Report No. 30379-CO Colombia Recent Economic Developments in Infrastructure (REDI) Balancing Social and Productive Needs for Infrastructure (In Two Volumes) Volume I: Executive Summary November 1, 2004 Finance, Private Sector and Infrastructure Unit Latin America and the Caribbean Document of the World Bank Table of Contents A COLOMBIA'S INFRASTRUCTUREININTERNATIONAL PERSPECTIVE . Financing for infrastructure............................................................................................. ..............55 B CONTRIBUTION OF INFRASTRUCTURETO ECONOMIC GROWTH .The infrastructure endowment ......................................................................................... 7 .................... Overview of infrastructure sectors................................................................................. 10 10 C INFRASTRUCTUREAND THE ACHIEVEMENT OF SOCIAL OBJECTIVES .Ademand-side perspective............................................................................................ .........14 17 Status of universal access .............................................................................................. 17 19 D LEGAL,REGULATORY AND INSTITUTIONAL FRAMEWORKFOR .Affordability of public services ..................................................................................... INFRASTRUCTURE .................................................................................................................. Cross-cutting regulatory issues...................................................................................... 24 24 27 E PUBLICAND PRIVATE FINANCINGOF INFRASTRUCTURE .Sector-specific challenges ............................................................................................. .................................. Public investment in infrastructure................................................................................ 31 31 34 F RECOMMENDATIONS .Role of private finance................................................................................................... ........................................................................................................ 38 3 DISCLAIMER This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed inthis paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map inthis work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance o f such boundaries. 4 Executive Summary 1. The purpose of this study is to provide an integral evaluation of recent economic developments in Colombia's infrastructure sectors. Specifically, the study covers the electricity, natural gas, telecommunications, water, and transport sectors. However, rather than treating each sector in isolation, the discussion i s organized around a number o f major themes that are of relevance for infrastructure as a whole. First, the study considers the extent to which the infrastructure sectors are advancing or hindering the competitiveness of the productive sectors of the economy. Second, the study looks at the social dimension of infrastructure, examining the extent to which infrastructure services are available to all citizens on an equitable and affordable basis. Third comes an analysis of the strengths and weaknesses of the institutional, legal and regulatory framework that governs the provision of infrastructure services. Fourth, the financing needs of the infrastructure sectors are assessed, considering the potential contribution of both public and private sources of finance. 2. This Executive Summary covers the major highlights of the study. A much more comprehensive treatment of all the issues will be found in the main volume of the report. Furthermore, the dozen background papers on which the study i s based can be downloaded directly from the following website: www.worldbank.ornlredico1ombia together with an atlas o f infrastructure maps. The study begins with a benchmarking exercise that seeks to place Colombia's infrastructure sectors in a broader international perspective. A. COLOMBIA'S INFRASTRUCTURE ININTERNATIONAL PERSPECTIVE Financing for infrastructure 3. Colombia has sustained one of the most stable flows of infrastructure investment among Latin American countries (Figure l(a)). From 1980-95, Colombia was investing 2-3 percent of GDP in the infrastructure sectors. This percentage surged to around 4 percent o f GDP during the mid-l990s, when the boom for private participation in infrastructure was at its height, falling back to around 3 percent with onset of the economic crisis in the late 1990s, and subsequently remaining at around that level. This makes Colombia one of the only countries in Latin America (alongside Chile) that has consistently maintained infrastructure financing above 2 percent of GDP. In comparison with East Asian peers, Colombian investment in infrastructure i s towards the lower end of the range found in countries such as Indonesia, Philippines and Thailand during the mid-1990s, but has proved to be much more stable over time. During the last decade, Colombia's infrastructure investments have been heavily skewed towards the energy 5 sector, reflecting the country's desire to increase the share of thermal plant in its generating portfolio, inorder to reduce vulnerability to hydrological crises (Figure 1 (b)). Figure 1:Total investment ininfrastructure I i r 7% I 100% 90% 80% 70% 60% 50% HTrarspoti 40% Iemm Td 30% .Energy 20% 10% 0% I 1 % 4 , ,., , , , , L * I I , ,, , ,I I I , ,& k& P k&@ k d k !tpp.! , , (a)Trend over time (b) Sectoral composition Source: Adapted from Easterly and Serven ,2003 4. Colombia succeeded in raising private finance for infrastructure without a simultaneous decline in public finance. The vast majority of Latin American countries experienced a sharp decline inpublic financing for infrastructure investments towards the end of the 1980s, falling from 2-4 percent of GDP to around 0-2 percent of GDP. Colombia alone has managed to sustain public finance for infrastructure at around 2-2.5 percent of GDP, although the long term trend shows a gradual decline (Figure 2a). Private financing of infrastructure in Colombia started comparatively late, but reached relatively highlevels by the mid-1990s (Figure 2b). Flows peaked at close to 2 percent in 1997, and then began to tail-off reflecting the economic crisis of the late 1990s and the loss of investment grade rating in 1998. Nevertheless, they have remained consistently above 1percent of GDP, experiencing a modest resurgence towards 1.3 percent of GDP in 2001/02. In contrast to other Latin American countries where the growth in private investment was largely offset by the simultaneous decline inpublic investment, Colombia sustained highlevels of both public and private finance. 5. In contrast to other countries, almost all of Colombia's private capital flows were invested directly in the sector, rather than being captured as fiscal revenues. . Overall, private financing contributed about 40 percent o f infrastructure investment in Colombia during the 1990s, in contrast to other middle income countries that financed over 50 percent o f infrastructure investment from private sources. Colombia i s also unusual among Latin American countries for emphasizing green field projects over asset divestitures in its privatization program. As a result, almost all (98 percent) of the private capital flows received resulted in infrastructure investment, rather than being captured as general fiscal revenue. This can be compared with only 40-60 percent of private capital flows resulting in infrastructure investment elsewhere in the region. Another striking contrast i s the relatively low level of private participation in the telecommunications sector inColombia, compared with other countries inthe region. 6 Figure 2: Breakdown of infrastructure financing by source I (a) Public finance (b) Privatefinance Source: Adapted from Easterly and Serven ,2003 The infrastructure endowment 6. As a result, Colombia presents high levels of access to basic household services compared to its Latin American peers. By regional standards, Colombia presents very high coverage of basic household services. In water and sanitation, Colombia's coverage levels are 10-15 percentage points ahead of the other countries in its income bracket, and ahead of those of richer countries such as Brazil and Mexico. As regards electricity, Colombia's coverage is about 10percentage points ahead of countries with similar income levels, but about 15 percentage points behind that of richer countries such as Brazil and Mexico. Furthermore, comparing Colombia against a set of 80 developing countries shows that the country's coverage levels are significantly higher than would be expected given its challenging economic, social and geographic conditions (Figure 3). This i s particularly true for road density and teledensity where Colombia's endowments are respectively 35 percent and 19 percent higher than might be expected. 7. Access to services in Colombia is relatively equitable across the income spectrum, but relatively inequitable between urban and rural areas. Service coverage expanded more rapidly during the last decade in Colombia, than in other Latin American peers, reaching an additional 2 percent of the population each year with water, sanitation andelectricity services, and an additional 3 percent of the population each year with telephony services. As a result, Colombia has achieved a relatively equitable pattern of access to household service across the income distribution, particularly with regards to telephony services. On the other hand, Colombia presents a relatively large access gap between urban and rural areas. In the case o f electricity and sanitation, coverage in rural areas lags 35-40 percentage points behind coverage inurban areas. 7 Figure 3: Deviation of infrastructure endowment from expectedlevel s 'i; 50 .-0 U g! -Q 0 Ua, EE P -50 .I- .-!= 3 -100 0 .- 5 U5 -150 4- C a eQ,a,-200 Source: Own calculations based on World Development Indicators. Notes: The figure reports percentage deviation between observed infrastructure endowments and those predicted by an econometric model that takes into account the country's GDP per capita, urbanization rate, urban growth rate, general population growth rate, population density and household size. The coefficients of the model are estimated based on a sample of 80 developing countries from around the world. 