 5.3.4 Valuation of concession assets 147 5.3.5 Informational requirements f 149 5.3.6 Regulatory accounting norms 150 V 5.3.7 Regulatory risk ................................................................................................. 153 5.4 Addressing financing i ........................................................... 153 5.5 Summary ........................ ........................................................... 160 Appendix .............................................................................................................................................................. 162 Appendix l a : Transparency Elements in Transport Concession Bidding Appendix 16: Transparency Elements in Telecom and Electricity Bidding Appendix 2a: Comparison of Transportation Competitive Biddings Stages............................ 164 Appendix 2b: Comparison of Telecom and Electricity Competitive Biddings Stages...........165 Appendix 3a: Obligations and Penalties: Transportation Concessions................................... 166 Appendix 3b: Obligations and Penalties: Telecommunications and Electricity 167 Appendix 4: Significance tests ................................................................................................................ 168 Appendix 5: Measuring consumer welfare gains ............................................................................... I75 Appendix 6: Calculating indirect effects with a CGE model........................................................... I 7 9 Appendix 7: Public perceptions study methodologies ................................................. I86 References .......................................................................................................................................................... 188 Listof Figures Figure 1: Population Expressing Dissatisfaction with Privatization in Latin America (percentage) ................................................................................................ Figure 2: Usage of Telephone Services in Urban Peru According to Figure 3: Population Expressing Dissatisfaction with Privatization (percentage) .......................... 2 Figure 4: Number of Private Participation Processes per Year Figure 5: Private Participation Revenues By Sector up to 199 Figure 6: Private Participation Process Progress (1991-1999) .......................................... Figure 7: Differences between privatizations and concessions............................................................. 7 Figure 8: Rate of Approval of the Privatization Process .................................... 8 Figure 9: Evolution of the number of lin Figure 10: Evolution of Basic Local Rate Figure 11: Evolution of Average Long D Figure 12: Number of Residential Electr .................................................... 13 Figure 13: Electrification Coefficient (%) Figure 14: The evolution of the index of Figure 15: Household's Electric Tariff an Figure 16: Average Cost per KWh Supplied to Final Consumers - Generation (Real Nuevos Soles)......... ........................................................................................... 15 Figure 17: Main Arequipa-Matarani road............................................ 19 Figure 18: Vehicle Traffic on the Arequipa-Matarani road ..................... 20 Figure 19: Monthly Toll Income (US$) on the Arequipa- Figure 20: Fares (US$) on the Arequipa-Matarani road ................... Figure 21: Evolution of Cargo by Type of Transaction (Tons): Matarani Port..... Figure 22: Evolution of Cargo by Type of Product (Tons): Matarani Port............ Figure 23: Average Tons per Ship: Matarani Port ....................................................................... Figure 24: Number of Ships Processed and Hours Stayed per Ship: Matarani Port................... 25 vi Figure 25: Average Cost per Dock Usage (US$/Ton) : Matarani Port ........................... 25 Figure 26: Retribution and Investment (Thousand of Soles) .............................................................. 27 Figure 27: Tons of Freight Processed (Thousand of Tons) ........................................... .27 Figure 28: Transport of Passengers (Thousands of Passe Figure 29: Transport of Passengers, Freight and Mail................ Figure 30: Dispute Resolution Mechanisms for Telefonica ................................................. Figure 31: Usage of Telephone Services in Urban Peru Ac Figure 32: Electricity Consumption in Urban Peru by Expenditure Quintiles Figure 33: Consumers Saving and Losses due to the Matarani Road Concession (Nuevos Soles) ................................................................................................... Figure 34: Average Matarani Port Fares by Type of Cargo (US$) .................. Figure 35 : Consumer Welfare Gains (Losses) of Matarani Port Concession (US$) - Price Effect........................................... ............................... 84 Figure 36: Concession Impact o Figure 37: Employment Evolution for Transport Firms (number of employees Figure 38: Evolution of State Return over Sales before and after private part Figure 39: Economy-wide Circular Flow of Goods and Services....................................................... Figure 40: Evolution of telecommunications indicators: Peru versus Latin America..................93 95 Figure 41: Evolution of electricity distribution indicators: Peru ver Figure 42: Changes in trends in Latin America........................................ Figure 43: How those without Basic Services Satisfy their Needs.......................... Figure 44: Perceptions of Privatization in the 1990s esults-Percentage of Respond ciated Privatization with... ........................................................................................... results of Privatization .............................................................................. Figure 47: Were Funds from Privatization Adequate Figure 48: What do you Think has been the Effect of Privatization on Direct and Indirect Employment?........................................................ Figure 49: What are Principal Advantages and Figure 50: How would you Qualify the Impor a Company's Effect on the Community? .............................................................................. 117 Figure 51: How would you Qualify the Impor Services Offered by Public Service Provi Figure 52: I f The State is Obligated to Sell a Company, which of the Following Conditions are Essential or Very Important?................................................................................................. Figure 53: Accountability in infrastructure services....................................................................... Figure 54: How the guarantee facility works ................................................................. 157 Figure 55: Structure of a Private Infrastruct & Support Facility............................ 158 Figure 56: Options for Risk Mitigation unde Figure 57: Options by Typology of Countries ...................................................... Figure 58: Social Accounting Matrix Structu Listof Tables Table 1: Infrastructure Privatizations/Concessions in Peru................ Table 2: Characteristics of State vs. Private Companies ................... Table 3: Report summary by secto .......................................................................... xxvi Table 4: Infrastructure privatizatio .......................................................... 3 Table 5: Amounts Involved in Priv ions) ................................................... 5 Table 6: Service quality indicators in the Telecommunications Sector ............................ 10 Table 7: Electric Industry Structure and Privatization Progress 12 Table 8: Efficiency Improvements in the Electric Sector ............................. 14 vii Table 9: Main Indicators of the Electric Sector, 1993-2002 16 Table 10: Bidding prices for the regional electric companies .............................................................. 17 Table 11: Number of Bidders that Participated in the Stages of Concession Processes ...........33 Table 12: Selection Criteria in Infrastructure Concessions and Privatizations 35 Table 13: Market structure for concession contracts in public transportation infrastructure .38 Table 14: Activities in the Lima Airport by Type of Operation ............................................................ 40 Table 15: Market structure According to Concession Contracts of Telecommunications and Electricity........................................................................................................................................................ 41 Table 16: Prices in the Electricity Industry and their Regulation 43 Table 17: Privatized Companies in Generation: Economic Groups - 44 Table 18: Main Risks and how they are Addressed in Transportation Concession Contracts*50 .................................................................................................................................. 51 Table 20: Defining Major Force Events in Concession Contracts: A comparison ...................... 52 Table 21: Defining Major Force Events in Electricity and Telecom Contracts ............................... 53 Table 22: Bonuses in the Concession Contracts in Peru ..... 57 Table 23: Incentives in Privatization Contrac 58 Table 24: Timing and Number of Renegotiati ....................... 61 Table 25: Selected Modifications to the Lima Airport Concession ..................................................... 63 Table 26: Timing and Number of Renegotiations in Telecom and Electricity Privatizations....64 Table 27: Summary matrix for Process chapter ....................................................................................... 69 Table 28: Performance Measures 70 Table 29: Privatized Companies and Data Availability 72 Table 30: Performance Changes after Privatization: Telecommunications and Electricity (Difference between Means and difference in Difference Tests) ............................................... 73 Table 31: Performance Changes after Concession: Transportation (Difference between Means and Difference in Difference Tests) 75 Table 32: Telephone: Changes in Consumer S 78 Table 33: Electricity: cChanges in Consumer Surplus by Expenditure Quintiles in Urban Areas ................................................................................................................................................................ 79 Table 34: Consumer Surplus Levels and Changes in Urban Areas, Including Effects of Greater Access .... Table 35: Fares of the Table 36: TUUA Fares Table 37: Diurnal Landing and Takeoff Fares (US$ without taxes) .................................................. 86 Table 38: Impact of Layoffs on Performance Indicators for Telecommunications and Electricity. 88 Table 39: Impact of Layoffs on Performance Indicators for Matarani Port and Lima Airport..90 Table 40: Non-financial Companies included in the Fiscal Analysis .................................................. Table 41: State Return over Sales after private participation - Non-financial Companies......91 91 Table 42: Impact on Macroeconomic Aggregates 94 Table 43: Summary matrix for Impacts chapter 98 Table 44: Predicted opposition to privatization i 03 Table 45: Level of Access to Services and Level of Satisfaction..................................................... 108 Table 46: Performance of telephone companies according to the attributes judged to be important for public service providers .............................................................................................. 109 Table 47: Performance of electricity companies according to the attributes judged to be important for public service providers. Table 48: Characteristics of State vs. Priva Table 49: Status of Preparation of ProInversion's Project Table 50: Potential positive impacts of infrastructure ser ... Vlll Table 51: Examples of Prequalification Criteria in Private Infrastructure Concession\Tr ansactions Table 52: Identification and Allocation of Risks Table 53: Performance Chang and difference in difference Tests) ..................................................................................................... 169 Table 54: Performance Changes after Privatization: Electrolima (Difference between Means and Difference in Difference Tests) 170 Table 55: Performance Changes after Privatization: Electroperu (Difference between Means and Difference in Difference Tests) .................................................................................................... 171 Table 56: Performance Changes after Concession: Matarani Port (Difference between Means and Difference in Difference Tests)...................................................................................... Table 57: Performance Changes after Concession: ENAFER (Rail Road Concession) Table 58: Performance Changes after Concession: CORPAC (Lima Airport Concessi Table 59: Price Elasticity of Utilities in Urban Areas, 1991, 1994 and 1997............................... 178 Table 60: Impact on GDP and Prices by Economic Activity (%) ...................................................... 184 Table 61: Impact on Real Wages by Type of Worker (%) ....................................................... 185 Listof Boxes Box 1: Public rejection of electricity privatizations in Arequipa Box 2: A failed privatization in the electrical sector Box 3: Transparency Problems in the Matarani Port Concession Process ...................................... 31 Box 4: Possible Reasons for the Price Paid by Telefonica Internacional de Espafia ....... 36 Box 5: Access problems of the South-Oriental Railway 39 Box 6: The TdP worker dispute 46 Box 7: Obligations and Penalties I IO case ..................................... 55 Box 8: OSPITEL's reasons for approving Addendum 1in the TdP contract ................................... 66 Box 9: Resolution of telecom disputes between TdP and private operators Box 10: Little difference in perceptions between those with access to basic services an those without Box 11: Making a deal with the community: Water supply and accountability in Indonesia 121 Box 12: Community action in rolling back environmental externalities in postwar Japan.....123 Box 13: Putting accountability mechanisms into large infrastructure projects: The Nam Theun 2 dam. Box 14: Grouping Box 15: The cost Box 16: Measures for Determining Compensation at the Termination of a Concession.........149 Box 17: Inform Box 18: Comm Box 19: Genera Box 20: Using insured bonds: lessons from highways in Latin America ....................................... 157 i x I Executive Summary The 1990s were characterized by a worldwide policy redirection toward private participation in infrastructure, a trend which was particularly pronounced in Latin America. That effort was part o f a broader set o f structural reforms aimed at economic modernization and liberalization. Bringing the private sector into infrastructure was driven by disappointment with ineffective state-operated utilities, the need for significant investments and scarce public funds, and the greater flexibility offered by technological change and regulatory innovation. In Peru, dismal macroeconomic indicators, record underemployment (86 percent), and hugely unprofitable public enterprises in the late 1980s and early 1990s provided particular impetus for private participation. As a result, Peru launched a private participation program in 1991, hoping that private investment would be a key driver o f economic growth. Since that time, many o f the largest state-owned enterprises (SOEs) in telecommunications, electricity, and transport infrastructure (roads, railroads, a port, and an airport) have been sold or concessioned. After over a decade o f experience, enthusiasm for private participation in infrastructure has cooled considerably in Peru and around the world (see Figure 3). The paradox i s that practically all o f the economic studies on the impacts o f private participation show considerable benefits. Of course, there were some shortcomings, including issues o f transparency, poor regulatory effectiveness, systemic renegotiation, employment issues, and lack o f programs to account for the poor. As o f now, governments are proceeding with the utmost caution in considering further private sector participation in infrastructure. Yet, the needs for infrastructure improvements remain large and public funds remain scarce. Thus the issue i s how to move forward. Figure 1: Population Expressing Dissatisfaction with Privatization in Latin America (percentage) . . . . I VEN MEX BRA CHC PER ECU COL ARG BOL SLV NIC HND PAN GTM Note: The 1998 results reflect survey respondents who disagreed or strongly disagreed with the statement, "privatizations of state companies have been beneficial for the country." The 2004 numbers are of those who were less satisfied or much less satisfied with public services after privatization, in terms of price and quality. Source: Latinobarometro surveys for 1998 and 2004 X This is the motivation for this report, looking backwards to move forward. The Government o f Peru has launched an aggressive second phase o f infrastructure concessions covering all sectors, including transport, rural telephony, gas, energy and water and sanitation. To help inform the second phase, this report provides an in-depth analysis o f the past program in Peru-the telecommunications, electricity, and transport infrastructure privatizationsiconcessions-by analyzing the process, design and impact o f the program as well as the public perceptions. While there are some key differences between privatizations and concessions, many o f the lessons from the privatization processes can be applied to the upcoming round o f concessions. Hence, the report highlights that concessions are the way o f the future-partly due to the public's reluctance to relinquish ownership o f State assets-while building on lessons from all private participation in the past. The report draws on a wide range of sources and analytical tools, including descriptive, contractual, econometric, and survey analysis. Indoing so, it aims to identify lessons for the on-going new phase o f infrastructure concessions. PRIVATE PARTICIPATION PROGRESS Peru's overall privatization strategy began in 1991 when laws were passed that established frameworks to sell public companies and encourage private investment in infrastructure. Sales accelerated rapidly and b y 1994, telecommunications and much o f the electricity sector had been privatized. Privatizations started to lose popularity in the mid-1990s, causing the second Fujimori government to look to concessions as a possible alternative. (Concessions grant a private firm the right to operate a defined infrastructure service, use the physical assets and to receive revenues from the service provision.) Transport concessions began in 1998, and in the next few years, railroads, one port, one highway, and the main airport (Lima) were concessioned. An exception i s the Arequipa-Matarani road which was concessioned in 1994. Table 4 shows a list o f the firms that are analyzed in this report as well as their privatizationlconcession dates. Although Peru's experience with private participation has not been perfectt, it i s considered to be one o f the best examples in the region, In fact, when compared with other countries in Latin America, the Peruvian experience comes out quite positive from an "institution building" and "attracting and mobilizing private capital" point o f view, especially given that Peru had investment ratings o f B to BB for most o f the period analyzed. Teleconznzurzicatiorzs.In 1994, the two State telephone monopolies (CPT and ENTEL) were sold to Telefonica de Espafia for over US$ 2 billion. Shortly thereafter, CPT and ENTEL were merged to create Telefonica del Peru (TdP). The terms o f the contract required TdP to nearly double the number o f lines installed in the country and in exchange, it was given a national monopoly over local, long distance, and international service for five years. Since privatization, coverage and service quality have improved dramatically, and prices have mainly fallen, with some exceptions having to do with new services. Electricity. Between 1994 and 1998, privatizations led to 14 private distribution and generation companies with commitments to increase total generating capacity by 560 MW (compared to an installed power base o f 4,379 MW in 1994). A t present, the privatized companies represent 64% o f the power generation capacity o f the National Electric System and 79% o f the distribution 'For instance, there have been concerns over h o w the government used revenues from privatization. x i service. Even though privatized firms have shown substantial improvements in coverage and service quality, public protests in Arequipa over the planned sale o f two electricity generation companies forced the government to shelve hture privatization plans. Sector Privatized entity Date Concession or Privatization Telecom CPT and ENTEL (became Telefcinica del Peru) Electroperu Privatization/ (became 8 different companies) 1995-1996 Concession Electrolima Privatization/ (became 5 different companies) 111 1994-1996 Concession Electro Sur Medio 1997 Privatizatiord Concession Electricity Electro Centro 11 1998 Privatization/ Concession Electro Nor Oeste 1998 Privatization/ Concession I I 1998 Privatization/ I Concession Electro Norte Medio 1 1998 1 Privatization/ Concession Arequipa - Matarani road 1994 Concession Roads Ancon-HuachePativilca road 2003 Concession Airports Lima Airport 200 1 Concession 0 Ports Matarani Port 1999 Concession &m Central Railroad 1999 Concession Railroads Southern Railroad 1999 Concession Southern Oriental Railroad 1999 Concession Roads. The Matarani-Arequipa road was concessioned by the Ministry o f Transport and Communications in 1994 as a one-off event outside o f the privatizatiodconcession institutional framework. Subsequently, the "Road Development Plan 1996-2005" proposed concessioning eleven separate road networks. The road networks grouped high-traffic highways with lower demand roads in a cross-subsidy scheme. The Red Vial No. 5 was prioritized as the first to be concessioned and the other ten road networks were put on hold. This cross-subsidy scheme was later found to be unfeasible due to fiscal constraints and only the Ancon-Huacho-Pativilca portion o f the Red Vial No. 5 was concessioned in 2003. Both road concession contracts required new investments which would be recouped from toll income. In May 2005, the government concessioned the IRRSA North corridor. xii Ports. In 1999, a 30-year Build-Operate-Transfer (BOT) concession was granted for the Matarani Port. It i s the only port concession in Peru so far. Although the 110 port concession was proposed that same year, there were not enough final bidders in its auction. Matarani i s the second most important port in Peru and i s situated on the Arequipa coast, south o f Lima. It handled 8.3 percent o f the country's ship traffic and 8.4 percent o f its cargo in 2001. In comparison, the main port o f Callao (Lima's port) handled almost 60 percent o f ships and 70 percent of cargo. The concession contract stipulated investment commitments and maximum rates that could be charged for certain port services over the first five years, among other things. Railroads. Peru has three main railroads-the Central, Southern and Southern Oriental-which together make up 93 percent o f the country's rail network. In 1999, the three railroads were combined into a single package for the concession bidding process, although the winner was allowed to divide them after the auction. The main objective o f the concession was to attract private investment to improve the sector, which was characterized by lack o f maintenance, low quality service and financial problems. The contract was for 30 years and the concessionaires were required to obtain an international quality certification by the fifth year. Airports. There are sixty-one airports in Peru and the only one that i s operated by a private firm i s Lima's Jorge Chavez International Airport. It i s the most important airport in Peru, representing 84 percent o f national freight and 63 percent o f passenger transportation in 2002. Although there were supposed to be other regional airports concessions, the poor profitability o f them resulted in the concession o f only the most attractive one: Lima. The concessionaire won the right to manage the airport, but the essential services-navigation assistance, aeronautic radio-communication and air traffic control-would still be under the administration o f the State- owned airport company. The 30 year concession was a BOT scheme with investment commitments o f US$1.2 billion over 30 years. CHARACTERISTICS OF THE PRIVATE PARTICIPATION PROCESS Awarding processes. Sound awarding processes and contract design are necessary to limit opportunistic behavior by firms and ensure that consumers enjoy an adequate share o f the performance benefits usually expected from privatization and concession processes. A good regulatory framework that is insulated from political pressure will make these processes more transparent and credible, enhancing public perceptions. Peruvian infrastructure awarding processes were transparent for the most part. However, in a few instances last minute changes in some contracts' conditions or extended contract closing time caused public faith in the process to suffer. For most privatizations and concessions, a good institutional framework was in place at the design stage with mechanisms that ensured access to information to all potential investors. However, more efforts could have been devoted to illuminating how government privatizationrevenues were used. Infrastructure concessions and privatizations should be awarded through competitive bidding rather than direct adjudication or bilateral negotiation. In Peru, all processes fulfilled this recommendation. Contracts were awarded after adequate bidder qualification stages, with the exception o f the Arequipa-Matarani road which, being one o f the first to be implemented, had an experimental character. X l l l ... The degree o f competition in most telecommunications and electricity awarding processes was reasonable. This was possibly explained by the market size involved in each process, and consequently, by their attractiveness to investors. Most processes offered the private operator ample and stable demand, through granting an exclusivity period in the most important services in telecommunications, or an exclusivity area in electricity distribution. On the other hand, the degree o f competition in the transportation infrastructure awarding processes was relatively low. Inthe awarding process, the design o f the selection criteria and the choice o f the reserve price are also very important. Most processes in Peru used a single and objective criterion-the maximum revenue or retribution to the State-to determine the winning bidder, with investment goals also imposed in some cases. Public perceptions were likely damaged by the prioritization o f government revenue over other criteria with more tangible benefits to the public services' users. Large differences between the final price paid and the reserve price may have been due to an overly optimistic estimation by the concessionaire (in the Lima Airport case) and a too low base price (in the TdP case). Very high bids may create concerns about high user charges once the concession starts or may be a reflection o f bidder's opportunistic behavior. Bidders might expect that once the concession i s awarded, a bilateral renegotiation would lead to a change in financial equilibrium clauses, resulting in higher tariffs or lower investment requirements. It may be argued that Telefonica's offer was based on variables that COPRI (the private investment promotion committee) did not take into account, such as a bonus for the monopoly period, special knowledge o f the South American market, or an interest in consolidating its leadership position in the region. For the Lima Airport, the high bid i s more difficult to explain, as it was not financially feasible. As a result, several significant renegotiations have been required. Contract design: unbundling and competition. In general, the privatization and concession processes in Peru established the market structure where service operation would take place. Introducing competition in infrastructure service provision was desirable for its effect on efficiency and innovation. In most cases, it was determined that certain specified services would be provided under monopoly conditions and some others would be opened to competition. Nevertheless, given the small size o f some markets, concessionaires were also .allowedto operate inthe competitivesegments. Although third party access and a non-discriminatory environment were mandated, in practice there have been some shortcomings. In the Matarani port, the Lima airport and the railways, no new operators o f transportation services have entered the market. A t least initially, the absence o f competitors did not result from opportunistic conduct by the concessionaire, but instead from efficiency reasons. If vertical unbundling had been strictly enforced immediately after the concession contract was signed, continuity in the provision o f certain services would have been endangered. Now, vertical integration prevails since most services are provided by firms closely linked to the concessionaire. While the original access model might be appropriate for the Southern Oriental railroad because it i s an attractive route with high demand, antitrust policy and regulation could be improved to xiv guarantee access by third parties. For example, the renting o f wagons to the concessionaire's operator could be limited and the allocation o f schedules could be better regulated. In contrast, vertical integration could be formalized in those railways that face strong inter-modal competition in order to make the concession more profitable. For the Lima airport, a fully vertically unbundled model would have been another option, given the large size o f the airport. In telecommunications, the main services-local, national and international calls-were designed to be provided inmonopoly conditions for five years, while other services were open to competition. The exclusivity period was supposed to provide the necessary favorable conditions for Telefonica del Perb to make the significant required network investments and to allow for tariff rebalancing. Given its large size and its control o f essential installations, TdP continues to preserve a high degree o f market power several years after the end o f the monopoly period, partly due,to high interconnection charges and a certain degree o f cross-subsidization in favor o f its subsidiaries. This calls for an important monitoring role for the regulator. The Peruvian Electricity Concessions Law (LCE) divided the industry into segments: generation (competitive) and distribution and transmission (both monopolies). Also the LCE determined two markets: a regulated market where distributors provide energy to users inside its concession area, and a free market, where distributors compete with generators to provide energy to large users. Vertical and horizontal integration i s allowed up to certain limits and subject to the antitrust agency's (INDECOPI) evaluation. This model i s currently very common both in developed and developing countries. Despite significant growth in the free electricity market, its development may have been restricted by limited access to distributors' facilities. The low levels o f competition may be due to the fact that conditions and access charges are not regulated, but freely agreed upon by the parties. Another problem is the high degree o f market concentration in the electricity sector resulting from horizontal or vertical integration. The implementation o f an ex-ante access regulation could facilitate the future market development. Furthermore, the roles played by OSJNERG (as the regulator in access controversies) and INDECOPI (as the authority in charge o f vertical and horizontal concentrations) are fundamental. Contract design: risk allocation. The experiences reviewed provide some positive and negative examples o f risk allocation. It was found that the government adequately ensured the stability o f contracts through contract-law and Stability Agreements. This scenario was reinforced in the electricity and telecommunication cases through the LCE and the Telecommunication Law which define the framework under which the sectors develop and determine many contractual tenns. The government also freed concessionaires from exogenous risks, including major force risks, in most cases. However, there could potentially be problems classifying and interpreting major force risks in certain industries as they are based on vague definitions. The government also assumed some commercial risks related to road demand forecasts. Another option to reduce uncertainty for the concessionaire without compromising government income could have been to endogenize the concession duration based on demand fluctuations, as has been tried in other countries. Commercial risks are also borne by the State in the telecommunications case through the establishment o f the exclusivity period. However, this case seemed justified to ensure the fulfillment o f investment commitments. xv There have been cases in which the State or other institutions imposed risks on concessionaires by generating uncertainty about the future. Examples include the lack o f a government decision on road tariff policy and the discontinuity in the electricity generation privatization program. Similarly, there has been uncertainty caused by Congressional actions in relation to the elimination o f the fixed part o f the telephone tariff (which would have changed the contract terms) and uncertainty created by the Constitutional Tribunal after its decision to restitute former TdP workers. Contract design: penalties and incentives. In most Peruvian processes, penalties are adequately specified in the contract itself or in the regulation framework to penalize failure to fulfill investment, retribution or quality obligations. Sanctions in concession contracts depend on the magnitude o f the fault. However, termination i s a possibility in most cases as a last resource mechanism in cases o f serious failures. Economic sanctions stipulated in the legislation conveniently link the magnitude o f the sanctions to the annual income o f the concessionaire in the public transportation and electricity sectors. However, for Telefonica, sanctions tend to be very low inrelation to revenue and provide weak incentives to influence behavior. Positive incentives for Telefonica involve the possibility o f prolonging the period to recover its investments. Every five years, the regulator (OSIPTEL) evaluates this option based on whether Telefonica has completed the terms o f its contract. For the electricity contracts, the only incentive i s the indefinite term o f the concession, which was deemed necessary to reduce risk to the company and guarantee optimal levels o f investment. Incentives have not been used very much in public transportation infrastructure concessions. In most cases, the main incentives have been concession duration extensions. Extensions were stipulated in the Matarani Port and Lima Airport contracts to encourage long term investments. For the Arequipa-Matarani road, the possibility o f extensions was offered as an incentive to avoid tolls increments. However, this mechanism was included in an addendum to correct a contract failure. For railroads, discounts on payments to the government were used to encourage investments, given the bad state o f the infrastructure. Contract design: certainty versus flexibility and the reasons for renegotiations. When a contract i s designed, it i s crucial to strike a balance between certainty and flexibility. Flexibility i s needed to adapt to unforeseen changes in the environment and to incorporate new technologies and innovation, while certainty is important to attract potential investors. Too much flexibility can be prevented by a suitable definition o f goals and obligations. Performance targets, for instance, let the concessionaires achieve desired goals while ensuring autonomy and use o f their know-how. Other specific targets tied to investment or technical requirements present more rigidity. For most o f the privatizations and concessions in Peru, the use o f such specific goals reduced flexibility and increased the need for renegotiations. Railways were the only case where performance targets were used. Allowing for a contract change based on mutual agreement between the parties i s an important and appropriate mechanism when the environment in which a contract develops i s subject to rapid changes. The telecommunications and electricity contracts allowed for that possibility in Peru. Renegotiation can be a useful instrument to address the inherently incomplete nature o f xvi contracts. However, when renegotiations are excessively used, doubt i s cast on the validity o f the contract and there i s an appearance o f opportunistic behavior by the concessionaire and/or the government, leading to a reduction in social welfare. In addition, if bidders believe that renegotiation i s feasible, it could cause them to make unreasonable commitments to win the bid, assuming that the commitments could be changed in the future. Furthermore, renegotiations tend to be bilateral negotiations between the operator and the government that are not subject to competitive pressures and lack transparency (Guasch (2004)). Despite the need for renegotiations to improve flexibility, transportation contracts in Peru have seemed too open to renegotiation. There has been more than one renegotiation in each transportation concession contract, with the exception o f the Matarani Port. Overall twelve addenda to the transportation concession contracts under analysis were signed between June 1995 and June 2003. Furthermore, the first renegotiation often took place in the first year o f the concession. This i s worrisome since the average time to renegotiate transportation concessions since the award date has been 3.1 years in Latin America and the Caribbean (Guasch (2004)). There were more than seventy modifications to the original Lima Airport concession contract in just three years. Most o f the important changes made it easier for the concessionaire to obtain loans to fulfill its investment obligations and changed the conditions o f the financial-economic equilibrium clause. In general, all the major modifications to this contract indicate changes to the "rules o f the game" after a short period o f time or failures in the original contract design. As a consequence, these addenda have been criticized by the public and the main unions o f the country, particularly in relation to the lack o f transparency in the renegotiationprocess. In contrast the Matarani Port concession contract had only one modification. Although this implied an important change, it seemed justified to make the original investment plan more reflective o f actual market requirements. This would indicate a failure o f contract design since the contract did not offer flexibility for such adjustments. In the telecommunications and electricity sectors, there have been only four addenda between 1998 and 2002, which i s fairly low in comparison to the number o f transportation infrastructure renegotiations. Similar patterns have been found for Latin America in general (Guasch (2004)). Moreover, most o f the concessions' first renegotiation took place in the third or fourth year o f privatization, which i s a relatively long period after the initial concession award. Both the LCE and the Telecommunications Law recognized renegotiations and covered many adjustment procedures, including changes to the concession area, tariff regime, and other obligations. The renegotiation processes were open and depended on the mutual agreement o f the different parties in all cases. TdP's concession contract had three main changes: i)the one year advance in the opening o f the telecommunications market to competition; ii)the two years postponement o f the application o f a productivity factor to price adjustments; and iii)the redefinition o f the local calling area. For some authors, there are clear indications that these changes mainly benefited TdP since they allowed the firm to enjoy two more years o f monopolistic rent, and productivity improvements were not transferred to users through lower tariffs. Other authors offer more positive interpretations. They argue that the advanced opening o f the market was socially beneficial and xvii generated significant user savings since users experienced an immediate one-time benefit from tariff reductions-the alternative being gradual reductions over three years. IMPACTS OF PRIVATE PARTICIPATION Firnz pevfornzance. The effects o f private participation on finn perfonnance have been largely positive. Telefonica del Peru has seen significant profitability and operating efficiency gains, even after controlling for a variety o f other possible explanatory factors. Results for electricity have been similar. Electrolima had significant improvements in profitability, operating efficiency, and leverage and Electroperu had significant gains in operating efficiency. Analyzing the impact o f transportation concessions on finn performance has proved more o f a challenge due to a lack o f data and comparable, non-privatized firms. Profitability, operating efficiency, and leverage ratios seem to have improved for the Matarani Port and railroads, but not all differences are significant. The Lima Airport concession i s the only one that does not show improvements. Sales efficiency i s the only indicator that shows a significant positive change, while most o f the indicators show performance declines. Possible explanations include the small amount o f post-concession data (only 2002) and the difficulty o f isolating the influences o f the private company (LAP) and the State-owned company (CORPAC) which operate together. However, a rudimentary analysis that compared LAP performance measures with combined LAP and CORPAC data showed that the L A P indicators were considerably better. Consumer welfare. While firm performance has improved in most cases, the effects o f privatization on consumer welfare are less clear cut. For telecommunications, usage o f long distance and international services has increased for consuiners in all expenditure quintiles, while local services use has increased slightly among the poorer quintiles (Figure 2). Local, long distance, and international rates have all fallen (to varying degrees), but the monthly fixed charge has risen substantially. Despite the fixed charge increase, all quintiles show a positive gain in consumer surplus if the tremendous increases in access are taken into account (there were around 5,000 households with telephones in 1993 and 150,000 by 1999). Quality o f service in terms o f percentage o f calls completed, time for repairs to be completed, and waiting time for line installation, have all improved substantially. In contrast to telecommunications, electricity consumers seem to have become worse off. In urban Peru, electricity consumption and consumer surplus have fallen for all expenditure quintiles. This can primarily be attributed to increases in the residential variable electricity tariff. Of course, privatization has also brought service improvements, including reduced time to respond to complaints, make repairs, and install meters. And the number o f customers nationwide increased by over 50 percent between 1994 and 2002. Transport concessions have had positive impacts on consumer welfare, for the most part. For the Matarani-Arequiparoad, gains from shorter driving times (measured in average labor income per hour) outweighed the consumer losses due to higher fares, producing a net positive impact. The impact o f the Matarani Port concession on consumers has been ambiguous: port fare prices increased, but ship turn-around times improved and import/export cargo traffic increased. xviii Figure 2: Usage of Telephone Services in Urban Peru According to Income Quintiles Local National 1991 1994 1997 1991 1994 1997 Yea, Year International 1991 1994 1997 Year Source: Peru Living Standards Measurement Surveys (1991, 1994 and 1997). The Central and Southern railroad concessions have probably not had a large impact on consumers because they face strong competition from bus and trailer transportation. Hence, they lack a monopoly's ability to increase prices. Passenger traffic i s also minimal on these routes. In contrast, the Southern Oriental railroad i s the only form o f public transportation to the Machu Picchu ruins, so it has a healthy tourist clientele. Although only one operator currently runs trains on this route, the concession contract stipulates that it i s open to competition. If competitors succeed in entering the market, consumers will likely benefit fiom better prices or service. Consumer impacts for the Lima Airport concession have been difficult to measure. Various fee changes have produced ambiguous results and it i s too early to see capacity improvements from new investments. Employment impacts. Employment has clearly fallen at TdP since privatization. However, a significant amount o f new employment has been created at companies that provide services to them. In most cases, these companies hired laid-off personnel from TdP and total employment in xix the telecommunications sector increased by 160 percent between 1993 and 1998. Privatized electricity companies have seen large employment drops, but it i s unclear whether laid-off employees were reabsorbed inthe sector. Employment has grown since the concession year for the Matarani Port and Lima Airport, whereas it has fallen substantially for railroads. The reason i s that the port and airport started a labor reallocationprocess and reduced employment prior to the concession. Government revenue. Because many state-owned enterprises were not profitable, privatization has led to an increase in recurrent government revenue. This analysis i s based on a set o f non- financial companies that were privatized in Peru, not just infrastructure. In line with this result, taxes paid by Telefonica have increased substantially since privatization. Nevertheless, there have been concerns about how one-time government revenues from the sale o f State assets have been spent. Ecorzonzy-wide impacts. We use a Computable General Equilibrium (CGE) model to estimate the economy-wide impact o f price changes in telecommunications and electricity. We estimate that an overall telecommunications price decrease o f 5.5 percent (a weighted average o f annual local, long distance, and international rate changes between 1994 and 1998) would increase the GDP o f the whole economy by 0.184 percent and create 19,000 new jobs (economy-wide). Similarly, a 5 percent decrease in electricity prices (the average annual decrease o f the variable commercial rate between 1994 and 1998) would increase GDP by 0.098 percent and increases employment by 10,000 new jobs. International conzparisons. Telefonica's coverage, labor productivity, and network digitization improvements have all broken with pre-privatization trends and outpaced privatized telecommunications companies in the rest o f Latin America. For electricity, the number o f connections has grown and distributional losses have fallen faster than both pre-privatization trends and Latin American privatization averages. PUBLIC PERCEPTIONS OF PRIVATE PARTICIPATION Public perceptions o f privatizationhave deteriorated significantly across Latin America since the late 1990s. To uncover more details about how this trend relates to infrastructure in Peru, the World Bank commissioned two studies. One relied on focus groups and the other on 1,808 household interviews. The two studies revealed similar results: Peruvians generally view the privatization process negatively. Because o f their precarious living conditions, survey participants' most serious worries related to the rising cost o f living and lack o f employment. Privatization was thought to adversely affect both by increasing tariffs and laying-offworkers. More than 80 percent o f those surveyed said private companies charge more than their State-run counterparts. Yet they also recognize that the private firms offer better services, are more efficient, pay better wages, and treat their employees better, while State-run companies generate more corruption. When asked whether State or private companies generate more jobs, surprisingly, there was no perceived difference between the two. This result is hard to reconcile with other survey findings that strongly associated privatizations with unemployment. A possible explanation i s that respondents were thinking about job generation in the medium to long-term, xx and not in relation to the labor shock associated with the transition to private control. Also notable was the fact that private companies were seen to treat their workers better (see Table 2). Though concessions were not fully understood, the concept did garner more support than privatization. State Companies PrivateCompanies N o answer Characteristics ( Y O ) (YO) (YO) Charges more 11 81 8 Offers better quality services 12 79 9 More efficient 25 66 9 Pays workers more 23 59 18 Treats workers better 29 49 22 Generates morejobs 42 42 16 I s more corrupt 59 26 15 These statistics reveal that there are many positive perceptions o f privatization that could be exploited. Better communication from the government regarding the reasons for and benefits o f privatization could be combined with more participatory processes. In other words, the government could educate citizens about new investments and tariffs adjustments required for infrastructure services to be sustainable. The filtering and watchdog role o f the regulator in limiting opportunistic firm (or government) behavior could also be better communicated and publicized. Armed with such knowledge and existing perceptions about private sector quality and efficiency, the public might support private participation in infrastructure. MOVING FORWARD Despite negative public and investor perceptions, there i s still a case for more private participation. Chapter 3 shows the impacts o f privatization in Peru, which were largely positive in terms of firm performance, government revenue, coverage, and quality. Similar results have been found throughout Latin America (World Bank and Inter-American Development Bank (2005)). Infrastructure has also been shown to improve output and growth, health and education levels for the poor, and reduce income inequality. With these benefits in mind, Peru has a full pipeline o f new infrastructure concessions and estimates o f the infrastructure deficit range from US$lO to 16 billion. Public financing will remain important and it can increase and leverage resources in the sector. However, public financing alone i s unlikely to be adequate. The report draws on private participation experiences in Peru and elsewhere to provide a blueprint for moving forward with the concessions process. The key elements-based on almost twenty years o f experience and over 1,500 episodes o f private participation-are the following: 0 Addressing social issues o Overalltransparency o A communications program o Community involvementand accountability:a bottom-upapproach o Accountingfor the poor and the losers 0 Improvingconcession contractstructure xxi Enhancingregulatorydesign Developingfinancialinstrumentsand tapping local markets Addressing social issues. To move forward with concessions, the concerns o f the public and negative impacts on the poor and laid-off workers must be addressed. Improving transparency, mounting an effective communications campaign, and involving local and regional stakeholders in private participation decision-making are all key elements to win public support and ensure more positive concession outcomes. Changing public opinion will require concerted public relations efforts that educate people about the benefits o f private sector participation, efforts which have been lacking in the past. Involving people at the community level can ensure that government, private investor, and end-user objectives are aligned and hold service providers accountable to their customers. Similarly, civil society and international organizations can play a role in holding infrastructure providers accountable and decentralization i s often a way to increase accountability and responsiveness to local needs. However, decentralization poses a number o f coordination challenges, both vertically (between central and local governments) and horizontally (among various subnational institutions). Improvements in infrastructure services have great potential to benefit the poor. However, for the poor to realize these benefits, specific attention needs to be paid to their needs and vulnerabilities throughout the private participation process. This i s necessary for both equity reasons and to improve public support for concessions. Connection and consumption subsidies, quality-price tradeoffs, flexibility in payment, and phasing-in periods for tariff adjustments are all measures that can help the poor. If a concession leads to employment losses, special consideration also needs to be devoted to laid-off workers. Measures to compensate the affected and to re-train them for reinsertion in the labor market ought to be considered. Examples include on-site tutorials on employment, micro-enterprise, or self-employment opportunities. Improving concession contracts.Sector performance, level o f conflicts, and public perceptions have been influenced by shortcomings in concession contract design. Though some mistakes have been made with concessions in the past, the experience o f more than 15 years makes it easy to avoid such mistakes in the future. Improvements over past designs should be a cornerstone o f the moving forward program in Peru. We draw on that experience, and in doing so, offer a blueprint for concession design that should significantly improve prospects for all concessions, enhance the awarding process, fine-tune incentives and penalties, reduce the incidence o f renegotiation, improve risk allocation, and facilitate dispute resolution. All o f these changes will have substantial positive economic consequences. In terms o f the concession awarding process, there should be two stages: one to prequalify interested parties on the basis o f experiences and a technical proposal if applicable, and a second one to solicit bids from the prequalified bidders using a single criterion for selection. There are three criteria that have generally been used to award concessions after prequalification has taken place. The first-minimum tariffs-have frequently been used, but they have some flaws. The salient choices for concession award criteria, based on efficiency, incentives, and effectiveness against renegotiation, should be either an annuity payment-canon or minimum subsidy for unprofitable concessions-or the least present value o f revenues (LPVR), when appropriate. xxii Broad and sweeping statements in concession contracts about financial equilibrium without reference points are undesirable and often have been the source o f conflicts and inefficiencies. Such clauses should not guarantee financial equilibrium without making reference to efficient operation and preserving the sanctity o f the bid. Financial equilibrium clauses should also specify the capital base that the firm i s allowed to earn a fair return on. To avoid opportunistic renegotiations, the concession contract should address as clearly as possible: i)events that would trigger tariff adjustments and the extent o f the adjustments; and ii)events that would trigger a renegotiation o f the contract with guidelines about the process and outcomes o f the renegotiation. Performance bonds and penalties should be adequately specified in the contract itself or in the regulation framework and penalize non-compliance-such as not meeting targets, failure to fulfill investment, retribution or quality obligations-with agreed clauses. Sanctions in concession contracts should depend on the magnitude o f the fault or non-compliance and on the extent o f reoccurrence. Enhancing regulatory design. The main objectives o f a regulatory framework are: i)to induce the regulated firm to operate at the lowest (efficient) possible cost; and ii)to closely align prices (tariffs) with costs allowing the firm to earn only normal profits. Usually there are other subordinate objectives as well that complement the main ones, such as increasing coverage and access, improving quality o f service, and addressing issues o f universal service obligations. To effectively determine tariff adjustments, the regulator must have the professional capacity, regulatory instruments, and accounting norms to determine the cost o f capital and evaluate the concession assets. Interms of determiningfuture tariffs, a number o f lessons have emerged. For instance: Arrangements and criteria for the re-adjustment and revision o f tariffs should be clear; 0 Re-adjustments should occur annually (or more frequently when permitted by law) in March or April based on inflation to end-December o f the previous year; 0 Extra-ordinary revisions o f tariffs should only be permitted in clearly defined circumstances, such as changes in the rates or calculation o f specific taxes and allowed costs, and should not cover the normal commercial risks o f providing the service such as changes in the cost o f labor or operational inputs. There are two different types o f regulatory regimes-rate o f return and price cap-with tradeoffs between each. Price caps provide incentives for securing efficiency gains, at least between tariffs reviews. They are low maintenance in the sense that they do not require high levels o f information about the concessionaire's operations (at least between tariff reviews). Yet, they induce a higher cost o f capital as a result o f their inherent riskiness and efficiency effects can be lost through rapid renegotiation. Rate-of-return regulation does not provide strong incentives to reduce costs and requires much higher information levels for the regulator. However, it generates a lower cost o f capital since the associated risk i s lower. Determining an accurate cost o f capital and valuation o f concession assets i s critical to ensuring that the concessionaire earns a fair return on investment. One way to value assets i s to rely on a full cost o f service approach that takes account o f the replacement value o f all o f the assets o f the xxiii concession and allows for the estimated depreciation o f these assets over standard periods rather than according to historical accounting conventions. The regulatory asset base would thus be equal to the replacement value o f the concession assets and would be quite separate from the book value o f such assets. Effective regulation requires good information about the operations o f the regulated firm. The regulator needs to periodically collect information about costs, revenues, prices, investments, financial data, and realized demand from the operator. Reporting requirements, including form and frequency, should be stated in the concession contract. The contract should also provide the regulator with subpoena powers to coerce the information from the operator and the right to impose significant and increasing fines in the event o f non-compliance. In order to make use o f data collected from the operator, the data should be standardized and the regulator must have the capacity to analyze it, ajob which can be quite complex. Operators face risks due to possible political interference and unpredictability in the implementation o f regulation, which can drive up the cost o f capital. To reduce this risk, regulatory frameworks and agencies should be created by legislation, rather than administrative procedures or presidential decrees. Laws are much harder to overturn or modify than decrees and contracts. Developing financial instruments and tapping local markets. Moving forward will require adjustments to project financial structures, Most o f the upcoming concessions require some form o f public financing support, such as equity, debt, credit support, or guarantees, because they are not financially viable. That requires innovative types o f private public partnerships (PPP). What i s emerging as a salient choice i s a PPP mode that entails concessions: the private sector makes the investments in the project and the public sector repays the investors in annual payments for the life o f the concession. Yet the long duration o f those projects and the perceived volatility of governments and the economies prompts the private sector to secure guarantees for the agreed government financial commitments. Peru presents increasing domestic financing options for infrastructure relative to the rest o f the region. Over the last few years it has deepened, expanded and improved its domestic capital markets and securities laws and regulation are relatively strong for the region. One option that could be further exploited i s the country's highly liquid Pension Funds. However, regulations tend to require investment-grade ratings for Pension Funds to invest. Guarantees from multilaterals-such as the World Bank Guarantee Facility-can provide that rating. In fact, Peru i s the first country in Latin America and the Caribbean to develop that instrument. Yet those types o f guarantees, while quite effective in lowering the cost o f capital and providing significant credit enhancement, might not be enough to bring in the sizable amounts required for infrastructure finance. That i s why consideration ought to be given to develop other instruments, such as well designed and appropriate infrastructure funds, which i s often not an easy task. An example being developed by the World Bank to facilitate the financing o f infrastructure projects i s the Private Infrastructure Financing & Support Facility (PIFSF) in Indonesia. In summary. To accomplish the changes mentioned above, adjustments will be required along the following lines: xxiv i) Transparency o f the transaction, transparency in the use o f the proceeds, or public funding, ifneeded; ii) An effective communication campaign motivating and defending the need for the project and for the private participation; iii) Communityinvolvementandbuy-in,requiringthat noprojectbedevelopedunlessit has the full support o f the affected community. In other words, it will be a bottom-up process unlike the previous top-down ones; iv) Explicit measures to account for the poor and for unemployment. Possibilities include social tariffs (but with improved implementation to minimize the leakages), social connections or a phasing out period if tariffs were to increase substantially, and reintegration or compensation for laid-off employees; v) A regulatory framework with clarity on guidelines for tariffs adjustments and more emphasis on regulatingby objectives rather than by means; vi) Stronger regulatory institutions with increased autonomy, both operationally, and financially; vii) The use and extent o f regulatory instruments, such as regulatory accounting and informational requirements, built into the contract; viii) Efficiency and benchmark measures to be used for regulatory purposes such as tariff adjustments; ix) Measurements o f the cost o f capital; x) Commitment to no opportunistic renegotiation and upholding the sanctity o f the bid; xi) More effective use o f performance bonds; xii) Allocation o f risks including reconsideration o f the balance between price caps and rate o f returnregulation; xiii) Conflict resolutionmechanisms, appeals and procedures; xiv) Overall concession design and incentive structures. xv) Increased utilization o f local financial markets. The need for increased and improved infrastructure remains. Public hnds remain scarce. Thus, the way forward requires some form o f private sector involvement, with adjustments made to the old model based on past experience from Peru and the rest o f the world, The Government o f Peru has already incorporated many key learnings in the early stages o f its second-wave concession program, but there remains scope for improvement. xxv L, U 0 U P n h P m E E .3 2 > X Y L % L X 0 '5 a erl 3 0'S 0 B t : -.. m 0, n 1. Private Participation Progress and Highlights 1.1 INTRODUCTION The 1990s were characterized by a worldwide, and quite particularly in Latin America, policy redirection toward private participation in infrastructure. That effort was part o f a broader set o f structural reforms aimed at economic modernization and liberalization. Bringing the private sector into infrastructure was driven by disappointment with ineffective state-operated utilities (low coverage and poor and inefficient quality o f service), the need for significant investments and scarce public funds, the promise o f private funding, and the greater flexibility offered by technological change and regulatory i n n ~ v a t i o n .In Peru, dismal macroeconomic indicator^,^ ~ record underemployment (86 percent), and hugely unprofitable public enterprise? in the late 1980s and early 1990s provided particular impetus for privatization. As a result, Peru launched a privatization program in 1991, hoping that private investment would be a key driver o f economic growth. Since that time, many o f the largest state-owned enterprises (SOEs) in telecommunications, electricity, and transport infrastructure (roads, railroads, a port, and an airport) have been sold or concessioned. After over a decade o f experience, enthusiasm for private participation in infrastructure has cooled considerably in Peru and around the world. Many developing countries have had to slow or stop the privatization process and SOEs still account for 14 percent o f telecommunications subscriptions, 40 percent o f electricity connections, and 89 percent o f water connections (Andres, Foster, and Guasch (2005)). The unwillingness to privatize or backlash against private sector participation appears to be associated with a rather negative perception o f the privatization process (as shown in Figure 3). The paradox i s that practically all o f the economic studies on privatization impacts show considerable benefits associated with the process (Chong and Lopez- de-Silanes (2005), Andres, Foster, and Guasch (2005), McKenzie and Mookherjee (2003), Birdsall and Nellis (2003)). That does not mean that there were not problems and issues and that the outcomes could not have been much better (among them were issues o f transparency, poor regulatory effectiveness, systemic renegotiation, employment issues, lack o f programs to account for the poor, etc). Yet, there appears to be a disconnect between the impacts and the perceptions o f the public. As o f now, governments are proceeding with the utmost caution in considering further private sector participation in infrastructure. Yet, the needs for infrastructure improvements remain large and public funds remain scarce. Thus the issue i s how to move forward. Economies from large-scale production and delivery have diminished in some activities, especially telecommunications and power generation. And regulatory innovation made unbundling possible. Unbundling promoted competition b y separating activities where economies o f scale are important (e.g. electricity transmission and distribution) from activities where it is less so (electricity generation). (World Bank (1994) and World Bank (2004a)). For the period 1987-92, the annual change in gross domestic product at constant prices was -4.9 percent, annual inflation was 733 percent (reaching a high o f 7,650 percent in 1990) and the real effective exchange rate depreciation for the period was -2.04 percent. Public firms accumulated losses o f more than U S $ 4 billion during 1989-1990 (Apoyo (2000)). 1 Figure 3: Population ExpressingDissatisfactionwith Privatization (percentage) 1or3 80 GO 40 2I! n VEN M E X B R A CHL PER ECU COL A R G ECL SLV NIL HND P A N G I M Note: The 1998 results reflect survey respondents who disagreed or strongly disagreed with the statement, "privatizations of state companies have been beneficial for the country." The 2004 numbers are o f those who were less satisfied or much less satisfied with public services after privatization, in terms of price and quality. Source: Latinobarometro surveys for 1998 and 2004 This is the motivation for this report, looking backwards to move forward. The Government of Peru has launched an aggressive second phase program o f infrastructure concessions covering all sectors, including transport, rural telephony, gas, energy and water and sanitation. To assist in the success o f this second phase, this report provides an in-depth analysis o f the past program in Peru-the telecommunications, electricity, and transport infrastructure privatizationsi concessions-analyzing the process, design and impact of the program as well as the public perceptions. The analysis aims to identify lessons for the on-going new phase. Table 4 shows a list o f the firms that are analyzed as well as the privatization dates. The telecommunications and electricity processes began in 1994, while those for transport began in 1998 (with the exception o f the Arequipa-Matarani road project). Water and sanitation i s not included because privatization processes in that sector are just beginning. Although Peru's experience with private participation has not been perfect, it i s considered to be one o f the best examples inthe region. This report is unique in that it draws on a broad spectrum o f sources to provide a comprehensive picture o f Peru's experience with private participation in infrastructure. First, it includes a descriptive survey o f each sector that experienced private participation (chapter 1). Next, it thoroughly reviews the privatizatiordconcession contract awarding process, contract design, risk allocation, competition, and renegotiations processes (chapter 2). Then, it uses econometric analysis to examine how private participation affected firms, consumers, employees, and government revenue (chapter 3). It also uses surveys and focus groups to understand public perceptions about private participation (chapter 4). And it offers some guidance for moving forward with the concession process (chapter 5). This rich variety o f sources and analytical tools distinguishes it from previous piecemeal analyses o f infrastructure inPeru. 2 Table 4: Infrastructure privatizations in Peru Sector Privatized entity I privatization Date Of I Telecom CPT and ENTEL I 1 (became Telefbnica del Peru) 1994 Electroperu 1 I (became 8 different companies) 1995-1996 Electrolima (became 5 different companies) 1994-1996 Electricity 1IElectro Sur Medio 1997 Electro Centro II 1998 II I Electro Nor Oeste I 1998 I I Electro Norte I 1998 I 1 Electro Norte Medio I 1998 I Arequipa - Matarani road 1994 Roads Ancon-Huacho-Pativilca 2003 E Lima Airport 2001 0 Airports %m Ports Matarani Port 1999 L t., Central Railroad 1999 Railroads Southern Railroad 1999 Southern Oriental Railroad 1999 1.2 NEWINSTITUTIONAL FRAMEWORKS In 1991, Peruvian legislation created new frameworks and government bodies to oversee private participation. Parallel frameworks were set up to sell public companies and to promote private investment in infrastructure.6 The Commission for the Promotion o f Private Investment (COPRI) was set up to sell public companies and began work in 1992, quickly privatizing telecommunications (1994) and various electricity companies. PROMCEPRI was established to promote private investment in infrastructure, but did not commence operations until 1996 when new laws were passed to facilitate concession^.^ PROMCEPRI began to promote private participation in transport infrastructure in 1996 with the creation o f a road network committee or CEPRI (special privatization committee). Originally, it was assumed that COPRI would close in 2000 after finishing the process o f selling public companies and PROMCEPRI would continue its work managing concessions. However, for political reasons, COPRI ended up absorbing Legislative Decree (L.D.) 674, passed in September 1991, oversaw the sale o f public assets and L.D. 758 (November 1991) oversaw private investment in infrastructure. 'In August 1996, L.D. 839 modified L.D. 758 and re-launched the process. In December 1996, a new legal framework to make concessions was established (Supreme Decree (S.D.) 059-96-PCM). 3 PROMCEPRI in 1998, along with its committees (CEPRIs). COPRI was later fused with other institutions to form PROINVERSION in 2002. Another important legal change was the Law o f Promotion o f Foreign Investment (L.D 662), which established the equal treatment o f national and foreign capital. This decree meant that foreign investment could occur in all economic sectors and carried out in any o f the managerial modes allowed by law. With this idea in mind, during 1992, diverse norms were instituted to facilitate the privatization process. Specifically, the State was authorized to grant, without limits, safeties and guarantees to protect acquisitions and investments. Foreign investors were also granted facilities to pay the taxes and debts o f former SOEs, and in some cases, these commitments were suspended till the end o f the process. In 1993, all the reforms previously adopted were raised to a constitutional level with the approval o f the new Political Constitution. The promotion o f free private initiative, the establishment o f equality between national and foreign investors, the fostering o f competition and equal treatment for all economic activities, and the potential to sign Stability Agreements between private investors and the State, were among the main changes. In addition, the State subscribed to several International Agreements for the protection o f foreign investment and conflict resolution through internationalreferees. Alongside the launching o f the privatization program, another set o f structural reforms were carried out. They were designed to promote market-based competition, free international trade, make labor markets more flexible, liberalize the financial system, eliminate price controls, and deregulate markets in various sectors. In addition, a law was passed which facilitated the nationalization o f foreign citizens who wanted to invest in Peru. All o f the reforms were complementary to the privatization program, as they allowed for more efficient performance o f private firms. 1.3 PRIVATE PARTICIPATION PROGRESS: ALL SECTORS Peru's overall private participation strategy started in late 1991. One o f the main objectives was to transfer the most important public companies in a relatively short period o f time. The initial tasks included definition o f the privatization methods, identification o f a priority list o f public enterprises to be privatized (depending on their specific weight, degree o f crisis, etc.), and the creation o f the individual privatization committees. The most common privatization practice adopted to transfer the control o f a firm was the public auction, because o f its transparent and competitive nature. The private participation process accelerated rapidly until the mid-l990s, after which time it tapered off. Table 5 shows the dollar amounts involved in each year's transactions and projected investment and Figure 4 shows the number o f processes per year. Between 1991 and 2002, 263 privatization and concession processes generated US$ 9.9 billion in revenue (included capitalizations) and US$ 11.45 billion in investment commitments. Telecommunications, mining, and electricity have generated the most transactional revenue and projected investment (Figure 5) and most sectors have made substantial privatization progress. Telecommunications and finance have been completed privatized, electricity has been about 65 percent privatized, while agriculture i s less than 40 percent complete (Figure 6). 4 Table 5: Amounts Involvedin Private Participations(US$Millions) TRANSACTIONS Options Rights, Projected YEAR orOfassets shares Concessions Small assets, Capitalizations Total investment Other ~~ ~ 1991 2.6 2.6 1992 207.5 1.4 208.9 706.0 1993 316.7 20.7 6.5 343.9 589.3 1994 2579.2 4.7 610.8 3194.7 2050.0 1995 1089.0 6.6 9.1 120.1 1224.8 70.1 1996 228 1.8 344.2 2.7 40.0 2668.7 2695.0 1997 447.1 99 8.8 126.4 681.3 706.2 1998 25 1.8 35.1 5.2 292.1 220.6 1999 286.3 10.9 3.1 300.3 166.6 2000 307 200 1 266 2002 355 ~ Total 7462.0 516.5 41.5 897.3 9845.3 7203.8 Source: Commission for the Promotion o f Private Investment(COPRI) for 1991-1999 data and Proinversion for 2000-2002 data. Figure 4: Number of PrivateParticipation Processes per Year 40 T 36 37 QE Source: Commission for the Promotion of Private Investment(COPRI) for 1991-1999 data and Proinversion for 2000-2002 data. 5 Figure 5: Private Participation RevenuesBy Sector up to 1999 (US$Millions) Telecoinmunicaiians Electricity n,ning Hjdrocarbon Industry Financing Fishery Transport I Tourism Agriculiure Other Projected Investment 0Transactio4 t - _~ _ _ _ _ _ _ _ ~ _ _ _ Source: Commission for the Promotion of Private Investment (COPRI) Figure 6: Private Participation Process Progress (1991-1999) Source: Commission for the Promotion of Private Investment (COPRI) Privatizations started to lose popularity in the mid-l990s, causing the second Fujimori government to look to concessions as a possible alternative. Concessions grant a private firm the right to operate a defined infrastructure service for a certain period o f time and to receive revenues from it. In contrast, privatizations sell State assets to the private sector for an unlimited operation period. In Latin America, concessions have typically been used for transport and water and sanitation, while privatization has been typical in the telecommunications and electricity 6 generation sectors.' See Figure 7 for a breakdown o f concession / privatization differences. Despite their differences, concessions and privatizations are part o f a continuum o f private participation and lessons regarding the awarding process, contract design, and impacts can be applied to both. Figure 7: Differences between privatizations and concessions DIFFERENCES P CONCESSION PRIVATIZATION Rights/ I I Transfer of right of Transfer of property Property exploitation and use o f the to the company company's physical assets I Duration I Limited/ 15-30 years Un1imited I I More Extensive Less Extensive I I II 1 I -1 Provision for Termination/ Generally non-existant Cancellationof the concession I 1 Government retains Non-existant proprietory rights Peru's concession framework was initiated in 1996, but transport concessions did not get under way until 1998.9The regulatory body (OSITRAN) and CEPRIs for the concession o f airports, ports, road networks, and mobile telephones bands (amongst others) were created and three railroads, one port, one highway, and the main airport (Lima) were then concessioned. However, approximately US$6,090 million o f infrastructure investment in this sector remains to be done (IPE (2003)) and there has been virtually no private participation in water and sanitation. After 1999, the process slowed down because o f political and social crises related to corruption o f high fbnctionaries, as well as external events such as the Russian crisis and "El Nifio." Despite the success in terms o f revenues and investment commitments, public approval o f the privatization process has been decreasing steadily (Figure 8). Several reasons for this have been suggested: lack o f adequate and realistic communication from the government regarding the scope and benefits o f the process; a perceived negative impact o f privatization on employment in the short and middle run; the expected rise in tariffs for public services; and public disapproval with the idea o f selling "strategic" enterprises." The following sections discuss each o f the infrastructure privatizations and concessions inmore detail. * In Latin America, 65 percent o f total private participation in infrastructure was in the form of concessions. This number was 98 percent for transport and 89 percent for water and sanitation. 'The Arequipa - Matarani road was concessioned in 1994, but it was not part of an institutionalized process, lo See Torero et.al (2004) for why this could be the case in the telecommunications sector. 7 .Figure 8: Rate of Approval of~the Privatization Process _ _ _ 1 . .. ... . . - . ~ I 91 I? 92.12 93 12 94 I? 95.12 96.12 97.12 98.12 99 I? I - Raw of Approbal of Privatization '. -___-- Source: Commissionfor the Promotionof Private Investment (COPRI) 1.4 TELECOMMUNICATIONS The Peruvian government sold both Compaiia Peruana de Telefonos (CPT) and Empresa Nacional de Telecomunicaciones (ENTEL) in 1994. CPT was the company that provided basic telecom services in the Lima area, while ENTEL was the national and long distance carrier, as well as the provider o f local telecom service in the rest o f Peru. Divestiture took place through an auction using a first-price sealed bid mechanism. Approximately 35 percent o f CPT and ENTEL common shares (the minimum required to give the buyer control o f the merger) were sold to a Spanish based venture, Telefdnica de Espaca. Telefonica was no stranger to acquiring telecom providers in Latin America, as it had already purchased the former Telefonos de Chile (currently known as CTC) and Argentina's ENTEL. The outcome o f the auction was overwhelming: Telefonica paid $2,004 million, while the second highest bid was $800 million, an amount closer to the US$536 million base price set by the government.'' Shortly after buyingboth companies, Telefonica de EspaEa S.A.merged them and created Telefonica del Peru S.A.(TdP). A t the same time, the government created the Supervisory Agency for Private Investmentin Telecommunications (OSPITEL) to regulate TdP. The privatization process o f CPT and ENTEL (now TdP) continued in the following years. In 1996, there were still 65 percent o f shares remaining, divided between minor shareholders (36.3 percent) and the Peruvian State (28.7 percent).I2 The latter decided to sell 26.6 percent o f the firm shares to small individual investors through a successful Citizen Participation System. The sale o f CPT and ENTEL to TdP i s considered a large accomplishment o f the government because o f the considerable amounts involved: it generated US$3.6 billion in state revenues and US$1.56billion ininvestment commitments. IIAccordingto IPE (1999),this was the highest price paidper line in Latin America. I'Telefhica de Espafia only purchased35 percent of the original shares. 8 As part o f the initial auction process, bidders had to agree to detailed investment requirements. For example, TdP was required to install 638,000 new lines in Lima (double the number installed at the time) and 559,600 in the rest of the country (1.7 times the number installed). In addition, there were goals related to quality and failure rates. In exchange, TdP was given a national monopoly over local, long distance, and international service and the provision o f lines for five years (initiall~).'~They were also given a 20 year operating license with possible five year extensions. Since the end o f the monopoly period, several competitors have entered the market, but TdP still dominates the field. One o f the main problems in the telecommunications sector was low coverage levels. Based on the country's GDP per capita, Peru should have had a telephone density o f 11 lines per 100 inhabitants, but had a mere 2.9 lines in 1993. This i s very low compared to other countries in the region, such as Argentina, Brazil, Mexico, Colombia, Chile, Bolivia and Ecuador. The limited extension o f the network was the result o f declining fiscal revenues, a debt crisis, and the fact that (subsidized) rates did not reflect the cost structure. These factors constrained plans for expanding the network, and the end result was a low level of telephone density. This led to a growing unsatisfied demand that manifested itself in the waiting time for a telephone: 118 months on average in Peru in 1993 compared to 17 months in Colombia and 11 months in Mexico. Since privatization, coverage has expanded considerably (Figure 9). It has risen from around 700,000 lines in service in 1994 to over 2 million in 2002. In 2002, there were 6.15 fixed lines per 100 inhabitants and TdP met its investment commitments by 1998. This may be the reason why TdP decided to advance the date for terminating the monopoly period. There was a small decline in the number o f lines in service after 1997, possibly as a consequence o f high fixed charges, the increase in cellular phone penetration, or even market saturation. Figure 9: Evolution of the number of lines installed and in service 3.E+06 7 Rivatization I I' 2.E+06 4 II ............................. _ , . * I , I .I- /' 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 .......No of fixed installed lines -Noof fixed lines in service - - No of mbil lines in service Source: Osiptel & Entel Peru. l3Although the monopoly, according to contract, was initially scheduled to expire inJune 1999,TdP shortened it to expire on August, Ist, 1998. 9 The quality o f services provided by the state-owned enterprises CPT and ENTEL was below international standards. In 1992, only 35-40 percent o f all phone calls were completed. This low efficiency was partly due to the small size and obsolete technology o f the network (only 41% o f the network was digital), which made it prone to congestion. Inadequate maintenance o f the cables was also a factor. Telephone cables have an expected lifespan o f 15 years, but by 1993 some cables had been in service for over 60 years. Now, the network i s almost entirely digital and virtually all calls are completed (Table 6). In addition, the percentage o f repairs completed in less than 24 hours has gone from 49 percent before privatizationto 99 percent in 2001, Table 6: Service quality indicatorsin the TelecommunicationsSector 1992 1995 1998 2001 Local MIIS completed*, YO nla 95% 100% 100% National long-distance calls completed', % nla 50% 54% 100% International long-distance calls completed*, % nla 91% 100% 100% Digitization of the network, YO 41% 77% 90% 96% * In Lima metropolitan area Source: OSIPTEL; Telefonica; CPT. The low levels o f investment by CPT and ENTEL are partially a reflection o f the companies' low earnings, since telephone service charges were below costs. This prevented them from generating the funds needed to expand the network or to improve service quality. Political, rather than technical, criteria guided the administration o f tariffs, generating a distortion in the tariff structure. The government subsidized certain services (local telephony) with high rates for international long-distance and other services. Tariffs differed substantially from those in other countries o f the region. The cost o f installing a telephone line in Peru (US$ 1,500 in 1993) was above the average for Latin America, and the rate for international long-distance service was extremely high. On the other hand, Peru had a very low basic monthly rate (US$ 2.00 in 1993). After privatization, the monthly fixed charge increased substantially, while local, long distance, and international rates fell by varying degrees (Figure 10 and Figure 11). Installation charges have now dropped to about US$ 170 and installation takes only about 10 days. Despite these improvements, TdP i s not a popular company-Peruvians feel that they pay the highest telephone rates in Latin America. 10 Figure 10: Evolution of Basic Local Rates (Real Nuevos Soles) 0.25 0.20 0.15 5 0.10 5V I I I 0----A -+e--+- +--A! 0.00 1990 1991 1992 1993 1994 1995 1996 1997 1998 19992000 2001 Monthly Fixed Charge -Rate per minute of local call Source: Comision de Regulacion de Tarifas de Comunicaciones & Osiptel. Figure 11: Evolution of Average Long Distance Rates (Real Nuevos Soles) Privatization 14.00 12.00 1::::2 10.00 ... -- 0.20 T 4.00 * -~ 2.00 0.00 I I , - 0.00 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Average rate per rninute of domstic long distance call -Rate per minute of call to USA - ILD Source: Comision de Regulacion de Tarifas de Comunicaciones & Osiptel. 1.5 ELECTRICITY In 1992, the government approved the Law o f Electric Concessions (D.L. 25844) which unbundled power generation from electricity distribution and transmission. Power generation was opened to competition, as opposed to transmission and distribution which are usually considered natural monopolies. Between 1994 and 1998, privatizations led to 14 private distribution and generation companies with commitments to increase total generating capacity by 560 MW (compared to an installed power base of 4,379 MW in 1994). At present, the privatized companies represent 64% o f the power generation capacity o f the National Electric System and 79% o f the distribution service. The government created two regulatory bodies in this sector: the Supervisory Agency for Private Investment in Energy (OSINERG), and the Commission of Energy Tariffs (CTE) which was 11 later absorbed by OSINERG. The privatization process in this sector i s not yet concluded because one o f the major generating enterprises, Central Hidroelectrica del Mantaro, and all o f the distribution enterprises in the south are not yet privatized (see Box 1). Yet, the electric sector has become the second largest generator o f revenues and investment commitments for the State: US$2.3 billion in revenue has been collected and around US$700 million has been engaged as investment commitments. Table 7 summarizes the structure o f the industry and privatization progress. State Owner Enterprise Activity Post Privatization Efective Control Year of Sale Price Privatization (US$ million) Electro Peru State Cahua Private 1995 41.8 ETEVENSA Private 1995 120.1 Electro Peru Generation EGENOR Private 1996 228.2 EEPSA Private 1996 59.67 EGASA State EGESUR State EGEMSA State Generation EDEGEL Private 1995 524.5 Privatization increased the power generation capacity by more than 25 percent in five years, tilting the balance from hydro to thermoelectric generation. Peru's heavy reliance on hydroelectric power makes the country especially vulnerable to climatic circumstances, like floods and droughts. A more balanced generation profile was needed, and thermo-electrical generation has been increased over 136 percent since 1996 to meet this demand. Privatization has also had an important positive effect on the quality of life of the population through increased access to electricity. For example, Edelnor's investments have added 225,000 customers in approximately 500 communities to the network (Figure 12). Overall, the electrification coefficient has risen from around 60 percent at the time o f privatization to over 75 percent by 2002 (Figure 13). 12 Figure 12: Number of ResidentialElectricityClients(Lima) 1 E+06 - 1 E+06 - ! ,' ......I 1 E+06 /- I Oectrolim d Melnor .......Edelnor + Luz del Sur - b- Luz del Sur Source: CTE Statistical Year Books Figure 13: ElectrificationCoefficient ("/o) 80% 1 Rivatization I I 65% - 60% - 50% T- , 1989 1990 1991 1992 199319941995 1996 1997 1998 1999 2000 2001 2002 Source: IPE (2003) Similarly, performance and customer service have improved, transmission losses have been cut by half, and informal connections have been significantly reduced. To provide a more customer- friendly service, distribution companies started call-lines to ease customer doubts. Distribution companies have made huge improvements in efficiency, which can be seen in the reduction of time required to obtain certain services (Table 8). In the case o f Edelnor, time to respond to complaints has gone from 27 hours to 1 hour, defects are repaired in 1 day instead o f 7, and meters are installed in 2 days rather than 45. 13 Table 8: Efficiency Improvements in the Electric Sector Electrolima Edelnor Luz del Sur 1990 1993 1994 2000 1994 2000 Distribution losses (YO) 11.18% 19.07% 16.16% 9.30% 17.22% 8.20% Collection effectivity (days) 157 93 111 67 112 55 Time to respond to querieskomplaints, hours 27 1' Time to repair defects, days 7 1' Time to install meter, days 45 2' In the pre-reform period, rates for final users varied according to usage: industrial, commercial, residential, public lightening, and general agricultural, among others. After the reform, energy tariffs were simplified, leaving only the residential (BT5) and commercial (MT4) fixed and variable rates. As Figure 14 shows, the residential and commercial fixed rates remained roughly constant during 1992-99. On the other hand, the residential variable rate increased dramatically between 1993 and 1994, was relatively stable between 1995 and 1996, decreased between 1997 and 1998, and increased again between 1998 and 1999. The commercial variable rate showed a slight increase because o f the need to ensure sector viability. Figure 15 complements this analysis, including years 2000-2003 and comparing electric with water tariff evolution. Water tariffs have increased substantially, while residential electric tariffs (BT5) have not grown since 1995, despite water being state-owned and electricity being privately operated. Figure 14: The evolution of the index of basic electric rates (August 1990 = 100) 420 i22 370 320 270 220 170 120 70 ~__I._BT5 (Residencial) Variable - - - MT4 (Commercial) Variable 14 Figure 15: Household's Electric Tariff and Water Tariffs - L i m a (Real Nuevos Soles) si 807 Si 50 ~ /-4------ z Si 30 4 -Household Water Tariff (30cubic meters) HouseholdBectric Tardf up to 100 Kwih --._HouseholdBeclricTartff Tariff up to 180 K w h HauseholdBectric up to 30 Kwih Source: INEI With respect to the average generation costs, shows that there i s a clear gain in efficiency for the case o f Electroperu (which i s still state-owned) although for other generation companies the results are not clear. Finally, summarizes the substantial growth and service quality improvements the sector experienced between 1993 and 2002, many o f which have been touched on above. For instance, the energy network grew from 24,370 kilometers to 36,170 kilometers between 1993 and 2001, an increase o f 48 percent, and employee productivity doubled. Figure 16: Average Cost per KWh Supplied to Final Consumers - Generation (Real Nuevos Soles) Rivatization I :I I 0 , - I 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 ....... Oectroperu -:e Cahua - rb Egenor Eepsa Source: CTE Statistical Yearbooks 15 Table 9: M a i n Indicators of the Electric Sector, 1993-2002 1993, 1994 1998 2002 Installed power (MW) 4,379 5,515 5,917 Per capita energy production (Kwh) 501 n.d. 812 Number o f customers (in millions), national 2.3 1 3.05 3.62 Energy connection cost (US$) * 250 n.d. 124 Loss o f energy in the distribution systems at national level, % 20.6 12.4 9.1 Per capita energy consumption (kWh) 404.4 566 n.d. Electric energy sales (GWh) 9,335 14,009 n.d. Energy network (thousands o f Kms) * 24.37 n.d. 36.17 Average days to install an electric meter * 45 n.d. 1 Productivity by employee Edelnor 509 1,066 n.d. Luz del Sur 426 1,000 n.d. * Data from 2001 Note: Privatized companies represent 64% o f the power generation capacity o f the National Electric System and 79% o f the distribution service. Source: Electricity Head Office; MEM,Statistical Yearbook 1998; Edelnor; Luz del Sur; CTE and Continental S A P , Evolution and Perspectives o f the Peruvian Electric Sector; IPE (2003) Despite these performance improvements, there has been considerable public opposition to continuing the privatization process. In fact, an attempt to privatize two companies in Arequipa in 2002 failed due to violent public protests. This event has frozen future electricity privatizations, despite the fact that a number o f important generation and distribution firms remain in public hands (see Box 1). What's more, four regional electricity companies in the north and center o f the country that were sold to a local group under very favorable credit conditions ended up reverting to State control (see Box 2) . Box 1: Public rejection of electricity privatizations in Arequipa Egasa and Egesur were established in 1994 as electrical generation companies in Arequipa as property o f the Fund for the Financing o f the State Entrepreneurial Activity (FONAFE). In 1996, the CEPRI o f the regional electricity companies included them in the privatization process, first separately, but later together.* Under that scheme, the State sold both companies in June 2002 via public bidding. Tractebel S.A. was the highest bidder and was awarded the bid for US$167.4 million. The population of Arequipa protested the awarding o f the bid to a Belgian company because o f the perception that the sale o f the State assets would bring increased unemployment and more expensive rates. Faced with strong popular opposition and violent protests, the mayor o f Arequipa filed suit with the Superior Court o f Justice o f Arequipa to stop the privatization process, claiming that the citizens' constitutional rights had been violated. The suit demanded that the city o f Arequipa be recognized as owner o f shares o f the electrical companies. The court ruled in favor o f the city and the government was forced to abandon all privatization efforts. President Toledo ultimately was forced to apologize and reshuffle his cabinet and the privatization program was then shelved (Economist Intelligence Unit (2005)). *Resolution Suprema No. 438-96-PCM. 16 Box 2: A failed privatizationin the electricalsector The regional electric companies were included in the privatization process in July 1992. The Special committee responsible for handling the process was created in 1996 and COPRI approved the privatization strategy in 1998. There were three main objectives o f this process: encourage participation o f domestic investors, improve service quality, and to increase coverage inrural and remote areas. Four electric distribution enterprises in the northern and central part o f the country were initially selected to be included in the process: Electro Nor Oeste S.A. (ENOSA), Electro Norte S.A., Electro Norte Medio S.A. (Hidrandina), and Electro Centro S.A. The State was selling 60 percent o f its shares in the enterprises in the following terms: 30 percent initially and the remaining 30 percent could be bought by the buyer within the 12Ihand the 24Ih month after the contract was signed. The price for the remaining 30 percent was the one offered in the winning bid plus 3 percent annually. There were very favorable credit conditions to purchase the initial 30 percent. The buyer could pay 10 percent in cash and the rest in 24 semi-annual installments. The interest rate was Libor + 2 and there was a grace period o f up to three years during which the buyer only paid interest. As collateral, the buyer put a letter o f credit for 20 percent o f the debt as well as the enterprises' shares. If the buyer did not exercise its option to buy the remaining 30 percent two years after the contract was signed, the State was free to sell them. Inorder to guarantee that the new buyer could control the enterprises before exercising its option to purchase the remaining 30 percent, a shareholder pact was devised to assure that the buyer would control the voting decisions o f this additional 30 percent. A trust fund with these shares was set up in COFIDE for that purpose. Inpractical terms, the new buyer was granted voting powers for 60 percent of total shares by paying upfront for only 30 percent o f them. Jose Rodriguez Banda S.A. offered the highest bid for all four companies and the contract was signed in December 1998. summarizes the bidding offers. Enterprise Base price Number of Winning bid Winning bid/ (US$ 000) bidders (US$ 000) Base price Electro Nor Oeste S.A. 12,956 2 22,885 1.77 Electro Norte S.A. 11,345 4 22,119 1.95 Electro Norte Medio S.A. 36,079 1 67,879 1.88 Electro Centro S.A. 26,275 1 32,690 1.24 As shown in the table, the winning bids were considerably above base prices and, as time would confirm, overly optimistic. In fact, the buyer did not exercise the option to buy the remaining 30 percent o f the shares within the period stipulated in the contract and hence the State regainedthe control o f the enterprise. This was not a smooth process. The private owner alleged that the State did not honor several o f its obligations in the contract and hence was no longer responsible for buying the 30 percent. After some legal disputes, there was a settlement that allowed the State to repurchase the shares, although at a lower price than the bidding price. With the proceeds o f the shares plus an additional disbursement, the domestic investor cancelled the initial debt for the 30 percent o f shares. There is no pending obligation from the private investors in this case. B y the end o f 2001, the four enterprises were back in full control o f the Peruvian State. 1.6 WATER AND SAR'ITATION Water and sanitation has seen very little private participation since 1990. Perhaps the only exception was a 27 year BOT concession to the Agua Azul Consortium (comprised of the Italian 17 f i r m s Acea and Impregilo and the Peruvian firm Cosapi) in 2000. The objective o f the concession was to optimally utilize surface and ground water in the Chillon river basin to be sold in potable form to Sedepal (Empresa de Sewicio de Agua Potable y Alcantarillado de Lima). To this end, Agua Azul agreed to operate wells and build and operate a water treatment plant that would provide Sedepal with water to service 800,000 inhabitants o f the districts to the north o f Lima. Sedepal agreed to pay a certain amount per month to Agua Azul in exchange for a predetermined volume o f water. Part o f the payment (80 percent) would be converted to dollars and adjusted annually according to the US consumer price index and part (20 percent) would be paid in nuevos soles, adjusted every four months according to changes in the Peruvian price index. The concession contract had clauses to handle times when Agua Azul delivered less water than agreed to or when Sedepal wanted to purchase additional water as well as various penalties for not fulfilling the contract terms. Agua Azul has been successful in tapping local capital markets and raising long-term financing on good terms, in large measure due to the guaranteed revenue stream from the State. T o date, the concession has been operating well and Sedepal has been fulfilling its payment obligations. While the Agua Azul concession has been successful, it is not the same as a traditional concession because 100 percent o f its revenues come from a contract with the State-owned Sedepal. (Apoyo and Asociados (2005)) Sedepal was the only water-service provider included in Peru's initial private participation program, but it still has not been sold or concessioned. Despite this, the government has tried to improve the company's services and coverage. Additionally, in 1992 the government created the Super-intendencia Nacional de Servicios de Saneamiento (SUNASS), the National Office for Services o f Sanitation, as the regulatory body for this sector. SUNASS i s responsible for controlling the quality o f the service, tariffs, regulation, new investment, and inter-sector coordination. While Agua Azul has been the only major concession so far, there are other water and sanitation concessions in Peru's pipeline. However, the lack o f experience to date means that the water sector i s not covered in this report's analysis o f processes and impacts (chapters 2 and 3). 1.7 ROADS In 1999, Peru had a road network o f about 78,127 km o f which only 13 percent was paved, 24 percent were improved (unpaved) and 63 percent were unimproved (unpaved). Compared to other countries in the region, Peru's roads were in poor shape. While Peru had 6,692 km o f paved roads per million square kilometers of territory, countries like Venezuela, Argentina, and Chile had a density five, three, and two times higher, respectively. According to the MTC's 2001 highway improvement plan, an estimated US$5,005 was necessary to improve the existing roads (not including expansions). Based on these circumstances, the government has looked to concessions to improve the road network. The Matarani-Arequipa was concessioned by the MTC in 1994 as a one-off event, A couple o f years later, the "Road Development Plan 1996-2005" proposed concessioning eleven separate road networks. The road networks grouped high-traffic highways with lower demand roads. However, this cross-subsidy scheme was later found to be unfeasible due to fiscal 18 constraints. The Red Vial No. 5 was prioritized as the first to be concessioned and the other ten road networks were put on hold. 1.7.1 Arequipa - Matarani Road The Arequipa - Matarani road i s approximately 104 kms long and connects Arequipa city and Matarani Port or Mollendo and 110 sea shore. In November 1994 the road was awarded as a concession to Graiia y Montero (a construction company) through CONCAR S.A. It was a BOT (Build, Operate and Transfer) contract that stipulated the rehabilitation, maintenance, and exploitation o f the highway through two tolls. The contract length was initially 74 months, but 55 months were added later and an additional 13 months were allowed if CONCAR did not raise the tolls. The concessionaire was initially required to invest US$6.6 million in road rehabilitation during the first six months, although this amount was later increased to US$8.2 million. In the end, the enterprise invested nearly US$ 10.5 million as a result o f mistakes in the measurement o f the asphalt used. According to the contract, after the first six months, only periodic road maintenance expenditures were required. Figure 17 shows actual maintenance expenditures. Figure 17: Maintenance expenditureson the Arequipa-Matarani road 250 1 n tft200 v) 3 'ij 150 -0 v) az 100 c +0 50 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 Source: CONCAR . As the road connects to the Matarani Port and sea shores, the traffic is composed o f heavyweight vehicles bound for the port (mostly from Bolivia), as well as lightweight vehicles bound for the shores. As shown in Figure 18, the vehicle traffic has grown since the concession from 1.2 million vehicles in 1995 to 1.33 million in 2002. The difference in traffic growth between heavyweight and lightweight vehicles i s explained mainly by the development o f the Matarani Port. The increase in heavyweight traffic explains why toll income has grown (Figure 19) despite fare reductions (Figure 20). The evolution o f fares was established inthe concession contract. 19 Figure 18: Vehicle Traffic on the Arequipa-Matarani road 16001 1400 +j1200 .-c!g 1000 + 800 'c1 c 600 v) 400 200 0 1995 1996 1997 1998 1999 2000 2001 2002 Light-w eighted vehicles BI Heavy-w eighted vehicles Source: CONCAR Figure 19: Monthly Toll Income (US$)on the Arequipa-Matarani road O i 10-93 10-94 10-95 10-96 10-97 10-98 10-99 10-00 10-01 10-02 10-03 -Light-w eighted vehcles - Heavy-w eighted vehicles Source: OSITRAN 20 Figure 20: Fares (US$) on the Arequipa-Matarani road 1.7 1,I4 -"-"'-.................... .___ ....,_ _ _ ..I. ........ 1.0 1 -,-...,. -,., 0.9 - .----.- .,.--_ -".11111.9" ~ 08-.---- - 7 ------r - -7 --- 7- - -- 1995 1996 1997 1998 1999 2000 2001 -Toll per light-w eighted vehicule (US$/vehicle) .I- Toll per heavy-weighted vehicle axis (US$/axis) Source: OSITRAN This increase in toll income implies higher government revenue-the concession stipulated that 12% o f toll income would be returned to the State through a Road Fund (Fondo Vial). This fund i s charged with the rehabilitation and maintenance o f the national road system. Those disbursements represented nearly US$ 1.9 million between 1995 and 2002. There i s also a regulation fare o f 1% o f gross income distributed to OSITRAN (the transportation regulatory body). Finally, to cover traffic risk, the contract specified that the concessionaire would be covered through fund contributions if traffic was lower than initially predicted. The Arequipa-Matarani road concession process was unique in that it was conducted by the Ministry o f Transport and Communication (MTC) outside o f the COPRI/PROMCEPRI institutional framework. Because o f this, the process was short on resources and experience and the contract had a number o f flaws. 1.7.2 Ancdn-Huacho-Pativilca Road In 2003, the Ancon-Huacho-Pativilca concession was awarded to NORVIAL S.A.for 25 years. In accordance with the national road development plan, it initially included two roads o f the Red Vial No. 5: a profitable one (Ancon-Huacho-Pativilca, 161 kms and 11,000 axles per day) and an unprofitable one (Lima-Canta-Unish, 220 kms and 1,500 axles per day).I4 The concession incorporated a cross-subsidy design with a co-investment scheme between the State and a private counterpart. But due to the lack o f fiscal resources, the project was divided and the unprofitable road taken out. In this sense, the auction changed from a "least State investment" auction to a "best retribution to the State" auction where the concessionaire must pay a portion o f its monthly toll income to the State. The contract followed a BOT (built, operate and transfer) scheme and its main objective was to construct, rehabilitate, and maintain road infrastructure and to exploit services. The contract guaranteed a minimum level o f traffic and a minimum annual income up to 10 years after the first construction stage disbursement. The concession contract established investments l 4 Urrunaga, et. a].200 I 21 obligations o f US$ 26.4 million in the first stage and US$ 35 million in the second stage. The contract also stipulated a maximum fare to be collected. This fare will change through the construction stage (stage l), after which time it will have a base rate o f US$ 1.50 and only be adjusted for inflation. 1.8 PORTS In May 1999, a 30-year BOT concession was granted for the Matarani Port. It i s the only port concession in Peru so far. Although the 110 port concession was proposed that same year, there were not enough final bidders in its auction. After this concession, the process was stopped by a law (27396), which stated that no process could be carried out until the National Port System Law was in place. Although the Law was approved in March 2003, further concessions have been delayed. The Matarani Port i s the second most important port in Pet4 and i s situated on the Arequipa coast, south o f Lima. It handled 8.3 percent o f the country's ship traffic and 8.4 percent o f its cargo in 2001. In comparison, the main port o f Callao handled almost 60 percent o f ships and 70 percent o f cargo. The primary influence zone o f the Matarani Port i s the southern departments o f Pet4 and the north o f Bolivia. Before the concession the port had 3 bollards for 35,000 DWT (Death Weight Tonnage) ships and a roll on - roll off bollard for rolling load ships. Moreover the port had specific equipment to unload grain and cereals (53,000 MT capacity) and to put on board minerals (600 to 1000 MT / hour). The concession was awarded to Santa Sofia Puertos S.A. and TISURi5 (Terminal Internacional del Sur) was charged with the operation o f the port. Santa Sofia Puertos won the auction with a US$ 9.68 million bid and an infrastructure and equipment investment commitment o f US$ 4.6 million over the first 5 years.16 The winner was also forced to pay 5% o f its gross income to the State and 1% to the regulator (OSITRAN). The contract specified maximum rates that could be charged for regulated port services over the first 5 years. These services are considered an essential part o f the facility and include: ship services (moor, unmoor and use o f bollard) and loading services (use o f docks and grain store). Towing and pilotage services are also regulated services but are not offered by the operator. Nonetheless, the contract allowed vertical integration between port services and the port operator. The main productive changes after the concession were related to grain storage capacity and were due to a new commerce strategy focused on grain export and import to P e h and grain transit to Bolivia. Other productive changes were due to pre-auction measures. The State, through ENAPU (Empresa Nacional de Puertos), reduced the number o f workers and invested around US$ 7.65 million (Alcazar & Lovaton (2003)) to clean up the enterprise and make it more attractive. Both actions increased the efficiency index, i.e. increases in the grain unloading IsTlSUR was formed from 99.99999% of Santa Sofia Puertos S.A. and 0.00001% of Calixto Romero Seminario. Both enterprises belong to Romero Group. Io The investment commitment for the first 5 years was later increased to US$ 5.7 million. Eventually, US$ 9.5 million in investments (subject to traffic levels) was required. 22 speed, reductions in logistic expenditures through a new grain wastage policy, quicker turn- around time for ships, more terminal occupation, and higher labor productivity.l7 Figure 21 shows how cargo traffic to and from Bolivia i s mostly responsible for the increase in the total cargo o f Matarani Port. This increase i s a result o f the new grain commerce strategy of TISUR and the lack o f response from the rival port, Arica.'8 Figure 21: Evolution of Cargo by Type of Transaction (Tons): Matarani Port 1,400,000 - Concession ! A I Coasting Trade Exports to R r l j Imports to R r u Traffic to Bolivia Source: Alcazar & Lovaton (2003). As mentioned before, the new commerce strategy emphasizes grain and cereals. This has coincided with an increase in Bolivian cargo which i s mostly wheat and soybeans. Figure 22 separates the shipments according to the type o f cargo and it i s possible to differentiate the increase o f grain and cereals import-export against other shipments such as minerals, container cargo, etc. One reason for the decrease in minerals cargo i s the suspension o f the Tintaya mine whereas the reasons for the reduction in container cargo were the transfer o f an important company (Gloria S.A.)to Lima and the drop off o f regional production. OSITRAN Inform 03 1-03-GRE-OSITRAN. "Evaluacion economica de la Concesion del Terminal Portuario de Matarani: Ai70 2002" Mayo 2003. ' * Forhrther details see Alcaza, Lovaton (2003). 23 Figure 22: Evolution of Cargo by Type of Product (Tons): Matarani Port 1,400,000 1,200,00011 1,000,000 800,000 600,000 400,000 200,000 0 1996 1997 1998 1999 2000 2001 2002 rn Grain and Cereals0 Minerals 0 Others Source: Alcazar & Lovaton (2003). With regard to output, quality and efficiency indicators, significant improvements took place. The number o f ships processed and the number o f tons per ship has increased during the last few years. The number of ships in 1990 was 125 and by 2001 it reached 273. Similarly, the tons per ship increased from 4 tons in 1998 to 6 tons in 2002 (Figure 23). Figure 24 and Figure 25 show how the average hours a ship remains in port (quality indicator) and the average cost per dock usage (efficiency indicator) diminished during the past years. Despite these improvements, the concession has been criticized for excessive vertical integration and market concentration. Figure 23: Average Tons per Ship: Matarani Port 6 00 5 00 4 00 3 00 2 00 100 - _ _ 1998 1999 2000 2001 2002 Source: Alcazar & Lovaton (2003). 24 Figure 24: Number of Ships Processed and Hours Stayed per Ship: Matarani Port 350 I T 70 3007I 250-L I I 10 I I . Ships processed -burs per ship Source: 1990 - 1997: CEPRI Matarani; 1998 - 2001: OSITRAN. Figure 25: Average Cost per Dock Usage (US$/Ton) : Matarani Port i I Concession I 3.50 i I Source: 1990 - 1997: CEPRI Matarani; 1998 - 2001: OSITRAN. 1.9 RAILROADS Peru has three main railroads-the Central, Southern and Southern Oriental-which together make up 93 percent of the country's rail network. Prior to the September 1999 concession, all the railroads were part of ENAFER, a state owned enterprise. The three railroads were combined into a single package for the concession bidding process, although the winner was allowed to divide them after the auction. A 30-year concession contract was awarded to the Ferrocarriles del Peni consortium, which then split into two operators: Ferrovias Central Andina S.A. who operated the Central railroad and Ferrocarril Transandino S.A. who operated the Southern and Southern Oriental railroads. 25 The main objective o f the concession process was to attract private investment to improve the sector, which was characterized by lack o f maintenance, low quality service and financial problems. T o begin the process, ENAFER was legally, financially and operatively restru~tured.'