8. However, Colombia lags behind its peers as regards its stock of paved roads, internet access and electricity generation capacity. Comparisons of productive infrastructure are conceptually more difficult than those made for social infrastructure, given that they reflect other factors that are more difficult to control for in a cross-country regression. Thus, energy use is strongly driven by the underlying structure o f the economy, and paved road density i s driven by the internal spatial distribution of economic activity. Nevertheless, and subject to these important caveats, Colombia's endowments in some specific areas of productive infrastructure appear to be relatively small in comparison with its Latin American peers. Indeed, based on a wider international comparison, endowments of electricity generation capacity, internet density and paved roads are a fraction of what would be expected given the country's economic, social and geographic conditions (Figure 3). However, i t i s important to note that the lag in electricity generation capacity does not indicate any problem of unsatisfied demand, but rather reflects the relatively low energy intensity of the Colombian economy. Furthermore, Colombia's relative position in these areas o f productive infrastructure has been deteriorating steadily since the 1960s,reflecting relatively low growth rate o f these infrastructure stocks. 9. Moreover, there is mixed evidence as to the relative quality and efficiency of Colombia's infrastructure service providers. Prices of electricity and telecommunications services in Colombia are towards the middle o f the range observed around the region. However, comparative data on efficiency indicators suggests that electricity and telecommunications utilities perform below the regional average. A number of recent international industrysurveys on the quality of transport infrastructure 8 rate Colombian ports among the best in the region, but provide a more pessimistic evaluation of the quality of road andrailroad services. 10. Therefore, Colombia needs to enhance its performance on productive infrastructure, while preservingits achievements inthe social sphere. On the basis of this comparative assessment, the challenge for Colombia would appear to lie primarily in reorienting its infrastructure investments towards the productive sectors, without jeopardizing its strong performance in social infrastructure. Given that the overall level of resources devoted to infrastructure will be difficult to increase, this shift will primarily have to be funded b y improvements in the efficiency of social investments, as well as new sources of finance for productive infrastructure. While the private sector can play an important role in addressing the deficits in energy and telecommunications, the key challenge of financing improvements in the road network will necessarily remain a predominantly public responsibility, given that the limited scope for toll roads has already been largely exploited. 9 B. CONTRIBUTIONOF INFRASTRUCTURE TO ECONOMICGROWTH Overview of infrastructure sectors 11. The following paragraphs provide an introductory overview of Colombia's infrastructure sectors as they look from a supply-side perspective. The electricity, natural gas, telecommunications, water, and transport sectors are each considered inturn. 12. The electricity sector underwent major reforms in the 1990s' but significant vertical and horizontal integration persist limiting the extent of effective competition. The electricity sector reforms of the 1990s aimed to vertically unbundled the power industry and led to full liberalization o f generation and retail activities. In addition, the private sector took over the operation o f about 55 percent of generation and 40 percent of distribution activities. Nevertheless, in practice, the extent of real competition has been limited due to the persistence of vertical integration, particularly among the largest public enterprises, and market power remains concentrated in a handful of larger players (Table 1). Furthermore, the coexistence of public and private operators in competitive segments of the market leads to private sector complaints of uneven treatment between public andprivate firms. (a) Generation. Colombia has 13.6 gigawatts o f generation capacity, with the two largest groups controlling more than 50 percent of the market. The portfolio has traditionally been dominated by hydroelectric schemes with very limited storage, making the country vulnerable to the El Niiio weather phenomenon. Following major rationing during the 1992/93 drought, Colombia succeeded in diversifying its generation portfolio by raising the share of thermal plant from 20 percent to 34 percent of total capacity, thereby withstanding the droughts of 1997/98 and 2002/03 without major economic loss. (b) Transmission. Colombia's integrated national power grid covers one third of the country's territory and serves 96 percent of the population. The transmission network i s generally in good condition. The main problem i s terrorist attacks on transmission towers, which were responsible for 0.4 percent of power demand going unsatisfied in 2002. The radial topology of the transmission network amplifies the disruption caused by such attacks, suggesting that strategic reinforcements could reduce the extent of the problem. More recently the security situation has been improving, with the number of attacks falling by one third duringthe last year. 10 (c) Distribution. Colombia has 30 distribution utilities. Notwithstanding considerable heterogeneity, on average these present deficient operational performance, linked to poor maintenance of distribution infrastructure. For example, average distribution losses stand at 25 percent, with considerable variation around the mean, and power outages at 12 hours per month. A significant number of utilities appear to be producing below minimum efficient scale. The poor performance of distribution companies often leads to payment arrears that ultimately jeopardize the financial health of upstream activities, creating pressure for government to bailout insolvent utilities. (d) Retail. There are 42 power retailers, although four of them control around 90 percent of the market and most are vertically integrated. Customers with more than 0.1 megawatts of installed capacity are free to negotiate prices and currently represent 30 percent of the market. Table 1:Distribution of market share inelectricity and telecommunications National Bogota Medellin Others Total holdings holdings holdings Electricity 0 Generation 32 20 21 27 100 Transmission 84 6 7 3 100 Distribution 17 25 20 38 100 0 Retail 21 24 23 32 100 28 18 11 100 16 25 0 100 Value added 19 23 39 100 0 Mobile 0 0 100 100 ource: Ownelaboration basedon sector d a. Note: In 2003 anew mobileoperator was formedas ajoint venture betweenBogotaand Medellin utilities 13. Colombia has successfully completed an ambitious plan to massivelyscale-up natural gas use both for electricity generation and for domestic purposes. Colombia has significant natural gas reserves at longstanding fields along the Atlantic coast, as well as more recent discoveries in the Andean foothills. The country i s close to completing an ambitious national plan, formulated in the early 1990s, to extend natural gas service to the economic heartland of the country through a rapid expansion of the transportation network based on public private partnerships. This has provided the basis for the substantial increase in thermal electricity generation noted above. It has also made possible a rapid penetration of natural gas in domestic markets in the country's major cities, where 20 gas distribution utilities have raised coverage rates to 77 percent o f the potential residential market in 327 municipalities. However, natural gas remains uncompetitive for the industrial sector, relative to coal and fuel oil. In addition, the government has recently announced its intention to export natural gas to western 11 Venezuela as a first phase of a plan to integrate gas transportation networks with Central America. 14. Telecommunications reform emphasized market liberalization over privatization, however market structures remain concentrated. Full competition was introduced into all telecommunications services, however market structures continue to concentrate power in a handful of dominant players (Table l), which control between 70 to 100 percent of the market for each of the four major services (local, national, international, and mobile). Both long distance and mobile markets operate under triopoly arrangements. There are 29 local telephony operators, although the national operator Telecom serves 43 percent of this market. Public enterprises continue to play a dominant role in all activities (except for mobile telephony), accounting for 64 percent of sector turnover. As was the case for the electricity sector, the coexistence of public and private operators complicates the creation of a level playing field for competition. 15. Following the reforms, Colombia's telecommunicationssector has undergone substantial modernization. Local exchanges have now been almost entirely digitalized. Following the liberalization o f international calls, three submarine wire and satellite connections were established providing multiple links to international networks, and carriers developed two extensive national optic fiber networks (one of them over power transmission lines). Mobile service was introduced to Colombia in 1994, and density had reached 11 percent by 2002 still well behind Guatemala (14 percent), Argentina (18 percent) and Chile (43 percent). Following the award of a third mobile license in 2003, service expansion has accelerated markedly such that mobile penetration i s expected to draw even with fixed line penetration during 2004. Internet use remains heavily concentrated in the three largest cities, and internet density at 4 percent lies well behind that of Peru (10 percent), Argentina (12 percent) or Chile (24 percent). Broadband services are readily available in Bogota at present (and to a lesser degree in Medellin, Cali and Bucaramanga) but only incipient in the other large cities, due to the presence of infrastructure bottlenecks. A key constraint i s that local telephony operators have exclusive rightsto provide broadband services over the exercise. 16. The water sector is fully decentralized with major differences in financial sustainability and service quality across large and small utilities. The water sector has been fully decentralized, with service provided by more than 1,300 utilities. The larger utilities present reasonable operational and commercial performance. More than half of the 26 largest utilities, equivalent to 13 percent of the urban market, operate with some degree of private sector involvement. A significant number of them have some form of private participation, often following the mixed enterprise model entailing public and private sector Board representation, private sector led operation, and public-private co- financing of investments. The smaller utilities present very weak operational and financial indicators, and provide water of deficient quality, in many cases without disinfection. A recent survey found that outside the 23 largest cities, 60 percent of water samples failed to pass potability tests. During the 1990s, sustained efforts were made to raise water tariffs closer to cost recovery levels, lifting the average domestic tariff from US$0.33 per cubic meter in 1990 to US$0.78 per cubic meter in 2001, at least for the larger utilities. Given that 88 percent of households in Colombia are metered, this large 12 tariff increase led to a strong demand reaction, reducing average household consumption from 34 to 19 cubic meters per month over the same period, and in some cases postponing the need to develop major new water resources. 17. Road is the prevailing means of transport for freight and passenger Wows, while maritime transportation dominates international trade. Some 79 percent of Colombia's ton-kilometers of freight are hauled by road, while 92 percent of passenger trips are made by bus. Regarding private vehicles, motorization rates remain relatively low at 55 per 1,000 inhabitants. The deteriorating economic and security situation has kept traffic volumes stagnant inrecent years. Interms of international trade, 96 percent of exports by weight leave the country through sea ports, the vast majority through privately owned ports handling bulk movements of agricultural produce and minerals. Roads. Colombia's primary road network i s managed by the national roads agency, INVIAS, while secondary roads have been decentralized to the departmental level, and tertiary roads have been partially decentralized to the municipal level. About 70 percent of the primary network and 15 percent of the overall network i s paved. Traffic flows are heavily concentrated with a quarter o f the primary network carrying two thirds of the total vehicle kilometers. Most o f the corridors that present sufficient traffic volumes to support toll-based concessions have already been contracted out to the private sector. The percentage of the paved road network considered to be in regular or bad condition rose from 22 percent in 1998 to 29 percent in 2003. Official estimates indicate that the cost of rehabilitating the network back to a pristine standard, i s in the order o f US$380 million. If current trends continue, by 2010 the percentage of the network classed as regular or bad would increase further to 46 percent with the rehabilitation backlog escalating to US$800 million. Trucking. In addition to the problems noted with the road network, Colombia's trucking industry presents significant deficiencies. The fleet consists largely of two-axle, gasoline powered trucks, that tend to be more than 20 years old. Modern management practices are a rarity leading to vehicle utilization rates of only 70 percent. The underlying cause i s the extreme informality and atomization of the industry, with most transport firms sub-contracting their services to single vehicle owner-operator micro-enterprises. Railroads. Colombia has two major railroads currently operated under concession to the private sector. The railroads present a cost advantage for the movement of dry bulk products and containers, and traffic (largely coal) has grown substantially during the 1990s. However, further rehabilitation must be completed before rail can make its full potential contribution to the movement o f freight towards the two coasts. 13 Inland waterways. Colombia has 7,000 kilometers of waterways with permanent navigation. The Magdalena river offers a third potential freight link between Bogota and the Atlantic coast, but would require significant investments indredging and multi-modal transfer facilities. Ports. Colombia has five regional port societies (SPRs) providing public access facilities. These were given in concession to the private sector during the 1990s, registering substantial improvements in performance. However, they now face key bottlenecks in sea and land access, as well as the need to upgrade their facilities to handle larger volumes of trade anticipated as a result of the forthcoming Free Trade Agreement (FTA) with the United States. Air ports. Although Colombia has 74 national and 513 municipal airports, the bulk of commercial transportation i s concentrated in a handful of the larger airports. A key problem i s the lack of infrastructure needed for the efficient management o f freight. Multi-modal transportation. Multi-modal transportation and modern logistics services are still incipient. There i s a lack of multi-modal operators, ando f facilities to permit transfer o f freight between modes. 18. The potential for regional infrastructure integration has only been partially achieved, leaving significant room for progress. Colombia's pivotal location between South and Central America, makes it well placed to act as an integrating link between sub-regional infrastructure networks. Significant advances have been made in the integration of electricity and telecommunication networks, while Colombia's new natural gas export project will pave the way for greater integration o f sub-regional transportation infrastructure. Inthe road sector a number of priority links have been identified under the Initiative for the Regional Integration o f Infrastructure in South America (IIRSA). However, the full benefits o f these projects will not be reaped until trucking operations can be effectively integrated across borders, avoiding the need for freight transfers between vehicles at the frontier. A demand-side perspective 19. Having said all of that, it is equally relevant to ask how the infrastructure sectors look from a demand-side perspective. Inorder to answer this question a survey was performed for 60 manufacturing firms that are large consumers o f infrastructure 14 services. This was complemented by in-depth qualitative interviews of firms participating inthree strategically important productionchains-paper, sugar and textiles. 20. About half of infrastructure services are consumed by the productive sector, where they represent important intermediate inputs.It i s estimated that infrastructure services account for at least 16 percent o f total production costs in Colombia, with considerable variance across economic sectors. Energy and transport are particularly important for manufacturers, while telecommunications are most significant for services. Concerns about the efficiency and reliability of infrastructure services lead manufacturers to self-supply about 25 percent of their infrastructure services by value; with 40 percent of firms runningtheir own electricity generators, and 20 percent operating their own fleet of trucks. Firms complain about significant quality problems in terms o f power supply interruptions, lengthy waiting lists for fixed line telephones, and considerable delays and security risks on road freight. They often take advantage of the opportunity to switch between competing suppliers, and report greater satisfaction with services provided by private sector operators. 21. Interviews with infrastructure intensive firms reveal some of the key service deficiencies from a production perspective. Firms wishing to circumvent public power supplies find themselves restricted by regulations that prevent sales of excess power over the grid, or even transfer of power across different stages of production by vertically integrated firms. Firms engaged in the early stages of production find road haulage services very costly, in part due to restrictions on nocturnal circulation of trucks due to security considerations. Firms active in later stages of production complain about deficient quality of road freight services that prevent them from responding with sufficient agility to real time demands placed by their customers. Exporters complain about the considerable costs, risks and delays associated with access to the port of Buenaventura. Firms reliant on agricultural raw materials, face significant coordination problems due to the scarcity and highcost of telecommunications services inrural areas. 22. The FTA will imply a considerable and abrupt incremental pressure on infrastructure, mainly on transport gateways linked to international freight trade. Colombia i s expecting to sign a FTA with the UnitedStates inearly 2005, which is likely to generate substantial variations in trade volume, composition and routes. According to various projections, exports are predicted to grow between 2-8 percent, comparable to one year's historic growth. Imports, on the other hand, are expected to grow by between 12-24 percent, comparable to three to five years o f historic growth. This asymmetry i s attributed to the fact that Colombia's existing import tariffs tend to be higher than the trade barriers currently faced by Colombia in the U S and other regional markets. The increase will involve products with an average unit value (US$ per ton) higher than the present one, as a result of the growing incidence o f imported manufactured goods, textiles and meat. The most impacted infrastructure services will be the international transport .gateways, mainly ports and airports, and eventually border crossings. Road traffic i s expected to increase with concentration in specific segments linking the international gateways with the major centers o f production and consumption. The trucking industry will face growing volumes; as well as heightened demand for high quality logistics services, particularly as regards container movements, freight facilities 15 and multi-modal coordination. Ports will also need to upgrade their terminal facilities in order to handle the higher volumes of trade. 