~ The auction winner was selected based on who would pay the State the highest percentage o f its gross income. The winning consortia offered 33 percent" as principal retribution for the use o f railways and complementary services and 50 percent for use of the State's locomotives and wagons. The contract followed a BOT format and gave the concessionaire rights to maintain and manage the railways in exchange for payments from train operators. In other words, the operator would be in charge o f passenger and cargo services and the concessionaire would collect rights o f way and wagon rents. The concessionaire was required to guarantee the existence o f at least one train operator, which could be a division o f the concessionaire (operating under a different trade name) or an independent firm. At the same time, the concessionaire was required to allow competing train operators access to its tracks. Right-of-way fares were regulated, while the others were supposed to be competitive or negotiated between the concessionaire and the train operators. Currently, the only train operators are related to the concessionaires: Ferrocarril Central Andino S.A. and Perurail S.A. The absence o f other train operators could be explained by low demand in the Central and Southern railroads and certain barriers to entry in the Southern Oriental railroad (highly profitable route). The contract was established for 30 years but it is renewable up to a maximum o f 60 years. During this time, concessionaires are obligated to maintain and rehabilitate the railways, but there are no investment targets. Instead, an international quality certificate i s required before the fifth year o f the concession (United States Federal Railroad Administration, FRA, Class 11). Nonetheless, the State introduced investments incentives through smaller retribution payments. The concessionaire does not have to pay retributions to the State during the first 5 years (and only half the retributions during the next 5 years) if investment levels plus maintenance expenses are higher than the retribution, which has often been the case (Figure 26). ''ENAFERhad US$50 million inlossesbefore the concession (APOYO (2002)). lo 24.75% for Ferrocarril del Centro and 37.25% for Ferrocarril del Sur y Sur Oriente. 26 Figure 26: Retribution and Investment (Thousand of Soles) 30,000 , 25,000 4 20,000 - 15,000 10,000 ~ 5,000 0 1999 2000 2001 2002 Retribution &I Investmnt Source: OSITRAN According to IPE (2003), the railway concession has had a positive impact through the increase in cargo traffic (Figure 27) and performance improvements, such as fewer derailments, less wastage, and more availability o f locomotives. For the Central railroad, the labor productivity indicator o f metric ton - kilometers per employee has reached 1.75 million, well above the World Bank standard o f .75 million. For the Southern and Southern Oriental railroads, labor productivity increased in terms o f passenger - kilometers transported. According to OSITRAN, there were 10 workers per 832 million passenger - kilometers in 1998, but by 2002, this number had dropped to 4. There has also been a positive impact on fiscal flows. Between 2000 and 2002, the state received US$ 16.8 million through taxes, retributions and regulation fees. In contrast, the State only received US$0.45 million in net revenue from ENAFER in 1994.2' Figure 27: Tons of Freight Processed (Thousand of Tons) 2,500 . Concession , 2,0001 El Central South rn Southern-Oriental Source: OSITRAN The Southern Oriental railroad goes from Cuzco city to Machu Picchu (tourist attraction), so it mainly transports passengers rather than freight.22 The opposite i s the case for the Central 2 'www.caretas.com.pe/l362/mdf/mdf.html 22 The Southernrailroad i s the only way to visit the Machu Picchu ruins, so it does not face competition from other modes of transportation. 27 railroad, which transports freight but not passenger^.^^ This i s due to the competition between bus and train transportation. The Southern railroad-which travels between Cuzco and Matarani Port-transports passengers and freight. The development o f this railroad i s tied to the growth o f the Matarani Port and consequently to the increase in Bolivian cargo. Nonetheless, the transport o f passengers and cargo has been affected by competition from highways. Highways have been a particular threat because o f the absence o f strict weight control for cargo trucks, the absence of tolls, and recent rehabilitations. Figure 28 shows the decline in passenger levels for the Central and Southern railroads. Figure 28: Transport of Passengers(Thousands of Passengers) 2'5001 Concession ~ I 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Ei Central 0 South EI Southern-Oriental Source: OSITRAN 1.10 AIRPORTS There are sixty-one airports in Peru and the only one that is operated by a private firm i s Lima's Jorge Chavez International Airport. The rest are operated by the Corporaci6n Peruana de Aeropuertos y Aviacion Comercial (CORPAC). The Lima Airport i s the most important airport in Perk-it represented 84 percent o f national freight and 63 percent o f passenger transportation in2002. Since the beginning o f the last decade, CORPAC was characterized by a lack o f investment resources and political interference, which affected flight fares and freight charges. This made it a good candidate for private participation. Although there was supposed to be other regional airports concessions, the poor profitability o f them resulted in the concession o f only the most attractive one: Lima. Thus, the concession was awarded to the Lima Airport Partners (LAP) consortium in 2001.24 L A P won the right to manage the airport, but the essential services- navigation assistance, aeronautic radio-communication and air traffic control-would still be under CORPAC administration. 23Passengers are only transported on Sundays and holidays. l4The consortium was originally comprised o f Fraport AG Frankfurt Airport Services Worldwide (42.75 percent), Bechtel Enterprises (42.75 percent), and Cosapi S.A.,a Peruvian engineering and construction firm (14.5 percent). 28 The 30 year concession was a BOT scheme that obligated L A P to invest, maintain and improve the airport capacity. L A P agreed to disburse US$ 1,214 million throughout the concession period: US$ 110 million before August 2005; US$ 90 million to expand the airport between 2005 and 2008; and US$ 1,105 million from 2009 to 2030 to construct a second landing strip (before 2015), 56 gate sleeves and a new control tower. Although there were some initial delays related to financing25 and contract addenda, the initial investment commitments have been fulfilled and new airport facilities were opened in February 2005. In fact, the airport has been recently described as the best in South America. The concession contract also regulated fares for services provided by the concessionaire, including the airport use fare (TUUA) charged to passengers, the landing and takeoff fare, and the fare charged to the fuel provider. Duringthe first 8 years, the fares are fixed as described in the contract, and then they will be set according to a productivity factor. Figure 29 shows the evolution o f passengers, freight, and mail at the Lima Airport. Although the number o f passengers fell after the concession, it i s hard know whether this was partially caused by the concession or by outside factors, such as the September 11, 2001 terrorist attacks in the United States. LAP was the winner o f the concession auction because it agreed to pay the highest percentage o f its gross annual income to the State as retribution: 46.5 percent. Based O n this retribution as well as a fee paid to the regulator, the State received US$ 38.3 million from L A P during 2001 and 2002. However, the high retribution percentage has forced it to modify various contract aspects, i.e. investment goals have been reduced for the first 42 months from US$ 100 million to US$ 80 million. Figure 29: Transport of Passengers,Freight and Mail 4,600 T Concession ; T 140,000 120,000 100,000 0 80,000 0 P L 60,000 0 0 40,000 I- 20,000 0 1996 1997 1998 1999 2000 2001 2002 Passengers +Freight and Mail Source: O S I T R A N and CORPAC There were some issues worth mention that affected the investment commitments. The concession contract allowed for modifications if requested by the financing entities. Before signing the contract, it was known that the financiers would request changes, but unfortunately, they only analyzed the contract after it was signed. Both OPIC and KfW requested changes to the contract before they would disburse the financing. The State delayed more than seven months before finally accepting the changes in 2002. As a result, the capacity o f the concessionaire to complete its investment commitments was affected. 29 2. Characteristics of the Privatization and Concession Processes This chapter takes an in-depth look at the characteristics o f the various privatization and concession processes, including how private investors were selected and various features o f contract design. It i s based on a detailed study o f contract documents, reports from regulatory authorities, and a wide variety o f other sources. This wide-angle lens allows us to examine issues that cut across all types o f infrastructure, including elements o f the competitive bidding process, criteria for winner selection, the structure o f competition in each sector, dispute resolution mechanisms, and other characteristics. 2.1 AWARDINGPROCESS Transparency i s key to ensuring credibility in privatization and concession awarding processes. Transparency requires a clearly defined authority to take charge o f the process and well publicized rules o f the game, including mechanisms to submit bids, criteria for prequalification, and an appropriate winner selection method. 2.1.1 Principlesfor transparency The process o f awarding contracts for the provision o f infrastructure services has generally been transparent in Peru. Almost all concessions and privatizations had an authority with defined hnctions driving the process. The CEPRIs (Special Privatization Committees created for each of the processes) and PROINVERSION (Agency to Promote Private Investment) were in charge o f promotion, schedules, provision o f rules, modifications, and supervision; these agencies also provided advance information to bidders. These regulatory agencies ensured access to information through "data rooms" and "white books" were issued for each process. A data room provides bidders access to statistics, maps, and related documents and contracts. Documents included information about the general characteristics o f the process, definitions, terms for questions and access to information, requirements for proposal presentation, and criteria for the winner selection. Publicly available white books maintained records o f all the information on the concession and privatization awarding processes from the beginningto the signing o f the contract. They included reports from different authorities, economic and financial documents prepared by consulting firms, and all official letters.16 Initially white books were only published when each process was completed, because authorities were fearhl that access to detailed information would generate excessive questioning and boycott the processes. Criticisms in this respect have induced PROINVERSION to publish in its website every modification o f the rules and contracts as well as comments received in each o f the stages o f the process. 20 The Arequipa-Matarani road concession in 1994 was an exception to the generally transparent concession processes in Peru. I t lacked both an appropriate institutional framework and a specialized authority -the process was driven by the Ministry o f Transport and Communications. I t also lacked both a data room and a white book and in fact, it is n o w very difficult to find information on the characteristics and results o f the different stages o f the project's bidding process. 30 External factors such as political and institutional instability, an incomplete regulatory framework, or an extended-contract closing time can affect the transparency o f the awarding process. For example, a change o f relevant authorities (Lima Airport) or a workers strike (railways) could alter the concession processes. The period between the selection o f the winner and the final signing o f the contract in most cases in Peru, was long with unclear explanations as to why. However, ideally this period should be short (except in specific justified cases) to avoid a bilateral negotiation between the selected bidder and the government that may result in changes in the original terms o f the contract and an advantage to the winner in relation to the other bidders. Last-minute changes in contracts created doubts about transparency in the processes, like in the cases o f TdP and TISUR (Matarani Port, see Box 3), and officials provided insufficient responses when the processes received public criticism or opposition. This criticism was based mainly on the level o f the public services tariffs (telecommunications and electricity), the workers' situations (TdP and railways) and the destination o f the funds. (See Appendix 1 for a review o f these issues for each concession and privatization). 1Box 3: TransparencyProblemsin the Matarani Port Concession Process A congressional committee investigated the lack o f transparency in the concession process o f the Matarani Port. The committee concluded that the contract awarded responded neither to a need for investment nor to a need for revenue collection, but that the decision was a political one and that the monopoly it created posed the risk that the concessionaire could abuse its position. The committee also questioned the irregularities associated with important changes in the terms o f the award o f the port project after the contract had been signed. Changes, made through a "Clarification," referred specifically to the definition o f the principal operator o f the port terminal. Whereas the original award contract named Stevedoring Service o f America as the operator o f the port terminal, this company was replaced by the firm Santa Sofia Puertos, with the explanation that it had been an error. In the same "Clarification," the word "canon" was replaced by the word "retribution". This change had some direct implications on the destination o f the payments by the concessionaire, since in the first case payments were made to the municipal government and in the second case to the central government. Since these "last minute changes" occurred after a concessionaire had been chosen and after the contract was signed, questions arose regarding the legitimacy o f the contract and the role o f the authority in charge o f the process. A final issue regarding transparency concerns government revenue from privatizations and concessions. The issue affects all processes and its complexities make it difficult to analyze. At the time o f the concessions, a legal framework established in 1996 regulated the destination o f State funds. These laws determined that the Treasury had to deposit funds resulting from privatizations or concessions in an open account in the Central Reserve Bank o f Peru. The deposits o f these funds occurred after deducting the costs o f the process, which included approximately 2 percent commission for the Private Investment Promotion Fund (FOPRI) and the promoting investment bank. According to a Congressional Investigative Committee for Economic and Financial Crimes, there were indications that these funds were used for different purposes such as paying the external debt, buying arms, and corruption. The Committee found an obvious lack o f transparency when in October 1999 three institutions-the Central Bank, the President, and the Minister of Economy-each disclosed a different amount of income from the concession process. Upon being asked by Congress, the executive refused to explain, saying the subject was secret. 31 2.1.2 Competitive bidding processes Peru has used competitive bidding in all o f its concession and privatization processes. A competitive bidding process-with winner selection based on transparent, objective, and quantifiable factors-discourages government discretion or opportunistic behavior by firms and serves as a market mechanism to select the best proposal typically at lower costs. The procedures are similar to those used in other countries and include pre-qualification, proposal presentation, evaluation, and winner selection. Through these stages, the government attempts to make bidders reveal information about risks, benefits, and market predictions. While bidding in telecommunications and electricity attracted more investors, competition in transport concessions in Peru at the awarding stage was not as strong as intended. Timing and transparency problems during the processes may have affected the outcomes. Transportation concessions occurred later when the country was entering a period o f political turmoil. Another explanation may simply be that the telecommunication and electricity processes involved larger markets and were potentially more profitable than transportation, thus attracting more bidders (in fact, the smallest electricity processes-Ede Chancay and Ede Cafiete-attracted fewer investors). Table 11 shows the number o f bidders that participated in each stage o f the processes. With minor differences among concessions and privatization processes, these stages typically included call for bids, pre-qualification based on legal or financial requirements, qualification based on technical requirements and submission o f economic offer. See Appendix 2 for a detailed description o f the different stages for each case. Some possible reasons why bidders withdrew- other than the natural filtering by stage-for the different processes include: Matarani Port: last-minute changes to the concession's termination clauses.27 Arequipa-Matarani road: high risk due to the short length o f the concession, lack o f institutional framework, and difficulties forecasting traffic levels.28 Railways: the strong opposition o f ENAFER's workers and changes in relevant authorities (including the Executive Director o f COPRI). Lima Airport: political context (e.g. Montesinos video scandals) leading to change o f ministers and members o f COPRI just before Stage I11 and uncertainty over proposed addenda to the contract introduced before signing. Edegel and Ede Cafiete: base price set too high. ''Accordingto '*Macroconsulta representative o f the concessionaire, Dionisio Romero of TISUR, April 9, 2002. (2000) 32 Table 11: Number of Bidders that Participatedin the Stages of Concession Processes Railways Etevensa EEPSA Numbers in parenthesis represent bidders that passed the stage. One bidder went for 110, another for Matarani and two ** bidders for both ports together. The eight firms were grouped into five consortia. *** All bidders retired becausethe base price was too high; ****the process restarted with a new base price. One bidder presentedoffer for Ede Chancay. Source: White books, Semana Economica and Macroconsult (2000). 33 2.1.3 Criteriafor winner selection The criteria for choosing the winning bidder should be based on transparent, quantifiable factors. The use o f multiple criteria, even with a well specified weighting function, i s not desirable because it i s more exposed to manipulation, corruption and contesting o f the award by the losers. Most processes in Peni, including the railways, the Matarani Port, Lima Airport and Anc6n- HuachePativilca concessions as well as the telecommunications and electricity privatizations, used a single and objective criterion to determine the winning bid. The only exceptions were the Arequipa-Matarani road concession and Etevensa electricity privatization. In the first case, selection involved discretion by the Ministry o f Transport and Communications (MTC) and was based on weighting several aspects related to economic (duration, financial guarantees and providing complementary services) and technical (working plans, administration o f the toll, and having an engineer) factors. Concession duration accounted for 70 percent o f the selection weighting criteria. The government preferred shorter duration contracts, hoping to gain from recurrent competitive bidding processes. Inmost cases, firms were first evaluated according to technical criteria. All the qualified bidders would then make an economic proposal which was the only determinant o f the final decision. This procedure, although more robust against potential corruption, did not allow for discrimination between qualified candidates with different technical capacity and did not filter out overly optimistic bids. Most o f the selection criteria had fiscal objectives-initial payments or retributions to the State-rather than efficiency or user benefit objectives. Table 12 summarizes the selection criteria and bidding results. A few o f the privatizations (Etevensa, EEPSA, and Electro Sur Medio) included investment requirements, while the Arequipa-Matarani road was decided mainly on concession duration. 34 Infrastructure Selection criteria Results Matarani Port Largest initial payment (over US$ 9.5 million) Us$9.6 million Ferrocarril del Centro: 24.75% Railways Largest percentage o f retribution to the state Ferrocarril del Sur and Sur Oriente: 37.25% Arequipa - Matarani road Average weighted factors ('I 6 years and 2 months (originally) Lima Largest percentage o f retribution to the state (over 20 percent) 46.5% Ancon-Huacho- Largest percentage o f retribution to the state Pativilca road (over 5 percent) 5.5% TdP Largest payment over US$546 million US$2,002 million I - .~ I I 1 Edelnor and Luz Largest payment over US$128 and US$129 1 US$176 million and US$212 del Sur million, respectively million, respectively Largest payment over US$373 million r-Edegel (US$273 million in cash and US$lOO million US$424.5 million and US$100 incertificates o f debt) million of certificate of debt, (21 Ede Chancay Largest payment over US$10 and US$8.2 US$10.4 million for Ede Chancay and Ede Cafiete million respectively US$8.6 million for Ede Cafiete (3) I1 Egenor Cahua I1 Largest Largest payment over US$ 21.12 million payment over US$ 175 million 1I US$ US$ 4 1.8 million (4) I 228 million sur 1 Largest payment over u s $ 3 5 million(5) 1 II US$ 51.2 million I I Investment in I I generation expansion + Etevensa shareholder participation million +potential to install additional generation capacity. The value o f I these had to be greater than U S 6 5 million. EEPSA Participation o f Electroperu greater than 33 I 59.17 percent L uercent. (" I ' I I (1) The weighted factors referred to certain requirements the concessionaire should have: experience, appropriate rehabilitation and maintenance plans, and technical staff, among others. (2) and (3) Correspond to the second round o f competitive bidding. (4) This sale was made through an installment plan (via the Entrepreneur Promotion Program since the small size attracted only national investors); the initial payment was US$ 8.4 million. (5) Half paid in money and half through investment on expansion o f the electrical frontier. (6) COPRI valued this participation o f 33% at US$59.4 million. This includes US$40 million to construct a plant o f 80MWs. Source: Call for bids, sale and concession contracts, and COPRI. Another important factor worth considering i s how far the final offer was from the minimum or reserve price. In principle, a reserve price i s necessary as a safeguard against collusion for a below-market offer and for public credibility. The reserve price should be set at the minimum justifiable level to identify qualified bidders, not at the government's best estimate of market value, Unreasonably optimistic bids could reflect opportunistic behavior with the expectation that ifthings do not turn out well, contract renegotiation would be possible. In Peh two cases 35 deserve attention. The winningbid for Lima Airport was over twice the minimum(see Table 12), implying that LAP could have trouble making the high periodical payments to the government. Telefonica del Peru's bid was US$2 billion, nearly 4 times the reserve price, while the other bidders, Peruvian Telecomunication Holdings and Telecomunicaciones Peruanas, only offered US$857 million and U S 8 0 3 million respectively. The winning offer implied a price per line o f US$8,773, higher than the prices offered in other Latin American countries. There i s debate over whether the price paid was too high, but one possible explanation i s that the reserve price was set too low (Box 4). Box 4: Possible Reasons for the Price Paid by Telefhica Internacionalde Espaiia The investment bank Morgan Grenfell, CEPRI's advisor for privatization, analyzed the Telefhica's offer inder three scenarios: a) by comparing it with the value o f telecommunication companies in other developing :owtries whose shares were traded on the stock markets o f their respective countries; b) by analyzing the 3ensitivity o f base prices; and c) by evaluating the offer after looking individually at the potential value o f :ach one o f the awarded concessions. From the analysis several points arose, listed here in respective order: a) relefhica could have made its proposal taking into consideration the high valuations o f Southeast Asia :ountries, like Hong Kong Telecom, Malaysian Telecom, and Telecom Asia o f Thailand. These companies were highly valued due to the high rates o f economic growth in the region with important potential increases in the demand for services. Telefbnica may have expected substantive improvements in the economic situation o f Peni and, therefore, in the demand for telephone lines and services; b) Telefbnica may have expected much greater levels o f income, cash flows and increases in productivity than a North American bidder would have. This case reflected differences in risk perceptions and in the horizon o f investment recovery among bidders; and c) the economic value o f the offer by installed line inPeru went from US$7,910 to US$5,884 when the potential value o f other granted concessions was discounted. This value was not very different from the stock-exchange capitalization for the lines that Telmex and Telefonos o f Chile had in 1992. CEPRI Telecom also argued that the highoffer o f Telefbnica de Espaiia could be explained by: i)the presence o f three auction bidders, whereas in other regional telecom privatizations there were a maximum o f two; ii) the premium o f 20 percent to 30 percent that, according to officials o f Telefhica, had been incorporated into the offer to ensure its win; iii)discrepancies over the time it would take to recover the investment; iv) greater knowledge by Telefhica o f the South American telephone market potential; and v) its strong interest in consolidating its leadership inthe region. The larger bid offers were attributable, not only to greater competition, but also to the clear vision o f the government regarding this process, made possible because o f the experience gathered by the Peruvian team from other Latin American privatization efforts. In addition, the model to evaluate the Peruvian companies should have been different from the one used in countries with a telephone density two or three times greater than the Peruvian one. It i s notable that the Peruvian case constituted the first experience o f privatization o f telephone services ina country with such low densities (2.94 lines for every 100 inhabitants in 1993). Source: COPRI (1999) and Campodonico (1999). In summary, selection criteria were simple and quantifiable, as recommended by theory, in most Peruvian concession and privatization processes. However, they prioritized fiscal objectives as in most cases the selection criteria were based on the largest payment over the specified reserve price. To compensate, investment goals were imposed in some cases. There are various explanations for the large differences between the final price paid and the reserve price, including an overly optimistic estimation o f the concessionaire (in the Lima Airport case) and a too low base price (in the TdP case). To prevent opportunistic behavior associated to excessive bids, the recent contracts negotiated by PROINVERSION demand that the guarantees provided 36 to ensure fulfillment o f the contractual obligations be proportional to the price offered by the bidder. 2.2 CONTRACT DESIGN This section analyzes the main design elements o f concession and privatization contracts for infrastructure services in Peru examining: 1) the duration o f contracts, 2) the market organization created by contracts and how they determine incentives for the main agents involved; 3) the appropriateness o f risks allocation; 4) issues related to bonus and penalties; 5) the balance between certainty and flexibility; 6) conditions and characteristics o f renegotiations; and 7) dispute resolution mechanisms. In the case o f transport concessions, contracts are the main basis for analysis. For telecommunications and electricity privatizations, both the sale and the operating licenses (concession contracts) are examined in an integrated way since the former contains the terms o f the sale and investment projects and the latter contains efficiency and tariffs conditions. 2.2.1 Duration of the contracts Contract duration should allow for a fair return on investment, while attaining the formal objectives stipulated in the contracts. In Ped, the duration o f the transportation infrastructure contracts i s reasonable given that construction and maintenance objectives need to be attained while improving operational efficiency. The duration o f the Matarani Port, railways and the Lima Airport concessions are 30 years, and the Ancon-Huacho-Pativilca road i s 25 years. In contrast, the 74-month Arequipa-Matarani road concession was considered too short.29 Regarding telecommunications privatizations, which had sale and concession components, the concessions were also for a fixed time. In particular, bidders had to sign 20-year concession contracts and commit to installing a certain number o f new lines.30In contrast, operating licenses were not part o f the privatization process o f electricity services. Instead, they were part o f an administrative procedure with the Ministry o f Energy and Mines (MEM) that determined the duration o f the contracts. 2.2.2 Unbundlingand competition Introducing competition in infrastructure service provision i s desirable for its effect on efficiency and innovation. Concession contracts for public transportation infrastructure established which activities would be subject to monopoly versus which had competitive conditions (depending on size and technical considerations, etc.) and mandated non-discriminatory environments free from cross-subsidies. Telecommunication contracts were based on a finite monopoly period for TdP and electricity privatizations unbundled the sector into competitive and non-competitive segments. Table 13 shows the market structure for transportation concessions and Table 15 for privatizations in the telecommunications and electricity sectors. 29The duration was a result of the process's selection criteria, discussed in the previous section. 3o638,000 new lines in Lima and 559,600 in the rest of the country. 37 Table 13: Market structure for concession contracts in public transportation infrastructure Provided incompetition Horizontal gration Integration by the concessionaire concessionaire) Complementary Essential activities**: Banned Railways services*: storage, (between hotels, advertisement, passengers and freight Allowed different self-services, and others.transport services. routes) Roads Management of roads Mutaruizi Port. The concession model used for the Matarani Port i s based on vertical unbundling, or the separation o f the activities o f port services. The benefits o f the model are achieved only if, in practice, mechanisms are put in place that promote effective competition when possible, and regulate the conditions that guarantee non-discriminatory access to third parties. A sole concessionaire who acts as a monopoly provides some o f the port services while others are open to competition. However, the concessionaire i s allowed to integrate vertically in some services to guarantee a profitable return for potential investors given the small size o f the market. Currently, the competitive segments are dominated by firms tied to the concessionaire, possibly because market size does not allow for competition between multiple operators. While several cases o f anticompetitive behavior have been investigated, all o f them have been discarded. Given this evidence, the merits o f the current model are a posteriori questionable vis- a-vis a single operator model that i s vertically integrated. In a vertically integrated model, the monopoly profits could be shared with the State in a transparent manner. However, the model chosen may have been the most appropriate at the time, given the strong opposition to the concession process displayed by the trade unions o f ENAPU's workers and by the established small port service providers. Railways. The concession model used in railways i s based on competitive access, but in practice vertical integration prevails. Concessionaires and operators o f both railway systems are strongly linked, and no new operators have entered the market. A t least initially, the absence o f competitors did not result from opportunistic conduct by the concessionaire, but instead from efficiency reasons. If vertical unbundling had been strictly enforced immediately after the concession contract was signed, it would have endangered continuity in the provision o f services. Concession contracts include mechanisms to prevent anti-competitive behaviors that would preclude access to new operators. An example i s the clause related to the bidding process for scheduling and to rolling stock rental. Yet these mechanisms are insufficient in eliminating predatory behavior by the concessionaire. In fact, the concessionaire has rented all o f its wagons for a 10-year period to its subsidiary operating company. These difficulties could be eliminated by placing limits on the concessionaire's rental agreement for rolling stock or implementing a bidding mechanism for this material. 38 In the case of railways, the appropriateness o f applying the same concession model to different railways routes facing different degrees o f competition i s also questionable. The Central and Southern railroads face strong inter-modal competition from roads and consequently low profits, making potential competitors reluctant to enter these markets. Therefore, vertical integration could compensate for low profits by allowing scale and scope economies through joint coordination o f traffic, transport and complementary services (Thomson (2002)). On the other hand, the Southern Oriental railroad i s dedicated mainly to tourists and provides the only transportation access to Machu Picchu. This route seems highly attractive to potential competitors (Box 5). The application o f an access model in this route seems more appropriate, as third parties have incentives to become operators and competition may bring about substantial efficiency gains. Box 5: Access problems of the South-Oriental Railway The firm Santuario Inca Machu Picchu Railway (FERSIMSAC) applied for a license to compete with Peru-Rail in the operation of transportation services in the southern oriental route. Initial discrepancies with the concessionaire related to some access conditions, resulted in OSITRAN`s intervention. OSITRAN proved that there was not a scheduling conflict and released an access order in FERSIMSAC's favor (Directive Council Resolution 003-2003- CDIOSITRAN), and six-month access was granted as a temporary measure until the signing o f the access contract with the concessionaire. Despite this, FERSIMSAC experienced problems in fulfilling all necessary requirements, including adequate wagons and crew, complying with security regulations and having an updated insurance policy. Despite the access right granted by OSITRAN, FERMISAC never began operations and its license expired. Source: OSITRAN Lima Airport. Part o f the Lima Airport concession contract was solely operated and part was open to competition, a division based on capacity criteria and economies o f scale. The contract gives monopoly control o f "principal operations"-aero-navigation services, landing and take off, emergency and security, etc.-to Fraport AG Frankfurt Airport Services Worldwide (FRAPORT). Non-principal operations-ground handling, commercial services, etc.-can be handled by the principal operator or by third parties (Table 14). To avoid anti-competitive behavior in non-principal operations, the concession contract contained open access and non- discriminatory conditions. Nevertheless, there are concerns that the principal operator can exercise an unfair amount o f control over competitive services. One solution would have been to separate the two areas entirely by not allowing the company providing monopoly services to offer competitive ones3' See Bosch and Garcia-Montalvo (2003) for a discussion of how allowing airport service companies to operate simultaneously as both monopoly and competitor obstructs, or at least delays, the startup of competition. 39 Table 14: Activities in the Lima Airport by Type of Operation Non-principalOperations PrincipalOperations (Performedby mainor secondary Commercial operators) Operations Servicestied to the TUUA: transportation Hotel Services tied to the airport usage of luggage(from terminal to outside the administeredby an tax (TUUA): emergency and airport) and of passengers(between operatorchosen security airport Services terminalandplane), deliveryof luggage, through an auction information,waiting rooms, lighting, etc. process ~ Landingand take-off(traffic Vehicularparking control, communications, plane Non-airportservicesfor airlines: counters, offeredby parkingup to 90 minutes, etc.) VIP halls, hangars, etc. concessionaireor by a third party Parking(additionaltime to 90 Publicity offered by a I minutes) Protocol Services* third party BoardingBridges Groundhandling Aero-navigationservices: track signals, platformlighting, Airplane fuel emergencyservices,etc.** Telecommunications. In the initial plan adopted for telecommunications, local and long-distance services-previously offered separately by CPT and ENTEL-were integrated. The telecommunications contract3* established exclusivity rights for five years for Telefonica del Penj (TdP) in local, long distance, and international fixed line service. TdP was the monopoly provider during that time, after which competitors were free to enter the market. The contract specified that tariffs would be rebalanced to reflect real costs-rather than rely on cross subsidies-and regulated during the exclusivity period. After a certain number o f years, a price cap adjusted by a productivity factor would govern tariff levels. The exclusivity period was designed to allow for increased investment in network expansion, a high priority o f the government given the underdevelopment o f telephony before privatization. The exclusivity period established in the contract with TdP was later shortened from five to four years, causing speculation as to why (see also Box 8 in section 2.2.6). A positive interpretation would be that the investment objectives were achieved ahead o f schedule and tariffs were already rebalanced, allowing for an earlier introduction o f competition. The early opening up o f competition for all services brought important benefits to consumers who received tariff reductions o f 30 percent for long-distance calling, 50 percent in connection charges, and 22 percent in local intra-"departmento" calls. ''This refers to the telecommunications concession contract (operating license), which was a component of the privatization (sale of share assets) contract. 40 Others suggested that TdP asked the government to shorten the period in exchange for modifying some clauses o f the contract. The most important modification was the delay in the application o f the price cap for two years until 2001. Consumers may have been worse off, as applying the price cap would have put hrther pressure on tariff reduction. Another modification was that calls within a "departamento" would be considered local calls. This reinforced TdP's dominant position in the local fixed line market due to high interconnection While these arguments are debatable, it i s clear that TdP has maintained a dominant position despite a complete opening o f the market. A report by the telecommunications regulator (OSIPTEL) for concession contract renewal found that TdP had not complied with the competition clause (Clause 11). The report outlines proceedings against TdP for anti-competitive practices and abuse o f its dominant position.34 These problems are more attributable to a weak regulatory framework and antitrust oversight than to flaws in contract design. Table 15: Market structure According to Concession Contracts of Telecommunications and Electricity Activities during the exclusivity Activities after the Deriod exclusivityDeriod Provided in Providedin monopoly competition Providedin conditions conditions competitionconditions Fixed national, Fixed national Other services international calls, Telecom and international from Category 11* services from Category Allowed Allowed calls and unregulated services** I1and unregulated services Subject to Subject to Electricity Distribution and Transmission Generation INDECOPI`s INDECOPI`s evaluation evaluation ` Category I1 includes submarine cable, asymme ic circuits, local leased c wits, national leased circuits, international leased circuits, frame relay circuits, infovia tariffs, unired tariffs, ISDN tariffs, telex service, ** telegraph service, additional services, terms o f use, and public telephones. Unregulated services include fixed-mobile, mobile-mobile, fixed-personal communications service (PCS), fixed-digital trunk service, PCS, trunk service (tariffs not yet available). Source: ITU (2001) for categories o f telecommunications services. Some have argued that the interconnection tariff constituted the main obstacle to developing competition in local teleph~ny.~'In Peru, interconnection charges are freely determined by agreement between service and network operators, subject to maximum daytime and night charges established by OSIPTEL. Companies that aspire to enter the market criticized the day tariff o f US$0.029 per minute, a charge very high compared to those o f other countries in the region. For example, in Chile the daytime rate i s US$ 0.017 and in Mexico, US$ 0.022. j3 Campodonico (1999) and Oscategui (2001). j4 For example, a proceeding (Resolution 052-2002-CCOiOSIPTEL) imposed a fine o f 1,097 UIT on Telefonica for not complying with essential obligations o f interconnections, unfair competition and abuse o f its dominant position. In other proceedings (Resolutions 039-2003-CCOiOSIPTEL y 019-2003-TSC/OSIPTEL), Telefonica was fined 25 UIT for abuse o f its dominant position in the modality o f unjustified refusal to trade. ''Oscategui (2001) y Ruiz Caro (2002). 41 OSIPTEL proposed a staggered reduction o f the interconnection charge, which was criticized by the sector's operators who demanded a more immediate reduction. Faced with the inability to solicit competition at the local level, the development o f competition has been limited to both national and international long-distance telephone services, and to cellular phones. TdP's control o f essential facilities has also hurt competition and accentuated its dominant position, creating barriers to entry to other companies in different markets. For example, the Peruvian Scientific Network (Red CientiJica Peruana) accused TdP o f wanting to remove it from the internet market and o f practicing price dumping by not charging its subsidiary supplier for network access. TdP will be using cross-subsidies in favor o f one o f its vertically integrated branches o f activity.36 Electricity. In order to introduce competition, the 1992 Electricity Concessions Law (LCE) divided the sector's activities into four segments: generation, transmission, distribution and commercialization. Generation companies compete to supply energy to distribution companies and large con~umers.~'Prices for energy supplied to distribution companies are negotiated, but subject to a maximum regulated by the Committee for System Operation (COES), while market forces determine the prices paid by large consumers. In contrast, the transmission and distribution sectors operate in monopoly condition^,^^ as they have control over an essential facility and have exclusivity rights to operate inside a concession area.39 Distribution and transmission activities are regulated as natural monopolies with maximum price ceilings set by the Energy Tariff Commission (CTE) (Table 16). 36 Macroconsult (2000). 37 According to the LCE, large consumers are those with demand for capacityo f 1 M W or more. 38 Distribution entities have an obligation to serve all consumers within the concession area. They must also maintain supply contracts with generators sufficient to cover their total supply requirementsover a 24-month period ''inA advance, and comply with pertinent environmental,technical and operating criteria. special case is the distribution concession for Lima, a market representing55% o f the total. Specifically, it was divided into two parts: Edelnor in the north and Luz del Sur in the south. The aim was to give regulators a benchmarking tool that would foster competition (Macroconsult (2000)). This implied that the State favored the benefitsof competition over possibleeconomies of scale of a single distribution company. 42 Table 16: Prices in the Electricity Industry and their Regulation Regulated activities Competitive activities I Price regulated by CTE Agreement betweenparties Sales from generation companies to large consumers Compensations to transmission operators and distributors Price regulated by COES* Capacity and energy transfers between generators according to the terms o f Sales from distributors to final users their contracts (called pool) Sales o f energy from generators to * ** COES tries to simulate competition according to generatc s costs. CTE and COES participate. Source: LCE One o f the main problems regarding competition in the electricity sector i s found in the market for free clients, in which distribution companies have incentives to limit entry. The number o f free clients was very limited in the second half o f the decade o f the 9Os, but has grown significantly since 2001. In 2002 there were 261 fiee clients, five times greater than in the previous five years. Yet despite this growth, the development o f this market may have been conditioned by restrictions distributors imposed on access to their facilities. As indicated in a report by the Commission on Free Competition o f the Institute of Defense o f the Competition and Protection o f the Intellectual Property (INDECOPI),40 the distribution networks' monopoly- like characteristics favor anti-competitive behaviors like the unjustified refusal o f access to the distribution networks, or the establishment o f high network access charges. As well, clients can be dissuaded to contract with other providers by threats o f service interruption or through loyalty awards granting them favorable treatment if services are interrupted. Though the LCE specifically obliges transmission and distribution firms to allow third parties access, the low levels o f competition observed may be due to the fact that conditions and access charges are not regulated but freely agreed upon by the parties. The implementation o f an ex-ante regulation o f both the tariffs and the conditions o f access could facilitate the future market development. Another problem i s the high degree o f market concentration in the electricity sector resulting from horizontal or vertical integration. In order to guarantee competition, there i s a 15 percent limit on the market share o f a company in the generation, distribution and transmission businesses. There i s also a cross-ownership restriction, which limits private companies to no more than a 5 percent market share in one segment. Despite these limits, the five largest private generation companies now belong to only three economic groups. Table 17 shows how the structure o f the market has changed since reforms began in this sector. The presence o f Electropeni - whose market participation i s still considerable - i s another factor o f distortion for 40INDECOPI (1999) 43 competition in the electricity sector and can discourage the entrance o f new private investors in the market4' If not addressed, high concentration and anti-competitive behavior may continue to inhibit development in the electricity sector. Legislation and the concession contracts have established conditions to control for anticompetitive behaviors. The role played by OSINERG as the regulator in access controversies, as well as INDECOPI as the authority in charge to approve vertical and horizontal concentrations, is also fundamental. Table 17: PrivatizedCompanies in Generation:EconomicGroups- 2001 Economicgroup Company Initial members Transactions Entergy Endesa de Chile Endesa Spain acquired Edegel Graiia y Montero 65% of Enersis BancoWiese Enersis acquired 30% of Workers Endesa de Chile Takeover bid, three times hdesa (Spain) Endesa Spain Etevensa Inv Creditos del Peru Cosapi Workers Endesa EEPSA EnergiaAndina Peru Electricity Fund DominionEnergy Marc Cox Chilegener acquired 49% luke Energy(EEUU) (1) Egenor Thomas Farrel of stocks to Dominion Workers Energy Duke Energy- Takeover bid Sipesa Vattenfall acquiredSipesa VRG Energy(EEUU) (2) Cahua Workers NRG acquiredVattenfall Skanska Bot-Nordic Power Invest NRG acquired Skansa Bot (1) Also has Termoselva (generationcompany) (2) Also has CNP Energia (generationcompany) Source: Aguilar (2003) 2.2.3 Risk allocation Identifying risks and allocating them between parties involved in a contract i s important because it determines incentives. Privatization and concession contracts must consider various types o f risk be they political, technical, commercial, or major force events.42In theory, risks should be borne by the party best able to assess, control, and manage them. This section analyzes how concession contracts in P e deal ~ with relevant risks. 4'IPE (1999), Macroconsult (2000). 42Major force risks are explained later as unpredictableevents that may occur over the life o f the contract. 44 There i s positive evidence on how risks have been managed in the contracts under analysis here. The State has assured contract stability through Contract-Law (Contratos-Ley) and Stability Agreements. As the name implies, contract-law requires that contract modifications be in the form o f a law. Stability Agreements are signed between investors and PROINVERSION43and generally help to protect and promote national and international investments by imposing barriers to unilateral modifications by the government. In the electricity and telecommunications sectors, this i s reinforced by the Electricity Concessions Law (LCE) and the Telecommunications Law. These laws define the framework in which the sectors develop and determine many contractual terms. In addition, the State protects concessionaires from exogenous risks (for instance, in relation to expropriations for roads or major force risks). Also, technical risks are generally well allocated. However, there have been negative side-effects o f risk management. For example, the State protects concessionaires o f transportation infrastructure services from commercial risks through mechanisms that compromise the level o f income the State receives. The exclusivity period in telecommunications contracts presents similar concerns, but i s commonly used worldwide to compensate for investment commitments. These mechanisms could be improved through other methods o f protection. In the case o f roads, a better option could be to endogenize the duration in the contract according to demand fluctuations. There are also cases where the State has created risk for private participants. For instance, indecision about public road tolls in one case and the discontinuity o f some privatizatiodconcession programs has created uncertainty for private agents. In the telecommunications sector, an order from the Constitutional Tribunal for TdP to re-hire workers and a Congressional threat to modify contractual terms has added risk to the privatization. 2.2.3.I Political risk Political risk i s a result o f uncertainty over government actions. Such actions could include legislative changes, changes in regulatory conditions, or the inability o f the government to meet its contractual obligations. Stability agreements help to mitigate general political risk by guaranteeing security in taxes, property rights, and the equal treatment between national and foreign investors. They control the scope o f economic activities o f the State. Such agreements were generally part o f all concession contracts in the transport sector, with the exception o f CONCAR (Arequipa-Matarani road) and the railways (although railway concessionaires have the right to ask for a stability agreement). T o avoid these risks in the telecommunications sector, contracts to sell shares and for concessions take the form o f contract law. As well, generation and distribution firms in the electricity sector had the opportunity to sign agreements on legal and fiscal stability, and on foreign currency free disposal, as stipulated by the LCE. Generally, the LCE reduces, ex ante, certain political and technical risks in concession contracts, establishing a framework that defines 43 Supreme Decree No048-98-EF has information about the scope of protection given to investors through legal stability agreements. Law No27342 - regulates legal stability agreements under Legislative Decrees No662 and 757. Law No 27514 - modifies the subscriptionto legal stability agreements. 45 certain key aspects o f the contractual relationship (criteria for determining tariffs, duration, obligations and contract expiration, among others). The LCE i s important because it guarantees certain homogeneity in the contractual terms o f different concessionaires and complicates the individual modification o f contracts. Contractual clauses that determine tariff rates and their revision are the most exposed to possible State intervention. In the telephone sector, for example, the legislature attempted in the beginning o f 2003 to modify the contractual terms o f the concessions related to the collection o f the monthly fixed connection fee, known as basic rent. Had this initiative been successful, it would have violated not only the legal stability o f concession contracts but also the autonomy o f the regulating institutions, as it would have regulated a matter o f exclusive competency o f OSIPTEL. An additional threat to the stability o f the tariffs in the telephone sector was the high price that TdP paid in the privatization process. Parts o f the population and some institutions voiced their concern that in order for the firm to rapidly recuperate its investment, it would have to drastically increase tariffs. This perception and the corresponding political pressure resulted independently from the contractual terms that established tariffs and the formulas to update prices. This threat to the normal development o f the privatization represented a political risk because o f the possible legislative changes that it could cause. Box 6: The TdP worker dispute A ruling by the Constitutional Tribunal ordering Telefonica del Peru to rehire workers it had laid off was another interference by the State that generated great controversy in the telecommunications sector. A t the time that CPT and ENTEL were privatized, they both employed over 12,000.' TdP's reduced its work force to approximately 5,600 workers in 1998. Some reasons for the reduction were the elimination o f duplicate hnctions, the outsourcing o f services and technological advances in the expansion o f the network. In following years, TdP continued to gradually dismiss employees, especially union workers. The workers went on strike and forced a collective bargaining negotiation, which ended with unsatisfactory results. The situation worsened with the dismissal o f 126 workers at the end o f 2000. The unions Fretatel Peru and Unitary Union o f Peru reacted by convoking more strikes andultimately the unions filed suit against TdP. The dismissals by TdP, which was one o f the companies emblematic o f the wave o f privatizations in the country, contributed to disturbing the political panorama. Though their actions were supported by Article 34 o f D.L.728, the Law o f Labor Productivity concerning dismissals (which authorizes companies to reduce up to 10 percent o f their personnel for economic or market reasons, with previous authorization from the Ministry o f Labor), it evoked an immediate reaction. InNovember 2002, the Constitutional Tribunal found in favor o f the union and the federation o f workers. The Court indicated that the law authorizing the workforce reduction violated the constitutional rights o f the workers and harmed their rights to unionize. The Constitutional Tribunal ordered TdP to return the dismissed employees to their old jobs and abstain from hture dismissals. In the transportation sector, the fees established in the contracts have been tied to the user toll charges on the nation's highway network. To attract private operators, the State had to increase its tolls so that cheaper State roads would not compete with concessioned routes. Historically, tolls charged on highways administered by the Peruvian government have not been sufficient to maintain the routes. The national average for tolls was US$0.71,44 while the basic tariff necessary 44Bonifaz, Urmnaga, and Vasquez (2001) 46 to cover maintenance costs was US$1.60 per 100 kilometers, according to the National Highway Maintenance System (SINMAC). This created a problem since unsubsidized road concessions are only viable if tolls can cover maintenance and investment costs. Although the State began to increase its tolls in 2001, NORVIAL postponed signing its concession contract (Ancon-Huacho- Pativilca road) in September 2002 until the MTC completed its commitment to raise tolls on the national road network, which it did in December 2002. By postponing the signing, NORVIAL was able to avoid the political risk associated with the government not fulfilling its promise^.^' In the electrical sector, government ambiguities have generated great uncertainties among investors over the government's commitment to privatizations, concessions and private investment in general. The delays in transferring important assets o f Electropeni and the ' suspension o f the privatization o f Egasa and Egesur in Arequipa, are examples o f such weaknesses and reflect the public rejection o f these processes (see Box 1 in section 1S). 