23. Finally, the infrastructure needed to improve the productivity and competitiveness of the Colombian economy i s estimated at US$2,600 million per year (or 3.2 percent of GDP). Those productive sectors expected to underpin Colombia's future economic expansion are more intensive in the use of infrastructure services, than the country's traditional agricultural and mineral exports. Inparticular, they will require a significant volume of reliable electricity supply, as well as cheaper and better quality transport services. Overall investment needs to provide the infrastructure to enhance the productivity and competitiveness o f the Colombian economy are estimated at US$2,600 million per year (or 3.2 percent of GDP). About 46 percent of the total (or 1.5 percent of GDP) corresponds to maintenance and rehabilitation expenditures, and 48 percent (or 1.6 percent of GDP) to investment in new infrastructure assets. The transport sector makes the largest demand on resources, accounting for 50 percent of the total (or 1.6 percent of GDP), followed by 30 percent (or 1 percent of GDP) for the electricity sector. The percentage of these investments that could be financed by the private sector varies substantially across sub-sectors, ranging from 18 percent in the transport sector to 100 percent in the natural gas sector. However, the overall share of the total that could potentially be financed by the private sector stands at 31 percent (or 1.0 percent of GDP). This could potentially increase to 40 percent (or 1.3 percent of GDP) should it prove possible to expand private participation in electricity generation and distribution. Within the overall investment agenda, highest priority should be given to maintenance and rehabilitation expenditures that conserve the condition of existing assets. Among new investments, those associated with alleviating strategic trade-related bottlenecks in the transportation network would appear to be particularly urgent. 16 C. INFRASTRUCTUREAND THE ACHIEVEMENT OF SOCIAL OBJECTIVES Status of universal access 24. Colombia has made substantial progress towards universal access, but significant challenges remain particularly in rural areas. Colombia i s very close to reaching universal access to electricity, water and (broadly defined) sanitation in urban areas. Coverage gains have been particularly large in two new service-natural gas and mobile telephonpwhose penetration has jumped some 20 percentage points in the last six years. However, significant inequities persist in access and quality o f services. Coverage deficits are heavily concentrated in rural areas, depressed Regions 1-4, and lower socio-economic strata of the population. Consequently, around four million people lack access to safe water and sanitation and two million lack access to electricity. To reach universal access in a reasonable time frame, the rate of coverage expansion for rural electricity, water, and sanitation services will need to accelerate significantly from historic average annual rates of 1-2 percent towards annual rates o f 2-3 percent. Nevertheless, Colombia i s well placed to meet the Millenium Development Goal for water and sanitation; at least at the national level, since the target is likely to be overshot inurban areas compensating for anticipated shortfalls inruralareas. 25. Moreover, many rural communities face significant isolation. More than a third of the rural population, and about half of small rural towns, lie more than two kilometers away from the tertiary road network. Indeed, the average rural household lives 2.5 kilometers away from an all season road. In the Amazonian region, environmental considerations mean that the road network i s almost entirely absent. However, there i s a substantial network of inland waterways and airports that provide a certain degree of accessibility for these remote populations. 26. Colombia has developed one of the most comprehensive policy frameworks for universal access to infrastructure services. These include a number of cutting-edge programs reviewedbelow, particularly inthe case of telecommunications. (a) Electricity. About two thirds of the unserved rural population i s within reach of the interconnected system (SIN), while the remaining third belongs to the non-interconnected area (ZNI). Colombia has established two special funds for financing rural electrification projects: the Rural Electrification Fund (FAER) to focus on the SIN, and the Non- Interconnected Zones Fund (FAZNI) to focus on the ZNI. Both are funded by a surcharge of US$0.40 per megawatt-hour of electricity sold on the wholesale market, and hence each have a potential revenue collection of US$18 million per year. However, due to fiscal constraints, the funds are only currently being allowed a combined total budget o f US$10 million, leading to a significant backlog (US$46 million) o f appraised investment projects that are unable to go ahead. Beyond these funding constraints, the 17 tariff formula for distribution utilities in the SIN fails to create adequate commercial incentives for service expansion. Moreover, in the ZNI, generation i s dominated by diesel plant, and the full potential for renewable energy sources has yet to be explored. (b) Natural gas. As mentioned above, Colombia has made great strides in expanding urban residential access to natural gas. It i s noteworthy that numerous measures were taken to ensure that this service would be accessible to low income households, including: cross-subsidies between transportation and distribution; prioritization of network expansion in low income areas; creation of exclusive service areas with low income coverage targets; and the incorporation of cross-subsidies in natural gas tariffs across socio-economic strata. Telecommunications. Colombia's rural telecommunications program, known as COMPARTEL, has been particularly successful at promoting universal access. The program is financed from a sectoral fund that collects some US$30 million per year, based on sector levies that vary from 0 percent of revenues (local telephony) to 5 percent of revenues (long distance and mobile telephony). The financing mechanism has been criticized for distorting relative prices o f competing telecommunications services. In common with the electricity funds, there have recently been fiscal restrictions that have prevented the full use of the revenues collected. Since its inception in 1999, the program has invested some US$140 million to provide public telephony services in more than 14,000 villages benefiting more than 5 million people, and community telecenters in more than 1,400 small towns benefiting almost 9 million people. This has been achieved by biddingout the construction and ten year operation and maintenance of the facilities to the private sector on the basis of minimum subsidy request. The key challenges in the medium term are to take internet connectivity into public schools, municipalities and other public institutions, which i s estimated to cost as much as US$46 million. There are also some concerns as to whether the rural public telephones and community telecenters will be self-sustaining beyond the initial ten year period, necessitating consideration as to whether further subsidy will be needed and on what basis. (d) Water and sanitation. There are no direct national programs aimed at accelerating access to water and sanitation in rural areas, given that this i s a direct municipal responsibility. However, the national government provides substantial indirect support through the creation of an enabling legal and financial framework. In particular, the following measures can be highlighted: the constitutional right o f access to water and sanitation that provides a basis for judicial claims by unserved communities; 18 regulatory measures to reduce connection charges and require them to be spread over a three year period; the existence of substantial fiscal transfers to municipalities that are earmarked to the water sector and skewed towards smaller and poorer municipalities; the official publication of urban coverage targets in the National Development Plan; and the existence of a number of programs to support smaller water utilities. One area where the national policy framework does not appear to be so favorable to the achievement of universal access goals for water and sanitation i s with regard to technical norms for network expansion, which restrict utilities to applying high cost conventional solutions. Inthis sense, it is noteworthy that in Colombia less than 1percent of the population are served by public taps, and only 2 percent b y latrines. (e) Transport. In urban areas, the average commute to work i s 30 minutes. The poorest households are most likely to walk, or face the longest journeys on public transport. Commuting times are coming down in Bogota following vehicle circulation restrictions, and the Transmilenio project. The combined effect has been to reduce commuting times in the city by 10-20 minutes each day on average journey times o f 45-65 minutes, with the largest time savings concentrated in lower income neighborhoods. The aim i s now to replicate BRT programs in secondary cities. In rural areas, the `Roads for Peace' initiative has had a significant impact on the maintenance o f 34,000 kilometers the tertiary road network. However, the program is now reaching its close raising questions as to the sustainability o f these achievements. 27. Overall investment requirements to meet universal access objectives in Colombia are estimated at around US$1,500 million per year (or 1.8 percent of GDP). About 50 percent of this investment (or 0.9 percent of GDP) relates to expansion of service coverage in urban areas, while 44 percent (or 0.8 percent of GDP) relates to expansion in rural areas. The transport sector makes the largest demand on resources, accounting for 39 percent of the total (or 0.7 percent o f GDP), followed by 23 percent (or 0.4 percent of GDP) for the water sector. The share of these resources that could be financed by the private sector is relatively low standing at 18 percent overall, and relates primarilyto expansion of fixed line telephony. Affordability of publicservices 28. Colombia has a unique and longstanding cross-subsidy scheme applying a common national framework to the electricity, natural gas, telephone and water services. The system dates back to the 1960s, and was subsequently enshrined in the 1991 Constitution, and more recently reformed by Law 142/94. The system makes use o f a zonal poverty criterion with a six level social stratification of all neighborhoods 19 throughout the country, according to the physical quality of housing and general amenities. Neighborhood classifications are undertaken by local authorities in line with broad national guidelines. Households in the upper strata have surcharges added to the price they pay for utilities, while those in the lower strata receive percentage discounts, with both being explicitly reported on the customers' bills. The absolute size of both subsidies and surcharges was substantially reduced under Law 142/94, initiating a process of rebalancing that i s complete for all but the water sector, where particularly large adjustments were necessitated. The scheme i s based on the principle of income redistribution at the national level, so that some utilities with a preponderance of higher strata customers generate surplus resources that are supposed to be transferred to others that face a deficit situation. There i s also provision for cross-subsidy resources within each sector to be complemented by government transfers inthe case of an overall deficit. 29. The cross-subsidy scheme appears to meet a genuine social need inthe water and electricity sectors, but is less justifiable for natural gas and telephohe services. Following cost recovery principles introduced by Law 142/94 utility tariffs have increased substantially in Colombia during the last decade, doubling the cost o f a subsistence basket of utility services. Nevertheless, utility services remain relatively affordable for the average household, which currently spends less than 5 percent of income on utility services (equivalent to US$20 per month), and even in the absence of cross-subsidies would spend only 8 percent of income on utility services (equivalent to US$35 per month). For first quintile households, however, the affordability of utility services is a more critical issue and the cross-subsidy scheme plays an important role (Figure 4). Thus, the average first quintile household currently spends 10 percent of income on utility services (equivalent to US$11 per month), but this would rise to 17 percent of income (equivalent to US$19 per month) in the absence o f the cross-subsidy scheme (Figure 5). The water and electricity service together account for more than three quarters of household spending on utilities, while the water sector accounts for more than three quarters of the cash value o f subsidies receivedby households. Figure 4: Average expenditure on household utilities by income quintile I I I 40 z 35 30 & 25 a 20 15 3 I O 5 0 Income quinble (a) As percentage income (b) Inabsolute terms Source: DANE, Encuesta Calidad de Vida, 2003 adapted from Melendez, 2004 20 Figure5: Weight of utility subsidy by incomequintile I 18% -' 50 g 12% 10% c 8% 0 Source: DANE, Encuesta Calidad de Vida, 2003 Adapted from MelCndez, 2004 30. In the case of natural gas, a connection subsidy would appear to have a greater impact than a consumption subsidy. Although the natural gas service has been included within the same cross-subsidy framework, the social issues that arise with respect to this service are significantly different. In this case, the regulated price of natural gas i s already about half of that of its closes domestic substitutes (LPG and electricity), and hence it i s automatically more affordable for low income households. The key problem, however, relates to the high capital costs associated with switching to the service. These are estimated to be around US$400 to cover the connection charge (around US$160) and interior conversion (around US$240), which present a significant obstacle for low income households. Calculations, therefore, suggest that subsidizing natural gas connections as opposed to subsidizing natural gas consumption may be a more efficient way o f promoting broader access to this service. 31. In the case of local telephony service, the growing number of options to conventional household service make the cross-subsidy scheme less relevant. The traditional social policy objective for telecommunications has been one of promoting universal household service through cross-subsidies. However, the enormous structural changes experienced in the telecommunications sector, have naturally generated a wide array of options for low income consumers including widespread public telephony, special calling plans for low level use, and increasingly low cost mobile services under prepayment schemes. The existence of these alternatives makes the case for cross- subsidization of traditional fixed telephony household services less relevant. 32. The current cross-subsidy system is not very successfulintargeting resources towards poor families. The subsistence consumption thresholds used to provide an upper bound on subsidy allocations are high relative to average consumption even at the upper end of the income distribution. As a result, subsidy allocations may be significantly higher than strictly necessary. The stratification system on which the cross-subsidies are based does not appear to be well correlated with the income and poverty status o f the household (Table 2). Thus, inthe case of water and electricity subsidies, about 50 percent 21 o f beneficiaries are non-poor households, and only around 30 percent of the cash value of subsidies given i s captured by the poorest. Indeed, the overall impact of the scheme on the distribution of income i s mildly regressive (reflected in the small positive values of the quasi-Gini coefficient). Surcharges, on the other hand, do appear to fall disproportionately on the bills of richer households. There i s evidence that the number of households classified in the first two most subsidized strata has risen from 31 percent to 63 percent of total households during the last decade. Given that strata classifications are performed at the local level, and the surplus cross-subsidy resources are supposed to be transferred out of the local jurisdiction, the system clearly embodies political incentives for downward revisions of strata classifications. Table 2: Targeting performance of stratification system TargetingErrors Inclusion 1 Gross subsidy distribution Surchargedistribution Exclusion Share to poorest 1 Quasi-Gini Share to poorest Quasi-Gini Electricity 51.2 1.5 31.4 +0.02 8.7 +0.60 Gas 62.9 1.4 - - - - Telephone 25.6 21.5 28.2 +0.08 1.9 +0.62 Water 51.1 0.7 34.0 +0.11 8.7 +0.61 Notes: Poorest defined as bottom 40 percent of the income distribution. Data do not permit calculation of value of natural gas subsidies received. 33. Moreover, the cross-subsidy system presents major financial imbalances undermining the financial sustainability of the utilities. In the water, electricity and telephony sectors the cross-subsidy system presents a financial deficit equivalent to 20 percent, 12 percent and 2 percent of sector turnover respectively (Figure 6a). In the case o f the electricity sector, this deficit i s covered b y government transfers through a national solidarity fund. However, in the case o f the telephony sector no such transfer i s made, while in the case of the water sector equivalent transfers are made to municipalities that often fail to reach the local water utilities. These deficits can be understood in terms of the fundamental imbalance that exists between the percentage of contributors and recipients in the general population (Figure 6b). Thus, while in the natural gas sector about 75 percent of customers are net contributors to the cross-subsidy scheme, in the water sector only 15 percent of customers are net contributors. These imbalances also have a strong geographical pattern, given that in Bogota only 40 percent of customers belong to the first two most subsidized strata, whereas in rural areas 90 percent of customers belong to this group. Financial deficits can be expected to deteriorate over time as service expansion initiatives incorporate more low income households, and as a result o f the National Development Plan Law that introduced a freeze on utility tariffs charged to lower strata households. 34. Financial deficits could be reduced by adjusting existing parameters, but targeting improvements will require an overhaul of the stratification system. The current legal framework gives significant regulatory discretion over a number of key parameters of the utilities cross-subsidy framework. It is estimated that reducing subsistence consumption thresholds could reduce the financial deficit o f the cross-subsidy system by almost 20%, but would not do much to improve targeting performance. Furthermore, limiting Stratum 3 subsidies to households that meet additional poverty 22 criteria could reduce the financial deficit by 30% and slightly improve targeting performance. However, more substantial improvements in targeting performance would require much more substantial changes in the eligibility criteria. One possibility would be to replace the stratification system with the SISBEN indicator used to target other social programs. However, the SISBEN system i s in the process of being reformed to make it more similar to the stratification system. While the earlier SISBEN indicator i s shown to have very strong targetingproperties, its replacement performs no better than the current stratification. Alternatively, a completely new multi-dimensional poverty indicator could be developed. Simulations show that one based on housing characteristics, educational attainment, energy consumption and ownership of luxury goods could significantly improve targeting performance, although it would likely entail substantial administrative costs. Figure 6: Financial imbalance of cross-subsidyscheme O S u r c h a r g s Subsidies PJRalance mComper6ated balance (a) Financial deficit (b) Imbalance of contributors and recipients Source: Own calculations based on data from Regulatory Commissions. 