2.2.3.2 Technical risks Technical risks in transport are related to preparation, construction, and operation problems (completion dates, quality, cost o f postponement and modification, etc). In most cases, they are assumed by the concessionaire with the assistance o f an insurance contract. Problems frequently arise at the construction stage when land expropriation i s necessary. For example, the Ancon- Huacho-Pativilca road concession agreed, as part o f the administrative procedures, that the State would be in charge o f expropriating and purchasing private lands before NORVIAL would begin construction. Although Peru has a suitable legal framework to carry out expropriations, these have still not been executed because both the layout o f the route and the legal situation o f lands to be expropriated are uncertain. A similar problem occurred in the Lima Airport concession. The contract's investment obligations included construction o f a second runway, which required expropriation o f surrounding areas. The government agreed to finish the expropriation by February 2006, but serious difficulties related to judicial processes caused the postponement of the runway construction from the 1lththe 14`h year o f the concession or five years after the to government would turn over the required areas. Thus, an addendum to the contract shielded the concessionaire from risk related to the government's ability to deliver the additional area. In the Arequipa-Matarani road the State carried the risk o f higher-than-agreed costs of rehabilitation. In that agreement, the M T C only recognized a maximum o f 10 percent additional costs and ifover-costs were to occur, the duration o f the concession contract would be extended. This arrangement was to keep the State fi-om gaining potential benefits fi-om a new competitive bidding. This risk allocation i s less than satisfactory, since the concessionaire i s a construction company and should be able to manage road rehabilitation risks. The main technical risks that dominate telecom and electricity industries are associated with investments, expansion o f the network and quality goals. These risks are generally assumed by TdP. While OSIPTEL monitors compliance o f the established minimum requirements, the State also receives bank guarantees o f approximately US$18 million. Most electricity sector contracts also require bank guarantees and/or guarantees o f project fulfillment for determined amounts. 45Another example is the Arequipa-Matarani road. In this concession, the revision o f the tolls i s tied to the variation in tariffs of the Panamericana Sur highway. 47 These conditions, with the exception o f Edegel, were stipulated in contracts to buyhell shares. On the other hand, where the operations o f the electric companies were difficult because o f lack o f access to public property (ground, subsoil and air), the risk fell on the State because the State has the power to establish servitudes46by force. 2.2.3.3 Commercial Risk Commercial risks occur due to uncertainties regarding demand levels, These risks should be allocated to the concessionaire or operator to provide incentives to bidders for accurately estimating the expected demand. In some cases, however, when uncertainty i s too high, the risk might be shared with the government. In Peru, a guarantee o f minimum income has been used to mitigate the possibility o f commercial risk. These clauses use a number o f methods that vary by sector to guarantee concessionaires a minimumpre-set level o f income. In the transport sector, the State shares risk by guaranteeing a certain income depending on traffic levels. If traffic levels are as expected or better, the concessionaire contributes a certain amount to a "Road Fund." However, if the traffic turns out to be less that expected, the concessionaire i s compensated through this fund. Inthe case o f the concession o f CONCAR, for example, the traffic level was overestimated in the initial years, a problem that was covered by funds from the Road Fund.47From 1996 to 2000, traffic was heavier than expected by 3 percent for light traffic and 21 percent for heavy traffic. Since then, there has been a reduction in the traffic and an overestimation o f the initial contract figures. The methods used to make predictions about future traffic and possible substitution o f highways by users are imp~rtant.~'The plan to award concessions for highways in Peru within the framework o f the Highway Concessions Program o f 1997 considered parameters and estimates established by MTC.49Since initially the concession o f Road Network No. 5 was part of this plan, the parameters and traffic estimates on which it was based may not have been appropriate because they were not based on an relevant surveys to evaluate preferences o f alternative routes. In the case of the Arequipa-Matarani highway, the awarding mechanism mainly based on the shorter concession term, could have caused bidders to overestimate demand. This scenario generates problems o f adverse selection because the State would be inclined to select those bidders ready to assume a greater risk. Traffic guarantees might have been necessary in Peru to attract private investment5', but other mechanisms could have been employed. For instance, in Colombia or Chile third generation concessions, the period o f concession i s automatically adjusted according to the amount o f traffic 46 Servitudes refer to the ability o f concessionaires to install aqueducts, electrical lines, distribution substations on private property, telecom systems, access routes, etc. 47 The Road Fund comprised a certain percent o f income o f the concessionaire. Depending on the expected traffic, the concessionaire would receive 88 percent o f the toll as repayment, and contribute 12 percent to the .fund. If the revenue from the tolls lessened because of reduced traffic, the concessionaire could reduce contribution to the fund up to 12 percent. Were that not sufficient, the M T C would have to compensate the concessionaire. 4RAnotherimportant element in Peru that determines the success o f schemes that guarantee minimum income is the above mentioned national policy on tolls. 49Bonifaz, Urmnaga, and Vasquez (2001). In many cases, these traffic guarantees were required by the so called "approved creditors" and served to reduce substantially the financial costs o f the concessionaire. 48 on the highway. This mechanism reduces the concessionaire's risk by ensuring an income, but limiting the State's co-financing by allowing that any overestimation o f demand be translated into a reduction o f the period o f concession. Other commercial risk clauses offer flexible mechanisms to alleviate acute fluctuations in income or costs or allow exclusivity rights for a certain period. For example, the Lima Airport establishes that if the concessionaire's gross income decreases or costs increase more than 5.5 percent with respect to estimated income reported to OSITRAN, it can propose mechanisms to overcome the economic imbalances. These include a request for tariff modification or contract duration extension. OSITRAN can accept or deny the concessionaire's request. In telecommunications, granting TdP the monopoly over basic services for a limited period, ensured them a certain commercial yield and, in some ways, eliminated commercial risk for the company until 1998. This may be an example where it would be appropriate to free the concessionaires from commercial risks. The objective was to improve telecommunications services, which demanded high levels o f investment. The exclusivity period might also have encouraged more investors to take part in the bidding process, thereby increasing competition for the market. Similarly, in electricity distribution, exclusivity periods guaranteed the concessionaire a relatively constant income off o f clients' fixed fees. Even if some commercial risks remained related to the clients' ability and willingness to pay, granting the exclusivity rights reduced the need for other kinds o f guarantees. 49 - s[r! e, 3e, 3e, v, e, B e, BE v, e, m x 3 9 B M mh c-' s v1 m Y 2 8 Y 0 20 Y a IA Y d Ld e, M 4 3 UL 0 3 2 ?e c) V m s.-.- U )I E Y h U )I I 5aJ U s x W E m .-.-mE E 0 U V zE 0 U I 0) c .e W0 v1 a2 2 )I m 56 E W 9e m e .d : -T, .I aJ .c * ++ +++ 2.2.3.4 Major force risks or "acts of god 'I Major force risks should be defined in each contract, as should the provisions necessary to deal with them. In all cases in Peru, concessionaires did not bear the burden o f these types o f risks. Major force risks are unpredictable events that may occur over the life of the contract and have dramatic effects. They could include war, strikes, riots, natural disasters, and epidemics. Table 20 defines such events for each concession contract in the transport sector and their effects on the contracts. Telecommunications and electricity concessionaires are protected against them, although definitions tend to be vague. For instance, in the cases o f Egenor, Cahua, EEPSA, Etevensa and Electro Sur Medio, risks are not specified, leaving the authority to determine what classifies as a major force event. As these are supposed to be extreme events, contract obligations are suspended in all cases (Table 21). Table 20: Defining Major Force Events in Concession Contracts: A comparison Arequipa- 4nc6n-Huacho- Matarani road Pativilca road Xailways Lima Airport Matarani Port Unpredictable Events out o f :vent, not ?vent out o f the control Zvent reasonably control o f the attributable to the Zontained in ifthe party, which )ut o f the control o f Iefinition concessionaire narties, that he Peruvian :awes the delay or m e o f the parties, in that cannot be prevents them to Zivil Code interruption o f spite o f its efforts to avoided. accomplish their 2bligations in the xevent it. 3bligation for 6 :oncession contract. months. Natural events: Wars, civil conflicts, strikes, War and War, strikes, natural Wars, civil 3vents that rainfalls, wars, archeological those disasters and others zonflicts, strikes, Aassify civil conflicts, :ontained in earthquakes or strikes, among discoveries, some the Peruvian similar. natural disasters, others, governments act, amone others. Civil Code among others. Ifthese events Ifthe events last more suspendthe than six months: Ifevents last more concession for temporary increase in than 6 months: more than 3 Suspension or tariffs, extension o f the i)the Statecan months, the termination o f the make a loan to the contract. contract or any other concessionaire, at Effects parties can Suspension appropriate measure. least to cover its in the agree to modify The concessionaire o f the operation costs; ii) :ontract the length o f the Ifthe events last more concession. can request an contract. that twelve months and OSITRAN can Otherwise, the extension o f the airplanes cannot take of authorize a tariff concessionaire construction and land: any party can increase; or iii)the may call the period. terminate the length o f the termination o f concession contract. concession can be the contract. extended * Peru's Civil Code, in article 1315, defines malo xce as an extraordinary and unforeseen event that impedes ~~~~ the execution o f an obligation or determinesit to be partially completed: late or defective. Source: Concession Contracts. Elaboration: GRADE 52 Table 21: Defining M a j o r Force Events in Electricity and Telecom Contracts TdP Electricity Generators Electricity Distributors and Electro Sur M e d i o (except Electro Sur Media) Unforeseeable extraordinary Definition Contained in the Contained in the Peruvian Peruvian Civil Code* events that the parties could not avoid** Civil Code* Events that War and those War and those contained in classify contained in the No definition o f events** Peruvian Civil Code the Peruvian Civil Code I The concessionaire is The concessionaire i s exonerated from the The concessionaire is accomplishment o f the Effects On the exonerated from the exonerated from the contract without sanctions. contract fulfillment o f the contract without fulfillment o f the contract The concessionaires can sanctions. without sanctions vary the conditions for I * I providing electricity for some time. Peru's Civil Code, in article 1315, defines major force as a case where there i s no blame, consisting o f an extraordinary and unforeseen event that impedes the execution o f an ** obligation or determines it to be partially completed, late or defective. N o t applicable for the privatization o f Edegel. Source: Contracts. Elaboration: GRADE 2.2.4 Penalties and bonuses Contracts should establish bonuses and penalties to provide adequate incentives for the operator to fulfill concession or privatization obligations. Penalties and bonuses should be clearly defined and based on measurable factors, A range o f penalties also makes sense, as applying only one severe penalty (for instance, termination) could be very harmful and lack credibility. On the other hand, penalties cannot be so insignificant that the operators ignore them or treat them as built-in operational costs, In Peru, concession contracts contain sanctions for misconduct on the part o f the concessionaire, such as missed retribution payments, unfulfilled investments commitments and deficient quality in service provision. In addition, other mechanisms o f risk control, like the provision o f guarantees, and other positive incentives are in place. This section will analyze some o f the most important elements o f fines and penalties that concession contracts use. Appendix 3 presents concessionaires' principal obligations and their respective penalties. 2.2.4.1 Retribution Penalties With respect to retribution payments, the Matarani Port, railroads and Lima Airport concessions establish very similar mechanisms for sanctions. In all the cases, the sanctions give concessionaires the possibility to correct deficiencies. Only repeated missedpayments are a reason for cancellation o f the concession. In the contract for the Ancon-Huacho- Pativilca highway, these conditions are clearly defined (the lack o f payment three consecutive times or any five times in a year); while in the other cases they are not specified. 53 In the telecommunications sector, failure to pay retributions to the State is considered serious and entails monetary payments determined by OSIPTEL. Ifthe concessionaire fails to pay the retributions, it i s penalized roughly $30,000-$50,000 (30 to 50 UITs)." These sanctions have been criticized as very low in relation to TdP's revenues, more than USSl billion per year. In addition, termination o f the concession contract i s possible if the concessionaire does not pay retributions after 20 days o f being notified by the Ministry or OSIPTEL. In electricity contracts, penalties for failure to pay retributions to the State vary according to the concessionaire's revenue and are set by the Directive Council o f OSINERG.52For example, a sanction on a company that produces more than 1,000 million Kwh could reach a maximum of approximately US$650,000. 2.2.4.2 Investment penalties For investment commitments, all transport concessionaires must pay a fine for failing to make required improvements. There i s a relationship between the amount o f the fine and the gravity of the deficiency in most cases. The Arequipa-Matarani highway i s an exception. IfCONCAR does not begin scheduled works on time or i s delayed or stopped without reason, the contract could be terminated and CONCAR immediately forced to pay the guarantee-of-completion penalty stipulated in the contract. This threat o f contract resolution appears to be somewhat extreme and does not contemplate gradual penalties to enforce the fulfillment o f obligations. In all cases except the Ancon-Huacho-Pativilca highway, the State can invoke its right to terminate the concession. This penalty o f suspending a contract i s applied only in the case o f non-compliance o f investment promises, since investments are one o f the main objectives o f a concession. In telecommunications, sanctions for investments depend on the magnitude o f the nonc~mpliance.~~ Inthis case, sanctions range from US$50,000 to US$500,000, which are quite low compared to the amounts involved. Contract termination only takes place as a last resort and only after recurrent violation^.^^ Inthe electricity sector, investors were subject to sanctions for delays in the implementation o f installations. These fines for the most part were U S 2 0 0 per MW not installed per day. In practice, the sanctions were never imposed since in all cases, the goals were achieved in the time expected. Further, if generation or distribution firms did not complete investment projects on schedule, the State would execute previously required guarantees. The MEM was to evaluate and determine whether investment goals were accomplished or not and decide about the continuity o f the contractual relationship. Partial liberation o f guarantees was possible as long as investment progressed as expected. *' At2004, an UIT was equivalent approximately to US$900. 52Through R.N. 028-2003-OSiCD. 53 Noncompliance is measured according to certain indicators, such as by population center without service, by percentage points under the set goals, by number of public telephone without installation, etc. 54 There are no State indemnities in the event that the Ministry decides to `terminate a contract in advance due to continued noncompliance on the part of a company of its investment obligations or of the provisions o f OSIPTEL. 54 Box 7: Obligations and Penalties in Practice:the Electro Sur Medio case [n 2002, CIDEF prepared a report on the development of Electro Sur Medio's privatizationprocess, which signaled cases o f noncompliance o f the privatization contract. There were two principal themes o f the investigation: one related to HICA's obligation, as buyer o f Electro Sur Medio, to invest its own resources without jeopardizing the patrimony o f the company; and the other regarding the clause stating that Electro Sur Medio could only have contracts with H I C A or its shareholders for administrative, operational or consulting services, for a value o f less than 2 percent o f the annual net sales. An audit in2001 requestedby OSINERG found that: - Electro Sur Medio had transferred $8.2 million to HICA. These transfers included advance payments for a contract for management and technical assistance that surpassed the limit o f 2 percent o f the annual net sales. - H I C A had acquired shares o f C. Tizon P. S.A., Amauta Industrial and Constructora Vasquez Espinoza S.A., -valued at US3.6 million, with financing guaranteed by class "B" stock o f Electro Sur Medio. Electro Sur Medio had granted guarantees for $12.5 million, in an endorsement o f letter of guarantee obtained by HICA, which disregarded its obligation to fulfill its investment commitments with its own resources and without jeopardizing the Electro Sur Medio patrimony. Electroperu confirmed the accuracy o f these findings and informed COPRI. In January 2002, COPRI evaluated the reports o f Electroperu and concluded that the only breach was the transfer o f finds o f Electro Sur Medio to H I C A and urged H I C A to return finds in excess o f the 2%. With respect to the other finding, COPRI indicated it implied a violation to Article 106 o f the General Law o f Societies and not a breach of contract. InApril 2002, Electroperu asked COPRI for authorization to proceed with the contract resolution for noncompliance o f H I C A Investments, as stated in Clause 16.1 o f the contract. Nevertheless, COPFU held its position on maintaining the contractual relationship. Source: CIDEF 2.2.4.3 Quality Penalties Transportation concession contracts establish some minimum requirements related to quality, but both the definition o f those requirements and the corresponding sanctions are often not clearly detailed. In the case o f the Matarani Port, the contract does not explicitly establish a sanction for a breach of the minimum operational criteria, but the general clause of contract termination or OSITRAN's Regulation o f Sanctions could be applied. Worse still, in the concession contract for the Arequipa-Matarani road, it i s only specified that the concessionaire must ensure good transit levels, but it does not define the parameters o f compliance or the sanction for noncompliance. Inthe case of railroads, investments are tied to reaching certain standards. Concessionaires must maintain the assets of the concession in good condition in order to reach the standards o f United States Railroad Administration (FRA) Class I1 within the first five years o f the concession, or else face the sanction of the contract expiring. In the Lima Airport concession contract, sanctions relate directly to the fulfillment o f the goals of investment, respecting the established quality criteria. In telecommunications, sanctions about quality are clearly defined and are either: i) economic sanctions and ii) in case o f recurrent failings, termination o f the contract. Sanctions o f the first type go from US$50,000 to US$500,000 and their applicability i s similar to that related to investment obligations. 55 For electricity generation and distribution, quality norms and sanctions are set by OSINERG." The magnitude o f the sanctions depends on the size o f the generators and distributors in terms o f production or sale o f energy, respectively. There are four scales o f sanctions where the maximum reaches 650 UITs and the minimum 1 UIT. 2.2.4.4 Guarantees infavor of the State and Related Sanctions Other mechanisms exist as incentives to contract compliance in addition to sanctions. There are clauses that include a guarantee for the State in case the concessionaire fails to fulfill his obligations. These clauses are very similar in all transportation concessions. The Regulation for the Collection and Application o f Infractions, Sanctions, and Rates56approved by OSITRAN i s applied to contracts o f transportation concessions. This regulation contains the framework for sanctioning, typifies infractiod7 and their corresponding sanctions, dictates procedures and determines competent organs for the imposition o f sanctions, collection o f fines and rates. Of course, the application and effectiveness o f these sanctions depend on the monitoring and enforcement capacity o f the regulator. OSITRAN uses mechanisms o f information verification with a certain degree o f technical independence to make its sanctions effective. During2001, eight sanctioning processes began and concluded in effective sanctions. In 2002, there were 11 such cases and seven began sanctioning administrativeprocesses. In telecommunications, guarantees are required to attain the concession's objectives. Failure in implementing a guarantee i s sanctioned by termination o f the contract. Nevertheless, the amount o f the guarantee at US$8 million, becomes clearly insignificant in relation to the concessionaire's annual income o f more than $1 billion. Contracts in the electricity sector included investment obligations which stipulated that the strategic investor had to maintain ownership o f at least 25 percent o f the capital o f the company for a minimum o f five years. This condition was justified because the strategic investor guaranteed compliance with the investment obligations in the sale o f share contracts. For this reason, economic sanctions were placed for noncompliance on the emissions o f such guarantees, which were between US$500,000 and US$10 million. These sanctions represented a significant percentage o f their annual income. When contracts did not stipulate sanctions, often regulating agencies would. In the telecommunications sector, OSIPTEL's Regulation for the Collection and Application o f Infractions, Sanctions, and Rates would apply. In the electricity sector, OSINERG Rules on the Tipification o f Infractions and Scale o f Finds and Sanctions i s applied. This last regulation establishes -as in the case o f transportation- a sanctioning scale applicable to the concessionaires, according to their annual income. Sanctions are applied depending on the "D.S.020-97-EM and R.N. 028-2003-OSiCD. "This regulation was passed under the Resolution of Directive Council of OSITRAN (006-CD/OSITRAN) in December 1999, and modified by Resolution 029-2001-CD/OSITRAN. "The Regulation has establishedfour categories of infractions, each correspondingto a sanction: extremely serious, very serious, serious, and light. 56 production or sale o f energy by generation or distribution firms, respectively. The maximum reaches 700 UITs and has a limit three times the amount due. 2.2.4.5 Incentives Contracts must establish, in addition to sanctions, incentives for concessionaires to hlfill their obligations and operate efficiently. Table 22 provides details for the different incentives offered in transportation sector together with their desired objective. Incentive Objective Discount in the retribution for rehabilitation Increase investment until FRA I1 [Railways and maintenance conditions are reached I Arequipa - Matarani road Extend time o f concession Not to increase tolls I Matarani Port I Extend time of concession I Increase investment I ELimaAirport I Extend time of concession I Increase investment In the case o f railway concessions, a mechanism is included to encourage investment. During the first five years, the total amount of the Principal Retribution and the Special Retribution need not be paid if the concessionaire effectively invests in rehabilitating or maintaining the railway (Clause IO. 1). Between the fifth and 10th year, concessionaries have the rights to not pay 50 percent o f the Principal and Special Retribution. This incentive seems necessary given the bad state o f the railway's assets before the concession and fits adequately into a scheme o f quality investment targets (FRA 11). T o effectively work as an incentive, this mechanism will need to result in higher quality standards than FRA I1 (which is required by the contract anyway) or in reaching the FRA I1 standard before the five years stipulated inthe contract (in 2004). In the case o f the Arequipa-Matarani highway, incentives were included to control toll increases. One o f them involves extending the term o f the concession if CONCAR does not increase the Maximum Tariffs established in Attachment 2 o f the contract (November 2000). The tolls have remained frozen from November 2000 to November 2002, when the last tariffs readjustment was made. As result, the concessionaire has managed to extend the term up to a maximum o f 13 additional months.58 Both for the Matarani Port and the Lima Airport, there i s also the possibility o f extending the duration o f the concession at the concessionaire's request. This mechanism, although reducingmarket competitive pressures, can generate incentives for long-term nonobligatory investments so that concessionaires will have more time to take control o f their benefits. Also in the sectors o f telecommunications and electricity, the contracts contain incentives to reach related objectives for each sector (Table 23). j8Inaddition to the 55-month extension awarded for others motives. 57 Table 23: Incentives in Privatization Contracts Incentive Objective TdP Evaluation o f Correcting bad performance o f the concessionaire contract renewal and assuring accomplishment o f the contract Electricity generation companies Undefinedlength for operations Encouraging investment Electricity distribution companies Undefined length for operations Encouraging investment In all cases, incentives exist and are tied to the duration o f the operating licenses awarded as part o f privatization. With regard to electricity contracts, the indefinite term o f the operating license would have attempted to reduce the company's risk by guaranteeing optimal levels o f investment. In the case o f TdP, the duration o f the contract allows for necessary investments to expand services.59 Additionally, TdP has the possibility o f prolonging the period o f recovery o f its investments through a performance evaluation carried out by OSIPTEL every five years. In 1999, TdP obtained an extension on its operating license from 2014 to 2019. However, in June 2004 a new five-year extension was not granted. Thus, this mechanism i s an incentive for the company to fulfill all the aspects o f the operating license contract. 2.2.5 Certainty and Flexibility In designing the contract, it is crucial to obtain a balance between certainty and flexibility. This balance depends on the possibilities offered in the contract to adapt to situations as they appear without changing the essence o f the contract and without granting excessive discretion to the entity in charge o f supervising compliance. Too much flexibility can be prevented by a suitable definition o f goals and obligations. Performance targets, for instance, let the concessionaires achieve desired goals while ensuring autonomy and use of their know-how. Most concession contracts for public transportation infrastructure have investment targets or technical requirements instead o f performance targets, except railways. The railways concessionaires had to reach the United States Railroad Administration (FRA)60Class I1 standards within the first five years o f the concession or their contract might be terminated. In the case of the Matarani Port, specific targets related to investment goals such as contamination control mechanisms, loading systems improvements, and new storage areas. However, initial investment goals had weaknesses that led to modifications according to more realistic freight traffic estimations. Investment goals o f approximately US$7.9 million were established in the concession contract for the Arequipa-Matarani road. CONCAR 59Macroconsult(2000). "The Federal Railroad Administration (FRA) is one of ten agencies within the U.S. Department o f Transportation in charge o f inter-modal transportation, as well as the timely and effective implementation o f the Transportation Equity Act and technological advancement. 58 should invest in rehabilitation, construction o f an alternate road, and periodic maintenance every four years. Due to miscalculations, investment cost more than was budgeted, and OSITRAN authorized a 24 percent budget increase.6' In both cases, the investment goals permitted a certain level o f flexibility to the concessionaire, even though they were specific. The Ancon-Huacho-Pativilca road contract had other provisions that allowed flexibility. Part o f the investment requirements was dependent on traffic, and the concessionaire could ask the supervisor to modify the investment program. The transportation contracts also introduced flexibility in the clauses on sanctions. The sanctions in the contracts for the Matarani Port and the Lima Airport for not making retribution payments or fulfilling investment goals incorporated an adequate level o f flexibility. But in other cases, there was either too much flexibility because established goals or applicable sanctions were not clearly defined (as in CONCAR), or extreme rigidity was introduced (as in the case o f the railroads). OSITRAN's Infractions and Sanctions Regulation significantly contributed to regulating this varied panorama. The telecommunications and electricity sectors are characterized by fast technological advance. In this sense, too much rigidity in the contractual terms can make i t difficult to incorporate new technologies and innovations. Sale contracts in the electricity sector had minimum requirements for investment, leaving to the concessionaires to decide on the know-how and the technology to use. In telecommunications, the concession contracts had a certain degree o f rigidity that forced the concessionaire to provide sometimes obsolete services (for example, telegraph). On the other hand, they also explicitly forced the concessionaire to incorporate technological advances for new services and quality improvements. This mechanism allowed flexibility since it did not impose any specific type o f technology as long as it was o f the latest generation. Other elements o f flexibility in telecommunications contracts are necessary because o f the sector's specific economic characteristics. Allowing for a contract change based on mutual agreement between the parties i s an important and appropriate mechanism when the environment in which a contract develops i s changing. The TdP's contract allowed for that possibility. An example was the modification o f the date that fixed telephone service would become open to competitors. The concession contracts were modified by mutual agreement between the parts (Section 4.5).62 Flexibility in the duration o f the contract i s also a possibility as seen in the contract o f TdP, which allowed for a renewal o f the concession every five years for five additional years subject to an evaluation o f compliance by OSIPTEL. In the electricity sector, the LCE and OSINERG's Regulation on Sanctions provided an important framework o f certainty for investors since they made explicit the terms o f the investments, obligations and consequences o f the breaches beforehand. In this case, the contracts left open the possibility o f being modified only by negotiation and agreement between the parties. ''Report 028-03-GR-OSITRAN, May 2003 "See D.S.021-98-MTC. 59 2.2.6 Renegotiations Renegotiation can be a useful instrument to address the inherently incomplete nature o f contracts. However, when renegotiations are excessively used, it casts doubts on the validity o f the contract and indicates opportunistic behavior by the concessionaire and/or the government, leading to a reduction in social welfare. In addition, ifbidders believe that renegotiation i s feasible, it could cause them to make unreasonable commitments to win the bid, assuming that the commitments could be changed in the future. Furthermore, renegotiations tend to be bilateral negotiations between the operator and the government that are not subject to competitive pressures and lack transparency (Guasch (2004)). Highrenegotiation rates have plagued Latin America, especially intransportation where 55 percent o f concessions have undergone renegotiation. T o ensure efficiency and avoid opportunistic behavior, contracts should: i) avoid ambiguities; ii) include clauses committing the government to not renegotiate except in the case o f well defined triggers and stipulating the process for adjustments; iii)require the concessionaire or buyer to pay fees when it requests the renegotiation (if the renegotiation i s done, the fee could be reimbursed); iv) when high or aggressive bids are offered, a detailed analysis should be done to dismiss unrealistic bids; and v) hire professionals to review, in a transparent way, claims for renegotiation (Guasch (2004)). 2.2.6.1 Transportation There has been more than one renegotiation in each transportation concession contract, with the exception o f the Matarani Port. Overall twelve addenda to the concession contracts under analysis were signed between June 1995 and June 2003. Furthermore, the first addendum often took place in the first year o f the concession (Table 24). This i s worrisome since the average time to renegotiate transportation concessions since the award date has been 3.1 years in Latin America and the Caribbean (Guasch (2004)).63 63Data i s from the mid 1980s to 2000. 60 Table 24: Timing and Number of Renegotiations in TransportationConcessions Concession I Addendum 1 GeneralDescription 1 1 Concession i s transferred from GyM S.A. to I Concesiones de Carreteras (CONCAR) June 21" 1995 Arequipa-Matarani Modification o f concession terms in response road (CONCAR) to the application of IGV to toll prices (extend November 17Ih I September 9Ih 1994 the duration o f the contract and other 2000 II IIcompensations) Merger betweenCORVAL and CONCAR and July 2nd2o01 increase o f payment to Fondo Vial Ferrocarril del Addendum 1 Complete the list o f assets I March loth2000 Centro (FVCA) December 6'h July 19Ih1999 Addendum 2 Anticipated devolution o f non obsolete goods 2002 I I I Ferrocarril del Sur I Addendum 1 ComDlete the list o f assets I \ March loth2000 and Sur-Oriente , (FETRANS) Exoneration o f access prices for wagons that January 1" 2003 July 19'h 1999 transport solid waste Nine modifications: New ways for the concessionaire to get financial support and April 6Ih2001 clarify arbitral procedures, among others* Eight modifications: the State should not Lima Airport Addendum 2 provide all the land necessary for the expansion July 25Ih2001 (LAP) February 14'h o f the airport, among others* 2001 Twenty modifications: New definition for Addendum 3 gross income; larger areas designated for September 30th I authorities, among others* 2002 Thirty four modifications: in the investment Addendum 4 1 plan, specially in relation to the second June 30th2003 runway, among several others* Matarani Port August 17'h 1999 Addendum 1 Modification of the investment plan July 26fh2001 In all cases, the renegotiation processes were formalized in the concession contracts. After an agreement was reached between the State and the concessionaire regarding the contract modification, it was submitted to OSITRAN for a technical opinion.64 Based on this opinion, the MTC then had to decide whether the contract revision will take place. In the Arequipa-Matarani road concession, Addendum 2 mainly dealt with two issues: a legal change imposing the application o f the sales tax (IGV) to toll prices and the payment o f the regulation rate (1%) to OSITRAN. In the first issue, the parties agreed that the concessionaire will pay the IGV tax amount while the concession was extended by 14 months, a period estimated to be necessary to compensate for the additional tax payment. O4This is in accordance with Law No. 26917 - Ley de Supervision de la Inversion Privada en Infraestructura del Transporte de Us0 Publico y Promocion de 10s Servicios de Transporte ACreo. 61 This addendum also established price caps on toll prices, price adjustment schemes, and insurance coverage and compensation for non-increasing tariffs. Although these issues should have been included in the original concession contract, at the time it was signed (1994), there was neither an institutional framework to promote concessions nor a regulatory agency. Therefore, this addendum i s a result o f a contract design failure, rather than o f opportunistic behavior. Other addenda are more questionable, including Addendum 3 o f the Arequipa-Matarani road contract. This addendum merged CONCAR and Corporacion El Ovalo S.A. (C0RVAL)-a Lima parking lot firm-with the approval of OSITRAN and MTC. The merger allowed them to enjoy.the tax benefits provided to mergers by the Peruvian legislation in fiscal year 2001. A t the same time, the payment to the Fondo Vial increased to 12%. The purpose behind those changes i s not very clear andjustified. There were more than seventy modifications to the original Lima Airport concession contract in just three years. Most o f the important changes made it easier for the concessionaire to obtain loans to fulfill its investment obligations and changed the conditions o f the financial-economic equilibrium clause. Changes were also made to ensure that certain creditors would be paid back in any circu~nstance.~~Other important changes included i)an exclusion o f some concepts from the payment o f the retribution rate to the State; and ii)the modification o f investment plans through changes in schedules and amounts o f money, particularly establishing that the construction o f the second runway would begin when the required additional lands were given to the concessionaire. Table 25 describes some o f these changes in more detail. 65Inparticular, LAP'S"approved potential creditors" - Overseas Private InvestmentCorporation (OPIC) and Kreditanstalt fir Wiederaufbau (KfW) demanded the modification of the concessioncontract to assure their loans would be paid back. The concessioncontract considered the possibility that approved creditors propose modifications immediately after it was signed. All the proposed changes could have been incorporated in a single addendum, but the regulator decided to proceed in several stages introducing first the least controversialdemands. 62 Table 25: SelectedModificationsto the Lima Airport Concession Issue Addendum The concession contract can be modified if it is required and appropriatelyjustified by the approved potential creditors to complete an EGP contract (Endeudamiento Addendum 1 garantizadopermitido) (clause 24 7) The definition of "Approved Creditors" (Acreedores Permitidos) is extended to include governmentagencies (clause 1.1). Addendum 2 A clause is added in which the State guarantees that Approved Creditors will receive all benefits establishedin the EGP contract. In this sense, obtaining EGP implies that the State and LAP assumed certain obligations with respect to the Allowed Creditors Addendum (clause 9.2). In case the concessionaire terminates the contract because of the State's fault, Approved Creditors will be paid back what is indebted, through a discount to the payment to the concessionaire. This repayment will include interest and commissions Addendum4 (before, it only included the amount indebted) (clause 15.5). In case the State decides to terminate the contract, Approved Creditors are paid back what is indebted, through a discount to the payment to the concessionaire. This repayment will include interest and commissions (before, it only included the amount Addendum 4 indebted) (clause 15.6). Approved creditors' opinion is required as part of the process of a contract renegotiation (before, only OSITRAN opinion was considered), as long as there is an Addendum 4 EGP contract (clause 24.7). Regarding the redefinition o f the financial-economic equilibrium o f the concession, the new scheme modified three aspects o f the original clause: i) Thesetoflawsandnormsthatwereexcludedforclaimingeconomicdisequilibrium.In particular, norms about tariffs and sanctions were excluded. ii)The minimum percentage of changes in incomes or expenses of the concessionaire required to apply this clause. Before, this minimumwas 15 percent o f variations o f the average o f the two previous years. After renegotiation, that percentage decreased to 5.5 percent for variations o f four consecutives periods o f three months. iii)Thetimethat should lapsebeforerequestingthe applicationofthis clause.Before, the concessionaire could only request it every two years; after the contract was modified, the concessionaire could request it at the end o f each period o f three months. Ingeneral, all the major modifications to the Lima Airport contract indicate changes to the "rules o f the game" after a short period o f time, failures in the original contract design or in the conditions offered by the concessionaire, or a combination o f the above. As a consequence, these addenda have been criticized by the public and the main unions o f the country, particularly in relation to the lack o f transparency inthe renegotiationprocess.66 ~ 66 The Asociacion de Servicios Aeroportuarios Privados (ASAEP) released a communication entitled "Transparencia para la libre competencia," where they asked that the contract renegotiation betweenLAP and the State be made public 63 The Matarani Port concession had only one addendum which modified the contract's investment plan to make it more reflective o f actual market requirements. A new plan for compulsory investment was established, increasing it to US$ 5.7 million for the first five years o f the concession. This addendum implied an important change in the rules o f the game established in the initial concession process. However, it also seemed justified because the original investment plan was not adequate. This would indicate a failure o f contract design since the contract did not offer flexibility for such adjustments. 2.2.6.2 Telecommunications and electricity In the telecommunications and electricity sectors there have been only four addenda between 1998 and 2002, which i s fairly low compared to the number o f transportation infrastructure renegotiations. Similar patterns have been found for Latin America in general. According to Guasch (2004), renegotiation was far less common in telecommunications and energy, partially as a result o f the higher competitiveness o f these sectors which significantly reduced the operators bargaining power. In most cases, telecommunications and energy operators were not the only service providers, giving governments more options for securing service from other operators-existing or new-in the event of a threat by an incumbent operator to abandon the business if renegotiations demands were not met. Most o f the contracts' first addendum took place in the third or fourth year o f the privatization, which i s a relatively long period after the initial award (Table 26). An exception i s Edegel, where the contract was amended in the first year to correct an insignificant error. Table 26: Timing and Number of Renegotiationsin Telecom and ElectricityPrivatizations Addendum GeneralDescription Renegotiation subscriptiondate i)Reductionoftheexclusivityperiodby one year; ii) Postponement of the Telefonica del Peru application of the Productivity factor until (TdP) 1994 Addendum 1 2001; and iii) Calls within a August 5th 1998 "departmento" or region defined as local calls (insteadof within cities or towns). Correction of the specification o f the Edegel (electricity number of stocks sold. The correction was November 30th generator) 1995 Addendum 1 insignificant as it impliedjust 1995 approximatelyUS$ 3.00 ElectroSur Medio Addendum 1 Extensiono f the operatingarea April 24`h 2000 (electricity distributor) 1997 Addendum 2 Extensionof the operating area July 24`h2002 64 The renegotiation processes were open and depended on the mutual agreement o f the different parties in all cases. For electricity, it was not necessary to adopt modification procedures for the concession contracts because they were general enough and based in the LCE. Both the LCE and the Telecommunications Law recognized renegotiations and covered many adjustment procedures, including changes to the concession area, tariff regime, and other obligations. For telecommunications, OSIPTEL i s charged with issuinga statement regarding technical aspects, tariffs, quality, and interconnection, amongst other things, for each modification that the Ministry considers.67 Addendum 1 o f Telefonica's concession contract advanced the opening o f the telecommunications market to competition. TdP's renegotiation was motivated by the accomplishment o f the investment commitments in a shorter time than initially planned. Two substantial contract modifications are included in the addendum: the delayed application o f the productivity factor and the redefinition o f what was considered a local call. For some authors, there are clear indications that this addendum mainly benefitedTdP. The concession contract established that the period o f limited competition would end in 1999. From that moment, the productivity factor was to be applied to price adjustments. However, the modification established that its application would be delayed until 2001. According to Oscategui (2001), this allowed TdP to enjoy two more years o f monopolistic rent, and productivity improvements were not transferred to users through lower tariffs. With respect to the definition o f the concession area and what a local call is, before the modification, a local call was defined as a call within an urban area. In other words, it did not necessarily include the whole province. After the change, a local call included everything inside a "Departamento". According to Torero, Schroth and Pasco Font (2003), this change implied an increase in the price of'local calls and explains the reduction o f growth in total consumer surplus since 1997. Although the real price o f local calls went up, at the same time, the effective price o f national long distance for calls within the Lima "departmento" went down. Other authors offer more favorable interpretations based on the social benefits o f the renegotiation. A report by OSIPTEL6* concluded that the advanced opening was very favorable for users. Considering only the tariff policy for fixed telephones, it found that the advanced opening generated accumulated user savings o f US$ 199 million (equivalent to US$160 per user) over a period o fthree years (the time duringwhich the application o fthe productivity factor was delayed). For OSIPTEL, the improvement i s explained by the single drastic reduction in tariffs at the time o f the advanced opening, instead o f waiting for a progressive reduction (see Box 8). `*"Libro "OSIPTELIS chargedwiththisfunctionbyDS.062-94-PCM,August1994. Blanco. El Proceso de Apertura del Mercado de Telecomunicacionesen el Peni". Estudiosen Telecomumcaciones.No. 10. 65 Box 8: OSPITEL's reasons for approvingAddendum 1in the TdP contract OSIPTEL prepared various scenarios to simulate the effects o f Addendum 1 to the TdP contract. The effects :valuated included the increase inthe local area and tariff variations on user welfare and profits o f established sompanies as well as potential entrants. With this objective, they estimated potential market participation, long distance traffic that would become local, interconnection charges, telephone demand elasticity, and other variables. One o f the alternatives evaluated was to continue with the original terms o f the concession contract. However, for OSPITEL, this meant that the size o f the local area, its respective tariffs, and national long distance tariffs would lack an adequate definition (from a technical perspective). This would make it more difficult to initiate the opening process inthese markets. TdP, on its part, was interested in reducing tariffs only one time and applying the productivity factor after a certain period o f time. Hence, it made a tariff proposal to this effect. Based on the results obtained and after reframing TdP's proposal, OSPITEL recommended to the Council of Ministers that the opening be advanced, given that users would experience an immediate one-time benefit from tariff reductions-the alternative being gradual reductions over three years. In addition, clarity was obtained inthe rules o f the game and in the attraction o f new participants to the fixed telephony market due to the increased local area definition. Source: "White Book. The Process o f Opening o f the Telecommunications Market in Peru." Studies in Telecommunications No. 10. 2.2.7 Dispute resolution mechanisms All the concession contracts under analysis include a dispute resolution clause incase some problems arise between the concessionaire and the government. Arbitration i s preferred over other mechanisms o f dispute resolution because arbitration rules tend to be more stable and because o f the credibility problems o f the Judiciary. Although all concession contracts propose to start with direct negotiation or negotiation through a "Centro de Conciliacion," if it does not work out, arbitration i s used. Arbitration mechanisms are quite similar across contracts, although the Arequipa-Matarani road and TdP contracts are exceptions. This mechanism i s standard and reduces the regulatory risk associated with uncertainty over dispute resolution. However, in some cases, the Judiciary i s also used to settle disputes. 2.2.7.I Transportation Contracts in the transportation sector only consider controversies between the concessionaire and the State while controversies between the concessionaire and other operators are managed by the regulatory agency OSITRAN. With the exception of the Arequipa-Matarani road, concession contracts have several arbitration stages. The first stage i s Direct N e g ~ t i a t i o nIf ~ . ~direct negotiation does not lead to a solution, the parties have to decide if the problem i s non-technical or technical. But the contracts do not define what i s considered technical; if the parties cannot agree in a certain time period, the 69 In the cases of railways and the Matarani Port some conciliatory institution could also intervene. For railways this institution was the Centro de Conciliacion y Arbitraje de la Camara de Comercio de Lima. 66 problem i s automatically considered non-technical. Technical controversies are solved by an expert on the issue and are all handled domestically. Non-technical controversies (in the case o f the Arequipa-Matarani road, controversies in general) are solved through arbitration "de derecho". In the cases o f the Ancon-Huacho-Pativilica road, Lima Airport, and Matarani Port, if the amount o f money involved in the controversy i s above a certain quantity, international arbitration i s used.70For railroads and the Arequipa-Matarani road, no distinction i s made and all controversies are subject to national arbitration. 1n.most cases, arbitration resolutions are not subject to appeal. Railroads are a partial exception, where if one party appeals, it has to present a security bond o f US$250,000 in favor o f the other party that would be executed if the appeal i s denied. In the Arequipa- Matarani road, appeals are also possible and the controversies go to the Judiciary for a final solution. This i s not recommended as arbitration loses many o f its good qualities (it i s faster and less costly, provides more security to investors, more confidentiality, and judges are required to have neutral specific experience backgrounds). 2.2.7.2 Telecommunications and electricity For Telefonica, controversies with both the State and with private operators are considered in its operating license contract (Figure 30).7' Controversies between TdP and other operators are subject to arbitration from OSIPTEL (Box 9). Figure 30: Dispute ResolutionMechanisms for Telef6nica e A controversy With the State With private operators I Direct negotiation Arbitration procedure Solution Arbitration in Camara 70 For the airport, all the non-technical controversies over US$5 million go to international arbitration 7 'The sale contract also considers arbitrage as a mechanism to resolve disputes between the State and TdP 67 Controversies with the State are first negotiated between the parties and if no solution i s reached, the controversy goes to arbitration in the Camara de Comercio de Lima, whose final resolution i s definite (no appeals in the Judiciary). For telecommunications, there i s no distinction between technical and non-technical controversies. As controversies between private agents are included in the operating license contract, it i s worth mentioning some relevant aspects o f the procedure. An OSIPTEL resolution determines the actual mechanism to resolve these kinds o f controversies, defining a conciliatory stage and two instances for resolutions. This mechanism has seen many changes since 1995, when the first rules were enacted. Box 9 illustrates how the mechanism works and makes some references to its main changes. Box 9: Resolution of telecom disputes between TdP and private operators The TdP concession contract specifies that conflicts between TdP and private operators regarding interconnection, tariffs and technical and economic issues must be resolved by OSIPTEL. The current rules were set in 2002 (Resolution o f Consejo Directivo No. Ol0-2002-CDIOSIPTEL), as a result of modifications to two previous resolutions (Resolucih de Consejo Directivo No. 001-95-CDIOSIPTEL in 1995 and No. 027-99-CDIOSIPTEL in 1999). The main modificationbetween the first and second version was the conflict resolution procedure. From 1995 to 1999, there were two alternatives: an administrative procedure (depending on OSIPTEL's norms) and an arbitral procedure (subject to arbitration laws). From 1999 to 2002, this distinction was eliminated. In botk cases, the Cuerpo Colegiado de Osiptel (CCO) was the first instance; ina second stage, the controversy would go to OSIPTEL`s chairman. The 2002 changes kept the main administrative aspects o f the 1999 rules, but modified the second instance fol ippeals. A Tribunal de Solucion de Controversias was created to perform that function. The resolution o f the rribunal is final. Disputes between the State and electricity generators and distributors are subject to arbitration. The procedure i s much simpler than in telecommunication and transport infrastructure contracts. When a controversy arises, the first stage i s to try to get a solution between the parties. If it does not work, the controversy goes to an arbitration process in Lima. In all cases an arbitration tribunal i s formed by three members whose resolution i s final (no appeals). The only case that allows for appeals i s when the final arbitration resolution does not follow the Arbitration Law (Law No. 26572).72 In this case, the party who appeals has to present a security bond (US$ 100,000 for generators and IJS$ 50,000 for distributors) that would be recovered if the appeal i s accepted by the Judiciary. It i s also worth mentioning that for Luz del Sur and Edelnor (distributors), both arbitration and the Judiciary channel are possible, depending on the parties' criteria. 72 This situation includes absence of notification, absence o f proof, resolution out o f time, and others related to formal procedures. 68 2.3 SUMMARY Table 27: Summary matrix for Process chapter Process characteristic Performance Awardingprocess Transparentbut relatively low competition in transport. Winning bid was muchhigher than reservepricefor telecom and airport Unbundlingand Well-designedon paper,but limitedinpracticein many cases. competition Risk allocation Contract-Law and Stability Agreements allocated and mitigated risks well. Political and commercial risks could have been handled better. Major force risks often ill- defined. Penaltiesand Appropriate for the most part, although penalties for Telefonica were miniscule incentives comparedto revenue. Renegotiations Provoked controversy in telecom and some transport contracts. Less frequent in telecom and electricity thanin transport. 69 3. Impacts of PrivateParti~ipation'~ In this chapter we analyze the direct and economy-wide impacts o f concessions and privatizations. Direct impacts relate to privatized firm performance, consumer welfare, and employment. Economy-wide impacts focus on how infrastructure price, quality, and access changes will impact sectors that use these services as inputs. A computable general equilibrium model i s used to capture the economy-wide impacts. The report also draws some comparisons with Latin America for telecommunications and electricity. 