35. Urban transport costs present a major burden for poor families, but existing government policies do little to assist them. Urban transportation expenditures represent some 24 percent o f first quintile income (or US$19 per month), providing these households with a low mobility level, equivalent to less than one journey per household member per working day. The government operates a number of policies directly or indirectly designed to make urban transport services more affordable. These include foregone tax revenues of US$70 million on diesel for the urban bus sector, and regulations requiring employers to spend US$293 million per year on transport subsidies for formal sector workers earning the minimumwage. However, only about 10percent of the beneficiaries of these policies come from the first two income quintiles, and they capture no more than 20 percent o f the value of these subsidies. 23 D.LEGAL, REGULATORY AND INSTITUTIONAL FRAMEWORKFOR INFRASTRUCTURE 36. Colombia's utility sectors rest on a solid legal foundation, rooted in the 1991 Constitution and developed extensively in the 1994 framework law. The 1991 Constitution emphasizes the central importance of utility services for public welfare, providing detailed guidance on their organization. This i s further developed in Law 142/94, which i s one of the most complete examples of utilities legislation in the region. ,However, sector-specific laws have only so far been developed for the energy sector, in spite of some attempts for telecommunications. Moreover, there have been some recent changes inlegislation that could weaken the overall legal framework inthe longer term. 37. The institutional framework comprises line Ministries, three Regulatory Commissions, and a cross-sectoral Superintendence. Line Ministers are primarily responsible for policy-making. However, they also preside both Regulatory Commissions and Boards of national state-owned enterprises, which leads to substantial conflicts of interest. There are three Regulatory Commissions dealing with each of the regulated sectors: energy, telecommunications and water. These operate by bringing together around a single table, key government representatives, such as Ministers, with technical figures, known as Expert Commissioners. Uniquely to Colombia, the task of regulation i s separated from that of supervision, with the latter being allocated to a cross-sectoral Superintendence (SSPD), whose responsibilities for enforcement extend across a broad array of areas. The SSPD i s also the competition authority for the utility sectors, notwithstanding the existence of a broader anti-trust agency. Cross-cuttingregulatory issues 38. The common regulatory framework established for the utilities faces a number of generic problems that affect all of the sectors to varying degrees. Following ten years of experience with the new regulatory model, certain weaknesses have become apparent. The government recognizes the need for some second-generation reforms to the legal andregulatory framework, and has initiated a formal review process. (a) Conflicts of interest. As noted above, the current system leads to significant conflicts of interest for the state in its role as policy-maker, regulator, and service provider. This makes it particularly difficult to achieve a level playing field for the competition between public and private operators that currently takes place in the electricity and telecommunications sectors. The most definitive resolution o f this problem would be to undertake further privatization of some of the major state-owned utilities; and the government has recently announced its intention to do this. One strategy that may be worth considering for those 24 utilities remaining in public hands would be to create a single holding company that exercises government ownership in all of the state-owned enterprises, thereby weakening the link with the immediate line Ministries. (b) Fuzzy boundaries. There i s an on-going debate as to where the boundary between policy-making and regulation should be drawn, that has led to highly visible conflicts between line Ministries and Regulatory Commissions. One of the underlying reasons for this is the absence of a more detailed definition of the precise responsibilities assigned to the policy-making and regulatory functions. In addition, there are no clear rules governing the delegation of functions between the Executive and the Regulatory Commissions, so that specific functions can be reassumed by the Executive at any time. The situation i s further exacerbated by limitations in the technical capacity of the Regulatory Commissions (which sometimes invites political intervention), and the lack of a consistent line from policy makers (which leaves regulators without the necessary guidance). While a definitive solution to this problem i s likely to entail a stronger institutional separation between the Ministries and the Regulatory Commissions, a helpful first step would be establish an ad hoc Commission to build consensus around a more detailed and exhaustive description of the functions accruing to policy-makers and regulators. (c) Human resources. The regulatory and supervisory institutions face a number of serious human resource problems, in particular with regard to the caliber of Expert Commissioners and Superintendents. One issue i s that the established profiles are too restrictive, since they require candidates to have experience o f working in the regulated sector but not during the year immediately prior to their appointment, thereby ruling out a large section o f the field. In addition, remuneration levels are not high enough to attract professionals with the necessary level of expertise, given that they lie below that o f senior managers inthe state-owned utilities that they regulate, and of other public sector employees with comparable levels of responsibility (such as the Board o f the Central Bank). Moreover, the fact that candidates are selected unilaterally by the President without any prior quality screening does not help to guarantee the best possible selection. Finally, notwithstanding the fact that the Regulatory Commissions and the Superintendence are self-financed by sectoral levies they remain subject to fiscal restrictions on their expenditures that prevent them from maintaining the minimumbase of human resources they need to be able to function effectively. (d) Appeals mechanism. The administrative appeal of regulatory decisions is not currently possible in Colombia due to the absence o f any administrative authority over and above the Regulatory Commissions. 25 Thus, the judicial channel is the only real mechanism available to regulated companies wishing to challenge regulatory decisions. However, judicial processes are extremely slow and unpredictable, as well as lacking specialized knowledge of the infrastructure sectors. Hence, in practice, there i s no meaningful recourse for regulated entities. In order to remedy this deficiency, a number of options are under consideration including expert arbitration, and the creation o f a specialized judicial chamber. Nevertheless, both options appear to run into conflict with the Colombian legal system. Another possibility would be to introduce an administrative appeals channel by creating a cross-sectoral supra-Commission composed of representatives of each of the three Regulatory Commissions with the explicit mandate of handling administrative appeals. The supra- Commission would be convened on an occasional basis to address regulatory appeals as they arise. (e) Superintendence. Since its inception, the SSPD has been plagued by numerous problems. In response to this situation, the current administration has launched a major initiative to improve the efficacy of the institution, and as a result many o f the problems described below are beginning to show signs of improvement. The SSPD's responsibilities include supervising performance of regulated utilities, acting as an anti- trust agency, intervening and liquidating non-performing utilities, and acting as an appeals body for consumer complaints. The SSPD supervises on behalf of other government agencies, in particular the Regulatory Commissions, but plays no role in designing the indicators that it i s required to monitor. As a result, the monitoring frameworks are often unrealistic, and place conflicting informational demands on the SSPD and the regulated companies themselves. There are also inconsistencies between the monitoring framework enforced by the SSPD, and those of the other internal and external auditing requirements placed on public enterprises. Moreover, the SSPD lacks the resources to exercise such extensive supervision of the more than 1,000 operators for which it i s responsible, and lacks also sanctions o sufficient weight to motivate compliance with its dictates. One o f the most challenging responsibilities of the SSPD i s the intervention and liquidation o f ailing public enterprises. A key problem in the past has been the absence of an early warning mechanism to allow remedial measures to be taken to avoid the need for intervention. The SSPD i s making progress with the development of such a mechanism, through a system of classifying companies according to the risk that they represent, and entering into special performance agreements with high risk cases. These reforms are positive and should be continued. However, the fact that there continues to be a need for the SSPD to intervene ailing enterprises i s essentially a reflection of underlying structural problems that have yet to be solved within the sectors themselves, due to weak managerial incentives and diseconomies o f scale. The ultimate solution to the problem o f intervention, therefore, i s to resume the agenda of structural reform. 26 (0 Anti-trust agencies. There are currently two competition agencies: the SSPD with responsibility for the utilities; and another for the rest of the economy, the Superintendence of Industry and Commerce (SIC). Furthermore, the Ministry of Communications has some responsibilities for anti-trust issues in the telecommunications sector. This creates conflicts and ambiguities ofjurisdiction. Moreover, none of these agencies has the human resources or technical capacity to discharge the anti-trust function effectively. It would therefore be desirable to concentrate anti- trust authority in a single specialized institution, such as the SIC, and focus efforts on building the capacity of that institution. The Regulatory Commissions and the Superintendence would continue to have a critical role to play b y being more proactive in using ex ante regulatory measures to prevent the concentration of market power, and in referring cases to the SIC when anti-competitive behavior has been detected. Nevertheless, addressing the substantial concentration of market power in the electricity and telecommunications sectors (recall Table 1 above) may ultimately require structural measures that go beyond the remit of the anti-trust institutions. Sector-specific challenges 39. In addition to these cross-cutting regulatory issues, the four utilities sectors present a wide range of specific challenges of their own. These range from the promotion of competition in electricity and telecommunications, to the elusive goal of cost recovery for the water sector. Electricity 40. The electricity sector is particularly vulnerable to anti-competitive practices. This is due to the persistent vertical integration of utilities, combined with substantial concentration of market power and the presence of major multi-utility operators. Market power i s 'further amplified by the limited participation of thermal plant under normal meteorological conditions, as well as transmission outages that tend to lead to further geographical segmentation of the market. The current legal framework contains many of the remedies to this situation, including the requirement for vertical unbundling, the upper limit of 25 percent on market share, and the requirements for accounting separation. These tools need to be more strictly applied in practice, and additional mechanisms need to be created. 41. Another key issue is the lack of incentives for the private sector to invest in the next generation of thermal plants. These stem from the problem of covering the 27 fixed costs of plants that are only occasionally dispatched, together with the fixed cost of the upstream gas transportation infrastructure required to fuel them. It also reflects the considerable instability of the regulatory environment, with an average of eight changes per year in the rules governing the wholesale electricity market. Indeed, a recent survey found that 80 percent of private investors in Colombia's electricity generation sector were highly dissatisfied with their experience and would choose to exit the country if possible. The government's recent decision to support the development of large hydroelectric plant further tilts the playing field against private sector finance of electricity generation. 42. A definitive solution must be found to the problem of dysfunctional distribution utilities, with their negative consequences for the public finances. A solution to this problem i s likely to entail some consolidation of the smaller utilities into more viable regional undertakings, a more effective intervention and liquidation mechanism for ailing enterprises, and greater private sector participation albeit under more limited arrangements such as management contracts in situations where private investment may not yet be feasible. Naturalgas 43. The relative novelty of the natural gas sector makes it an arena where strategic policy questions are still being worked. A central question i s the extent to which the state should be involved in further promoting the growth of the sector into areas such as industry, transport and export markets. As a result, the sector has been particularly susceptible to conflicts between policy-making and regulatory bodies, as evidenced by the controversy between the line Ministry and the Regulatory Commission over the issue of gas exportation. Telecommunications 44. The central issue in the telecommunications sector is the removal of a number of existing barriers to competition. The pattern of reform in the sector, with immediate liberalization and limited private participation, made it difficult for new private operators to compete with well established public incumbents. In addition, the high price attached to licenses for new entrants, together with extensive government bailouts o f failing public enterprises have further prevented the establishment o f a level playing field between public and private operators. 45. The sector also needs a new legal framework that eliminates the artificial dichotomy between fixed line telephony and recent competitors, such as mobile and internet. The current legal framework treats fixed line telephony in the same way as water and electricity services, while providing a completely different framework for new services such as mobile and internet. This situation i s at odds with the reality of growing 28 technological convergence and thereby distorts competition between these alternative services. A draft of a single unifying telecommunications law designed to address these problems already exists and needs to be finalized and implemented in a manner consistent with the framework of the imminent FTA. This legal harmonization should be accompanied by a corresponding simplification of the institutional framework overseeing the sector, so that all of the relevant functions are concentrated in the line Ministry, the Regulatory Commission and the anti-trust agency, removing telecommunications from the jurisdiction of the SSPD. Water 46. The highly decentralized nature of the water sector complicates the process of regulation and supervision from the center. As a result there i s a need to adopt a differentiated and phased approach that groups utilities into three or more size-based clusters and develop appropriate regulatory mechanisms for each case. Given that more than half o f the Colombia population is served by 40 large utilities, significant improvements could be achieved by concentrating regulatory efforts there, based on a traditional framework of natural monopoly regulation. In the medium-sized utilities, a simplified approach that focuses on a subset of more critical parameters and issues would be desirable. Inthe case o f the large fringe o f small-scale utilities, it i s even questionable whether regulation (as opposed to technical assistance) i s the appropriate form of engagement. 47. Significant progress has been made inreaching financial sustainability, albeit over a much longer period than originally anticipated. This i s a major achievement given the large and regressive nature of the tariff increases that were required, ranging between 100-300 percent for the lowest socioeconomic strata. However, this progress has been mainly confined to the larger operators, with smaller utilities subject to political interference and hence more reluctant to embrace the goal of financial sustainability. More recently, the provisions of the National Development Plan Law have frozen the tariff convergence process among the lower socioeconomic strata. There has been significant controversy surrounding the methodology for tariff determination, particularly regarding adjustments to be made for cost efficiency and the appropriate rate of return on capita1. 48. Further efforts need to be made to develop a framework for regulating quality of service. The absence o f a well defined framework for regulating quality of service i s a serious gap, particularly given the deficient quality of water in smaller towns. Moreover, it prevents the SSPD from effectively playing its role in this critical area of consumer protection. 29 Transport 49. The development of the legal, regulatory and institutional framework in the transport sector lags behind that of other infrastructure services. It i s moreover rendered particularly complex by the huge diversity o f market structures and institutional arrangements existing across different modes. However, important advances have been made during the last decade. In particular, sectors such as ports, toll roads and freight railroads have moved towards a modern regulatory paradigm based on concession contracts administered by regulatory entities. Results in the ports sector, in particular, have been particularly impressive. There is clearly scope for extending some of these lessons to transportation services, such as trucking and buses, which are still governed by a traditional regulatory framework and characterized by extreme informality. The Trunsmilenio experience provides an interesting model for the bus sector. The trucking industry would benefit from restructuring followed by greater liberalization of its regulatory regime. Finally, it will be important to build increasing capacity for integrated transportation planning within a unified framework that fully takes into account the complementarities and relative strengths of the different transportation modes, ensuring a level playing field between them and facilitating multi-modal use of the transport system. 