3.1 IMPACTS ON FIRMPERFORMANCE T o estimate the impact o f privatizations and concessions on firm performance, we look at measures o f profitability, efficiency, investment, output, employment, and leverage. Table 28 shows these measures as well as the predicted changes due to private participation. The analysis builds on frameworks from L a Porta and Lopez de Silanes (1999) and Megginson, et.al (1994). Table 28: Performance Measures Performance Measure Proxies Predicted Realtionship P(l) Profitability Return on Sales (ROS)=Net Income /Sales ROSA>ROSB Return on Assets (ROA) = Net Income/ Total Assets ROAA'ROAB Return on Equity (ROE) = Net Income/ Equity ROEA>ROEB P(2) Operating Efficiency Sales Efficiency (SALEFF)= Sales/Number of Employees SALEFFA>SALEFFB Net Income Efficiency (NIEFF)=Net Income/Number of Employees NIEFFA>NIEFFB P(3) Capital Investment Capital Expenditure to Sales (CESA)=Capital Expenditures/Sales CESAA>CESAB Capital Expenditure to Assets (CETA)=Capital Expendituresnotal Assets CETAn>CETAs P(4) output Real Sales (SAL)=Nominal SalesiConsumer Price Index SALA?SALB P(5) Employment Total Employment (EMPL)=Total Number of Employees EMPLAz Smirnov State Return on Sales (SROS) -0.3502 0.0394 -3.76 *** 14 15 *** 0,000 0000 *** (0.866) (0.037) Level of Significance. * IO%, **5%.*** I% 21 t-test for Ho about difference beween means. Unequal N's I t Standard Error in Parenthesis 41Shapiro-Francia test for normality. Ho: variable is nonnally dishibuted. 3 ' Test of equality: r' =(x, -x2)s-'(x, -x,>' Where x be a Ixk maaix o f the means and S be the estimated covariance matrix. 91 To complement these results, Figure 38 shows a non-parametric analysis o f the evolution o f state return over sales (SROS) o f privatized and concessioned enterprises. The analysis confirms the previous result and allows us to identify the trends o f SROS before and after private participation. Again, there is a considerable difference between government revenue when the enterprise was state owned and when it was privately managed. Although we do not have data that i s disaggregated by infrastructure sectors, we know that Telefbnica has paid considerably more in taxes after privatization. According to COPRI, it paid 163 percent more taxes in 1998 than in the year prior to privatization (1993). Figure 38: Evolution of State Return over Sales before and after private participation _ _ _ Before Privatization After Privatization ,04701- I 1985 1990 1995 2000 2005 Year Note: Uses smoothing including lowess-KSM. 3.5 ECONOMY-WIDEIMPACTS: TELECOMMUNICATIONS AND ELECTRICITY We use a Computable General Equilibrium (CGE) Model to estimate the economy-wide impact o f the privatization process. CGE models attempt to represent the circular flow of goods and services in the economy (Figure 39 shows a simplified example). They model all production activities, factors, institutions, macroeconomic components, and incorporate many economic linkages. These models are data-intensive and are constructed from combined national accounts and survey data that are first compiled into a Social Accounting Matrix. See Appendix 6 for a more detailed description o f the Social Accounting Matrix and CGE methodology. 92 Figure 39: Economy-wide Circular Flow of Goods and Services v, . Households Rest o f the - goods Imports markets p o Intermediate d s labor Government Source: Maddala and Miller (199 l), some additions with T o estimate the economy-wide impacts o f the telecommunications and electricity privatizations, we introduce the followingprice changes into the model: 0 Telecommunications: an overall price decrease o f 5.51 percent. This represents a weighted average o f the annual rate o f decline o f a 3 minute local (4.5 percent decline), domestic long distance (10.4 percent decline) and the international long distance (12.9 percent decline) call between 1994 and 1998.97 0 Electricity: an overall price decrease o f 5.09 percent. This i s the average annual decrease rate o f the variable commercial rate between 1994 and 199fi9* We multiply these price changes by the price elasticity o f each service to estimate the output change in each sector. The CGE model shows that the decrease intelecommunicationprices increases the GDP o f the whole economy by 0.184 percent, while the decrease in electricity prices increases GDP by 0.098 percent (Table 42). If we discount for GDP increases in the telecommunications and electric sectors, the indirect GDP impact would be 0.141 percent ($US63 million) and 0.073 percent ($US 32.6 million), re~pectively.~~Similarly, the telecommunications price change creates a .215 percent increase injobs (19,030 new jobs) while the electricity price change creates a .113 percent increase (10,002 new jobs). Ifwe consider an average annual wage o f Si. 2,860 in 1994, then the job increases generate approximately Si. 54.4 million and Si. 28.6 million (respectively) innew annual wages.'" O7 '*Theweights correspond to the relative importance o f each type of call in fixed telephone traffic. Variable commercial rate in Lima province which represents the greatest share of sales in Peru (67% o f the sales o f the distribution companies). ''Constant 1994 prices and an exchange rate o f 2.2 (equivalent to Si. 139 million and Si. 72 million). looAt constant 1994 prices. 93 Table 42: Impact on Macroeconomic Aggregates (%) 5.5 1% decrease in 5.09% decrease in telecom prices electricity prices GDP 0.184 0.098 GDP (indirect) 0.141 0.073 Absorption* 0.180 0.100 Consumption 0.190 0.100 Exports 0.240 0.100 Imports 0.190 0.080 Labor 0.2 15 0.113 *Absorption = Consumption + PublicExpenditure + Investment. The CGE model also allows us to estimate the impact o f the price changes on wages and different economic sectors (see Table 60 and Table 61 in Appendix 6). For instance, the main beneficiaries from telecommunications price changes are urban men (either non- skilled or skilled) who run their own business. For electricity, urban skilled wage earners (either male or female) have the largest wage increases. Mineral extraction and petroleum refining, machinery and equipment, and construction are examples o f sectors that would particularly benefit from the telecommunications changes. Machinery and equipment, construction, and textile products would be some o f the main beneficiaries from the lower electricity rates. 3.6 INTERNATIONAL COMPARISONS To understand how the impacts of private participation in telecommunications and electricity distribution in Peru compare to the rest of Latin America, we draw on Andres, Foster and Guasch (2005). They compared trends for Latin America and Peru before privatization with trends after privatization. The idea i s similar to the difference-in- difference approach used section 3.1 in that it attempts to control for factors that influence the variables in question but are unrelated to the privatization. For instance, a variable could show seemingly impressive performance improvements after privatization, but such increases could be consistent with pre-privatization trends, implying that privatization may not be responsible for the improvements. In most cases, post-privatization trends showed a marked improvement over pre-privatizationtrends. The following results were found for telecommunications: 0 Peru reported higher increases in coverage after the privatization than the average o f the rest o f the region. Note that at the moment o f privatization, Peruvian's coverage was significantly behind than the rest o f the biggest countries in the region. While Peru's teledensity in 1994 was 3.3 lines per 100 inhabitants, Argentina, Brazil, Chile, Colombia and Mexico had 14.5, 8.0, 11.3, 9.3 and 9.2, respectively. (Figure 40A) 94 Peru reported a higher drop in the number of employees. Lines-per-employeewere remarkablylow. On 1993, it was half of similar countrieswithout privatizations (for example, Brazil and Colombia) and less than one-third o f those with privatizations (for example, Argentina and Chile). (Figure40B) Consistentwith the previousindicators, the increasein labor productivity was above the region average. (Figure40C) Improvementsin network digitization outperformedthe region (Figure 40D) Prices in dollars for a 3-minute call and monthly residential service were above regionaltrends (Figure40E and F) Figure 40: Evolution of telecommunicationsindicators:Peruversus Latin America - z - 0 ______------- N 5- 8- .--c 8 - Fixed Telecommunications n Fixed Telecommunications Labor Productivity' Total Connections per Employee I 0 -5 0t 5 -5 0t 5 -cPeru -- All Trend before -cPeru -- All Trend before Trend after Trend after 95 E l Fixed Telecommunications Fixed Telecommunications Price of a 3-minutes Call (in Dollars) :- Monthly Charge for a Residential Service (in Dollars) --A%-----.\ /--- -.- 51-<-- 0 N --_c__ 0 , 0 - 5 .5 t -CI Peru -- 0I 5 All Trend before----- --cPeru -- All Trend before Trend after Trend after A similar analysis for electricity distribution was done with the following main conclusions: Output measures in Peru had important improvements in levels across periods, but, like in the rest o f the region, these results were driven by the firm specific trends. Hence, after this correction the effects o f change in ownership vanished. However, we also found that trends were higher in Peru, explained, in part, by the lower developed in coverage with respect to the average in the rest o f region (Figure 41A). The fall in employment was higher in Peru, also leading to higher labor productivity effects (Figure 41B and C). Reductions in distributional losses were greater in Peru than in the rest o f the region. Moreover, while the initial average conditions were worse in Peru, the improvements in this indicator were such that after three years o f private ownership, their average was below that o f the rest o f the privatized companies in the region (Figure 41D) Finally, contrary to the rest o f the region where an increase in the average retail price of electricity was found (as measure in real local currency), prices in Peru initially rose, but have since fallen below pre-privatization levels (this effect i s less pronounced when measured in dollars). However, contrary to the rest o f the region, Peruvian prices experienced a dramatic fall before privatization (Figure 41E and F). 96 Figure 41: Evolutionof electricity distributionindicators:Peru versus Latin America Electricity Distribution - Avg Number of Employees f - wi I 0 I?- ?-AI ,// i 0 - s i I -5 0 5 t -c-Peru -- ALL Trend before Trend after I I Electricity Distribution - n Electricity - Distribution - 0 .5 0t 5 -5 0 5 t --.c-Peru -- ALL Trend before peru ALL Trend before Trend after -- Trend after Electricity Distribution - Electricity Distribution - Avg Retail Price per MWH (in Real Local Currency) Avg Retail Price per MWH (in Dollars) n-lc -5 - 0t 5 -5 0t 5 Peru -- ALL Trend before --cPeru -- ALL Trend before Trend after Trend after Note: Data is normalized to 100 at t=O, the first year the firm was privately owned. There are two "Trend before" lines because one is shifted to allow easier comparison with the "Trend after" line. Also, trend lines are b a s e d on "All" Latin American countries in the sample Source: Andres and Guasch (2005). Figure 42 takes a slightly different look at Peru and Latin America comparisons. I t measures absolute changes in several performance indicators during and after the transition to privatization, although several o f the variables control for firm-specific time trends. The results in Peru are largely similar to the rest o f Latin America. The number o f subscribers 97 has gone down for electricity but up for telecommunications. Similarly, output has gone up for telecommunications but down for electricity, although the drop occurred earlier for Peru. Employment has fallen at privatized finns and labor productivity has gone up. Distributional losses have fallen, and quality and coverage have improved. While prices have risen for Latin America and Peru in telecommunications, on average, electricity prices have fallen in Peru while they have risen elsewhere. Electricity Distribution FixedTelecommunications LatinAmerica I Peru Latin America Peru Transition Post Transition Tnnsihon Post Transihon Transihon PostTransihon Transition PostTransiuon Quality I I I 1 - 1 1 I ICoverage NoData I I I I I I1 (*) -4I I I I I INote:(*) these variables were reportedIafter the firm-specific time trend consideration.I I I I I I I I I Source: Andres, Foster, and Guasch (2005) and Andres and Guasch (2005). 3.7 SUMMARY Table 43: Summary matrix for Impactschapter I1IType of Impact I Outcome I Firmperformance Positive for telecom, electricity, and some transport. Transport suffered from measurement limitations. Consumer welfare 1 Positive for telecom, mixed for electricity, and mainly unclear for transport. I Employment Negative for TdP (but positive for telecom sector), electricity, and railroads. Positive for port and airport due to pre-concession adjustments. Government revenue Positive for private participation overall (not just infrastructure), and positive for 1Itelecommunications. ______~ ~ F r a g e and quality Positive for telecom and electricity. (Economy-wlde Positive GDP growth andjob creation (estimated) for telecom and electricity. I International Positive for coverage, labor productivity, and digitization for telecom and positive comparisons for coverage and distributional losses for electricity. 98 4. Public perceptions of privateparticipation Public perceptions o f privatization have deteriorated significantly across Latin America since the late 1990s (Figure 3). To uncover more details about how this trend relates to infrastructure in Peru, the World Bank commissioned two studies.'" First, the studies sought to uncover what Peruvians expect from providers o f public services and how and when they feel the process should ideally take place. Second, they sought to understand how Peruvians felt privatization affected their lives and what public perceptions were about past privatizations efforts. The two studies used different methodologies, although both were conducted in cities around the country (see Appendix 7). The first was qualitative and relied on 16 focus groups o f people with and without access to privatized services. The second was a quantitative survey o f 1,808 persons in their homes. The separate studies revealed similar results: The Peruvians polled generally viewed the privatization process negatively, especially with regard to its impact on daily life. Participants indicated that the precarious economic conditions in which most live mean their most serious worries are the rising cost o f living and lack o f employment. Privatization was thought to adversely affect both o f these concerns by affecting both layoffs and increased tariffs for services. While the public feels that private companies have a series o f virtues absent from State-run enterprises, such as efficiency and the ability to offer better quality service, the fear o f the above mentioned adverse effects i s stronger than any perceived benefits offered by privatization. The three sections o f this chapter are divided into perceptions about the ideal privatization process, the actual one, and potential improvements for future privatizations. 4.1 SOCIAL CONTEXT The social context inwhich many Peruvians live serves as the backdrop to their perceptions o f how a public service provider should fulfill those needs. Here, we examine the most urgent and basic needs o f the studies' participants as well as their current levels o f access to public services. 4.1.1 Evaluation of Problems in Daily Life The surveys revealed that lack o f jobs i s perceived as the main problem existing in Peru today. Many people lack stable employment and there i s the perception that the situation has deteriorated over the past several years to the point that the majority o f the people polled said they have not had a permanent job for quite some time. Thus they feel compelled to accept temporaryjobs with poor working conditions or jobs that pay less than enough to cover their basic needs. There were also many references to contractors' abusive conditions, mistreatment and arbitrary pay. A second important problem i s the cost o f living and the general economic situation. Since good jobs were considered scarce, those interviewed said the general economic situation their families faced had also deteriorated. The cost o f basic necessities had increased, while Io' The World Bank contractedAPOYO Opinion y Mercado to carry out the studies. 99 incomes were decreasing. These two main problems led to a sense o f desperation among the Peruvians polled. Other problems considered less urgent were government inefficiency in solving citizens' basic problems and government corruption. In Lima and Iquitos, the focus groups also felt that Peruvians are too conformist and do not react to solve their many problems. Among the focus groups, concerns referred to the cost and/or lack o f basic services such as electricity, water and sewage systems as well as increased criminal and juvenile gang activity because o f deteriorating economic conditions."* A general feeling among the focus group participants was that they had few rights. They felt they were neither protected nor supported by the institutions o f the central or local governments. Though the concept o f a strong State responsible for managing citizens' problems was widespread, the perception o f social injustice was also strong. Those interviewed did not have a precise concept o f what they could demand or what they could do to resolve their problems. The perception that their precarious economic conditions could improve was almost nonexistent. 4.1.2 Access topublic services A considerable majority o f those surveyed have basic services in their homes: 95 percent have electricity while 84 percent receive water through the public water and sewage system.'03Over half o f participants had telephone service, with access for Lima residents being far higher than other regions. In the case o f electricity, there i s little difference between access in Lima and cities in the interior, though the difference in the Oriente, where only 85 percent have electricity, is significant. Lima i s at a disadvantage regarding water service. Only 82 percent o f those surveyed have service in their homes, compared to an average o f 88 percent in other provinces. Regarding public transportation, 96 percent o f those surveyed had used public roads, 40 percent had used airports, 7 percent had used seaports, and 12 percent had used river ports, but this figure increased to 70 percent for residents o f the Oriente. Generally speaking, northern Peru offers the best services, followed by Lima and then the southern cities. In contrast, coverage i s much lower in cities o f the central Sierra and especially the Oriente regions. Those who do not have access to either water or electricity live in more precarious economic conditions and rely on alternative methods to receive services. Water, considered essential, i s acquired from water sellers especially in Lima, and also from cisterns and wells. In the provinces, there also exist natural springs. Those who do not have electricity use candles and kerosene lamps. In Lima and Trujillo, participants frequently hook up to a neighbor who has the service and share costs (Figure 43). 102Focusgroups were designed so that half o f participants lacked access to basic services. I O 3These percentages are a result o f the quantitative study involving 1,808 participants, which is more representative o f the entire population than the focus groups. 100 Figure 43: How those without Basic ServicesSatisfy their Needs 4.2 PERCEPTIONS OF THE ACTUAL PRIVATIZATION PROCESS Recent experience with privatization has left Peruvians with both positive and negative perceptions. On a five point scale where 1 means "very negative" and 5 "very positive," the privatization process received a 2.7 on average. Over half o f respondents qualified it as neutral or somewhat positive. Only two percent o f participants thought privatization was very positive for the country, and 40 percent felt it was negative (Figure 44). When asked who benefited from privatization, the perception was that the government received the most benefits and the workers o f a company that was privatized benefited the least. A quarter of those polled felt privatization was beneficial for the country as a whole while 24 percent said it was damaging to the country.lo4 IO443 percent thought the government benefited from privatization, 30 percent felt users benefited and 22 percent workers benefited. On the contrary 61 percent o f participants felt privatization hurt workers and 53 percent felt it hurt users, while only 35 percent felt it hurt the government. 101 Figure 44: Perceptionsof Privatization in the 1990s 1 50 ......................................................................................... 31 I¶TOTAL Lima 0 Interior .-....... ......-........ Very Positive Somewhat Not p o s h or Somewhat Very negatiw No answer positiw negatiw negatiw Source: Survey conductedfor the World Bank. Basedon 1,808 interviews. The most popular reason given for the recent privatization trend was that government authorities sought personal gain (47 percent o f respondents). This perception o f corruption was stronger in the south and the Oriente. The second reason was that State-run companies were losing too much money (38 percent). Others attributed privatization to i)the need to raise money to pay foreign debts (23 percent); ii)pressure from the World Bank and IMF (20 percent); iii)the need to modernize the country and undertake public works (20 percent and 16 percent, respectively) and iv) pressure from foreign governments (8 percent). A separate micro-level analysis by Boix (2005) examined the impact o f individual characteristics on attitudes toward privatization in Peru in 1998 and 2003. H e uses a regression analysis to determine what the effect would be o f changing each variable from its minimum to its maximum (while holding all o f the other variables at their respective means). The results are shown in Table 44. For instance, the likelihood that a respondent opposed privatization was almost 50 percent higher if s/he thought that the economic situation o f the country or hisiher family was bad rather than good in 1998.And opposition to privatization fell for individuals with access to a sewage system and for people who read newspapers. However, somewhat paradoxically, opposition to privatization rose with the socioeconomic status and salary. Not surprisingly, opposition rose with lack of trust in the president. 102 1 Table 44: Predictedoppositionto privatization in Peru Changefromminimumto maximum value 1998 2003 Education -0.083 -0.021 Size of town -0.044 Self-employed 0.046 0.007 Employed in Public Firms 0.098 0.03 1 Not working 0.119 0.044 Retired 0.063 0.080 Housewife 0.030 -0.001 Student 0.089 0.012 Country in bad economic situation 0.445 0.164 Family in bad economic situation 0.485 0.000 Subjective assessment of salary 0.000 0.164 Lacks pride inbeing Peruvian 0.141 Socioeconomic status 0.202 0.000 Level of interpersonal trust -0.024 -0.167 Reads newsoaDers -0.052 -0.084 Trade is not beneficial 0.046 0.080 Has telephone 0.008 0.014 Has car 0.003 -0.05 1 Has access to water 0.086 -0.056 Has access to sewage system -0.097 0.060 Does not trust the President 0.116 0.000 President indispensable 0.194 -0.023 4.2.I PositiveAspects of Privatization On the whole, positive perceptions o f privatization were fewer than negative ones. Participants did not discuss the positive aspects spontaneously or explicitly and interviewers often inferred them from participants' answers. Those who did have access to basic services had more positive perceptions than those without access. In the latter category, the State was seen as having the responsibility to provide basic services while private companies only try to maximize profits and would not provide services as freely as the State might. Modernization (of public services and the State), better quality o f services, and wider access were the principal positive outcomes o f privatization expressed by those surveyed (Figure 45). For example, clients o f Telefonica de Peru and electricity services (in Lima) said that the time it took to receive access to services was cut in half. Those in the interior 103 provinces (e.g. Iquitos and Trujillo) especially felt that private companies offered better customer service. Lastly, a minority o f participants in Iquitos, Trujillo and Arequipa answered that working conditions had improved even though the number o fjobs decreased. Fourteen percent said nothing positive about the process. Figure 45: Positive Results-Percentage of Respondents who Associated Privatization with.. . Modernizationof public senices 37% Quality of senkes improvsd Modernizationof the State Cowage of senice increased State stopped losing money More devslopment State became more efficient None I 0% 10% 20% 30% 40% Source: Survey conducted for the World Bank.Based on 1,808 interviews. 4.2.2 NegativeAspects of Privatization A general consensus among the population interviewed agreed that privatization was more negative than positive. However, caution should be used in interpreting these results. In many Latin American countries, "rejection o f privatization" has been a comfortable vehicle to reflect frustration over the lack o f success o f socioeconomic policies (e.g. growth, employment, lack o f access to public services, inflation) that are not necessarily related to the consequences o f asset divestiture. That said, overall, over 60 percent o f those surveyed disagreed with privatization. The negative perceptions included government corruption, unemployment, higher tariffs, and others (see Figure 46). 25 percent in the South and 20 percent in the Sierra said there were no positive outcomes to privatization. Of those who favored privatization, half felt it improves services; but only 14 percent of the sample believed it created jobs; 11 percent thought the State was a bad administrator and 10 percent thought of it as an investment in the country. 104 Figure 46: Negative results of Privatization Corruption 46 Rise in unemployment 41 Funds obtained were misspent Services are more expensive 36 State lost patrimony Services are in foreign hands Monopolies still existed but were now privately run I I 0 25 50 Percent Note: Survey respondents were asked to choose the principal negative results of privatization. Source: Survey conductedfor the World Bank.Based on 1,808 interviews. Government corruption: Almost half o f those polled felt that the public has not benefited from the sale o f public companies. On the contrary, there i s the perception that profits were used for political ends and personal gain by officia1s.lo6 When specifically asked whether the funds garnered from the sales were reinvested adequately by the State, an overwhelming 88 percent o f participants thought funds had been misspent (Figure 47).Io7 Of those, most felt that part of the profits fed corruption; over half thought they had helped the reelection campaign of former President Albert0 Fujimori and some felt the government bought unnecessary equipment or carried out unnecessary public works projects. IO6With regards to corruption, 53 percent said that the privatization process was very corrupt while 41 percent felt that while there were isolated acts of corruption, they did not affect the whole process.Only 2 percent felt that the process was transparent and free of corruption. IO7This 88 percent differs from the 37 percent of respondents who said that privatization funds were misspent found in Figure 46 because of the way the question was phrased. The results from Figure 46 are based on what respondents thought were the principal negativeresults of privatization, whereas the 88 percent shown in Figure47 is based on a questionthat directly asked about how funds were spent. 105 Figure 47: Were Funds from Privatization Adequately Invested or Misspent? Howwere funds misspent? 76 Corruptionitheft Fujimori Unnecessary Unnecessary reelection equipmeni public works " 20 0 Social Encourage Infrastructure i lSecurity Defense n sectors production Source: Survey conductedfor the World Bank Effect on nsorkers; Massive layoffs were expected from foreign companies trying to minimize costs. Even if those interviewed acknowledged that layoffs constituted eliminating inefficient personnel or correcting an oversized bureaucracy, it was considered an unjust process. Only participants who had had a personal experience recognized that companies may offer incentives to employees to resign, but even then, they were thought insufficient for the needs o f the workers. On average, 41 percent felt unemployment was a negative aspect o f privatization.I O 8 Those surveyed also thought that workers for private companies have less rights, which translates into less stability and less benefits and that private companies seek to hire younger workers whom they can pay much less. There i s the feeling that under privatization there i s no regulatory framework; not only do workers have less protection, but so does the general population. In addition, there were concerns that companies hire foreigners who would take the place o f equally qualified local workers. The survey also asked if privatization created morejobs directly or indirectly. Indirect jobs are associated with other businesses that grow to provide services to private companies. Io'This figure rose to 55 percent in the interior 106 First, 63 percent o f all surveyed considered that presently there are fewer jobs.'09Only 15 percent felt privatization resulted in more direct employment. The disparity i s less when asked about the creation o f indirect jobs: 40 percent believe that indirect jobs diminished, while 32 percent believe they increased (Figure 48). Figure 48: What do you Think hasbeen the Effectof Privatization on Direct and Indirect Employment? I (0Interior 0 Lima ETotalI 10Interior EILima #Total 1 More jobs 5I ' 9 i 7 17 8 No effect 17; No effect 16: 620 No answer 0 25 50 75 100 0 25 50 75 100 Source: Survey conductedfor the World Bank. Basedon 1,808 interviews. Higher tariffs: The perception existed that foreign companies would not be subject to regulations and would arbitrarily raise prices o f services in order to recuperate their investment. This view was more prevalent among those who lacked access to basic services. Thepresence offoreign capital: The perception o f denationalizing the country or selling it to foreign interests was viewed very negatively, and seen as one reason the country remained poor. Foreign investors were viewed as profit seekers who exploit national resources and take profits abroad. It also was perceived that the country was incapable o f taking care o f its own necessities. Residents in the country's interior were more nationalistic than in Lima. 4.2.3 Quality of service loo58 percent o f the sample thought there were few jobs in Lima and 72 percent in the provinces. 107 Many participants expressed some level o f satisfaction with the quality o f basic services (Table 45). In total, more than half o f those who received water and electricity said they were satisfied with the services they receive; for the telephone system the level o f satisfaction was slightly less than half. The highest level o f satisfaction for the electricity and telephones was seen in the North while consumers in the Lima metropolitan area were the most satisfied with water and sewage services. Interms o f transportation services, both access to and satisfaction with highways were relatively high. Access to airports i s only 40 percent, but satisfaction among airport users was the highest o f all services at 59 percent. Access to sea and river ports was very limited, as were satisfaction levels o f their users. Surprisingly, perceptions o f people who had access to basic services were not significantly different that those who did not (see Box 10). Also, those who had personal experiences with privatization directly, as in the loss o f their own job or that o f a family member, did not answer more negatively than those who had not. Table 45: Level of Access to Services and Level of Satisfaction Level of Satisfaction Access to V e r y or V e r y or Service somewhat unsatisfied Electricitv Water and Sewer 84 52 31 Fixed telephone 52 49 38 Highways 96 56 26 Airports 40 59 22 River ports 12 42 36 Sea ports 7 44 33 Box 10: Little difference inperceptions between those with access to basic services and those without The overall perceptions o f the problems with privatization and its impact on daily life were not significantly different among the groups who had access to basic services and those who did not. In the latter category, there existed more the idea o f the State as having the responsibility to provide basic services. (Both groups criticized government-imposed charges for installations o f services, and considered that the price o f installation should be included in the cost o f the service.) Another exception was the perception that foreign companies would not be subject to regulations and would arbitrarily raise prices o f services in order to recuperate their investment. This view was more prevalent among those who lacked access to basic services. Comparing basic services pre- and post- privatization, the majority thought that the quality o f services had improved, but tariffs were also higher. The majority o f participants thought telephone service had im roved or remained the same after privatization, but were divided on electricity provision.' O Though privatization brought recognized improvements, the 'lo Comparing the services offered by private companies to those previously run by the State, 41 percent o f interviewees recognize an improvement in telephone services; 22 percent said service remained the same and 2 1 percent felt service i s worse. Inprovision o f electricity, 37 percent said there has been an improvement, 35 percent said the situation is the same, and 25 percent said services are worse. 108 perception that prices are too high for the services offered predominates in the case o f electricity and telephones. In contrast, a majority o f water users felt charges were appropriate for the service received.' " Though a major complaint against privatization was high prices, 41 percent said they would pay more if services were better, while only 12 percent would prefer lower prices even if the quality o f the services were affected. Survey participants were also asked to rate telecommunications and electricity firms according to the attributes that were considered important for companies offering public services. Table 46 shows that telephone companies (which are 100 percent privatized) perform fairly well in terms o f technical measures, including trained technicians, 24 hour service, and good quality of service. They also score well on coverage improvements and okay (surprisingly) on employment generation. However, people tended to think that telecom firms handled complaints badly, did not charge a fair price, abused the user, and did not offer flexibility for late payments. Results were very similar for electricity firms, although fewer respondents thought that they offered adequate payment schemes (Table 47).'12 Performs well Performs badly or very well Performs okay or very badly 1Trained technicians I 51 I 28 I 7 I 24 hour service 49 31 9 Coverage improvements 49 31 9 Installations in good condition 45 35 8 Quality service 38 39 12 Courteous uersonnel 34 41 14 ~~~ 'I'Of those surveyed, 59 percent of users of fixed telephones thought prices were too high and 55 percent o f clients o f electricity services 58 percent o f those receiving water services-even State-run services-consider the rates appropriate for the service they receive "'79 percent o f electricity distribution i s privatized, so some respondents are likely referring to the performance o f public companies 109 Table 47: Performance of electricity companiesaccordingto the attributesjudged to be important for public service providers Source Survey conducted for the World Bank Based 1,808 interviews 4.3 LESSONSTHE FUTURE FOR The studies analyzed in this chapter revealed a dichotomy between the perceptions o f State vs. privately run companies. More than 80 percent o f those surveyed said private companies charge more than their State-run counterparts. Yet, they also recognize that private firms offer better services, are more efficient, pay better wages, and treat their employees better, while State-run companies generate more corruption. Interestingly, the public does not perceive any difference between private and State-run companies regarding the crucial theme o f job creation, even though this was noted as a serious drawback to the privatization process when the question was approached from a theoretical standpoint. A possible explanation i s that respondents were thinking about job generation in the medium to long-term, and not in relation to the labor shock associated with the transition to private control. Also notable was the fact that 49 percent said private companies treat their workers better, while only 29 percent said State companies do (Table 48). These statistics reveal that there are many positive perceptions of private participation that could be exploited. Better communication from the government regarding the reasons for and benefits o f private participation could be combined with more participatory processes. l3In other words, the government could educate citizens about new investments ' and tariffs adjustments required for infrastructure services to be sustainable. The watchdog 'I3While it i s clear that the communication process was deficient during the 1990s, it i s also worth noting that the majority o f the population was in favor o f privatization in the early 1990s, rendering a communication campaign less critical. 110 role o f the regulator in limiting opportunistic firm behavior could also be publicized. Armed with such knowledge and existing perceptions about private sector quality and efficiency, the public might voluntarily opt for more private participation. Table 48: Characteristicsof State vs. Private Companies State Companies PrivateCompanies No answer Characteristics (010) (YO) (YO) Charges more 11 81 8 Offers better quality services 12 79 9 More efficient 25 66 9 Pays workers more 23 59 18 Treats workers better 29 49 22 Generates morejobs 42 42 16 I s more corrupt 59 26 15 Inthe absence of such a communication campaign and participatory processes, the negative perceptions surrounding private participation will likely outweigh the perceived benefits. The poverty in which many live means they are often left out o f the debate on the importance o f the quality o f services because they are faced with more immediate concerns, such as the rising cost o f living and employment. I t i s understandable then that participants indicated their preference that companies offering public services stay in the hands o f the State, since many perceive that private companies charge more for services and lay off workers, and that privatization i s linked with government corruption and officials intent on personal gain at the expense o f the general population. The majority o f Peruvians view their relationship with the State as a utilitarian one in which the government i s seen as a "grand employer," who provides stable jobs with good benefits. Given their more immediate concerns with employment, an inefficient, corrupt or bureaucratic structure i s preferable to a more efficient privatized one because it provides more opportunities. The relationship o f privatization to the daily needs o f people i s especially evident when proposing the privatization o f potable water services. Water was considered essential whereas participants implied they could survive without electricity or telephones. The Peruvians polled expressed a fear that privatization might translate into higher costs for water service. There have existed in some provinces, like Iquitos and Arequipa, strong local movements opposing such proposals based on the belief that water i s the basic resource for subsistence living. Hence, mechanisms that provide information to the community and involve them in decision-making are crucial."4 The level o f resistance would depend on which province a private participation proposal was being considered in; some provinces have politicized privatization much more than others and the possibility o f social mobilization against it i s greater. Arequipa and Iquitos are two provinces in which much resistance would be likely, as in Cuzco. In fact, attempted electricity privatizations in Arequipa in 2002 were abandoned due to fierce public opposition (see Box 1 in section 1.5). Meanwhile, Lima and Trujillo would likely react with more caution. I14Such communication and participatory mechanisms are being developed as part o f new water concession processes. 111 Concessions: A ViableAlternative Though knowledge about concessions was limited, the majority was against the idea. The study participants understood less about the concept o f concessions than they did about privatization. After an explanation o f the general meaning, the majority opinion was negative. This opinion was based on the assumptions that a private company would have a rental agreement over a service or company, whereby the company could act in its own interests, raising the price o f services (with the idea o f recuperating their investment quickly) and not worry about how they would leave the service or business once their contract ended. Participants perceived that the concession process was without regulations or controls and thus allowed a firm to do as it pleased, which could result in the service or company returning to the State in worse condition. On the other hand, some thought concessions were positive for the country, but only exclusively if there was an agreement that the State would not lose important assets. Though concessions were not fully understood or supported, the concept did gamer more support than privatization. If the State were compelled to give some business to the private sector, 72 percent preferred it be given in form o f concessions."' But when asked how agreeable they were to concessions in general, only 38 percent agreed. Il6 The perceived advantages o f concessions were that a company would remain ultimately under State control. Some said the concessionaire would comply with a set of government-imposed conditions that would defend the interests o f the public and the State. The disadvantages o f concessions included problems arising from the lack o f controls over concessionaires. Some thought the State would earn more from privatizing. Because o f the limited duration o f a concession, some thought concessionaires would impose high tariffs to obtain a quick return on investment, and could evade regulations or conditions (Figure 49). Future concession efforts should keep in mind the many negative perceptions that accompany the process. An attempt to alleviate concerns about the most sensitive issues through public awareness campaigns could go a long way in dismantling some o f the negative perceptions. However, increasing public knowledge would take a significant amount o f time since negative perceptions from past experiences are deeply rooted and will be hard to reverse. ' I 5While only 13 percent considered privatization a better choice. Nine percent of the sample said the State should maintain control of State-run services, no matter what circumstances existed (this figure rose to 19 percent in the southern cities). ' I o15 percent agreed with privatization, while 44 percentand 61 percent,respectively,disagreed. 112 Figure 49: What are PrincipalAdvantages and Disadvantagesto Concessions? No answer 0 25 50 75 100 0 25 50 75 100 Source: Surveys conductedfor the World Bank.Basedon 1,808 interviews 113 5. Moving forward: A Road Map for the Peru Concession Program Public and investor interest in private sector participation in infrastructure services has fallen since the late 1990s in Peru and all over Latin America, causing efforts in those types o f programs to stall. Part o f the reason for this waning interest has been macroeconomic conditions. The East Asian, Russian, Brazilian, and Argentine crises all contributed to investor disaffection with emerging markets, as did continuous conflicts and contract disputes among operators, regulators and government. Associated increases in the cost of capital are making those types o f investment less attractive. Public perceptions o f private sector participation continue to be an issue. According to Boix (2005), this i s because the public has little information about the mechanisms that lead to growth and therefore attribute disappointing growth results to prominent policy measures, such as private sector participation. Public perceptions have also been hurt by price increases, layoffs, unrealistic customer expectations regarding service levels (driven by politicians overselling the promise o f privatization), and accusations o f corruption. Perceptions o f loss o f control to foreigners and squandered government revenues from privatization have also damaged the credibility o f privatization. Despite negative public and investor perceptions, there i s still a case for more private participation. Chapter 3 shows the impacts o f privatization in Peru, which were largely positive in terms o f firm performance, government revenue, coverage, and quality. Similar results have been found throughout Latin America (World Bank and Inter-American Development Bank (2005)). Infrastructure has also been shown to improve output and growth (Calderon and Serven (2003), Chong and Lopez-de-Silanes (2005), Andres, Foster, and Guasch (2005)), affect health and education levels for the poor (Galiani, Gertler, and Schargrodsky (2002)), and reduce income inequality (Calderon and Serven (2004)). Public financing alone i s unlikely to be adequate to fully realize these benefits. According to the Peru's Ministry o f Economy and Finance (Direccion General de Programacion Multianual del Sector Publico), over US$lO billion i s required to fill the infrastructure gap in Peru. To rehabilitate the road network, US$3.7 billion i s needed, while US$1.8 billion i s required to increase the electrification coefficient from the current rate o f 76 percent to 100 percent. Sanitation improvements are estimated at US$5 billion. The Instituto Peruano de Economia (IPE) estimated the infrastructure gap to be much larger, at approximately US$18 billion. IPE estimated US$2.3 billion for telecommunications, US$6.9 billion for transport (including roads, ports, and airports), USS5.6 billion for electricity (considering a growth scenario similar to Chile's per capita consumption), and US$4.1 billion for sanitation (based on requirements similar to the coverage and development o f Chile). Peru's pipeline o f infrastructure concessions alone has estimated project costs o f US$1.6 billion (see Table 49). Private sector money i s not the only necessary ingredient for future infrastructure development in Peru. Also critical i s to design transactions, processes, oversight and policies that address the negative public perceptions, and instill confidence in both investors and the public, while providing an adequate regulatory framework. Moreover, changing public opinion to be more receptive to private sector participation i s a challenging 114 task. It will require concerted public relations efforts that educate people about the benefits o f private sector participation, efforts to improve transparency and involve affected stakeholders, and initiatives to account for the poor. The lack o f such efforts was a significant shortcoming o f past private sector participation in Peru. But securing public support will not be enough; eventually results on improved sector performance need to be shown. To improve sector performance, improving concession designs and establishing credible regulations and enforcement are essential, Appropriate new financial instruments such as guarantees from international bodies (for example the World Bank Guarantee Facility), domestic measures to maintain macroeconomic stability, and the development and tapping o f local capital markets all can encourage private investment and reduce the cost o f capital. We draw on private participation experiences in Peru and elsewhere to better understand how Peru's upcoming concessions can be improved. The main sources, aside from the Peruvian experience reported here, are Guasch (2004), Andres, Foster and Guasch (2005) Chong and Lopez-de-Silanes (2005), Birsdall and Nellis (2003), and McKenzie and Mookherjee (2003), among others. Based on this rich variety o f sources, this Chapter provides answers and a blueprint on how to move forward with private participation in Peru. Although the focus i s on Peru, these lessons can also be applied to Latin America and the rest o f the world. The first key step-already taken by Peru-is to choose the mode o f private participation. To address the public concerns about relinquishing ownership o f state assets, Peru has chosen the concession mode. For the concessions program to be successhl, a number o f things need to be addressed. These are issues that were often not handled adequately in the past, meaning that private participation has not lived up to its potential. They are also based on lessons from almost twenty years o f experience and over 1,500 episodes o f private participation. Issues fall into four key areas: 0 Addressing social issues o Overalltransparency o A communications program o Community involvement:a bottom-upapproach o Accountingfor the poor and the losers 0 Improvingconcession contractdesign 0 Enhancingregulatorydesign 0 Developingappropriatefinancial instrumentsand tapping the local markets 115 I 5.1 ADDRESSINGSOCIAL ISSUES This section looks at issues that can improve public perceptions o f private participation, increase community involvement and accountability, and address the needs o f the disadvantaged groups, including the poor and those specifically hurt by private involvement. Improving transparency, mounting an effective communications campaign, and involving local and regional stakeholders in private participation decision-making are all key elements to win public support and ensure more positive concession outcomes. T o address negative public perceptions, it i s first necessary to understand what citizens currently believe and what they expect from providers o f public services. The survey described in chapter 4 offers some insights on this score. Such knowledge can inform the design o f future concession contracts and provide the foundation for an effective communications campaign. Important Attributes of Public Service Providers Given Peruvian's concerns about economic conditions and employment, it i s not surprising that respondents considered employment generation to be the most important characteristic o f a company that provides public services: 77 percent felt it was indispensable or very important (Figure 50). Also important was that public service providers contribute to the economic development o f the country and support social services. In relation to service characteristics, the most important concerns were that a company would: i)not take advantage o f users; ii)offer fair rates; iii)provide quality service; and iv) good customer service (Figure 51). Figure 50: How would you Qualify the Importance of the FollowingAttributes in Relation to a Company's Effect on the Community? Generates employment Contributes to economic development Widens coverage of services Supports social services for the community 0 25 50% 75 100 Indispensable 0Very Important I I Source: Survey conducted for the World Bank. The graph shows only those who responded"indispensable" or "very important." Basedon 1,808 interviews. 117 Figure 51: How would you Qualify the Importance of the Following Attributes in Relation to Services Offered by Public Service Providers? Does not abuse the user Fair rates Customer Service Quality of services Trained technicians Installations in good conditions Ways to file claims Prior notice for cuts in service Good treatment of personnel 24 hour service Flexibility for late payments Needs-based pay packages E/Indispensable 0Very Important Y O Knowledge about and goalsfor theprivatization process Moving on to views about privatization, participants were asked how knowledgeable they were about the process. Though not all those surveyed had a clear concept o f the meaning o f the word privatization, most were aware o f the process once the word was explained. Where participants had only an elementary education level or where access to media outlets was minimal, understanding o f the term generally decreased. The following are some of the general perceptions about the privatization process: 1. It i s a process promoted by the government. 2. Peruvian companies are not allowed to participate. 3. It consists o f selling private and public companies to foreign investors 4. The sale to foreign investors i s not regulated. This last perception implies that Peruvians believe that the private company who takes over management of a State-run entity i s able to do as it pleases. The participants had no concept that the sale involves a contract, regulatory processes, nor specific norms to control private companies. These perceptions regarding regulations for the privatization process partly explain the negative perception surrounding privatization, since it i s believed that foreign enterprises would increase prices and do as they please with their employees. The survey also asked about public awareness o f the State's role in regulating public services. Many o f these institutions were well known to participants while others were largely unknown.'I7 `"Among these, the Ministry o f Transportation and Communications was the best known, among 87 percent of those interviewed. Some 83 percent knew o f the Institute o f Defense to the Consumer and of Intellectual 118 With both these perceptions on privatization and those on the ideal attributes o f a service provider in mind, let us now look at the circumstances under which Peruvians might find privatization an acceptable option. Participants found two reasons why privatization might be necessary: On one hand, State-run companies might be sold because the government seeks funds to implement its policies, or government officials seek to benefit from the profits o f the sale. This perception could be representative o f the government's lack o f communication with the public on why it has privatized in the past or how the profits o f such a sale were subsequently used. On the other hand, privatization o f a mismanaged SOE might be necessary to avoid bankruptcy. This response showed some awareness- principally among those who have access to services-that there may be a rational reason to privatize. In this sense, participants spoke about the bureaucracy inherent in State-run companies and how the inefficiency it creates has become an obstacle to development for certain sectors. Faced with the possibility o f bankruptcy, participants acknowledged that the only course o f action for the State might be to sell to foreigners who have more efficient management models and are more oriented toward production. Nevertheless, a little over half o f those participating in the studies felt there was no need to privatize infrastructure services because the State had sufficient resources to carry out the public works needed in the country.'I8 Given a hypothetical situation in which a mandatory project must be carried out when the State does not have resources, 53 percent felt that reducing expenses in other areas to make funds available for the project was an acceptable solution. Only 26 percent felt the work should be assigned to a private company, either through a concession or through full privatization.' l9 With privatization as the only option, those interviewed were most worried with the effect o f the sale on employment and on prices o f services. Transparency in the sale was also a top consideration, as was how profits o f the sale were used. Many felt the State should reinvest in the same area as the privatized service and should maintain shares in the private company so that it could continue to be involved in the decision-making process. Aside from creating jobs, private investment should also introduce innovative technology and offer wider coverage o f services (Figure 52). Property (Indecopi), 70 percent knew OSTNERG (energy regulator), 59 percent knew OSIPTEL (telecom regulator), 49 percent knew SUNASS (sanitation regulator), and 24 percent knew OSITRAN (transportation regulator). Opinions were generally positive o f all institutions with the exception o f OSITRAN and SUNASS. "* When asked if the State has sufficient funds to carry out public works that the country needs, on average 56 percent answered yes (52 percent thought so in Lima and 62 percent in the interior). IIoIn Lima, 30 percent favored bringing in a private company to perform the service, while only 19 percent in the interior provinces agreed. 119 Figure 52: If The State is Obligatedto Sell a Company, which of the FollowingConditions are Essential or Very Important? 1 Supervision of tariffs to avoid 73 abuse Peruvians :: maintain part ownership J i I I I I I I I 0 25 50 75 100 0 25 50 75 I00 Source: Survey conductedfor the World Bank. Basedon 1,808 interviews. The above survey data made it clear that measures must be taken to address the public's concerns about private participation if future efforts are to succeed. 5.1.1 Overall transparency A commitment to full transparency i s essential to begin the process o f regaining the confidence o f the public. That should start for example by placing on the internet all relevant material for the concession, including data, procedures to be followed, bidding documents, pre-qualifications requirements, the contract, amendments, etc. Whenever possible, transparent measures should be employed in decision-making processes and a consultation structure should be set up. For example, participatory processes that solicit local input could ensure public support and limit the impression that private participations were engineered by politicians seeking personal gain. Other essential elements to improve transparency in the awarding process are: allocating concessions through competitive bidding rather than through bilateral negotiations and using a single, quantitative bidding criterion to select the winner. Finally, transparency and periodic accountability inhow State 120 revenues from private participation are spent could limit concerns over whether resources are squandered. Holding public hearings over the life o f the concession, when appropriate, can give stakeholders the opportunity to express their views. It i s highly desirable to institutionalize public hearings and consultation processes, particularly prior to major decisions such as tariff adjustments and reviews, where all interested parties can have their say, present information, and influence decisions. 5.1.2 Conznzunicationsprogranz The government should design a communications programs to secure stakeholder support. It should be focused on the different relevant stakeholders and it should make clear the reasons for private participation in the project. It should describe the benefits and costs and identify the beneficiaries, losers, and consequences o f not moving forward. A communications campaign could also ensure that cost recovery principles, phasing-in periods, and social tariffs are well understood. Preferably, the government ought to use as messengers and communicators affected stakeholders, affected communities and sub- national governments. See Box 11 for an example o f how communication with the community in Indonesia allowed local water companies to implement much-needed improvements and tariff adjustments. Box 11:Making a deal with the community:Water supply and accountabilityin Indonesia Most drinking water in Indonesia is supplied by local government-owned water companies (PDAMs). Audits o f a number o f PDAMs have found serious problems, including low coverage levels, high levels o f unaccounted-for-water, non-potable water, and severe financial difficulties because o f low tariffs and mismanagement. While the local councils demand increased coverage and improved service, they often do not provide the PDAMs with sufficient resources to be able to achieve these objectives. A U.S.Agency for International Development (USAID) project assisted PDAMs to break the vicious cycle by developing corporate plans focused on using excess capacity, reducing water losses, increasing productivity, reducing staffing ratios, and enabling tariff increases. Emphasis on customer service is an important part o f the plans. By communicating the PDAM's plan for improving service and focusing on customer service, the PDAh4s gained the support o f the community, which in turn enabled the local councils to grant the much-needed tariff increases. For example, PDAM Banjar in South Kalimantan was in poor condition, with revenues barely covering out-of-pocket costs, water losses reaching 40 percent, and water provided only 12 hours a day. A plan was developed to enable the P D A M to improve services. But the success o f the plan depended on increased revenue from tariffs, and the local council would not approve the tariff increases. The P D A M therefore took the plan to the community, explaining the objectives and why the tariff revenues were needed. To cement its promises, the P D A M entered into a contract with the community, a contract which required real improvement in customer service. By gaining the support o f the community, the P D A M was able to obtain the tariff increase from the local council and has been able to decrease water losses, increase supply, improve water quality, and redesign their billing and collection system to virtually eliminate long lines and waiting periods. Source: Parton (2003) and ADB, JBIC, and World Bank (2005). 121 5.1.3 Conznzunity involvement and accountability To avoid some o f the problems o f the past, such as the failure in Arequipa, and secure public support, it i s essential that the affected communities in any proposed concession be involved from day one. Efforts to involve the affected communities were seldom done in the first phase o f private sector participation in Peru. The process was mostly top-down, and it needs to be bottom up, with strong buy-in from directly affected stakeholders. ProInversion i s already moving in that direction. The level o f involvement needs to go beyond the signing o f a letter of agreement or understanding between ProInversih and a sub-national government. Representatives o f the communities, affected user groups, and the sub-national government ought to be part o f the process and decision-making. This i s exceedingly important, especially for the possible water sector concessions in the government pipeline. Involving citizens in the process makes providers more accountable. This i s the so-called "short route" o f accountability-citizens exercise their client power over providers-laid out in the 2004 World Development Report (WDR) (see Figure 53). In contrast, the long route i s where citizens influence politicians and policymakers through the political process, depending on their level o f voice. In turn, policymakers influence service providers through implicit or explicit contracts or compacts. Hence, involving citizens and clients directly from the start can make providers more responsive. Figure 53: Accountability in infrastructure services Source: World Bank (2004b) The 2004 WDR also sets out a useful framework for analyzing and understanding the role and influence o f different interests in service provision. Accountability i s a set o f relationships among service delivery actors with five features: 0 Delegating. Explicit or implicit understanding that a service (or goods embodying the service) will be supplied. 0 Financing. Providing the resources to enable the service to be provided or paying for it. 122 Performing. Supplyingthe actual service. a Having information about performance. Obtaining relevant information and evaluatingperformance against expectations and formal or informal norms. a Enforcing. Being able to impose sanctions for inappropriate performance or provide rewards when performance i s appropriate. When appropriate accountability mechanisms are in place, community action can have a big impact, as was the case inpostwar Japan (see Box 12). Community consultation should always be at the center o f infrastructure projects, especially ones that bring significant costs to communities that do not enjoy many o f the benefits. Power generators, sewage treatment plants, and dams often have to be built in somebody's backyard. Provision for adequate compensation to affected communities i s a generally endorsed principle for all infrastructure projects o f this nature. But ensuring that all costs are compensated, and risks mitigated, is difficult even when strong accountability mechanisms are in place. In such cases, official lenders and donors may also have an important role to play. The mechanisms established around the Nam Theun 2 dam project in Lao PDR are examples o f how official lenders and donors have attempted to mitigate risks (see Box 13). Box 12: Community action in rolling back environmentalexternalitiesin postwar Japan In the years immediately following World War 11, Japan embarked on a period o f extremely rapid industrial growth, which, while highly successful in conventional economic terms, brought with it major environmental externalities. The consequences of air and water pollution were exemplified in the 1960s by the well- documented cases o f Yokkaichi Asthma and Minamata Disease. The Japanese response to such environmental problems originated at the municipal and local levels, where citizens exerted pressure on elected officials to take measures against offending industrial enterprises. A series o f voluntary agreements were made between industrial enterprises and local governments, which in many cases did not have any explicit environmental jurisdiction or responsibility. National legislation tended to follow later, and national standards for air and water quality invariably remain lower than those contained in the approximately 40,000 voluntary agreements now inplace inJapan. As far as domestic air and water quality is concerned, there is universal agreement that Japan has been an exemplary case, and a number o f factors are widely cited as reasons for this. These include a free press that publicized environmental issues; universal literacy, with a strong emphasis on technical education, which enabled citizens to understand the impact o f environmental degradation on their own health and well-being; and a democratic system in which local officials were compelled to take citizens' complaints seriously, yet were powerful and efficient enough to address complex technical issues and take measures to address pollution. Source: Aoyama (1994) and ADB, JBIC, and World Bank (2005). 123 Box 13: Puttingaccountability mechanismsinto large infrastructureprojects: The NamTheun 2 dam N a m Theun 2 (NT2) is a major hydropower project currently under consideration inthe Lao PDR. The project would entail an investment o f $1.2 billion to be undertaken by the private sector. Its goal is to generate more than 1,000mW o f electricity, almost all o f which would be exported to neighboring Thailand. But the project comes with a number o f risks arising from the structure o f the investment and the accountability environment. One risk i s the highly unequal allocation o f costs and benefits-revenues go to the national government, while costs are borne by communities and the environment in the project area. Another risk i s that the substantial benefits accruing to the government will not be effectively translated into benefits for the country due to governance weaknesses. The design o f N T 2 attempts to address these issues in a number o f overlapping ways: first, in assigning joint responsibility to the concessionaire and the government-through provisions written into the concession agreement-to finance and implement measures to compensate for impacts and costs suffered in the project area; second, in making changes to the project design to mitigate the risk o f negative externalities, where feasible; third, b y bringing international financial institutions into the financing o f the project, effectively making continued donor support conditional on the government's commitment to hlfilling its responsibilities in terms o f the concession; and fourth, through the establishment of mechanisms to earmark project revenues for poverty reduction, while improving the public finance management system through which they-and all expenditures-must be delivered. Active participation o f the international community has helped encourage local public consultation and disclosure to ensure that affected people are hlly informed o f the project and that their views are taken into account. More than 200 consultations and workshops were conducted in project-affected areas to ensure that affected people would participate in the measures to protect them from the impact o f the project. A total of $89 million has been designated as capital and operating expenditures for environmental and social mitigation and compensation. These obligations are defined, and costed in the concession agreement, which has been signed by the government and the private sector concessionaires. Mechanisms have also been developed to address weak accountability arrangements in the public finance management system, in particular, to facilitate more effective and transparent targeting o f N T 2 revenues toward poverty reduction goals. Proposed revenue and expenditure management arrangements will focus on the development o f the government's core public expenditure management system and will avoid building parallel systems through which to channel N T 2 revenues. Under this approach, N T 2 revenues will be channeled through the central treasury account to finance expenditures on eligible poverty reduction and conservation programs. These programs will be required to meet management and reporting standards. Expenditures and program impact will be monitored through a program o f expenditure reviews and expenditure tracking surveys. In this way, the implementation o f N T 2 revenue and expenditure management arrangements will support broader public expenditure management reforms. Source: World Bank staff and ADB, JBIC, and World Bank (2005). Civil society"' Civil society can play an important role in accountability o f infrastructure institutions through parliaments or through consumer participation in regulation. Civil society organizations and NGOs can provide small-scale infrastructure services, act as watchdogs against corruption and vested interest, and play an advocacy role for more sustainable infrastructure policies and services. Advocacy NGOs face difficult choices between This section draws on ADB, JBIC, and World Bank (2005). 124 representing the interests o f specific groups or issues and representing the interests o f society at large. H o w effectively and accountably they make those choices can have a significant impact on development outcomes. Decentralization i s often a way to increase accountability and the responsiveness o f infrastructure service provision to local needs. However, decentralization poses a number o f coordination challenges, both vertically (between central and local governments) and horizontally (among various subnational institutions). Decentralized governments have sometimes been isolated within their own jurisdiction. This is problematic because most network infrastructure has inter-jurisdictional backbones. Isolation can mean secondary or tertiary infrastructure lacks connections to primary infrastructure-in a sense, it goes nowhere. Some municipalities may be too small to achieve the scale necessary to deliver infrastructure efficiently. In competing with each other, municipalities may duplicate expensive infrastructure facilities, when such facilities, in fact, could have been shared. Avoiding these pitfalls depends critically on inter- jurisdictional cooperation-on filling inthe missingmiddle. Higher tiers o f government need to encourage lower tiers to collaborate where primary infrastructure requires such collaboration. Matching grants to induce decentralized governments to participate in such investments, and institutional mechanisms to encourage cooperation in infrastructure planning will play a major role. Central governments also have to ensure that they maintain sufficient capacity to monitor, manage, and coordinate in a manner that i s in line with policy and regulatory frameworks. The inadequacy o f such systems i s a frequent cause o f suboptimal service delivery and confused authority. 5.1.4 Accountingfor thepoor and the losers'22 Improvements in infrastructure services have great potential to benefit the poor (see Table 50). However, for the poor to realize these benefits, specific attention needs to be paid to their needs and vulnerabilities throughout the private participation process. This i s necessary for both equity reasons and to improve public support for concessions. Special consideration also needs to be devoted to those negatively affected by private participation, such as workers who could lose their jobs. First, an assessment o f how concessions would affect disadvantaged groups should be conducted. Based on that assessment, a strategy should be developed to accommodate disadvantaged groups, including the poor and laid-off workers. Insufficient attention to this was the cause o f the so-called backlash in many countries. ''IThis sectiondraws on ADB, JBIC, and World Bank (2005). '22This sectiondraws on World Bank (2006) and Estache, Foster, and Wodon (2002). 125 Table 50: Potentialpositiveimpactsof infrastructureservices on the poor Sector Direct Impactson the Poor Indirect Impactson the Poor Mainly for lighting,TV, radioat low Reducedenergy costs for enterprises levels o f income. encouragingemployment creationacross Electricity a wide range o f activities. Heating, cooking, appliances for self- Improvedhealthand other services employmentat higher levels o f (refrigeration,lighting,etc.) income. ImprovesICT access. Limitedimpact at low income levels. Reducedenergy costs for enterprises Pipedgas Heating, cooking at higher levels of encouragingemployment creation income. (limited range of activities). Reducedtransport costs and improved Access to employment and markets. market access for enterprises and service Roads Access to services(health, providers, loweringcosts of serving education). remote communities; general economic development. Reducedcosts and improvedmarket Railways Limited access for enterprises; general economic development. Urban Mass Transit Access to employment opportunities. Employmentcreation from more efficient labor markets. Reducedtransport costs for enterprises Ports Limited (e.g.bulk commodities like agriculture), expansion of business opportunities; general economic development. Reducedtransport costs for enterprises (highvalue, low bulk commodities, and Airports Limited services), expansionof business opportunities; general economic development. Better communicationaccess, aiding Informationand migration, informationon Employment creationthrough improved Communications opportunities, access to knowledge knowledgeof markets,reduced Technology (ICT) andpotentialengagement inwider management supervisioncosts, access to communities. wider knowledgebase. Water Improvedhealthoutcomes, time savings. Limited Improvedhealthoutcomes (e.g. reduced Sanitation Improvedhealthoutcomes. pollution by non-poor households and others). To expand access to the poor, policymakers can set service obligations or create incentives for providers. One way i s to specify universal service obligations, as i s common in the telecommunications sector. While this i s a worthy social objective, it may not be practical inthe short runwhen starting from low access rates. That is why service obligations should include details on time frames and how the obligation i s to be financed when customers are unable to pay. Defining connection targets i s another way to promote access. Targets are easy to monitor and can be enforced by financial penalties. Of course, connection targets can be met only if customers are able and willing to take up the service-and this depends on their ability to overcome impediments including title requirements and income and liquidity constraints (given the lumpiness of connection charges). 126 Infrastructure subsidies are another option and they can be justified not only by poverty reduction, but also by environmental protection. Although they would enjoy the environmental benefits, people often won't pay the full cost o f sanitation, mass rapid transit, or renewable energy. In cases in which those benefits are external to consumers, subsidies may be needed to realize the benefits. Clean water or rural roads may have an important impact on poverty, but they may not be affordable by the poor. Such projects may require subsidies. And reform programs that help the poor or the environment may not be politically sustainable without subsidies for those with the power to derail the reforms. Similarly, transitional subsidies sometimes may be worth considering during short periods o f economic crisis. But subsidies can become open-ended and addictive, their fiscal impact can explode, they can undermine financial discipline and blur accountability, and they can postpone much- needed reform. Subsidies need to be employed with great care. Subsidies should be a last resort after costs have been minimized through competition, regulation, appropriate technology and service standards, or public enterprise reform. Subsidies can be minimized through transparency, making them contingent on performance, or through subsidy bidding processes. In many countries, new connections are subsidized to meet access objectives and keep providers solvent. New connections can be subsidized from charges to existing users, particularly if the group o f existing users i s much larger and wealthier than potential new users. The water and sanitation concessionaire in Buenos Aires adopted this type o f cross- subsidy after renegotiating an initial contract that charged onerous connection fees to the poor. Government financing for connection subsidies i s also an option, as i s offering credit to consumers for connection purchases. In Colombia the law requires that connection charges for poor customers be spread over at least three years. Although connection subsidies can be promising, especially when coverage i s low, their targeting performance depends on whether they actually induce households to connect to the system, an issue little studied to date. Also, if subsidies are not covered by cross-subsidies, the sustainability o f funding needs to be taken into account. A complement to connection subsidies are consumption subsidies, either through means- tested transfers financed out o f general tax revenues or through lifeline tariffs, which require a transfer from those with high levels o f consumption to those with low levels.123 When considering lifeline subsidies, care must be taken to set a threshold that is high enough to gamer political support, yet low enough that the poor are the primary beneficiaries. For instance, evidence from Honduras suggests that their electricity subsidy i s too high--83.5 percent o f residential customers benefit from the subsidy (those consuming under 300 kWh monthly). Given that lifeline tariffs often have poor performance targeting, alternatives include geographical targeting that attempts to identify low income areas, and means-testing that evaluates the economic situation o f individual households. Means-tested vouchers for purchasing services are another subsidy option. They are similar to means-tested tariff subsidies with added flexibility for the user to select "Lifeline" tariffs mean that rates increase after surpassing a certain consumption threshold (the level deemed necessary to meet basic household needs). 127 a service provider. (Wodon, Ajwad, and Siaens (2005), Komives and others (2005), and Estache, Foster, and Wodon (2002)) Given the liquidity constraints o f the poor and the possible seasonality o f both use and income, introducing flexibility in payment i s likely to help expand access. Increasing the frequency o f billing i s one option. Prepayment devices, which facilitate budgeting for low- income households, are another. On the downside, prepayment could lead to frequent "self- disconnection." Utilities could also allow customers to choose fiom a menu o f tariffs with different combinations o f fixed (standing) and variable charges. Allowing lower fixed charges in exchange for slightly higher variable ones could benefit small-scale consumers. Enabling consumers to make certain quality-price tradeoffs by encouraging lower quality services i s also likely to be beneficial. This can be done by allowing lower standards o f formal provisioning in certain poor areas or by encouraging a vibrant network o f informal providers that can either operate independently or through subcontracting arrangements with the formal provider or concessionaire. Water provides an example. A study using data from 47 countries shows that informal providers such as point-source vendors (kiosks) and mobile distributors (such as tanker trucks and carters) systematically charge more than networked providers, both public and private (Kariuki and Schwartz (2005)). But informal providers can offer a valuable service because many poor users cannot afford connection charges or monthly lump-sumbills or live in areas inaccessible to utilities for legal or technical reasons. Recognizing that private connections for all households may not be a feasible goal in the near-term, the government and concessionaire could work with informal providers. For instance, kiosk services could be improved by subsidizing kiosk connections, increasing competition between kiosks, and introducing performance measures and quality standards (Gulyani, Talukdar, and Kariuki (2005)). Given that many o f Peru's upcoming concessions are not profitable, there will likely be pressure to move tariffs toward cost-recovery levels to make the concession more ' financially viable. Hence, measures should be taken to mitigate the impact o f any such price adjustments on the poor. Besides the subsidies mentioned above, phasing-in periods can reduce the shock associated with tariff adjustments. Price increases associated with private involvement have also hurt public perceptions. For that reason, tariff adjustments could take place well in advance o f the concession and it could be made clear to the public that such adjustments were necessary to reduce costly subsidies and increase resources for the truly needy. When a concession might lead to possible employment losses, measures to compensate the affected and to re-train them for reinsertion in the labor market ought to be considered. Examples include on-site tutorials on employment, micro-enterprise, or self-employment opportunities. In summary, it i s essential to have a strategy and an implementation plan to account for disadvantaged groups. 128 5.2 IMPROVINGCONCESSIONCONTRACT DESIGN Sector performance, level o f conflicts, and public perceptions have been influenced by weak and deficient design o f concession contracts. Those deficiencies have been usually due to a hurried process, vested interests, and questionable advice combined with the limited experience and resources o f governments in designing concessions. Though many mistakes have been made with concessions in the past, the experience o f more than 15 years makes it easy to avoid such mistakes in the future. Improvements over past designs should be a cornerstone o f the moving forward program in Peru. W e draw on that experience, and in doing so, offer a blueprint for concession designs that should significantly improve prospects for all concessions, reduce the incidence o f renegotiation, and improve sector and economic performance. Before a concession contract can be designed, i t i s first necessary to decide on the appropriate grouping o f public assets. For instance, when considering private participation in airports in Mexico, the government weighed three different types of groupings: individual airports, groups o f airports, and a single network o f airports. Box 14 describes some o f the concerns with each structure and why Mexico divided its airports into four groups. Mexico's experience i s generally considered a model for the region. Box 14: Grouping of airports for private participation in Mexico Mexico was considering the following three types o f groupings for private participation in airports. 1. Independent airports 2. Groups o f airports 3. A single network . . . . ' The independent airport scheme was not desirable for the following reasons: It did not encourage the participation of quality investors and operators and did not insure the participation o f private capital in all o f the profitable airports; The regulation o f a large field o f operators implied growing costs and difficulty in establishing general criteria for administration; Any economies o f scale would be lost; From the perspective o f the users (inparticular the airlines) and the service providers, a large number o f operators would make it difficult to adequately manage commercial and legal relations. The single network scheme also had significant shortcomings, including: Problems with legal interpretations; No guarantee that the public interest would be protected in terms o f social, economic, and national security concerns; Required strong regulation and did not offer sufficient elements to compare performance between airports; Difficult for the market to assimilate and difficult to sell politically. 129 Inthe end, the group of airports scheme fulfilled the most objectives of the authorities as well as the interests o f other key stakeholders. Four regional airport groups were formed with the following motivations: Promote operating efficiency at the global and regional level; Offer sufficient elements to compare performance between airports; Permit ties between airports with different specialties and potential for development; Establish an adequate balance between national and international passengers; Take into account expectations o f industry participants, including operators, airlines, service providers, users, and investors; Include all o f the airports that did not require operating subsidies; Benefit from administrative efficiencies in a regional grouping; Take advantage o f efficiencies from better coordination in marketing and regional promotions, e.g. the Ruta Maya. Source: Garcia-Lopez Loaeza (2005) 5.2.1 Concession awarding process Direct adjudication and bilateral negotiation to award concession contracts should be avoided except in exceptional circumstances-such as when there i s only one candidate for a concession. The norm should be to award concessions through competitive processes, meaning auctions undertaken inthe most transparent possible manner. A variety o f criteria-the competitive factors that interested operators bid on-have been used to allocate concessions in auctions settings. The most common are minimum tariff, minimum duration o f concession, minimum subsidy, maximum amount offered for the rights to operate the concession, largest investment value, minimum total revenue, largest number o f retained workers, and "best" overall proposal. Often it i s a combination o f any o f those mentioned. The most common criteria, particularly in the water sector and toll roads, have been awarding the concession on the basis o f either the lowest tariff or a points system combining a technical evaluation and the proposed tariff. Inprinciple the use of multiple criteria-even with a well specified scoring formula-is not desirable since it tends to lack transparency and it i s very susceptible to manipulation, corruption, and the contesting-by losers-of the award, inducing delays and protracted conflict. The use o f a single criterion should be the norm. While not all criteria are equally desirable, the chosen competitive factor can have a large impact on the success o f the concession. Although there i s room for adaptation to specific settings and needs, there are general principles that should rule the choice. The criterion should be sufficiently robust to account for the intrinsic uncertainty o f a concession which i s granted for a very long period. It i s often senseless to use a criterion that i s likely to be modified in the near future. The points system combining a variety o f factors and technical evaluation i s deficient and should be avoided. Technical proposals are almost useless for any contract that i s going to last for 25 to 30 years and are subject to manipulation and arbitrary decisions since the evaluation can be highly subjective. No concessionaire i s likely to follow its original technical proposal for more than six months, and, indeed, no government should require it. 130 An adequate process should take into account the following elements: There should be two stages: one to prequalify interested parties on the basis o f experiences and the technical proposal if applicable and a second one to solicit bids from the prequalifiedbidders usinga single criterion for selection. 0 Prequalification should set up eligibility procedures with strong technical and financial criteria (see examples in Table 51). The winning bid should be selected solely on the basis o f the financial proposals (single criterion, highest canon, or fee) submitted by bidders who have met the technical qualifications, which may include a requirement to submit a technical proposal for meeting the requirements o f the contract. The role and potential impact o f financial advisers and investment banks hired to implement concession transactions cannot be downplayed. Often they are the ones designing or advising on the concession. Careful attention must be paid to the incentives in their contract arrangements to take concessions to the point o f sale. Their success (which i s linked to their fees or commissions) can be measured in a number o f ways, including the number o f prequalifying and bidding firms for the contract, the winning amount (or minimumtariff) for the contract, and the amount of committed investment secured in the contract. The adviser response to those objectives i s often to make the contract and terms as attractive to investors as possible, minimizing their risks and reducing penalties and sanctions, even if it i s to the detriment o f long term efficiency and sector performance or if it i s likely to induce subsequent conflicts. Those objectives may or may not be stated by the government or grantee o f the concession. In any event, the government should be aware o f the implications o f those objectives and o f the incentives o f advisers to maximize success. 131 Table 51: Examples of Prequalification Criteria in Private Infrastructure Concession\Transactions Sector Country Transaction Prequalificationprocedure TechnicalCriteria FinancialCriteria Electricity Peru Lima electricity Qualificationat time of bidding; Customersand energy Minimum total value distribution bidders must exceed a score of sales per worker, total o f assets and net worth privatization 80 percent against six weighted customers,and energy quantitative technical and sales financial criteria Argentina Electricity A guarantee to cany out the Consortia to include Minimum asset value distribution bidding processwas required of qualified operator with of bidding companies; concessions bidders at the time of minimum experience proven increaseof at prequalification and ownership in least I O percent in asset consortium value in three years grior to bidding Transport Mexico Concessioning Registration through written Demonstratedlegal. Demonstrated financial of rail freight statement of interest; technical, and capacity lines authorizationo f registered administrative capacity parties by the Ministry of Communications and Transport basedon unifonn criteria Hungary BOT for toll Invitations for pre-qualification Capacity of bidders to Capacity of bidders to road basedon approved preliminary design, build, maintain. finance road without design plans evaluatedby expert and operatetoll road state aid assessment committee Water Argentina Buenos Aires $30,000 fee for prequalification Minimum population of Minimum requirements concession documents largest city and for total annual billing aggregate population and net share capital; served by bidder consortium shareholding distribution regime Bolivia La Paz Qualification processto take Consortia must include Minimum net worth concession place at saine time as economic water operator with and maximum debt-to- bids presented minimum experience equity ratio o f operator and extent of service Natural Mexico Concessioning Registration of interested Documentation of Documentation of gas of distribution bidders and meetings between technical and financial capacity regulator and prospective administrative capacity bidders to clarify information prior to technical bids; small registration fee Source: Ker t al 1998. There are three criteria that have generally been used to award concessions after prequalification has taken place (second phase). Minimum tariffs have frequently been used, but they have some flaws. The salient choices for concession award criteria, based on efficiency, incentives, and effectiveness against renegotiation, should be either an annuity payment-canon or minimum subsidy for unprofitable concessions-or the least present value o f revenues (LPVR), when appropriate. The following sections describe these three criteria in greater detail. Minimum tar@? Minimumtariffs, while in principle appealing, are "soft" anchors for concession awarding, and thus only used with special care and conditions. They are vulnerable since it i s a parameter that, at least every so often, automatically appears at the table for modifications and review, even in the best o f circumstances, and at that opportunity ifnot before, it can be subject to modifications, compensation, and rent extraction. Tariff bids have the major disadvantage that the winning tariff will almost always be less than the long run marginal 132 cost o f providing the service and are likely to be changed very quickly-mostly through renegotiation or review. This can only be avoided if the concession specifies a lease payment for existing assets that i s carefully calculated to reflect their value under the concession. In practice, this is all but impossible to achieve. As a consequence, some or most investment required to serve additional customers, will be unprofitable for the concessionaire, so that there may be constant problems in ensuring the concessionaire meets its service targets. There i s also a strong likelihood that one or more parties may bid a rather low tariff as a "loss leader," with the objective o f securing the concession and recouping any short term losses by re-negotiating a tariff increase at the first possible opportunity. This strategy will be particularly likely ifthe concessionaire i s able to move from an initial tariff to the new one immediately without requiring a phased adjustment averaged over several years. Also, minimumtariff criteria have little "lock-in" effect. Since winning firms do not pay anything at the transfer o f the concession, they have little to lose if they were to walk away form the concession, if, say, petitions for renegotiation o f contract were to be denied by the government. Maximum Transfer Fee /Minimum Subsidy During the second stage (after prequalification), the awarding criteria should be based on the maximum amount (or minimum subsidy) for the rights to operate the concession, having established a duration commensurate with the life o f the assets (20-30 years), tariff structure (maximum, price caps, rate o f return), and regulatory regime. The operator submitting the highest amount (or minimum subsidy in the event o f unprofitable concessions) obtains the concession. That amount o f money i s placed into an interest accumulating trust, which could be invested in risk free assets, to be disbursed to the government in annual amounts throughout the life o f the concession. While the first payment could be larger than the rest, it should not exceed 15 percent o f the total amount. The government could borrow against these payments, but only for the disbursements matching its elected period. This scheme has three attractive features. First, it uses a non-soft criterion to award the concession, which i s more difficult to renegotiate and i s consistent with the transfer for the use o f the existing assets and the right to operate the concession. Second, by forcing the operator to pay a single amount up front, it generates a lock-in effect, increases the commitment o f the operator, and grants increased leverage to the government in the event o f the operator's performance noncompliance. Third, it provides for increased ownership (of the concession decision) for future governments, because they will also benefit from annual payments. This component has a drawback, compared with a scheme where an operator makes annual payments, in the sense that it i s financially more expensive, because the amounts o f funds needed at the start are clearly greater than in a scheme with a simple canon, paid annually 133 by the operator. However, arguably, the benefits o f increased commitment are bound to be greater than these increased financial costs. Minimum Present Value of Revenues An interesting alternative developed by Engel, Fischer, and Galetovic (1998, 2000) is the use o f the minimum present value o f total revenue (LPVR) as the criteria to allocate certain types o f concessions. Interested operators bid on the present value o f total revenue to be received and the one submitting the lowest value gains the concession. Once that present value o f total revenues i s received the concession ends. The government sets up maximum tariffs and a discount rate that can be fixed or variable. This scheme has multiple advantages. Perhaps the most important is that it preempts renegotiation and rent seeking by opportunistic operators. This i s important because renegotiation o f concession contracts i s pervasive in Latin America. Almost 50 percent o f all concession contracts are renegotiated within three years o f the award o f the concession, and at least half o f the operators are opportunistic to the detriment o f the users o f the service. Another advantage i s the automatic compensation the operator received if factors such as demand and tariffs adversely affect revenues. The operator can then run the concession for additional years until the agreed present value o f revenues i s secured. Thus, it preempts needs and request for demand or traffic guarantees, with its fiscal implications. In addition, it provides clear and transparent compensation in the event o f the need to terminate the contract, which i s the residual value o f the LPVR. In the event o f a significant change, say a request from the government for more investment, that amount can be tagged to the original LPVR bid as the appropriate compensation. Other advantages are that it provides incentives to operate at optimal costs because any gains are fully captured by the operator, weakens incentives to submit frivolous offers and lessens the need to evaluate and forecast demand, thus reducing preparation o f bid costs. In addition, it facilitates oversight by regulators and considerably limits regulator's discretion, because oversight i s basically limited to assessing and accounting for the flow o f revenues. It also lessens incentives to request non-ordinary tariff increases, and, in the event o f cancellation o f the concession, it facilitates settlement structure because the salient parameter for compensation i s the remainder (yet uncollected) o f the present value o f revenues. Perhaps the main disadvantages are first that the length o f the concession i s uncertain, which can impact financing and second, agreement on the proper discount rate. When to Select the Transfer Fee or the LPVR Criteria to Award Concessions Which one o f these two should be selected depends on the characteristics o f the concession. The key characteristic i s whether the quality o f service provided by the concessionaire has a strong effect on demand or not. Because the LPVR in fact insures the operator against changes in demand, it i s only appropriate in settings where the concessionaire can do little to influence demand, and where objective quality standards can be set, measured, and enforced. Typical settings then are roads and highways, landing strips in airports, water reservoirs, and so on. Chile has used the LPVR scheme to award a highway concession. Alternatively, in settings where the operator through the provision and quality o f service 134 can significantly influence demand, LPVR criteria are not appropriate, and the annuity canon should be used. Typical settings for the latter are water and sanitation concessions, port concessions, network concessions, and so on. 5.2.2 "Financial equilibrium" clausesfor the operation of the concession Quite often concession contracts or regulatory frameworks, when ordinary or extraordinary tariff reviews are called for, state or make reference to the principle o f adjusting variables so as to secure the financial equilibrium o f the concession, usually in a forward looking manner.'24While in principle this i s appropriate and consistent with the spirit o f regulation, care should be exercised in how it i s stated. Broad and sweeping statements without reference points are undesirable and often have been the source o f conflicts and inefficiencies. Such clauses should not guarantee the financial equilibrium without making reference to efficient operation and preserving, the sanctity o f the bid. The risk o f an opportunistic bid should be hlly borne by the operator. Similarly, financial equilibrium clauses should specify the capital base that the firm i s allowed to earn a fair return on. The capital base should not necessarily include the transfer fee when the concession was awarded by that criterion. This i s often a key argument for renegotiation and also for the tariff review process. If the whole transfer fee i s allowed to enter into the capital base, it distorts the competitive bidding and reduces the value to the country. It then becomes more like a loan to the country-to be repaid through higher tariffs-than a purchase price indicative o f superior efficiency. In other words, the firm would be allowed to earn a fair rate o f return on any amount paid, which i s not the essence o f competitive bidding. The capital base should be linked to the book value o f the assets, rather than the transfer fee. Similarly, accumulated profits should not be allowed to be part o f the capital base, a common problem in Latin America, where it has been allowed. However, when they are invested under the appropriate guidelines, then they could enter into the capital base. Another element that needs to be very clearly stated in the financial equilibrium clause o f the contract i s the period o f application. This refers to the period o f time over which the financial equilibrium i s evaluated. In principle, i t could range from one year to the life o f the concession. Both o f those extremes are inappropriate-a three to five year period i s more suitable. If that period i s not clearly stated operators will choose the shortest period when the financial results have been deficient, and the longest period when the financial results have been very good. The choice o f the relevant period has been a source o f conflict where it was not properly specified. Finally, the principle o f financial equilibrium should be an ex-ante consideration, not an ex-post market outcome, in the sense that it should not bail the operator out for adverse outcomes from normal commercial risk. 5.2.3 Renegotiation clauses and triggersfor renegotiation 124Financial equilibrium refers to the ability of private investors to earn a fair return on their investment, given that regulation constrains the actions that concessionaire's can take (such as setting tariffs). 135 The intrinsic nature o f concessions contracts in infrastructure means that any contract could be incomplete. And renegotiation can be an efficient-albeit second best-instrument to address that issue, so it should not be ruled out on principle. But it can be framed so as to dissuade frivolous claims and support the valid ones. The concession contract should address as clearly as possible: i)events that would trigger tariff adjustments and the extent o f the adjustments; and ii)events that would trigger a renegotiation o f the contract with guidelines about the process and outcomes o f the renegotiation. The principle i s that small, exogenous changes that impact the financial equilibrium of the firm should not require adjustments, while large ones may. It i s also advisable to require that renegotiations be undertaken in the most transparent, open manner as possible, preferably by appointing a neutral and professional commission to review the claim and advise on the outcome. Sanctity of the bid contract. When facing petitions for renegotiation, the sanctity o f the bid contract must be upheld. The operator should be held accountable for its submitted bid. The financial equation set by the winning bid should always be the reference point, and the financial equilibrium behind that bid should be restored in the event o f renegotiation or adjustment. Renegotiation should not be used to correct for mistakes in bidding or for overly risky or aggressive bids. This i s another reason for the superiority and desirability o f transfer fees over minimumtariffs as award criteria for concession. 5.2.4 Penalties and incentives Performance bonds and penalties should be adequately specified in the contract itself or in the regulation framework and penalize non-compliance-such as not meeting targets, failure to fulfill investment, retribution or quality obligations-with agreed clauses. Sanctions in concession contracts should depend on the magnitude o f the fault or non- compliance and on the extent o f reoccurrence. And clearly, termination should be a possibility in all cases as a last resort mechanism in cases o f serious and reoccurring non- compliance. Economic sanctions stipulated in the legislation should link the magnitude o f the sanctions to the annual income o f the concessionaire in its operation. And the magnitude should be sufficient to have deterrent power. For example, an initial performance bond for a concession contract should not be less than (a) 2 percent o f the total value o f the contract and (b) 20 percent o f the estimated annual revenue o f the concession inits first year inorder to provide a strong enough incentive (deterrent). 5.2.5 Concessionlength andflnancing The combination o f long-lived assets and a high degree o f specificity has particularly important consequences for the length o f the concession contract. Ifthe residual value o f an asset at the end of a concession period i s highly uncertain, concessionaires will tend to write off any assets acquired during the concession over the remaining term. If the term i s short and concessionaires have to fund investment in those long-lived assets, they might demand substantial subsidies in exchange for the guarantee o f a given level o f service. The effect o f a shorter contract can be mitigated if assets are acquired at less than the replacement cost (that is, used assets) and if the government promises to repurchase them at 136 a fair market value should the concession be lost to the investor. Nevertheless, this need not mean that short-term concessioning i s not viable or that it i s less attractive than long-term concessioning. In fact, each approach has advantages and disadvantages. Short-term concessioning allows inore frequent competition and therefore maximizes the incentive to increase efficiency. An example o f short concessions i s the distribution and transmission in the power sector in Argentina. The regulatory agency usually fixes the tariffs for five years. At the end o f this period, an international bid i s called for the sale o f the control package o f the concessionaire. The incumbent puts his "reserve price" into a sealed envelope. Bids are accepted from all interested parties. If the incumbent i s the higher bidder he keeps the concession with no payment. It the incumbent i s out-bid he receives the highest bid (net of any debt to the state) and the bidder gets the concession. Shorter term concession contracts coupled with competitive rollovers at the end o f a contract can be a powerful efficiency inducing device, as long as the firm i s compensated for incurred investments. Long-term concessioning not only minimizes some o f this incentive but also fosters a relationship that i s more akin to that o f regulator and regulated than a true business contract. Long-term concessioning, however, also maximizes the opportunities for shifting responsibility to the private sector. It encourages more innovation and cost-efficiency than a short-term contract. Which o f these factors will weigh more heavily with policymakers will d,epend on the specific local circumstances. For example, on a network where the assets have been recently renewed, the attraction o f frequent competition might outweigh the benefits o f long-term concessioning. In parts o f the network where track and rolling stock are in urgent need o f renewal, a long concession period mightbe attractive. 5.2.6 Investment commitments Governments have found it politically attractive to require investment amounts in contracts as an indication o f success. But investments commitments are an indirect means to improve sector performance, increase coverage, and attain related objectives. It i s more efficient to state directly the outcomes those investments ought to produce, e.g. a new water treatment plant, access road, etc., when applicable. Outcomes are usually easy to measure and non- controversial. They should have clear technical and quality specifications or specific indicators, such as coverage rates, quality standards, or technical achievements. A timing schedule should accompany those specifications and include appropriate increases in the targets over the life o f the concession. T o link outcomes and investments, the concessionaire should be required to produce an action plan (after the concession has been awarded) regarding how much investment will be required to meet the scheduled targets. However, such a plan i s only indicative and non-binding, essentially to ensure consistency and feasibility in achieving the targets over time. 5.2.7 Better allocation of concession risks A key element o f concession design i s the identification o f risks and their proper allocation. Since concession design aims at establishing financial equilibrium for the concessionaire, inadequately assigned risk would raise both the cost o f capital and tariff levels. The 137 experiences reviewed in Peru provide some positive and negative examples o f risk allocation. The government adequately ensured the stability o f contracts through contract- law and Stability Agreements and freed concessionaires fkom exogenous risks, including major force risks, in most cases. However, there could potentially be problems classifying and interpreting major force risks in certain industries as they are based on vague definitions. The two principles guiding risk allocation are: i)the party that i s responsible or has more control over the risk factor should bear the risk and ii)the party that i s more able to bear the risk (less risk averse) should be assigned the risk.Under those criteria the major risks and their allocation in a concession i s shown in Table 52. 138 What is the Risk? How Does it Arise? How Should it be Allocated? Desigddevelopment ri Design defect Design fault in tender specifications Public sector to bear risk Contractor design fault Liquidated damages to be paid by contractor; once liquidated damages are exhausted, erosion of project company's returns Construction risk Cost overrun Within constructionconsortium's control Contractorto bear risk through fixed-price (inefficient construction practices, and so on) constructioncontract plus liquidated damages; once liquidated damages are exhausted, erosion of project company's returns Outsideconstruction consortium's control: Insurer risk if insurance is available; once insurance changes in the overall legal framework proceeds are exhausted, erosion of project company's (changesof laws, increased taxes, and so on) returns Outside construction consortium's control: Public sector to bear risk actions of government that specifically affect the project (delays in obtaining approvals or permits, and so on) Delay in completion Within construction consortium's control (lack Liquidated damages to be paid by constructor; once of coordination of subcontractors, and so on) liquidated damages are exhausted, erosion of project company's returns Outside construction consortium's control Insurerrisk, if risk was insured; once insurance (force majeur, and so on) proceeds are exhausted, erosion of project company's Failure of project to Quality shortfall, defects in construction, and so meet performance on liquidated damages are exhausted, erosion of project criteria at completion Operating cost risk Operating cost Change in practice of operator at project Project company to bear risk overruns company's request Operator failure Liquidated damages to be paid by operator to the project company; once liquidated damages are exhausted, erosion of project company's returns Failure or delay in Public sector discretion Public authorities to bear risk obtaining permissions, consents, and approvals Changes in prices of Increasedprices Allocation of risk to the party best able to control, supplies manage, or bear it (supplier, project company, or users) Nondelivery of Public sector failure Public authorities to bear risk supplies on the part of public authorities Revenue risk Changes in tariffs In accordance with the terms of the contract Project company to bear risk (for example, indexation of tariffs leads to 1 reduced demand) Government breach of the terms of the contract Public sector to bear risk Changes in demand Decreaseddemand Shortfall in quantity, Operator's fault or shortfall in quality liquidated damages are exhausted, erosion of project leading to reduced company's returns demand 139 1What is the Risk? How Does it Arise? How Should it be Allocated? I Project company's fault Liquidated damages to be paid by the project :ompany to public authority Financial Risk Exchangerates; Devaluation of local currency; fluctuations Project company to bear risk (hedging facilities might interest rates be put in place) Foreign exchange Nonconvertibility or nontransferability Public sector to bear risk; in case of contract termination, compensationto be paid by government Force majeure risk Acts of God Floods,earthquakes, riots, strikes, and so on Insurer risk, if risk was insured; otherwise, risk to be Changes in law Changes in general legal framework (taxes, environmental standards, and so on) could bear risk when changes are fundamental and completely unforeseeable; for example, switch from free market to central planning) Changes in legal or contractualframework Public sector to bear risk directly and specifically affecting the project company Breach or cancellation of contract; majeure expropriation, creeping expropriation, failure to borne by public sector; in case of contract obtain or renew approvals Operator's fault Liquidated damages to be paid by the operator; once incidents liquidated damages are exhausted, erosion of project l------ Pre-existing environmental liability I I Source: Kerf el a1 (1 998). 5.2.8 Addressing Termination of the Concession Any concession contract should address both how and under what conditions the concession can be terminated. It should also demand a performance bond to insure incentives for compliance and collection in the event o f non-performance. Inparticular, the contract should contain: 0 The circumstances under which the contract can be terminated and the compensation payable on termination, referring, as appropriate, to the provisions o f the law; 0 What happens at the end o f the concession period including (a) the possibility o f renewal o f the concession, (b) transitional arrangements if a new operator takes over the concession, and (c) the basis for calculating compensation for un-depreciated assets. 5.2.9 Arbitration rules and dispute resolution 140 Disputes between the operator and the regulator and government are standard and bound to arise under any contractual agreement. To reduce the regulatory risk associated with dispute resolution uncertainty, it i s advisable that an arbitration mechanism be used that i s perceived to be neutral and independent. Thus, the contract should contain provisions for arbitration and require that the rulings o f an arbitration be implemented on a provisional basis, with interim remedies in place even where they are appealed in the courts, since the appeal process can be quite long. The contract should contain clauses that establish dispute resolution arrangements prior to recourse to the courts, such as: Bindingarbitration may not be legally or politically acceptable, but it is possible for the two parties to agree in the contract to the appointment o f a panel o f experts at the beginning o f the concession to whom disputes may be referred by either party; The panel o f experts would normally be appointed by the Sector Agency and would include specialists in technical, economic, and legal aspects o f the sector; In disputes about technical issues the parties could agree to refer the questions to a single adjudicator rather than the panel in order to obtain a rapid decision; and Inthe event o f an appeal to the courts, the contract would state that the decision o f the panel o f experts should be implemented by the concessionaire on a temporary basis until a final court judgment. 5.3 ENHANCINGREGULATORYDESIGN The main objectives o f a regulatory framework are: i)to induce the regulated firm to operate at lowest (efficient) possible costs; and ii)to closely align prices (tariffs) with costs allowing the firm to earn only normal profits. Usually there are also other subordinate objectives that complement the main ones, such as increased coverage and access, improved quality o f service, and universal service obligations. T o effectively secure those objectives, the agency needs significant professional capacity and the appropriate information and regulatory instruments. 5.3.I Determiningfuture tarqfs The most common and appropriate structure for tariff setting in developing countries i s a price cap regime. Under such a regime, a very simple initial tariff structure i s applied for the first five years o f the contract, subject to annual re-adjustment for inflation in a given month o f each year. Tariff revisions should normally occur at five-year intervals and must follow a formula that applies to the average tariff that i s billed by the concessionaire, while the circumstances under which an extraordinary tariff revision i s permitted should be narrowly defined. Then the issue i s the process, guidelines, and regulatory discretion when adjusting the tariffs at the five-year interval. Much o f the experience and literature on the subject 141 assumes that tariffs will be set on a forward-looking basis under which the regulator attempts to estimate the level o f revenues-and, hence, tariffs-that will be required to cover the operating and depreciation costs o f an efficient operator together with a return on capital on the regulatory asset base. This approach allows a large degree o f discretion to the regulator in fixing many o f the critical parameters that enter into the calculation-for example, the cost o f capital, improvements in operating efficiency, and the cost o f investment programs. Implementing a major revision o f tariffs i s a severe test for any regulatory agency, especially when it has limited past experience and guidelines to rely upon. Thus, investors may have reasonable reservations about how the regulatory discretion will be used, which will be reflected in a higher premium for regulatory uncertainty. They may, therefore, strongly prefer the alternative o f relying upon a backward-looking formula for revising tariffs. Under this arrangement, the concession contract would specify, in more or less detail, a method o f determining the average tariff for the next period using data on costs, assets, etc., at the end o f the previous period. There could be scope for some discretion by allowing the regulator to determine the value o f X in the usual RPI-X re-adjustment formula for price cap regulated concessions. In the longer term, backward-looking mechanisms for price revision will come to be seen as excessively rigid. Further, there will be pressure to increase the degree o f regulatory discretion to reflect changes in circumstances and political priorities. This i s but one aspect o f the broader reality that concession contracts are never static but are subject to a continuous process o f re-negotiation. However, it i s first necessary to build up a stock o f confidence and case-law which gives concessionaires a reasonable basis for believing that they will be treated reasonably by the regulators in the exercise o f their discretionary powers. Key contract provisions about tariff revisions are: 0 Arrangements and criteria for the re-adjustment and revision o f tariffs should be clear; 0 Re-adjustments should occur annually (or more frequently when permitted by law) inMarch or April based on inflation to end-December o fthe previous year; 0 Tariffs should normally be revised at intervals o f five years using a pre-defined formula based on the reasonable costs o f providing the service, including a return on assets employed in providing the service calculated according to the capital asset pricing model, with a provision to implement revised tariffs gradually over the following five years; 0 Extra-ordinary revisions o f tariffs should only be permitted in clearly defined circumstances, such as changes in the rates or calculation o f specific taxes and allowed costs, and should not cover the normal commercial risks o f providing the service such as changes in the cost o f labor or operational inputs. 0 Frivolous and opportunistic renegotiation o f tariffs and terms should be strongly discouraged via specific penalties inthe event o f a ruling by a neutral body. 5.3.2 Regulatorystructure: rate of return vs. price caps 142 When designing a concession contract a choice needs to be made regarding the regulatory regime. The two salient choices are rate-of-return regulation and price cap regulation. Under rate-of-return regulation, the concessionaire's returns can be adjusted each year, keeping the rate o f return roughly constant. Thus investments inthe firm are subject to little risk, particularly the market-related risk that investors worry about. Ifreturns in the market as a whole rise, the regulated utility's returns won't rise much (though they can rise a little in the period before the regulator requires a price cut). And in the event o f a market downturn, the firm's profits will not fall below the contractually agreed, long-term target. Since the regulator will adjust tariffs to induce the agreed rate o f return, firms subject to this type o f regulation tend to have a lower-than-average cost o f capital. Price cap regulation has a different effect. The regulator does not set a target rate o f return in the short run.Therefore, profits from the regulated firm can vary from period to period and are free to vary with the returns in the market. Although returns can vary more under price cap regulation than rate-of-return regulation, returns for firms under price cap regulation are still less risky than non-regulated firms since their concession contracts tend to have financial equilibrium clauses (either implicitly or explicitly). The risk that affects a concessionaire's cost o f capital can be measured by a parameter called the firm's beta, which measures the relative risk o f the firm's equity compared to the market as a whole. Its value can depend on the type o f regulation used, and the higher the beta, the higher the risk o f the investment or project. Betas for infrastructure firms or projects are lower than 1, an average for all firms. But f i r m s subject to price cap regulation have higher betas and higher costs o f capital than firms subject to rate-of-return regulation in both utilities and transport operations. This has been shown by Alexander, Estache and Oliveri (2001) and Alexander, Mayer and Weeds (1996), among others. So, it should be expected that investors will demand a higher return on investment in a firm subject to price cap regulation. Establishing the values for each o f these items i s relatively straightforward when developed capital markets exist and companies are quoted on a stock exchange. Approximations have to be used in most less- developed countries. The average asset beta in infrastructure (which accounts for the leverage in the capital structure o f the projects) i s around 0.7 for high-powered incentive regimes, such as price caps regulation and around 0.3 for low-powered incentive regimes such as rate o f return regulation. There are tradeoffs among the two types o f regulatory regimes. Price caps provide incentives for securing efficiency gains, at least between tariffs reviews. They are low maintenance in the sense that they do not require high levels o f information about the concessionaire's operations (at least between tariff reviews). Yet, they induce a higher cost o f capital as a result o f their inherent riskiness and efficiency effects can be lost through rapid renegotiation. Rate-of-return regulation does not provide strong incentives to reduce costs and requires much higher information levels for the regulator. However, it generates a lower cost o f capital since the associated risk i s lower. In practice, both regimes tend to converge, and the level o f convergence depends on the frequency o f tariff reviews. The shorter the period between tariff reviews, the higher the convergence. As shown in Guasch (2004), price caps strongly increase the probability of 143 renegotiation. And renegotiations often increase the number o f cost components with automatic pass-through to tariffs, leading toward more hybrid regimes. For operators, requesting automatic pass-through rules for as many categories as possible was a rational strategy to mitigate risk and offset the increased cost o f capital. Besides automatic pass-through rules, renegotiations also aimed to increase the rate o f return to keep it consistent with the increasing cost o f capital or address issues that were overlooked at the bidding stage, Slowing down investment, reducing service obligations or increasing direct or indirect subsidies are all examples o f such efforts, particularly in the water and transport sectors in Latin America. This i s why a common outcome o f most renegotiations was a decrease inthe level and pace o f investments. Efforts by operators to mitigate the risk o f price cap regulation through renegotiation have often de-legitimized price cap regimes. Indeed, price cap renegotiations have often caused both outcomes and agreed terms to swiftly change. This suggests that if cost-plus regimes had been initially adopted, renegotiation may have been reduced. The way price caps were handled in practice in Latin America raises some difficult questions. Would a less incentive based regime have resulted in more and better investment? Had the region created earlier, stronger and better regulatory institutions, would the outcomes have been better? Was the problem the choice o f the regulatory regime or are we trying to blame everything on one o f many factors that contributed to the high renegotiation rates? Finally, could the high incidence o f renegotiation have been avoided? The answers to most o f these questions boil down to an understanding o f how price caps and the cost o f capital interact in high risk, weak governance environments. Weak regulatory capacity and weak effective government commitment to improve that capacity in Latin America meant that price caps alone did not yield the expected user benefits. Price caps provided incentives for operators to quickly secure efficiency gains, but many o f those gains were captured by governments or firms, rather than shared with users. To make matters worse, users often faced higher tariffs (relative to a rate-of return regime) as firms attempted to make up for the higher cost o f capital. Users were also hurt by the more frequent incidence o f renegotiations under price cap regimes. Renegotiations tended to delay or reduce investment since firms could not get immediate rewards from investments through tariff adjustments (either the existing tariffs already accounted for expected investments, or tariffs would only be adjusted at the next tariff review period a few years down the road). Ultimately, the ease and fast renegotiation o f contracts (before the usual five year review) may eventually lead to the adoption o f new regimes which could result in fairer tariffs, better access, and stronger commitment to fair returns to investors. This seems to be happening through the adoption o f hybrid regimes which will retain some o f the incentive effects o f the price caps while introducing cost recovery guarantees. Such regimes could ultimately reduce tariffs because they would reduce uncertainty and hence the cost o f capital. In sum, the 1990s Latin American experience shows that price caps alone will not do much for users. As a result, one could argue that rate-of-return regulation should be favored 144 despite the higher information requirements, particularly for concessions that require significant amounts o f investments. 5.3.3 Cost of capital and how it should be determined Any regulatory mechanism, be it a price cap or rate-of-return, requires that the concessionaire achieve financial equilibrium, meaning enough returns to reward the capital investments with its associated risks. That implies determining the cost o f capital and then the tariffs that will generate sufficient returns to cover that cost or rate o f return. Thus, it i s important to properly assess that cost o f capital, which often i s sector specific. If the true cost of capital i s underestimated (which has been a common practice), then the amount o f money or investment received from the concession will be overestimated. In practice, this bias i s partially offset by a tendency to underestimate the scope for reducing average operating costs and improving revenue collection. In consequence, it is important to establish a set o f principles for determining the cost of capital to be used in the revision o f tariffs. The usual approach i s to follow the capital asset pricing model. The cost o f capital represents the required rate o f return that investors might expect on a project. Since capital has usually two components equity and debt, the cost o f capital can be written as: Cost of Ci\pihl - - (Required rate of return on debt) x (Debt YOin the project) + (Required rate of return on equity) x (Equity O hin the project) Since interest expense typically i s tax deductible, the cost o f capital can be calculated either on a before-tax or an after-tax basis. It i s important to understand that the tax rate that i s relevant i s the one that applies to project sponsors. One can think about the required rate o f return on debt (that is, the borrowing cost) as having a number o f risk factors, each o f which commands a premium that must be paid to investors in order for them to bear that particular risk.'2SA very important one i s regulatory risk. Once the operator has to make a large commitment o f resources-either in purchasing the right to use existing assets or through investments in new fixed assets- it will be concerned by the risk o f effective expropriation as a consequence o f changes in regulations, tariffs, or contract terms which prevent it from earning an adequate return on the capital that it has allocated to the business. The consequence o f this regulatory uncertainty i s that investors will add a substantial premium to the rate o f return that they expect to achieve when bidding for concession contracts for water and sanitation services. This will be There is some question about whether these risk factors are separately priced in financial markets or whether they can be diversified away in well-designed international portfolios. The structure and model used here leave open the possibility that such factors may not be priced by allowing them to take on values o f zero. For a discussion o f factor models, see Ross, Westerfield, and Jaffe (1999). 145 reflected both in the value o f the existing infrastructure and in the tariff required to sustain investment in expanding infrastructure. liequired rate of return 011 dcht - Risk free borrowing ratefor specified time horizon + Premiumfor country risk Premiumfor currency risk + + Premium for project or sector risk Premiumfor + regulatoryrisk Similarly, we can think about the required rate o f return on equity investment as being equal to a risk-free rate plus a premium for the higher risk faced by equity relative to debt, as well as all four risk factors above. The equity risk premium i s a function o f how risky a specific sectoral investment i s relative to equity markets overall. This i s the beta factor (see Alexander, Estache and Oliveri, 2001). Thus: Requiredrate of return on equitj - Riskfree borrowingrate for specified time horizon + Equity risk premium(adjustedby project beta) + Premiumfor country risk Premiumfor currency risk + + Premiumfor project or sector risk Premiumfor + regulatory risk While in many cases the risk premiums required would be similar for debt and equity, this will not always be the case. For example, regulatory lags in approving pricing decisions may have a greater effect on equity holders, since creditors have a prior claim. This formulation also allows one to evaluate the effects o f changing risk premiums and guarantees on the cost o f capital. Financial return and cost of equity. In order to measure the overall return which shareholders in a specific project earned on the capital they invested in that project, and then determine ifthat return i s appropriate given the risk they took, one can compute the internal rate o f return they made on their investment (IRR) and compare it with the cost o f equity (CE)in the country and sector o f investment (see Box 15). 146 Box 15: The cost of equity The cost o f equity is a measure o f the appropriate return that investors should expect on equity investments in a specific country and sector, given the level o f risk o f such investments. The formulas used to estimate the cost o f equity (C,) are usually based on the Capital Asset Pricing Model (CAPM) developed by Gordon and Shapiro, where: CE (Rf)+ (MW)*p+ CRP+SRRp Where: Rf= Risk free rate MRP= Market Risk Premium (3 = Sector beta CRP = Country Risk Premium SRRP=Sector and Regulatory Risk Premium Each o f these parameters corresponds to a level o f return necessary to compensate for some specific risks: The risk free rate is the minimumreturn that can be earned on a risk free investment. I t is generally measured as the average interest rate on the U S T-Bill over a long historical period. The market riskpremium is the additional return that must be earned on equity investments over risk free investments to compensate for their additional non-diversifiable risk. It is generally measured as the average excess returnon the US stock market (measured using returns on the S&P500 for instance) above the risk free rate over a long historical period. Beta is a measure o f the nondiversifiable risk o f stock market investments ina specific industry. It i s usually estimated by specialist firms based on many financial, operational and strategic characteristics o f each industry. The market risk premium i s multiplied by beta as investors are only compensated for risks that cannot be diversified by an appropriate portfolio management.* The country risk premium is a measure o f the extra risk taken when investing in a specific country. It is generally measured on the basis o f the country's Moody or other credit rating compared to the USrating, comparing the average spread onbonds ofthat country with equivalent spreads inthe US over a long historical period. The sector and regulatory risk premium i s a measure o f the risk o f government non-compliance with agreed regulatory terms or o f unilateral changes by government on the regulatory framework. It is generally measured by an index capturing the historical volatility o f regulatory changes and non- compliance, and b y the degree o f independence o f the regulatory agency. Often it is also measured by surveying existing and potential operators. It can amount inthe range o f 2-6 percentage points o f the cost o f capital. If the overall return earned by project shareholders on their investment is lower than the cost of equity measured in this way, it means that they would have been better off investing their money in alternative investments given that they earned too little compared to the risk they took. *Note that individual firms can also have a Beta, as was mentioned insection 5.3.2. 5.3.4 Valuation of concession assets What clearly should not be used for the value o f the concession in the capital base-from which the operator i s allowed to earn a fair rate o f return-is the value paid at the bidding stage, regardless o f depreciation method. Doing that would take away the efficiency- competitive angle o f the auction, by allowing a rate o f return on whatever price was paid 147 for the concession. It i s standard practice to use the depreciated book value o f assets (after adjustment for inflation) as the capital base in the calculation o f tariffs. Further, depreciation peri.ods for most infrastructure assets are much shorter than the physical life o f the assets. The effect i s a large but unrecognized transfer from current customers to future ones, especially when a utility undertakes a large program o f investments. Current capital charges-depreciation and return on capital-are high but these fall as the book value of assets after depreciation declines. Thus, even if the real cost o f providing infrastructure services i s essentially unchanged, the level o f tariffs will steadily fall. This i s both inequitable and inefficient. Future consumers are likely to be richer, so that there i s no sense in requiring current consumers to make a disproportionate contribution to the cost of infrastructure. More important, the decline in tariffs will cause them to fall below the real cost o f providing services which will undermine the incentive o f the concessionaire to maintain or replace infrastructure, Sooner or later, tariffs will have to be raised sharply to correct this, which may be unpopular and difficult to justify unless systems have visibly deteriorated. Part o f the solution lies in the adoption o f more realistic depreciation periods for infrastructure assets. Current depreciation follows conventional tax and accounting rules, but regulators can easily require that capital values for setting tariffs are computed in a more appropriate manner. Relying upon regulatory decisions alone could, however, lead to difficulties in agreeing on the basis for paying compensation at the end of a concession contract or in the event o f earlier termination, since such compensation will normally include a payment equal to the un-depreciated value o f capital assets. Thus, it would be better if concession contracts were to specify clearly the basis on which depreciation should be calculated. This is only a partial solution, because the real problem flows from the fact that the structure o f assets and investments o f the typical concession will be far from that for an infrastructure operation in some kind o f steady growth equilibrium. A better alternative would, therefore, be to rely upon a full cost o f service approach that takes account o f the replacement value o f all o f the assets o f the concession and allows for the estimated depreciation o f these assets over standard periods rather than according to historical accounting conventions. The regulatory asset base would thus be equal to the replacement value o f the concession assets and would be quite separate from the book value o f such assets. One merit o f this approach i s that it would avoid the inconsistent accounting treatment o f concession assets. Despite the fact that the concessionaire does not own the assets and cannot mortgage or sell them, they are included in the balance sheet as if they were conventional fixed assets. A clear separation should be made between (a) the assets owned by the concessionaire, and (b) the assets that are used to provide services and should be included in the determination of tariffs. Similar issues apply to asset valuation at the end o f the concession when there is a clause allowing for compensation for not fully depreciated assets. Box 16 illustrates several options to secure that valuation. 148 Box 16: Measures for DeterminingCompensation at the Termination of a Concession The following are five asset-valuation methods, presented in order o f increasing sophistication, that could be used to determine the amount o f compensation to be paid to the concessionaire for sunk investments at the termination of the concession. Historical cost. This is the traditional accounting method o f valuation for the purposes o f financial reporting. It takes the cost o f the asset when it was purchased and depreciates it over a certain period of time. As a measure o f current value, it can be misleading because it ignores inflation and thus tends to undervalue assets. Inflation-adjusted historical cost. Historical cost can be adjusted to take inflation into account by increasing book value according to either a measure o f the general inflation rate, such as the CPI, or a measure more closely related to the assets involved. Depreciated replacement cost. An alternative is to consider what it would cost to buy the equivalent asset now-or, since similarly degraded second-hand assets may not be readily available, what it would cost to replicate the investment now, less an estimate o f the asset's depreciation in value since investment. A problem with the historical cost and depreciated replacement cost is that they do not consider changes inthe value o f assets brought about by changes intechnology. Optimized depreciated replacement cost (0DRC)-or modern-equivalent-asset (MEA) value. This is a refinement o f depreciated replacement cost. I t i s the cost o f replacing the asset with the cheapest asset that does the same job (the optimal asset). For example, if a new pipe-making material has been put on the market since the pipes in a water concession were laid, the optimized replacement cost i s the cost of replacing the pipes using the new, cheaper material. As before, the cost o f the new pipe must be depreciated to account for its deterioration. ODRC solves the problem o f changing technology, but, like its predecessors, has the effect o f compensating concessionaires according to some measure o f the cost o f investment. Concessionaires could thus be compensated even for making investments that were economically undesirable - that is, investments with benefits that fall short o f their costs, even when the costs are as low as possible. Optimized deprival value (0DV)-or market value. The method o f optimized deprival value attempts to take into account value as well as cost: the O D V is the minimum o f the ODRC and economic value, where economic value is the maximum o f the net present value (NF'V) o f future earnings and disposal value, and disposal value is the amount the asset could be sold for. All together, this implies that: ODV = min [ODRC, max (NPV o f future earnings, disposal value)] To avoid incentive problems, the estimate o f future earnings must be based on an estimated future tariff that is independent of the bids made when the concession is re-awarded. In principle, ODV accounting may generate compensation payments that give concessionaires the right incentives. But determining the ODV o f the concessionaire's assets is difficult, requiring assessments o f technology, the concessionaire's expected cash flows, and its cost o f capital. The choice o f accounting rule must of course take into account the practicality, as well as the theoretical advantages o f the options. In addition, it should be noted that ODRC and ODV subject the concessionaire to certain risks that do not arise with the simpler measures o f value. As a result, they may raise the cost o f the concessionaire's capital. Source: Kerfet a1 (1998). 5.3.5 I nformational requirementsfor adequate regulation Effective regulation requires good information about the operations o f the regulated firm. The regulator needs to periodically collect information about costs, revenues, prices, investments, financial data, and realized demand from the operator (see Box 17). Reporting requirements, including form and frequency, should be stated in the concession contract. 149 The contract should also provide the regulator with subpoena powers to coerce the information from the operator and the right to impose significant and increasing fines in the event o f non-compliance. As obvious as these requirements might seem, there have been plenty o f concession contracts that have failed to provide those details and rights and have led to protracted conflicts between operator and regulator. Without an explicit mandate legally grounded in the concession contract, experience has shown that the operator i s often unwilling to provide the relevant, properly formatted information. The most common argument used by operators i s that information i s proprietary and confidential and its submission might damage the firm's competitive edge. I =Information Requirements for Operators The concession has to supply to the regulator the following documents for each service activity: Regulatory accounting: 1. Income statement* 2. Balance sheet* 3. Cash flow statement* Additional documents: 1. Report o f accounting procedures used in the preparation o f the financial statements, with a detailed breakdown o f the cost drivers. 2. Report o f detailed transactions with related parties. 3. Report containing a detailed explanation o f current results and the deviation with the budgetary figures. 4. Report o f the use o f assets and investment plans*. 5. Sources o f financing o f current and future investments. 6. Operational statistics*. 7. Information on the methodology used to set unregulated prices o f monopolistically supplied (or quasi monopolistic) activities*. 8. Auditor's report. 9. Declaration o f director's responsibility.* 10. Any other informationthat the concessionaire considers valuable for a better understanding o f the information supplied. Note: The elements above with an * symbol should be presented conforming to a specific format provided by the regulator and the contract should make reference to it. 5.3.6 Regulatory accounting norms Along with the information requirements described above, effective regulation requires data that i s standardized and the capability to analyze such data. Both can be a problem in developing countries. Quite often the concession contract does not specify the format in which the data i s to be provided to the regulatory agency and there i s a lack o f proper definition o f the different variables, depreciation and amortization rules, treatment o f assets, the accounting methodology (whether to use current cost accounting (CCA) or historical cost accounting (HCA), renewal accounting etc.), what constitutes a valid investment, what prices are to be used to value investments, and what can be considered as costs and profits. When the regulator requests financial information from the firm, it often gets data that i s 150 difficult to work with, rendering the task o f regulation even more challenging and increasing the possibility o f conflicts. Even when data and financial information i s provided in a friendly format, particular care should be taken when analyzing it. The firm i s likely to interpret and manipulate the data to its best advantage, not surprisingly, resulting in artificially low rates o f return on capital to make the case for tariff increases. Hence, there i s a need for proper regulatory accounting, sound analysis, and norms that are spelled out in the concession contract. Typical questionable practices-facilitated by the lack o f proper regulatory accounting standards- used by concessionaires to increase their rents are shown in Box 18. Box 18: Common questionable actions that need to be addressed through regulatory accounting Excessive management fees-often have been equivalent to half o f the firm's net profits. Contracting subsidiaries or related companies to provide services or equipment at prices significantly above market rates. Inflatedinvestments proceeds. Transfer o f accumulated profits into the regulated capital base. Transfer o f capital in non-regulated areas of the firm into the regulated capital base o f the firm. Valuation o f pre-privatized assets at replacement costs. Using, when convenient, past performance asjustification for demands for future higher tariffs. Regulatory accounting should be guided by a set o f principles that make the gathered information a reflection o f the financial reality o f the regulated firm and useful for the regulator to perform his duties. For example, the information should assess bona fide and relevant costs, disentangle tasks performed by the operator and revenues that are subject to regulation from those which are not, allocate costs appropriately, assess the market value o f investments, and so forth. Box 19 shows the general principles that should govern an effective regulatory accounting system. The challenges o f applying regulatory accounting norms can be quite complex, as shown by an example from the United Kingdom. An exercise that restated the accounts o f regulated industries to reflect economic value principles found dramatic differences from the regulated companies' stated numbers (Carey and others (I 994)). The recalculated rates o f return differed from those stated in the accounts for the following reasons. First, the value o f the pre-privatization assets was reduced to acquisition cost, and the depreciation charge associated with these assets was correspondingly reduced. Subsequent additions were left at indexed replacement cost since these assets are to be remunerated at the cost o f capital. Assets in non-regulated businesses were not written down (even if the market valued them below replacement cost) as the companies concerned did not think that a write down to below replacement cost was necessary. 151 Box 19: General principles for regulatory accounting n e regulatory accounting theory stipulates that the information used by the regulator should accurately reflect the reality o f the firm. Consequently, regulatory accounting should establish principles that ensure: (i)theaccuratenessandveracity oftheinformationcollectedand(ii) provide general guidelines in cases that are not explicitly covered by the accounting norms or regulator's criteria. More specifically, regulatory accounting should apply the following general accounting principles: Causality: revenues, costs, assets and financial obligations should be allocated according to each activity that generates them. Objectivity: the allocation should be based in objective principles and should not imply any undue benefit to any organization or individual. Transparency: The methodology o f assignation should be clear. The accounts should be clearly distinguishable. Coherence: the allocation criteria should be constant from one year to another, however, in case o f changes (for example, in the accounting method), the company should provide the necessary revisions and explanations for such modifications. Materiality: an accounting departure i s considered material if its omission and misrepresentation has a potential o f altering the financial position or the nature o f the company's regulated and unregulated services. Neutrality: internal transfers o f costs and revenues should be transparent and inaccordance with an applicable standard cost. Sufficiency: the information should comply with the requirements established according to the norm. Disaggregation: the costs imputed to services should be previously assigned to each particular activity that generates these services. Second, accumulated profits should not be allowed to enter into the capital base, as they were included in the firms' financial statements and disallowed by the regulator. Profits are liquid assets, already earning a return, and are risk free. Only when those profits are properly invested they should enter into the capital base. Third, the depreciation charges were evaluated on the basis o f financial rather than operating capital maintenance: thus reductions in asset value which arose because o f differences between general and asset specific inflation were charged to the profit and loss account. Also, variations in the depreciation charge which arose because o f short-term fluctuations in the real value o f assets or because o f changes in methods o f indexation were eliminated. Assets were indexed at the rate o f general inflation and a correction was made to account for the difference between specific and general inflation that arises due to technological change. The above arejust some o f the ways that different interpretations o f variables can affect the imputed rate o f return significantly. This example comes from the U.K. where expertise in regulatory accounting is quite extensive. In Latin America where those skills are less prevalent, the possibilities for misuse and manipulation are even greater, thus the need arises to clarify the rules o f the game, to watch out for those practices, and to develop regulatory accounting expertise within the regulatory agency as soon as possible. 152 5.3.7 Regulatory risk A key risk that needs to be addressed is regulatoryrisk, associated with the implementation o f regulation. It refers to unpredictable actions on regulatory matters that could adversely affect the financial equilibrium o f the operator. The premium on the cost o f capital as a result o f regulatory uncertainty-as distinct from country risk-in Latin American countries has been estimated to be between 2 and 6 percentages points (Guasch and Spiller (2002)). The impact o f that increase can be quite significant. For example, an increase in the cost of capital from 15 percent to 20 percent for a $100 million dollar project with a 25- year life would require additional payments to investors o f about $5 million per year. Increases in the cost o f capital then translate into higher required tariffs; lower annual fees, canons, or transfer fees; or higher required subsidies, if applicable. Another way to see the impact i s that an increase in the cost o f capital o f 5 percent would imply, when offering a concession, a reduction o f about 30 percent in the net present value o f the initial payment and the concession fee realized by selling the company. Alternatively, the overall level o f tariffs would have to be about 25 percent higher for the first five years o f the concession to realize a fixed net present value for the company. Thus, the cost o f regulatory uncertainty will be directly reflected in the proceeds that can be realized by awarding private concessions or in the tariffs that must be charged in order to cover the debts that have accumulated by the existing operators. Reducing regulatory risk requires an adequate regulatory framework, a credible commitment to the stability of the regulatory framework, and predictability, with the usual caveats, o f regulatory decisions. All that requires a strong legal grounding-preferably on a law-of the regulatory framework, a significant amount o f autonomy o f the regulatory agency, and a hands-off approach by the government in regulatory decisions. The importance o f regulation with the strongest possible legal grounding cannot be emphasized enough. Time after time, attempts have been made, particularly by incoming administrations, to question existing concessions and to dismantle regulatory setups by previous administrations-often for political rather than technical reasons. Such efforts significantly increase regulatory risk, translating into higher tariffs or lower transfer values. Argentina (water), Bolivia (various sectors), Brazil (water, electricity), Panama (electricity), and Peru (various sectors) are among the countries where such outcomes have occurred, interfering with budgets, salary scales, and the like. For example, in Minas Gerais, Brazil, a new governor reversed a contractual agreement that gave operating control o f electricity distribution companies to new minority foreign owners that had purchased 33 percent o f the distribution companies' shares. Similarly, new mayors in other Brazilian municipalities have contested the concession terms o f water and other companies. Such interference argues for the creation o f regulatory frameworks and agencies in legislation rather than administrative procedures or presidential decrees. Laws are much harder to overturn or modify than decrees and contracts. 5.4 ADDRESSINGFINANCINGISSUES 153 Ultimately, there are only two sources from which infrastructure needs can be funded: consumers (via user charges) and taxpayers (via subsidies).'26 Financiers-whether in the private sector or in the development community-can change the requisite time profile o f taxes or user charges, but eventually their contributions have to be repaid or remunerated (and if they aren't, the consequences will generally rebound on consumers or taxpayers at some later point). When consumers pay for infrastructure. Charging consumers for use o f infrastructure services i s common. The challenge i s deciding the degree to which their contributions cover costs. The ability o f infrastructure providers to cover costs varies by sector. In the water sector, for instance, most utilities are unable to fully recover operational and capital costs. Non-cost-reflective tariffs may arise for a number o f reasons: they may reflect excessively high costs o f inefficiently run services, or they may reflect costs that are high for good reason. Sometimes tariffs are low for political reasons (as we saw above), and sometimes they are kept low to protect the poor. There i s general consensus now that consumption o f services by the poor can be subsidized, although there are considerable challenges in targeting the benefits o f below-cost consumption. When taxpayers pay for infrastructure. Subsidizing infrastructure from taxes raises micro issues, including ensuring that subsidies are channeled to expenditure with the highest returns, ensuring transparency, designing exit strategies, and balancing the emphasis on investment and maintenance. Subsidizing infrastructure from taxes also raises macro issues, including whether too much subsidization o f infrastructure threatens fiscal stability, or whether too little endangers economic growth and poverty reduction. When the private sector $names infrastructure. The private sector has invested approximately US$372 billion in Latin American infrastructure since 1990. Starting from roughly US$9 billion in 1990, private investment increased rapidly during the late 1990s. It peaked at US$70 billion in 1998 and dropped back to about US$15 billion by 2003.'27The key issue i s not whether financing should be public or private, but how the public and private sectors share the risks and rewards in a way that works for both sides. Financing and ownership are secondary. When oficial lenders and donors jnance infrastructure. Inpurely monetary terms, the role o f official lenders and donors has never amounted to more than a small percentage o f total infrastructure needs overall (although this varies considerably by country). Official financing fell temporarily after 1997, but it i s again on the rise, as the contribution o f infrastructure to poverty reduction-indeed, o f growth to poverty reduction-has been reappraised. In most infrastructure sectors there are activities in which private sector interest i s likely to be limited, and others in which private interest needs external support. Official lenders, donors, and development agencies have an important role to play in both cases. The challenge i s to maximize the role o f those relatively small amounts o f official financing-for example, by stimulating experimentation and innovation; supporting efficiency gains; mainstreaming environmental and social considerations; attracting private '" The following paragraphs draw on ADB, JBIC, and World Bank (2005). World Bank,Private Participation in Infrastructure (PPI) database. 154 investors to share risks with the public sector; and building effective institutions to plan, coordinate, and regulate infrastructure services. Covering costs through user charges i s a critical long-term objective. In the short term, user charges might be legitimately constrained by a variety o f factors or large investment needs might require upfkont financing to be recovered gradually from user charges. Sometimes those financial shortfalls can be filled by the private sector, but sometimes private financing will be insufficient, unavailable, or unacceptably expensive. Even where the private sector comes in, it often requires risk-sharing with the public sector. In cases in which the private sector cannot or will not provide all the financing or bear all the risk, investments with adequate economic rates o f return should be allocated fiscal space.'28 Adequacy will depend in part on competing claims from non-infrastructure expenditures and from the need to keep fiscal deficits low. It will also depend on the veracity o f the claim that user charges or private financing cannot fill the gap; sometimes it requires fiscal tightening to induce sector agencies to make reforms and seek other sources o f funds. Of course, this does not mean that more fiscal space for infrastructure should be a country's first step. In several cases, fiscal tightening for macroeconomic stability and debt sustainability would take higher priority. In most cases, the possibility exists for stronger promotion o f private financing in infrastructure and for higher user charges. And there can be opportunities for cost reductions, or better management and maintenance o f existing assets. In some cases, strengthening public expenditure management should come before more public expenditure. If adequate institutions and controls are not inplace, countries can easily veer from underspending to overspending. But if and when those difficult preconditions are met, governments should allocate fiscal space based on long-run growth objectives and in pursuit o f fiscal solvency. Infrastructure spending on worthwhile projects can create a virtuous circle: more growth, more fiscal revenue, more fiscal space. The challenge is to select the right projects-and put in place the policy and institutional frameworks that actually make them worthwhile. To move forward on the concession program in Peru, financing issues have to be addressed. In contrast to the past program o f concessions analyzed in this report, most of the projects in the concession pipeline are not fully financially viable, thus they require some form of government financial contribution through public-private partnerships (PPP). Specifically, the government concessions the project to the private sector, which in turns finances the project, and is repaid partly through user fees when applicable and partly through an annual government contribution (when the project i s not financially viable). The latter i s often chosen as the award criteria, so the operator requesting the minimum annual government subsidy gets the concession. There are two interrelated parts to the issue o f financing projects. The first i s the availability o f public funds and how to best leverage them. The second i s the private financing sources and environment-domestic and foreign-for the operator to secure the 12' The term fiscal space covers all forms of fiscal support, including guarantees and other contingent liabilities, as well as direct expenditures. 155 financing o f the project. Appropriate, non-distortive interventions to lower the cost o f capital for the investors are desirable, since that would reduce annual government obligations, allowing increased leverage o f public funds and lower user fees, if applicable. Scarce public funds and existing private sector risk perceptions mean that additional efforts are required to secure strong leverage o f public funds and net additional capital from both private and public sources. One option i s to tap domestic markets, an option which i s particularly promising in Peru relative to the rest o f the region. Over the last few years, Peru has deepened, expanded and improved its domestic capital markets and securities laws and regulation are relatively strong for the region. It i s one o f the markets with better "securitization" regulatory frameworks and, together with Brazil, a market where structured finance transactions are frequently placed in the domestic markets. And local savings via pension funds and insurance companies continue to grow between US$1-2 billion per year. Several infrastructure projects and local utilities have tapped into the domestic capital markets, including electricity and gas transmission, toll roads, construction companies, water companies, and electricity generation and distribution, among others. One promising option i s issuing insured bonds in local capital markets (see Box 20). Perhaps the only pending item in the capital market development agenda i s that long tenors are still only possible in local US dollars (there i s still a high dollarization o f the local financial system). T o complement Peru's domestic capital market development, a World Bank Guarantee Facility has been approved to facilitate access to local financing (see Figure 54). The facility provides guarantees on government financial commitments for infrastructure projects that satisfy a number o f requirements and safeguard^.'^^ Institutional investors in Peru, such as pension funds, mutual funds, and insurance companies have strict restrictions on potential foreign investments, as no more than 10 percent o f their portfolio can be invested abroad. In addition, they are traditionally not willing to invest in assets which are rated below AA. As a result, they are highly liquid and currently hold a substantial amount o f their portfolio in government assets (25 percent). These funds are therefore looking for highly rated domestic investment opportunities to diversify their portfolio and invest their liquidity. Discussions with local rating agencies indicate that the World Bank guarantees would have a substantially positive impact on ratings and make it possible for Peru's institutional investors to participate in some o f the infrastructure pipeline projects' financing. Even with the credit-enhancement resulting from the guarantee, not all projects in ProInversion's pipeline will be able to issue bonds on the capital markets. Some may remain unattractive to capital market investors for other reasons. They could however find financing from banks at improved conditions compared to the terms they would find without the guarantee. The credit enhancing role o f the World Bank guarantees will be all the more important since most o f the projects in the current pipeline are not financially viable and therefore would not otherwise attract private investments. IZ9Peruis the first country in LatinAmerica and the Caribbeanto developa guarantee facility. 156 IBox 20: Using insured bonds: lessonsfrom highwaysin Latin America I Insured bonds guarantee payments will be made according to a set schedule. They are generally sold in local capital markets and insured by international companies with the highest credit rating (AANAaa). They have been used to pay for large infrastructure projects-for example, highways in Chile and Mexico-and do not increase public debt. Instead, they rely on the cash flows from the infrastructure project to repay investors. Given the long term o f insured bonds (often 30 years) and denomination inlocal currency, they are a good fit with the revenue stream o f large infrastructure projects. For Peru to take advantage o f insured bonds, it is important to learn from other countries' experience. T o create conditions that are favourable for insured bonds in infrastructure concessions, the State should: Promote legal changes that facilitate the participation o f financial insurers, such as: o Dispute resolution mechanisms between private investors and the State that are based in international arbitration; o Clear compensation mechanisms for contract modifications, such as additional work requested by the State; o Clear compensation mechanisms for termination of the concession, with or without cause attributable to the concessionaire. Share risks with the concessionaires through things like subsidies, minimum guaranteed revenue, exchange rate insurance, variable terms, etc. Project design for insured bonds is also critical to their success, including: Minimumsize: the structure o f insured bonds requires a minimumsize o f roughly US$lOO million in investment; Projects with investment instages: it is preferable that a project's total investment (for a minimumof US$lOO million) is carried out at the beginning because it i s complex to structure and insure financing for later stages, which would introduce uncertainties in the whole project; Tariff adjustments: it is recommended that tariff readjustments are based on local inflation. 1Source: Project Finance Associates (2005) Figure 54: How the guarantee facility works While quite effective at lowering the cost o f capital and providing significant credit enhancement, the guarantees mentioned above might not be enough to bring in the sizable 157 amounts required for infrastructure finance. Consideration also should be given to developing well-designed and appropriate infrastructure funds (which i s not an easy task). An example is the Private Infrastructure Financing & Support Facility (PIFSF) being developed by the World Bank in Indonesia to facilitate the financing o f infrastructure projects. The PIFSF would offer a range o f financial products required to attract private long-term capital into commercially viable projects. T o do so, it will focus on innovative ways to un-bundle and mitigate risks for investors in infrastructure projects, promote the development o f local long-term bond markets, and develop expertise in structured financing for infrastructure. The PIFSF could provide a layer o f credit enhancement to banks, non- bank financial institutions (NBFIs) and capital markets that directly finance projects. Examples o f credit enhancements could be in the form o f partial risk and partial credit guarantees, liquidity facilities to help lengthen financing maturities, and backstop facilities for securitizations, etc. It i s expected that with these sorts o f products, the facility would attract other private investors into infrastructure. The facility could also have other financial institutions as investors and, in turn, the facility could provide financing directly to infrastructure projects in the form o f long-term loans, guarantees, etc. The facility will complement the role o f the existing financial institutions and banks in the financing o f infrastructure projects. Clearly, the products offered by the PIFSF should be expected to evolve gradually as institutional capacity, the private infrastructure environment, and financial sector reforms in the country evolve. Initially, it would likely be a source o f long-term financing and a provider o f credit enhancements for risks that could be mitigated in the market. Over time-as the necessary financial sector policy reforms are implemented and as the role o f the private sector in infrastructure increases-more sophisticated products could be offered by the facility. The range o f product offerings would be market driven. 1 Figure55: Structure of a PrivateInfrastructureFinancing& Support Facility -I I Additionally, there are complementary measures that can be implemented to mitigate existing risks as shown in Figure 56 and Figure 57, so as to reduce the cost o f capital in a non-distortionary way. 158 Figure 56: Options for Risk Mitigation under PPPs Short / mediumterm measures Longterm measures Government-relatedrisks * Expand coverage o f World Bank Establisha transparent, independent, Partial Risk and MIGA guarantees; and private sector-friendly legal and * Help design solutions for existing regulatory framework; loss-making businesses resulting from government contract breach; Structure new projects with minimum initial private equity injection, for instance establishing a WBIF to act as passive project equity holder. Foreignexchange risk Establish foreign exchange liquidity Develop local capital markets and facility; localcurrency financing instruments; * Participate to project financing, Ensure institutional investor preferably with local currency debt, regulations (pension funds, insurance expanding the range o f countries the companies) allow adequate Bank can lend to indomestic currency participationin infrastructure projects, (possibly using the WBIF to raise local (while still safeguarding the interest o f currency funding). pensioners or policyholders) Help achieve stable macro-economic environment and currency. Source: Sirtaine (2005) Figure 57: Options by Typology of Countries Countries with good track record Countries with poor track record indealing with private investom indealing with private investors Expandcoverage of World Bank Partial Structure new projects with minimum initial Countries with Risk and MIGA guarantees; privateequity injection, for instance - stable currencies Participateto project financing with local establishingaWBIF to act as passiveproject currency debt, expanding as appropriatethe equity holder; range of countries the Bank can lend to in Participateto project fmancing with local domestic currency. currency debt, expandingas appropriatethe / range of countries the Bank Where necessary, help can lend to in domestic design solutions for existing currency (possibly using loss-makingbusinesses the WBIF to raiselocal currency funding). Countries with resultingfrom government volatile currencies Expand coverage o f World contract breach Structurenew projects Bank PartialRisk and MIGA with minimum initial guarantees; private equity injection, for Establishforeign exchange liquidity facility. instanceestablishingaWBIF to act as passiveproject equity holder; *Participate to project financing with local currency debt, expandingas appropriate the rangeof countries the Bank can lend to in domestic currency (possibiy using the WBIF to raise local currency funding). *Establishforeign exchange liquidity facility to cover remainingFX exposure 159 5.5 SUMMARY The need for increased and improved infrastructure remains. Yet, public funds remain scarce. Thus, the way forward requires some form o f private sector involvement, with some adjustments made to the old model. Also, most o f the nearly twenty projects in Peru's infrastructure concession pipeline are not fully financially viable, requiring some form o f public sector contributions. T o move forward successfully, a number o f adjustments are required on process design and financing, particularly to secure elusive social and political support. That will require improvements in: i) Transparency o f the transaction, transparency in the use o f the proceeds, or public funding, if needed; ii) An effective communication campaign motivating and defendingthe need for the project and for private participation; iii) Community involvement and buy-in, requiring that no project be developed unless it has the full support o f the affected community. In other words, it will be a bottom-up process unlike the previous top-down ones; iv) Explicit measures to account for the poor and for unemployment. Possibilities include social tariffs (but with improved implementation to minimize the leakages), social connections or a phasing out period if tariffs were to increase substantially, and reintegrationor compensation for laid-off employees; V I A regulatory framework with clarity on guidelines for tariffs adjustments and more emphasis on regulating by objectives rather than by means; vi) Stronger regulatory institutions with increased autonomy, both operationally and financially; vii) The use and extent o f regulatory instruments, such as regulatory accounting and informational requirements, built into the contract; viii) Efficiency and benchmark measures to be used for regulatory purposes such as tariff adjustments; ix) Measurements o f the cost o f capital; x) Commitment to no opportunistic renegotiation and upholding the sanctity o f the bid; xi) More effective use o f performance bonds; xii) Allocation o f risks including reconsideration o f the balance between price caps and rate o f return regulation; xiii) Conflict resolution mechanisms, appeals and procedures; xiv) Overall concession design and incentive structures. Moving forward will also require adjustments to project financial structures. As mentioned, most o f the remaining projects require some form o f public financing support, such as equity, debt, credit support, or guarantees, because they are not financially viable. That requires innovative types o f private public partnerships (PPP). What i s emerging as a salient choice i s a PPP mode that entails the private sector making the investments in the 160 project and the public sector repaying the investors in annual payments for the life o f the concession. Yet the long duration o f those projects and perceived political and economic risk prompts the private sector to secure guarantees for the agreed government financial commitments. Implementing special Public Private Infrastructure Funds are another alternative to facilitate the financing o f infrastructure projects. Further development o f local capital markets so as to tap the significant liquidity o f Pension Funds, and thus eliminating foreign exchange risks, are also measures that have to be considered. The alternative pursued by some countries would be developing foreign exchange liquidity facilities. The Government o f Peru has already incorporated many o f these recommendations in the early stages o f its second-wave concession program. It has also secured an Infrastructure Guarantee Facility with the World Bank to provide guarantees against government default on financial commitments to private sector investors. 161 i r 0 Q) 0 0, 0 hl zb a, 7 nb n 2 n .I- E LL a, 8 E Q 0a, 3 Q 0 0a, 8 0 8 8 0 hl a, ? b Q) n ? .I- v) E ->. 3 ?0 3 a 9 7 Z E E? E E P LL -LLP L L m .-C 0 .-S .I- .I- .-v) C .I- a, 'C m Q P Q Q m C m L0 a 0 +5 2 5 a m W E a t -8 -8 - C m m -m C m e 0 P 7J 0 7J U U C m C m kP sm ke sm U U L 0 L 0 0 a, C 8C .-8 a v) .-a, .I- m wx W Q) Q) r W Q) : 6 S C Y m v) 7 zm S 3 v) Q, Y 0 W 0) 51b 51 n b E Q, a, 51 c -a S 88 E E 62 n - - E! e LL a, LL LL c 0 z L s a a, c O P -a, , N .-5 v) v) v) v) .- -m v) Q 0 .-C0 : .-mm U * Y Y .-E Y .-C0mm m m C n Q 0 .-.-a0 C m m .-E0 - 780 v APPENDIX4: SIGNIFICANCE TESTS The two-tailed Wilcoxon signed-rank test and the Hotel test are used to test for significant changes in firm performance variables after privatization. However, both tests are based on the assumption that the distributions are normal. Since the sample o f years for each individual firm i s small, the central limit theorem cannot be applied and it i s necessary to verify the normality o f the series. Therefore, the Shapiro-Francia test for normality i s used. When the Shapiro-Francia test rejects the null hypothesis o f normality, a non-parametric test called the Kolmogorov-Smimov (K-S) statistic i s used. The K-S statistic tests the equality o f the empirical hazards functions o f the pre and post-privatization performance indicators. The test evaluates the closeness o f the distributions jlp"-p"' and lpo>i-p,ii by computing the least upper bound o f all pointwise differences I , k - p r ~ ~ (x)- , @ r ~ ~ - p r ~ ~ (.) 1. The KS statistic can be written as: The null hypothesis (H, , y - p ' ~ ~ = picprn) is accepted if~ p " " + p " 'i s sufficiently close to ~ p r e - p r z l , in other words if the value o f D is sufficiently small or smaller than the critical value at a certain significance level. Table 53, Table 54, and Table 55 show the results o f these tests for Telefonica, Electrolima, and Electroperu respectively. Concession results are shown for Matarani Port (Table 56), railroads (Table 57), and the Lima Airport (Table 58). 168 c .-.-.- e Y N C a e L v1 e % c U 1 1 1 535 h h h : B 1 0 r; Hi * * * I ( * * ** ** O 60 ZO ZN m o ma N m m 0 0 E! 8 8 3 3 0 3 0 0 3 *I * * ** X * *I m 8 w 0'8o W % 0 0 1 - 0 P s r n P - 0 0 0 0 0 0 0 0 0 U N APPENDIX 5: MEASURING CONSUMERWELFAREGAINS T o assess consumer welfare gains from privatization, we follow the methodology applied by Mckenzie & Mookherjee (2002) and Torero & Pasco-Font (2001). Price changes affect welfare through changes in total expenditure, assuming that the level of consumption i s fixed during the analysis period. This i s not an unreasonable assumption, as electricity and telecommunications are fairly inelastic sectors. The sectors also have restricted competition, making it easier to attribute price and access changes to privatization. It would be necessary to include other sectors in the analysis to calculate the change in the consumption level and isolate the access and price effects completely attributable to the process. Based on household survey data, the first step establishes the expenditure o f every household in each one o f the markets under study. Then following the Banks, Blundell & Lewbel (1996) first order approximation to the change in utility, from the Mckenzie & Mookherjee (2002) study, we have that the change in utility could be expressed as the change in prices. __ =-(A A U 1% P, >w,o (8) XO Where "x," i s the total household expenditure in period "t", "wj{," i s the household initial budget share on service "j", "pj" i s the price o f service "j" and "U" i s the household utility. We can see that the impact o f price changes will be larger the greater the proportion o f expenditure in the household budget. The second approximation o f Banks, Blundell & Lewbel (1996) gives us: A U A l o g p j alogwj -(Alogp,)wjo 2 alogp, ~ 1 (9) XO This approximation allows some quantity response to the price change through the estimation of elasticity: dlog w, / dlog p, . The greater the effect o f "p" over "d',the larger the impact o f a change in prices on welfare. The effect o f prices over the household expenditure will be taken from the estimation o f the Engel Curve, resembling the Mckenzie & Mookherjee(2002) work. Explicitly the elasticity will be the estimation o f the ``y~''factor. To calculate the equation we will include demographic control variables ``2'' and household per capita income "x / n". This evaluation allows us to quantify the welfare effects for the household that had the service in both periods o f analysis. T o include the welfare effect Mckenzie & Mookherjee (2002) proposed calculating a virtual price in the initial period. This virtual price will represent "the lowest price under which a household would have chosen to consume zero units o f the service prior to privatization if they had had access to the service in 175 q u e ~ t i o n . " 'The idea i s to estimate the price that makes the Engel Curve equal zero and ~ ~ represent the access welfare as the difference between the virtual price and the actual price. Since there i s no level o f consumption prior to privatization, the post privatization level will be used to estimate the welfare effect. An additional task i s the introduction o f the two step Heckman (1979) selection correction because the Engel curve i s calculated only for those households that access the service and this could create a selection bias. Torero & Pasco-Font (2001) also use the change in prices, assuming a static consumption level, to estimate the welfare gain o f price effects. Using this methodology, we proxy for changes in consumer welfare with the calculation o f consumer surplus. Thus, we will assume that the level o f consumption o f the service under study i s fixed. We attribute changes in consumer welfare to changes in the level and structure o f tariffs. This procedure enables us to isolate the effect o f price fluctuations from changes in consumption patterns. Formally this i s given by: where G = consumer surplus in the public service q* = quantity consumed pl = initial price p2 = final price. In the Waddams, Price, and Hancock (1998) study, q* i s the intermediate consumption between that o f the two reference periods. Assuming a constant q over time could induce some problems, as it implies that the quantity consumed i s not affected by changes in prices. In other words, we assume the service demand to be inelastic. But, as mentioned by Waddams, Price, and Hancock, given that the goods under study have l o w elasticities, this methodology can give important information about the distribution o f welfare gains/losses arising from price changes between different households. A serious limitation in applying this methodology to Peru is that it does not take into account changes in the level o f access. Although this i s a reasonable assumption in a country like the UK where most o f the population has access and where this methodology was originally applied, this is misleading for Peru where a significant share o f the population had no access to services. This fact makes measuring the increase in public service access extremely relevant. In Peru, a major change resulting from the reform o f the utilities i s that a large share o f the population gained access to these services. This definitely has a positive impact on the welfare o f households and should be included in our calculations. Therefore, Torero & Pasco-Font (2001), tried to consider the effect o f increases in access. They estimate a system o f demand equations to measure the impact o f price changes on the utility o f the households, taking into account the utilities under study (electricity or telephone) as well as food, clothing, and other durable and non-durable goods. This i s an Mckenzie & Mookherjee (2002). Page 25 176 attempt to measure the substitution effect associated with the changes in the tariff structure o f the services under study. Given the need to control for changes in access, they propose using a two-stage Heckman estimation procedure. In the first stage, they model the decision whether or not to connect to any o f the services. For this, a probit model i s used: ~(sp)= Pilnb,,,) +pysec +PIp2cover + + ~3age P4age2 + + Pjnatong + p,worn + p,su + pasnu lln(inc) where P(sp) = probability ofhaving access to the service Pins = installationprice as a perpetuity(for telephones) cover = coverage ofthe service inthe district age = age of the household head natong = 1 ifthe mother tongue o f the householdhead i s a native language worn = 1 ifthe householdhead is a woman su = 1 if the household head has higher university education snu = 1 ifthe household head has higher non-university education sec = 1 ifthe householdheadhas secondary education inc = monthlyhousehold income. From this equation, the Inverse Mills ratio i s obtained and used to correct for the access problem. This ratio i s included in the second-stage estimation, i.e. demand estimations, to obtain price elasticities and the consumer surplus for the three services. The equation for each service is: Ln(qi) =PO+ Plln(pi) + Paage + P3age2 + Pdhsize + Pjold + Psnatong + p,worn + + pysnu + /3l~worn215+Pihln(inc) +pl7irrn + Plosec + @//cas + Plrhoras + Pl3horascon + P14rnale21.5 where q, = quantity consumed o f service i PI = price o f service i age = age o fthe household head hsize = number ofmembers inthe household old = number of members over 65 years natong = 1 ifthe mother tongue o f the householdhead i s a native language worn = 1ifthe household head is a woman cas = 1 ifthe householdheadis married su = 1 ifthe household head has higher university education snu = 1 if the household head has higher non-university education see = 1 ifthe householdhead has secondary education horas = number ofhours worked by the household head 177 hornscon= number o f hours worked by the spouse male215 = numberofmales between 2 and 15 years in the household worn215 = numberof females between 2 and 15 years of age inthe household inc - monthly household income irm - - Inverse Mills ratio. These equations estimate price elasticities and consumer welfare associated with each o f the services in each period o f time. This i s given by: 'se =-4 a PI where Wj = welfare associated with the consumption o f good i 4; = estimated quantity consumed of the good i P I = price elasticity o fthe good i. This procedure calculates the welfare gains from price changes before and after reforms for different household groups. Table 59 shows the price elasticities for the services under study. Specifically in the case o f telephone, the elasticity i s bigger than that shown in previous studies for Peru, particularly for local calls (see Pasc6-Font et al., 1999; Torero et al., 2000).'31This can be explained because in this case, the elasticity i s estimated for the domain in the survey that covers all urban Peru. The problem with this level o f aggregation i s that i t includes an enormous variance in terms o f penetrati~n,'~~which could explain the higher e1asti~ities.I~~With respect to electricity, the elasticity i s close to unity for the three periods under study. Table 59: Price Elasticity of Utilitiesin Urban Areas, 1991, 1994 and 1997 Telephone Local LDN LDI Electricity 1991 -0.93 1 -2.148*** -5.009 *** -1,001 *** 1994 -1.357 ** -2.779*** -3.090 *** -0.961 *** 1997 -1.129 -3.300*** -2.853 *** -1.033 *** Notes: Level o f significance: ****95%; *** 99%: The LSMS 1991 survey does not include information on communities inthe jungle. Source: LSMS (1991, 1994 and 1997). For the results o f the analysis, see section 3.2. 1 3 'The price elasticity in these two studies for local calls fluctuates between -0.28 and -0.47 based on information from Lima and other cities with high penetration rates. 1 3 *The lowest penetration i s 0.4 lines per hundred habitants in Huancavelica and in Lima, the province with the highest penetration, it is 13.5 per hundredhabitants. ' 3 3Similar studies like the one done by Martins-Filho and Mayo (1993) found too high elasticities, between 1.05 and -1.55, and there are other several studies that found elasticities between -0.446 and -1.734 (see Pasco Font et al., 1999 for a review o f other studies). 178 APPENDIX 6: CALCULATINGINDIRECT EFFECTSWITH A CGE MODEL T o run the Computable General Equilibrium model we had to use a Social Accounting Matrix (SAM) for Peru. A SAM i s a representation o f a country's productive structure and income distribution channels, and usually contains entries for productive activities, commodities, factors, institutions, capital account, and the `rest o f the world' (see Figure 58). In this study we use the 1994 S A M for Peru made by Segura and Garcia (2004). We use this 1994 SAM because it i s one o f the most complete available and because the privatization o f the telecommunication sector took place in that year. The privatization o f the electric enterprises also began in 1994. The matrix i s based on the typical structure o f macro SAMs, as reported in Logfren et. al. (2001). The S A M was constructed using information from the Input-Output Table made by the Instituto Nacional de Estadistica e Informatica (INEI) for 1994 (base year o f the national accounts); the 1994 Living Standard Measurement Survey (LSMS) carried out by the Instituto Cuanto which contains representative information on expenditures in every good or service and on income sources by type o f household; and the Peruvian Central Bank (specifically data not available in the Input-Output Table such as government transfers, revenues, debt payments, and factor rents received from abroad). It has accounts for institutions, government, firms and 3 types o f households: rural, and urban with skilled head and without skilled head.'34Similarly, it has 22 economic activities which cover the extractive sector (agriculture and livestock, fishing, and mining), the manufacturing sector (consumption, intermediate and capital goods), electricity and water supply, construction, trading, transport and telecommunications, and other services including the government. The factors considered are capital and 12 types o f labor classified according to gender (male, female), geographic location (urban, rural), qualification (skilled, non-skilled) and type o f employment (own account, wage earner).'35 Taxes are divided into 3 categories: direct taxes to institutions, sales taxes, and import duties. Finally, there are accounts for the external sector, transaction costs (trade sector), and the saving-investment account. Total annual expenditure o f households does not include payment o f direct taxes, transfers, or house renting, but includes self-consumption (valued at sale prices). Labor income o f every household member includes monthly income from the main and secondary activity (if there i s one), allowances, payments in kind, and other labor income, but excludes self- consumption. The sum o f all the expenditures and incomes are distributed by type o f household according to their relative importance in the LSMS. Wages by economic activity and type o f worker are also distributed according to the information in this survey. 134A person is considered skilled ifhe has more than 9 years o f study. 13'Inrural areas, labor is not subdivided by type of employment due to representativeness (there are not many female wage earners in these areas). 179 -r ;Ii U e &3E E $ G c I I I V a s a u o x q T -I- T -t General Equilibrium Model The model includes quantity and prices o f imports, exports, domestic output, production sold internally, and the level o f activity o f each product. The sources o f income and the expenditure decisions o f the institutions, factors, the external sector and the trade sector are also modeled. Similarly, it has a set o f restrictions which are used to balance the SAM and to close the model. I t also incorporates a real wage equation (at purchase prices) which allows one to modify the factor markets and vary both quantities and nominal wages. However, the model does not include taxes on the economic activity and value-added taxes (although the sales tax IGV, which i s included in the model, could be considered a value- added tax). The model also requires various elasticities. The consumption elasticities were obtained from the estimations o f the INEI (1997), which provides information on elasticities by type o f good and income level (high, medium and low income).136The substitution elasticity between factors in the output function, and the elasticities for the Armington function (which adds together national and imported goods) and the CET function (which adds together domestic output and exports) for each type o f good, were obtained from the Logfren et. al. (2001) base model. It i s assumed that the substitution elasticity between value-added and intermediate demand for the supply o f each good i s 0.6, and that the elasticity in the aggregated function o f a particular good produced by different sectors i s 4.j3' Finally, as in other studies, it i s assumed that the government deficit i s flexible (fixed direct tax rates), external savings i s fixed (flexible exchange rate), and investment i s flexible (fixed marginal propensity to save o f non-governmental institutions). Inthe labor market, it i s possible to have unemployment. In the capital market, the quantity used is constant but can move freely between industries, changing the nominal ea8ings but keeping constant the rate o f return in each sector. Indirect effects of the privatization process In order to estimate the economy-wide impacts o f the privatization o f the telecommunications and electric sectors in Peru, we consider that both processes decrease the price o f each corresponding service. In the case o f the telecommunication sector, this i s consistent with the re-balancing period after the privatization process that gradually reduced costs for the three types of calls (local, domestic and international long distance) and the installation charge, although it also increased monthly service charges (Torero et. al. (2000)). In the case o f the electric sector, this i s consistent with the evolution o f the index o f variable (basic) electric rates, in particular the commercial ones (CTE Statistical Yearbooks (1994 and 1998)). 13` Only the most representative goods (and their corresponding elasticities) of every economic activity in the S A M were considered, according to their relative importance inhousehold consumption. 13' Both values are similar to those used in other studies. 182 Since in the S A M the telecommunication sector i s included along with the transport sector, we first have to evaluate the impact o f a price change in the former over the output o f both activities and then simulate this output change in the CGE model. The same reasoning applies for the electric sector which i s grouped along with water supply. In this context, we work with the following scenarios: (1) Telecommunications sector: an overall price decrease o f 5.51 percent. This represents a weighted average of the annual rate o f decline o f a 3 minute local (4.5 percent decline), domestic long distance (10.4 percent decline) and the international long distance (12.9 percent decline) call between 1994 and 1998.'38 (2) Electric sector: an overall price decrease of 5.09 percent. This i s the average annual decrease rate o f the variable commercial rate between 1994 and 1998.'39 Ifwe multiply these price changes by the price elasticity o f each service, we can estimate the output change in each sector and, consequently, in the joint transport and telecommunication output and in the joint electric and water supply output. The following equations summarize this procedure: A%Q,(,) = Elasticity x A%<(,) ................................................ (1) A%Q = (R,(.,,)Ao/Q,(,,) +(Rl(,))A%Ql(,) ................................ (2) With equation (1) we obtain the percentage change in the output o f the telecommunication (electric) sector due to a price change in the service.'40 Equation (2) indicates that the percentage change in the joint output o f the transport (water supply) and telecommunication (electric) sector (Q) i s equal to the separate change in each sector multiplied by their relative importance in the total output and Rt(e,). However, if we keep constant the output o f the transport (water supply) sector, the output change in both sectors i s equal to the change in the telecommunication (electric) sector adjusted by its weight in the total output, as shown in (3). The price elasticities used correspond to those estimated by Torero and Pasco-Font (2001) and are equal to -1.357 for telephone service and -0.961 for electricity. We use estimates from INEI (2000) o f both the relative importance o f the telecommunications sector in the aggregated output o f the transport and telecommunication sector (20 percent) and the relative importance o f the electric sector in the aggregated output o f the water supply and electric sector (76 percent). 13'The weights correspond to the relative importance o f each type o f call in fixed telephone traffic. 13'Variable commercial rate in Lima province which represents the greatest share o f sales in Peru (67% o f the sales o f the distribution companies). I4ONote that we are only considering a change in the level o f consumption o f the service due to a tariff decrease. 183 Using these values, a 5.51 percent decrease in the price of the telephone service equals a 1.5 percent increase in the output of both the transport and telecommunication sector; while a 5.09 percent decrease in the electricity service equals a 3.72 percent increase in the aggregated output o f the water supply and electric ~ector.'~' Table 60 and Table 61 show some o f the CGE model outputs for wages and different economic sectors. See section 3.4 for a discussion o f the results. Table 60: Impact on GDP and Prices by Economic Activity (YO) Telecom Electricity GDP Prices GDP Prices Agriculture, livestock, hunting, forestry 0.098 0.116 0.034 0.071 Fishing, fish preservation 0.034 0.064 -0,009 0.052 Extraction of minerals, petroleumand gas, petroleumref. 0.256 0.018 0.099 0.045 Nutritionalproducts, milk products 0.126 0.056 0.038 0.053 Flour elaborationand baking, sugar industry 0.135 0.050 0.054 0.041 Drinksand tobacco 0.178 -0,052 0.066 0.008 Textile products 0.138 0.046 0.114 0.009 Apparel, articles of leatherand footwear 0.148 0.030 0.060 0.031 Furniture of wood and metal, rubber and plastic, mineral products 0.222 0.034 0.161 -0.008 Paper, impressionsand edition 0.125 0.056 0.098 -0.017 Chemicalproducts, medicines 0.155 0.038 0.109 0.006 Machinery and equipment 0.234 0.026 0.132 0.018 Electricity and water services 0.150 0.073 na na Construction 0.231 0.051 0.128 0.034 Commerce 0.155 0.007 0.068 0.052 Transport and telecommunications N a N a 0.066 0.051 Financial services and insurance 0.190 0.006 0.098 0.027 Restaurantsand hotels 0.111 0.086 0.055 0.045 Home services, house renting 0.090 0.094 0.030 0.065 Health 0.120 0.061 0.078 0.022 Education 0.092 0.096 0.052 0.043 Other goods and services (includes government) 0.140 0.037 0.054 0.025 na: Not applicable. ~ ~~ 1 4 'Since in the CGE model the level o f output i s endogenous, we need to consider these changes as changes in the efficient parameter of the output function of each sector (which is defined as a Cobb-Douglas in the model). 184 Table 61: Impact on Real Wages by Type of Worker ("/o) Telecom Electricity Male-urban-nonskilled-wage earner -0.016 0.026 Male-urban-non skilled-own account 0.070 0.002 Male-urban-skilled-wage earner -0.036 0.074 Male-urban-skilled-own account 0.034 0.005 Male-rural-non skilled 0.020 -0.031 Male-rural-skilled 0.015 -0.032 Female-urban-non skilled-wage earner 0.005 -0.005 Female-urban-non skilled-own account 0,006 0.000 Female-urban-skilled-wage earner 0.008 0.037 Female-urban-skilled-own account 0.023 0.008 Female-rural-non skilled -0.059 -0.042 Female-rural-skilled 0.006 -0.009 185 APPENDIX7: PUBLIC PERCEPTIONS STUDYMETHODOLOGIES METHODOLOGY STUDY ONE Type of Study:Qualitative Technique:Focus groups Sample:Women and men between 25 and 40 years old with or without access to services o f a private company Number of Dynamics:Upon request of the client we appointed sixteen focus groups from Dec. 16,2003 to Jan. 25 2004. The groups were as follows: 15 I cusco I Men IDIE 25-54 I No access to basic services 16 I Cusco 1Women ID/E 1I 25-55 I No access to basic services 186 METHODOLOGY STUDY TWO Type of Study Quantitative Technique Personal in-home interviews. Locality Men and women older than 18 and residing in Peru's principal cities. Kind of Sample Stratified by city o f residence, NSE, with random selection o f blocks by computer. Within each block a systematic selection o f homes was made. Size of the Sample 1,808 samples divided by region as follows: Northern city o f Trujillo, Chiclayo and Piura; central city o f Huancayo; southern cities o f Arequipa, Cuzco and Tacna; Oriente cities o f Iquitos and Pucallpa Measurements: The results were considered by socioeconomic level in Lima and by town size according to the actual sample distribution Margin of Error: +/- 2.3 percent for the totality o f the sample, assuming a confidence interval o f 95 percent Instrument: A structured and pre-codified questionnaire. The questionnaire was prepared by APOYO Opinion y Mercado and revised and approved by the World Bank. FieldWork: The questionnaires were done between March 12 and 28 2004 187 References Abdala, M. 1992. "Distributional Impact Evaluation o f Divestiture in a High-Inflation Economy: The Case o f Entel Argentina." Boston, United States: Boston University. Doctoral dissertation. Adam, C., W. Cavendish, and P. Mistry, 1992. Adjusting Privatization: Case Studies From Developing Countries. London, United Kingdom and Portsmouth, United States: J. Curry and Heinemann. Aguilar, Giovanna. 2003, "El Sistema Tarifario del Servicio Publico de Electricidad. Una Evaluacion desde el punto de vista de 10s Usuarios." Lima: PUCP Documento de Trabajo 224. Aharoni, Y. 1990. "On the Measurement o f Successful Privatization." In: R. Ramamurti and R. Vernon, editors. Privatization and Control o f State-Owned Enterprises. Washington, DC, United States: World Bank. Alcazar, Lorena and Pierina P . 2001. "La Regulacion y el Manejo de Controversias de 10s Sectores de Telecomunicaciones y Electricidad: un analisis institucional comparativo". Documento de Trabajo 5. Instituto Apoyo. Alcazar, Lorena and Rodrigo Lovaton. 2004. "La concesion de 10s puertos en el Ped: LQuien gano y quien perdio con la concesion del puerto de Matarani?. Consorcio de Investigacion Economica y Social (CIES). Grupo de Analisis para el Desarrollo (GRADE)/Instituto Apoyo. Lima. Andres, Luis, and J. Luis Guasch. 2005. "Evaluating Peruvian's Concessions." World Bank. Washington, DC. Processed. Andres, Luis, Vivien Foster, and J. Luis Guasch. 2005. "The impact o f privatization in f i r m sin the infrastructure sector in Latin American countries." World Bank. Washington, DC. Processed. Aoyama, Shunsake. 1994. Japan 's Experience in Urban Environmental Management. Metropolitan Environment Program: World Bank and UNDP. APOYO CONSULTORIA. 1999. "Situacion y Perspectivas del Sector Electrico". Servicio de Asesoria Empresarial. APOYO CONSULTORIA. 2004. "Comentarios a la revision del factor de productividad correspondiente a1 regimen de formula de tarifas tope para Telefonica del Peni S.A.A.: Segunda aplicacion, 2004-2007" Junio del 2004 Apoyo and Asociados. 2005. Project Finance: Consorcio Agua Azul S.A.- CAA, Peru Risk Analysis. Lima, Peru: Apoyo and Asociados. Available on line at www,aai.com.pe. Arellano, M., and Bond, S. 1991. "Some Tests o f Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations." Review o f Economic Studies: 58, 277-297. 188 Arellano, M., and Bond, S. 1998. "Dynamic Panel Data Estimation Using DPD98 For GAUSS." Mimeographed document. Asian Development Bank, Japan Bank for International Cooperation, and World Bank. 2005. Connecting East Asian: A New Frameworkfor Infrastructure. Barber, B. and Lyon, J. 1996. "Detecting Abnormal Operating Performance: The Empirical Power and Specification o f Test Statistics." Journal o f Financial Economics 41: 539-99. Betancor, O.and R. Rendeiro. 2000. "Airports." In: Privatization and Regulation of Transport Infrastructure. Guidelines for Policymakers and Regulators" World Bank Institute. Birch, M'.and Haar, J., editors. 2000. The Impact o f Privatization in the Americas. Miami, United States: North South Center Press. Birdsall, Nancy, and John Nellis. 2003. "Winners and Losers: Assessing the Distributional Impact o f Privatization." World Development 31(10): 1617-1 633. Boardman, A. and Vining, A. 1989. "Ownership and Performance in Competitive Environments: A Comparison o f the Performance o f Private, Mixed, and State-Owned Enterprises." Journal o f Law and Economics 32: 1-33. Bonifaz, Jose Luis. 1997. "Concesiones de Carreteras: Recomendaciones para el cas0 peruano". Mimeo preparado para el Seminario Internacional sobre Concesiones, organizado por la Universidad del Pacifico y el Programa de Post-Grado de ILADES/Georgetown University, realizado en Lima, Peni, el 2 de diciembre de 1997. Bonifaz, Jose Luis. 2001, "Distribucion Electrica en el Peni: regulacion y eficiencia". CIES. Universidad d4el Pacifico. Bonifaz, Jose Luis, Roberto Umnaga and Jennifer Wakeham. 2001. "Financiamiento Privado e Impuestos: el cas0 de las redes viales en el Peni". Proyecto Mediano del Consorcio de InvestigacionEconomico y Social (CIES), Lima. Bonifaz, Jose Luis, Roberto Umnaga, and Jessica Vasquez. 2001. "Financiamiento de la Infraestructura en el Perk concesion de carreteras." Lima: Centro de Investigacion Universidad del Pacifico (CIUP) Documento de Trabajo 37. Borcherding, T., Pommerehene, W. and Schneider, F. 1982. "Comparing the Efficiency o f Private and Public Production: The Evidence from Five Countries". Zeitschrift fir Nationalokonomie 2: 127-56. Bortolotti, B., Fantini, M. and Siniscalco, D. 2000. "Privatizations and Institutions: A Cross-Country Analysis." Forthcomingin Journal o f Finance. Bosch, A., and J. Garcia-Montalvo. 2003. "Free and Nondiscriminatory Access to Airports: A proposal for LatinAmerica." IADB.Washington, DC. Processed. Bourbakri, N. and Cosset, J. 1998. "The Financial and Operating Performance o f Newly Privatized Firms: Evidence from Developing Countries." Journal o f Finance 53: 1081- 1110. 189 Bouquet, Frank and Alain Fayard. 2001. "Road Infrastructure Concessions Practice in Europe". Boycko, M., A. Shleifer, and R. Vishny, 1993. A Theory of Privatization. Cambridge, United States: MIT Press. Boylaud, Olivier and Giuseppe Nicoletti. 2000 "Regulation, Market Structure and PerformanceinTelecommunications"Working Paper No. 237. OECD. Caillaud, B., R. Guesnerie, P. Rey, and J. Tirole. 1988. "Government Intervention in Production and IncentivesTheory: A Review of Recent Contributions." RAND Journal of Economics 19: 1-26. Campodonico, Humberto. 1999. "La Inversion del Sector Telecomunicaciones del Ped en el period0 1994-2000." Santiago, Chile: CEPAL Serie ReformasEconomicas 22. Campos, Javier and Pedro Cantos. 2000. "Railways". In: Privatization and Regulation of Transport Infrastructure. Guidelines for Policymakers and Regulators" World Bank Institute. Cannock, Jeffrey. "Regulacion y Procesos de Liberalizacion del Sector Telecomunicaciones:el cas0 peruano". V I Seminario AHCIET de Gestion Empresarial Carey, A,, M. Cave, R. Duncan, G. Houston, and K. Langford. 1994. "Accounting for Regulation in UK Utilities." Center for the Study of Regulated Industries, Institute of CharteredAccountantsin England and Wales, London. Carlton, Dennis and Jeffrey Perloff. 2000. "Modern Industrial Organization" Reading, MA: Addison Wesley. Caves, R. 1990. "Lessons from Privatization in Britain: State Enterprise Behavior, Public Choice, and Corporate Governance." Journal of Economic Behavior and Organization 13(2): 145-169. Chong, A. and Calderon, C. 2000. "Causality and FeedbackBetweenInstitutional Measures and EconomicGrowth." Economicsand Politics 12(1): 69-82, Chong, Alberto, and Florencio Lopez-de-Silanes,eds. 2005. Privatization in Latin America: Myths and Reality. Washington, DC: StanfordUniversity Press, World Bank. Congreso de la Republica del Ped. 2002. "Balance de la inversion privada y privatizacibn 1990-2001: objetivos, resultados Penh". Javier Diez Canseco, compiler. Fondo Editorial del Congresodel Ped. Lima. COPRI. 2000. "Evaluacion del Procesode Privatizacion. Sector Electricidad" Cuaderno de Trabajo. COPRI. 1999. "Evaluacion del Proceso de Privatizacion. Sector Telecomunicaciones." COPRI, Cuaderno de Trabajo. Peru. Processed. Crampes and Estache. 1997. "Regulatory Trade-offs in the Design of Concession Contracts" Preliminary Version. 190 Crandall, R. 1989. Efficiency Gains from Divestiture. Washington, DC, United States: Brookings Institution. Crepon, B., E. Duguet,and J. Mairesse. 1998. "Research, Innovation, and Productivity: An Econometric Analysis at the Firm Level." NBER Working Paper No. 6696. Cambridge, United States: National Bureau o f Economic Research. D'Souza, J. and Megginson, W. 1999. "The Financial and Operating Performance o f Privatized Firms Duringthe 1990s." Athens, United States: University o f Georgia. Donahue, J. 1989. The Privatization Decision. New York, United States: Basic Books. Eckel, C., Eckel, D. and Singhal, V. 1997. "Privatization and Efficiency: Industry Effects o f the Sale o f British Airways." Journal o f Financial Economics 43: 275-298. Economist Intelligence Unit. 2005. Peru Country Profile 2005. London, UK: Economist Intelligence Unit. Engel, Eduardo, Ronald Fischer and Alexander Galetovic. 1997. "How to Auction an Essential Facility When Underhand Integration I s Possible". Discussion Paper No. 1353. Cowles Foundation for Research in Economics Yale University. New Haven, Connecticut. Estache, Antonio, Vivien Foster, and Quentin Wodon. 2002. Accounting for Poverty in Infrastructure Reform: Learning from Latin America's Experience. Washington, D.C.: World Bank. Estache, Antonio and Gines De Rus. 2000. "The Regulation o f Transport Infrastructure and Services: A Conceptual Overview". In: Privatization and Regulation o f Transport Infrastructure. Guidelines for Policymakers and Regulators" World Bank Institute. Estache, Romero and Strong. 2000. "Toll Roads". In: Privatization and Regulation o f Transport Infrastructure. Guidelines for Policymakers and Regulators" World Bank. Galal, A., Jones, L., Tandon, P. and Vogelsang, I.1994. Welfare Consequences o f Selling Public Enterprises. Washington, DC, United States: Oxford University PredWorld Bank. Gallardo, Jose. 2000. "Privatization de 10s Monopolios Naturales en el Penk Economia Politica, Analisis Institucional y Desempeiio". Documento de Trabajo 188 Pontificia Universidad Catolica del P e d (PUCP). Lima Garcia-Lopez Loaeza, Eduardo. 2005, "Experiencias Exitosas de Participation del Sector Privado en Infraestructura: Estudio de cas0 # 1, Aeropuertos Regionales de Mexico." Lima, May. Gomez-Lobo, Andres and Sergio Hinojosa. 2000. "Broad Roads in a Thin Country. Infrastructure Concessions in Chile" Greene, W. 2000. Econometric Analysis. 4th Ed.New York: Prentice-Hall. Guasch, J. Luis. 2003. "Granting and Renegotiating Infrastructure Concessions -Avoiding the Pitfalls". The World Bank and University o f California at San Diego. 191 Guasch, J. Luis and Pablo Spiller. 1996. "Managing the Regulatory Process: Concept, Issues and The Latin America and Caribbean Story Book". Guasch, J. Luis, Jean-Jacques Laffont and Stefane Straub. 2002. "Renegotiation o f Concession Contracts in Latin America" The World Bank. Washington D.C. ' Guasch, J. Luis. 2004. Granting and Renegotiation Infrastructure Concessions: Doing It Right. Washington, DC: World Bank. Guasch, J.L. and Spiller, P. 1999. Managing the Regulatory Process: Design, Concepts, Issues, and the Latin America and Caribbean Story. Washington, DC, United States: World Bank. Gulyani, Sumila, Debabrata Talukdar, and R. Mukami Kariuki. 2005. "Water for the Urban Poor: Water Markets, Household Demand, and Service Preferences in Kenya." Washington, DC: World Bank Water Supply and Sanitation Sector Board DiscussionPaper Series 5. Hachette, D. and Luders, R. 1992. La Privatizacion en Chile. Oakland, United States: Institute for Contemporary Studies PressKentro Internacional para el Desarrollo Economico. Hansen, L. 1982. "Large Sample Properties o f Generalized Methods of Moments Estimators." Econometrica 50: 1029-1054. Heckman, J., and Hotz, J. 1989. "Choosing Among Alternative Nonexperimental Methods for Estimating the Impact o f Social Programs: The Case o f Manpower Training". Journal o f the American Statistical Association 84: 862-880. Heckman, J., and Smith, J. 1995. "Assessing the Case o f Social Experiments". Journal o f Economic Perspectives 9: 85-110. Heckman, J., H. Ichimura, and P. Todd. 1997. "Matching as an Econometric Evaluation Estimator: Evidence from Evaluating a Job Training Program." Review o f Economic Studies 64: 605-654. Heckman, J., Ichimira, H., Smith, J. and Todd, P. 1996. "Characterizing Selection Bias UsingExperimental Data." Econometrica 66: 1017-1098. Heckman, J., R. LaLonde, and J. Smith. 1999. "The Economics and Econometrics o f Active Labor Market Programs." In 0.Ashenfelter and D. Card, editors. Handbook o f Labor Economics, Vol. 111. Elsevier, Amsterdam: North-Holland Press. INDECOPI; 1999. "Competencia en el Mercado de Clientes Finales de Energia Electrica N o Sujetos a Regulacion de Precios". Documento de Trabajo No. 03 h e a de Estudios Economicos. Lima. INDECOPI. 2000. "Informe sobre las condiciones de competencia en la prestacion de 10s servicios de transporte de carga y pasajeros del ferrocarril Sur - Oriente". INFORME No 003-2000-INDECOPI/CLC. INDECOPI, BID and CAF. 1999. "Analisis de Competencia Sector Puertos". Lima, 192 Instituto Nacional de Estadistica e Informatica INEI. 1997. "Elasticidad y demanda de 10s principales bienes y serivicios". Instituto Nacional de Estadistica e Informatica INEI. 2000. "Agregados Macroeconomicos de la Oferta y Demanda Global 1991-99", Coleccion: Cambio de Aiio Base de las Cuentas Nacionales. IPE. 1999. "Las Privatizaciones y Concesiones". Serie Estudios IPE. 2003. "La Brecha en Infraestructura. Servicios Publicos, Productividad y Crecimiento en el Peni". Instituto Peruano de Economia (IPE) and Asociacion de Empresas Privadas de Servicios Publicos (ADESEP). Lima. ITU.2001. "Effective Regulation case Study: Peru". Jones, Stephen. 2004. "Contribution o f Infrastructure to Growth and Poverty Reduction in East Asia and the Pacific." Oxford Policy Management. Background paper for Connecting East Asia: A New Framework for Infrastructure, World Bank, ADB, and JBIC. Joskow, Paul. 1998. "Regulatory Priorities for Infrastructure Sector Reform in Developing Countries". In Annual World Bank Conference on Development Economics. Edited by Boris Pleskovic and Joseph Stiglitz. Washington D.C. Joskow, Paul. 1999. "Introduciendo la Competencia en Industrias de Redes Reguladas: De las Jerarquias a 10s Mercados en el Sector de la Electricidad". Documento de Trabajo No. 173. PUCP. Kapur, A. 1995. "Airport Infrastructure: The emergingrole o f the private sector". Technical Paper No. 313 The World Bank. Washington D.C. Kariuki, Mukami, and Jordan Schwartz. 2005, "Small-Scale Private Service Providers o f Water Supply and Electricity: A Review o f Incidence, Structure, Pricing and Operating Characteristics." World Bank - Energy and Water Department, Bank Netherlands Water Partnership, Public-Private Infrastructure Advisory Facility. Washington, DC. Processed. Kikeri, S., Nellis, J. and Shirley, M. 1994. "Privatization: The Lessons From Market Economies." World Bank Research Observer 9: 241-272. Klein, Michael. 1996. "Competition in Network Industries". Policy Research Working Paper 1591. The World Bank. Komives, Kristin, Vivien Foster, Jonathan Halpern, and Quentin Wodon. 2005. Who Benefits from Utility Subsidies?: Water, Electricity, and the Poor. World Bank. L a Porta, R. and Lopez-de-Silanes, F. 1999. "Benefits o f Privatization: Evidence from Mexico". Quarterly Journal o f Economics 114(4): 1193-1242. Laffont, J.J. and Tirole, J. 1993. A Theory o f Incentives in Procurement and Regulation. Cambridge, United States: MIT Press. 193 Laffont, Jean-Jacques and Jean Tirole. 2001, ((Competition in Telecommunications D.The MIT Press. London. Lexecon. 1997. "Essential Facilities: The rising tide". Competition Memo November 1997, London. Logfren, H., R. Lee, and S. Robinson. 2001, "A Standard Computable General Equilibrium (CGE) model in GAMS", TMD Discussion Paper N. 75, Trade and Macroeconomics Division, IFPRI. Lopez-de-Silanes, F. 1997. "Determinants o f Privatization Prices." Quarterly Journal o f Economics 112: 966-1028. Lopez-de-Silanes, F. and Zamarripa, G. 1995, "Deregulation and Privatization o f Commercial Banking: Pre- versus Post-Performance." Review o f Economic AnalysidRevista de Analisis Economico. 10 (2): 113-164. Lora, E., and Marquez, G. 1998. "El problema del empleo en America Latina: Percepciones y hechos estilizados." Document prepared for the meeting o f IDB Governors, Cartagena, Colombia. Luders, R. 1991. "Chile's Massive SOEs Divestiture Program, 1975-1990: Failures and Successes." Contemporary Policy Issues 9 (4): 1-19. Mackenzie, G. 1998. "The Macroeconomic Impact o f Privatization." IMF Staff Papers 45 (2): 363-373. Macroconsult. 2000. "Determinantes de 10s arreglos contractuales en la participacih privada en infraestructura: el Cas0 Peruano." Washington, DC: Inter-American Development Bank Research Network Working Paper R-390. Maddala, G. S. and E. Miller. 1991, "Microeconomia", McGraw-Hill. Marquez, G. and Pages, C. 1998. Ties That Bind: Employment Protection and Labor Market Outcomes in Latin America. Document prepared for the meeting o f IDB Governors, Cartagena, Colombia. McKenzie, D., and D. Mookherjee. 2003. "Distributive Impact of Privatization in Latin America: An Overview o f Evidence from Four Countries." Economia 3(2): 161-21 8. Megginson, W. and Netter, J. 2001. "From State to Market: A Survey o f Empirical Studies on Privatization." Journal o f Economic Literature 39: 321-389. Megginson, W., Nash, R. and van Randerborgh, M. 1994. "The Financial and Operating Performance o f Newly Privatized Firms: An International Empirical Analysis." Journal of Finance 49(2): 403-452. Ministerio de Energia y Minas. Anuarios Estadisticos. Varios numeros. Marllgaard, Peter and Claus Kastberg Nielsen. 2003. "The Competition Law & Economics o f Electricity Market Regulation" Centre for Industrial Economics Discussion Papers. University o f Copenhagen. 194 Oscategui, Jose. 2001. "Informe sobre Competencia y Regulacion en 10s Servicios Publicos de Telecomunicaciones." Lima, Peru: PUCP Documento de Trabajo 195. OSIPTEL. Estudios en Telecomunicaciones 3.Gerencia de Politicas Regulatorias. OSIPTEL. 1996. "Marco Legal de las Telecomunicaciones". Temas de Telecomunicaciones NO. 2 OSIPTEL. 2003. "Procesos de Privatizacion y Apertura de las Telecomunicaciones en America Latina. Un Analisis Cornparativo". Informe No. 001-GPR/2003. Gerencia de Politicas Regulatorias y Planeamiento Estrategico. OSITRAN and COPRI. 2001. "El Proceso de Concesion de la Infraestructura Ferroviaria en el Peni". Lima OSITRAN. Boletin Trimestral. (Several numbers) OSITRAN. Economics evaluation for TISUR, CONCAR, NORVIAL, LAP, FETRANS and FVCA. (Several years) Parton, Bennett. 2003. "Saving Utilities Duringthe Asian Financial Crisis." P3 News 3 (3). Pasco-Font, A., J. Gallardo, and V. Fry. 1999, "Household demand for telephone services inPeru", Osiptel -Telecommunication Studies No4. Project Finance Associates. 2005. "Seminario: Experiencias con la Participacion del Sector Privado en Infraestructura." Lima, May. Ramamurti, R. and Vernon, R. 1991. Privatization and Control o f State Owned Enterprises. Washington, DC, United States: World Bank. Ramamurti, R., editor. 1996. Privatizing Monopolies: Lessons From the Telecommunications and Transport Sectors in Latin America. Baltimore, Untied States: The Johns Hopkins University Press. Reyes, Jose. 2002. "Garantias en carreteras de Primera Generacion. Impact0 Economico" Archivos de Economia. Direccion de Estudios Economicos del Departamento Nacional de Planeacion de Colombia. Rubin, D. 1974. "Estimating Causal Effects to Treatments in Randomized and Nonrandomised Studies." Journal o f Educational Psychology 66: 688-701. Rubin, D. 1977. "Assignment to Treatment Group on the Basis o f a Covariate." Journal o f Educational Studies 2: 1-26. Rubin, D. 1979. "Using Multivariate Matched Sampling and Regression Adjustment to Control Bias in Observational Studies." Journal o f the American Statistical Association 74: 318-328. Rufian, Dolores Maria. 2002. "Politicas de Concesion Vial: Analisis de la experiencias de Chile, Colombia y Peni". Serie de Gestion Publica 16. Instituto Latinoamericano y del Caribe de Planificacion Economica y Social (ILPES). Santiago de Chile. 195 Ruiz Caro, Ariela. 2002. "El Proceso de Privatizaciones en el Peni durante el period0 1991- 2002" Serie Gestion Publica 22. ILPES. Santiago de Chile. Saavedra, J. and Maruyama, E. 2000. "Rigidez o flexibilidad del mercado laboral peruano: Un analisis econ6mico de la legislacion laboral." Working paper No. 28. Lima, Peru: GRADE. Saavedra, J., and Torero, M. 1999. "Labor Market Reforms and Their Impact on Formal Labor Demand and Job Market Turnover: The Case o f Peru." Working Paper R-394. Washington, DC, United States: Inter-American Development Bank, Research Department. Saenz, R. 1992. "An Overview of Privatization in Chile: The Episodes, the Results, and the Lessons." Santiago, Chile: CIEPLAN. Mimeographed document. Saldarriaga, Karin and Ruth Yonamine. 2003, "Revisih del Procedimiento de Autorizaciones de Concentraciones para el Mercado Electric0 Peruano". Segura, A. and J. Garcia. 2004. "Analisis del impact0 de la apertura comercial sobre la pobreza y desigualdad", in E. Ganuza, S. Morley, S. Robinson, and R. Vos: LQuien se beneficia del libre comercio? Promocion de exportaciones y pobreza en America Latina . y el Caribe en 10s 90, Programa de las Naciones Unidas para el Desarrollo PNUD. Shapiro, C. and Willig, R. 1990. "Economic Rationales for the Scope of Privatization." In: E. Suleiman and J. Waterbury, editors. The Political Economy o f Public Sector Reform and Privatization. London, United Kingdom: Westview Press. Shaw, L. Nicola, Kenneth Gwilliam and Louis Thompson. 1996. "Concessions in Transport". Discussion Paper. Transportation Water Urban Development (TWUD) The World Bank. Washington D.C. Shleifer, A. and Vishny, R. 1994. "Politicians and Firms." Quarterly Journal o f Economics. 46: 995-1025. Shleifer, Andrei; 1998. "State versus Private Ownership." NBER Working Paper No. 6665. Cambridge, United States: National Bureau o f Economic Research. Shukla, J. and Guasch, J.L. 1999. "Peru: Private Participation in Infrastructure. Overview o f Achievements, Priorities and Opportunities." Washington, DC, United States. Mimeographed document. Sirtaine, Sophie. 2005. "New Approaches to Attract and Finance Private Sector Participation in Infrastructure." World Bank. Washington, DC. Processed. Sinn, H.W., 1992. "Privatization in East Germany." NBER Working Paper No. 3998. Cambridge, United States: National Bureau o f Economic Research. Steiner, Faye. 2000. "Regulation, Industry Structure and Performance in the Electricity Supply Industry" Working Paper No. 238. OECD. Tamayo, Gonzalo. 2000. "Analisis de Competencia en el Sector Ferroviario". Lima. 196 Thomson, Ian. 2002. "La Privatizacion Ferroviaria en America Latina y el Caribe". In: Temas de Diwlgacion FerroviariaNo. 1, Asociacion Latinoamericana de Ferrocarriles. Tirole, Jean. 1990. "La Teoria de la Organizacion Industrial" Barcelona. Ariel. Torero, M. and Pasco-Font, A. 2001, The Social Impact o f Privatization and Regulation o f Utilities in Urban Peru. Helsinki,Finland: Wider. Torero, M., E. Schroth, and A. Pasco-Font. 2000. "The Impact o f the Privatization o f Telecommunications in Peru on the Welfare o f Urban Consumers." Lima, Peru: Mimeographed document. Torero, M., S. Chowdhury, and A. Bedi. 2003. "Telecommunications, Infrastructure and Economic Growth, A Cross-country Analysis", in M. Torero and J. Von Braun: Information & Communications Technology for Economic Development & Inclusion o f the Poor, Cluwer Publishers, Forthcoming. Torero, Maximo. 2001. "Logros y retos en el sector de telecomunicaciones: Un balance a seis afios de la privatizacion en el bienestar de 10s consumidores urbanos de telefonia fija" Documento de Trabajo No. 33. GRADE. Torero, Maximo. 2003, Peruvian Privatization: Impacts on Firm Performance. IDB Working Paper, Washington D.C. Torero, Maximo, Enrique Schroth and Alberto Pasco Font. 2003. "The Impact o f Telecommunications Privatization in Peru on the Welfare o f Urban Consumers" Torero, Maximo. 2005. "Peruvian Privatization: Impacts on firm performance." In Alberto Chong and Florencio Lopez-de-Silanes, eds., Privatization in Latin America: Myths and Reality. Washington, DC: Stanford University Press, World Bank, Inter-American Development Bank. Torero, Maximo, Enrique Schroth, and Alberto Pasco-Font. 2004. "The Impact o f Telecommunications Privatization in Peru on the Welfare o f Urban Consumers". Economia 4( 1). 2004. Trujillo, Lourdes and Gustavo Nombela. 2000. "Seaports". In: Privatization and Regulation o f Transport Infrastructure. Guidelines for Policymakers and Regulators" World Bank Institute. Vickers, J. and Yarrow, G. 1988. Privatization: An Economic Analysis. Cambridge, United States: MIT Press. Vining, A. and Boardman, A. 1992. "Ownership versus Competition: Efficiency in Public Enterprises." Public Choice 73: 205-39. Waddams Price, Catherine, and RuthHancock. 1998. `Distributional Effects o f Liberalizing UK Residential Utility Markets'. Coventry: University ofWanvick. Mimeo. Wodon, Quentin, Mohamed Ihsan Ajwad, and Corinne Siaens. 2005, "Targeting Utility Subsidies: Lifeline or Means-Testing?" World Bank. Washington, DC. Processed. 197 World Bank. 1998. "Concessions for Infrastructure. A guide to their design and award". World Bank Technical Paper No. 399. WashingtonD.C. World Bank. 2001, "Port ReformToolkit" WashingtonD.C. World Bank, and Inter-American Development Bank. 2005. "Infrastructure in Latin America: Recent developments and key challenges." Paper presented at the Diagnosis and Challenges of Economic Infrastructure in Latin-America and the Caribbean. Washington, DC. June 6. World Bank. 1994. World DevelopmentReport 1994: Infrastructure for Development.New York: Oxford University Press. World Bank. 1995. Bureaucrats in Business: The Economics and Politics o f Government Ownership. Washington, DC, United States: Oxford University PresdWorld Bank. World Bank. 2004a. World Bank Policy Research Report 2004. Reforming Infrastructure: Privatization, Regulation, and Competition.New York: Oxford University Press. World Bank. 2004b. World Development Report 2004: Making Services Work for Poor People. New York: Oxford University Press. World Bank. 2006. World Development Report 2006: Equity and Development. New York: Oxford University Press. Yan, B. Chung, S. and Roberts, M. 1998."Productivity and The Decision to Export: Micro Evidence from Taiwan and South Korea." NBER Working Paper No. 6558. Cambridge, United States: National Bureauof Economic Research. Yan, B., Chen, X. and Roberts, M. 1997. "Firm Level Evidence on Productivity Differentials, Turnover, and Exports in Taiwanese Manufacturing." NBER Working Paper No. 6235. Cambridge, United States: National Bureauof Economic Research. The word processed describes informally reproduced works that may not be commonly availablethrough libraries. Data Sources Banco Central de Reserva del Peni (BCRP). 1986-2000.Annual Report. ----. Web Page: http:Iibcrp.gob.pe. Cementos Lima S.A. 1988- 1998.Annual Report. Lima, Peni: CementosLima, S.A. Comision de Promocion de la Inversion Privada (COPRI). 1996. White Book. Cementos Lima S.A.Peni. ----. White Book; CEPREL- Electrolima S.A.Peni 198 ----. White Book; Electropeni, S.A. Peni ----. White Book; Electro Sur Medio, S.A. Peni ----.White Book; Empresas Regionales de Electricidad: Electro Norte, S.A., Electro Norte Medio, S.A., Electro Noroeste, S.A.and Electro Centro, S.A. Peni ----.Web page: http://www.copri.org. Cornision de Tarifas Electricas (CTE). 1984, 1985, 1986 - 1989, 1990 - 1991, 1996, 1999. Annual Report. Lima, Perk ----. 1992-1993, 1994, 1995, 1996, 1997, 1998, 1999.Statistical Yearbooks. Lima, Ped. Cornision Nacional Supervisora de Empresas y Valores (CONASEV); Web Page: http://www,conasevnet,gob.pe. Compaiiia Peruana de Telefonos S.A.. 1985 - 1987, 1990 - 1993. Annual Report. Lima, Ped. Electrolima, S.A. 1985 - 1988, 1990 - 1994, 1997. Annual Report. Lima, Ped: Electrolima, S.A. Electropeni, S.A. 1985 - 1999.Annual Report. Lima, Ped: Electropeni, S.A. Entel Ped, S.A. 1990, 1991, 1993.Annual Report. Lima, Perk Entel Pek, S.A. ---, 1985 - 1992.Statistical Yearbooks. Lima, Pen?: EntelPed, S.A. Sedapal, S.A. 1997. "Historia del abastecimiento del agua potable de Lima 1535 - 1996". Lima, Ped: Sedapal, S.A. Sedapal, S.A. 1984 -1 999. Annual Report. Lima, Ped: Sedapal, S.A. Superintendencia de Banca y Seguros (SBS). 1986 - 2000. WeeklyNotes. Lima, Peni. Superintendencia Nacional de Servicios de Saneamiento (SUNASS). 1998. "Indicadores de Gestion 1996-1 998". Lima, Perk SUNASS. ----. 1996 - 1999.Annual Report. Lima, Peni: SUNASS. ----.Centro de Docurnentacion, Web Page: http://www.sunass.gob.pe/cendoc.html. Telefonica del Ped, S.A. 1994 - 2000. Annual Report. Lima, Ped: Telefonica del Ped, S.A. 199 Other Data Sources: From the Commission of Energy Tar@: The Statistical Yearbook (1994-1 998). A yearly publication that details regulation o f electricity rates and economic results for the Peruvian electricity market. . CTEInforms (June 1999-November 2000). Monthly publication with news about regulation, markets, agents and other current topics o f interest in the electricity and hydrocarbon sectors in Peru and all around the world. El Informativo (June 1996-November 2000). Periodical publication containing technical articles, market information, evolution o f rates, company news, statistics and sector news. Commercial Information (First Quarter, 2000). Publication containing results from the processing and analysis o f commercial information provided by the electric sector companies. Yearly Report (1994-1 999). Electric Sector Operations (January 2000-November 2000). Monthly publication containing information on production and demand o f electricity, prices and other information related to the operation of the sector. From OSIPTEL: The Transformation o fTelecommunications inPeru: 1995 Report. Regulation and the Telecommunications Market: 1996 Report. The Openingo f the Telecommunications Market: 1997 Report. Consumers and Telecommunications: 1998 Report. Five years in the telecommunications market: 1999 Report. Institutional Report, 2000. Technical Studies (various titles). Studies in Telecommunications (various titles). Bulletins (various titles). 200 201