30 E.PUBLICANDPRIVATEFINANCINGOFINFRASTRUCTURE 50. Investment in infrastructure,from both public and private sources, peaked at 4.4 percent of GDP in 1997, falling back to 2.6 percent of GDP by 2003. Total investment in infrastructure averaged almost 3 percent of GDP during the 1990s (Figure 7). Public investment declined by 34 percent from its peak of 2.4 percent in 1997, as a result o f the macroeconomic crisis facing the country. Private investment dropped by 50 percent from its peak of 2.0 percent in 1997, reflecting Colombia's loss of investment- grade rating, linked to investor concerns about macroeconomic stability and the security situation. Investment in the energy sector peaked at 2.3 percent of GDP in 1996, while investment in the telecommunications sector peaked at 1.6 percent of GDP in 1998. Investment in the transport sector has been more volatile, cycling between 0.5 percent and 1.3 percent of GDP, while water sector investment has generally been below 0.5 percent of GDP. Figure 7: Breakdown of infrastructure investment ~~ (a) By sector (b) By financing source Source: DNP Notes: Transport includes roads, railways, ports and airports, while energy includes both electricity and gas, excludes mining and oil. Publicinvestmentininfrastructure 51. The water and, in particular transport, sectors continue to have a major relianceon routine fiscal transfers to finance investmentprograms. However, inthe energy and telecommunication sectors, public investment i s largely self-financed through user charges. Although fiscal liabilities for the bail out of failing state enterprises continue to arise on an occasional, but repeating, basis. (a) Electricity. The fiscal burden associated with the electricity sector has fallen substantially, but dysfunctional public enterprises continue to make significant demands. As o f 1990, the electricity sector accounted for one 31 third of Colombia's public debt stock, a figure which has fallen to less than 5 percent of the public debt stock today. Indeed, the only regular fiscal transfer to the electricity sector i s the payment to cover the deficit of the cross-subsidy system currently amounting to US$77 million per year. Nevertheless, there i s a substantial volume of occasional expenditures associated with particular liabilities and financial crises, amounting to almost US$200 million per year on average. Over the last decade, these have included bailout operations to bankrupt public enterprises (US$250 million per year), payments to support various privately funded hydro and thermal generation projects (US$200 million per year), and `loans' to allow loss-making distribution utilities to cover their obligations to the wholesale market (US$32 million per year). (b) Natural gas. There are no fiscal transfers to the natural gas sector, but major public investments are undertaken b y the state-owned enterprises Ecopetrol and Ecogas. An implicit subsidy of US$l55 million exists to the Build-Operate-Maintain-Transfer contracts (BOMT) used to develop the gas transportation system. (c) Telecommunications. Public enterprises finance a falling, but still significant share of investment in the telecommunications sector, and have generated significant liabilities for the state. While there are no fiscal transfers to the telecom sector, public enterprises continue to invest US$360 million per year, primarily in fixed line telephony. The share of public investment in the sector total has fallen from 99 percent in 1992 to 36 percent in 2002, still more than double the regional average of 15 percent. However, the state also bears major liabilities associated with the recent bailout of the national telecommunications incumbent, Telecom. In addition, there are major claims of uncertain magnitude against guarantees provided under a number ofjoint venture contracts with the public sector. (d) Water. Transfers from central government to municipalities remain the main source of investment financing. in the water sector, amounting to US$278 million in recent years. Rules governing the determination of fiscal transfers ensure that poorer municipalities receive three times as much per capita as the larger wealthier municipalities. In spite of this advantage, depressed tariffs and deficient management in the smaller water utilities means that these still only manage to invest about one third as much per capita as do their larger counterparts. Furthermore, diseconomies of scale in the smaller utilities mean that they achieve more limitedresults from a given volume o f investment resources. 32 (e) Transport. Public resources devoted to the transport sector have fallen by 50 percent in real terms from a peak of US$1,590 million in 1996. About 80 percent of these resources are absorbed by the roads sector, most of which i s undertaken at the central government level. Almost half of the budget of the national roads agency i s absorbed by emergency works and payment of traffic guarantees under first generation toll road contracts, leaving US$24,000 per kilometer for maintenance plus rehabilitation of the national road network. At the sub-national level, there i s huge variation (of ten to one) in the amount devoted to road maintenance across different departments. However, the average value of US$2,000 per kilometer per year i s well below the prudent benchmark o f US$lO,OOO per year for preventive maintenance of paved and unpaved roads, storing-up major liabilities for rehabilitation. 52. Public investment in infrastructure is not always efficient, but tries to be equitable. Colombia has developed rigorous procedures for evaluating and screening public investment projects, although these are sometimes circumvented for political reasgns. About 80 percent o f public investment in infrastructure i s channeled through the major national and municipal enterprises, while of the remaining 20 percent a growing share i s being executed by sub-national governments. The efficiency of public investment i s therefore often compromised by the inefficiency o f the state-owned enterprises that are largely responsible for delivering it, particularly in the case of electricity distribution and smaller water utilities. Wherever possible, the state makes conscious efforts to improve the equity with which public investment resources are distributed across regions and socioeconomic groups. However, this does not prevent investments from being highly concentrated in the larger urban areas, because of their much greater capacity for generating tariff revenues. 53. The instability of public finance is a major concern, promptingthe search for more reliable budgetary mechanisms. Budgetary allocations for infrastructure (particularly transport) tend to be unpredictable, moreover inefficiencies in public contracting processes can lead to huge gaps between initial budgetary appropriations and ultimately executed investments. In order to circumvent these problems there has been growing use o f `future appropriations' and `earmarked sector funds', with the former currently absorbing about 40 percent of the available fiscal space for infrastructure spending. These mechanisms succeed inproviding greater budgetary stability, although at the cost of reduced flexibility. 54. Substantial contingent liabilities incurred in early private sector experiences, have prompted the development of a sophisticated accounting framework. Inorder to attract the first wave o f private investors, Colombia provided guarantees resulting in payments of up to US$4 billion over the period 1993/14. In retrospect, these guarantees appear poorly designed and perhaps overly generous. However, at the time, they seemed necessary to overcome private sector perceptions o f risk, and the resulting investments clearly brought substantial social benefits. Following the establishment of a regulatory framework, significant private investment took place with substantially reduced 33 guarantees. Moreover, a prudent fiscal accounting and budgeting framework has now been established, which should serve to limit their future accumulation. 55. There is evidence that public investment has borne the brunt of fiscal adjustment in Colombia. Given the country's burgeoning pension liabilities, it i s likely that there will be further pressure for adjustment inthe coming years. However, cutbacks in public investment may ultimately prove fiscally counter-productive due to negative consequences for economic growth and hence ultimately for tax revenues. The IMF has been evaluating whether commercially-run public enterprises might be excluded from the fiscal accounts, although most of Colombia's public enterprises do not yet meet all of the relevant criteria. Since the public enterprises currently make a positive contribution o f 0.9 percent o f GDP to the fiscal balance, this would necessitate off-setting reductions elsewhere inthe budget. Role of private finance 56. Colombia has succeeded in attracting US$14.5 billion of private investment in infrastructure; half of it from international investors. Green field projects and concessions have played a more important role than conventional asset sales. Private capital flows have declined substantially following the loss of investment grade in 1998. Investors express serious concerns about regulatory and judicial uncertainty, as well as the conflicts of interest inherent in the state's double role as regulator and service provider. They remain nevertheless optimistic about the future, and see a growing role for bonds (as opposed to bank debt) as a means o f financing infrastructure projects. 57. In the process, the country has accumulated a diverse range of experiences with private participation across the infrastructure sectors. Private participation has been most intense in the energy sectors, and conspicuously absent in the telecommunications sector. Electricity. Private participation ininfrastructure has been concentrated in the energy sectors, where the bulk of the asset sales also occurred. Over the period 1993/01, there were US$1.8 billion o f asset sales (primarily in the generation sector) and a further US$3.1 billion of capitalization (primarily in the distribution sector). Future private investment in thermal generation i s jeopardized by the volatile regulatory environment. Moreover, greater efforts are needed to prepare failing public distribution utilities for private sector participation. Natural gas. The private sector played a pivotal role in the rapid expansion of the natural gas market in Colombia, with 50 percent o f the new transportation network developed under B O M T contracts. In the distribution sector, existing utilities were sold off and green field concessions awarded for network development innewly served towns. 34 Telecommunications. There were a number of unsuccessful attempts to privatize the major telecommunications operators in the 1990s, so that private participation i s largely combined to mobile and value added services. Factors currently constraining further private participation are the complex regulatory environment described above, the highcost of licenses for new entry, and the difficulty of competing with major public incumbents. Water. By regional standards, Colombia has seen a significant amount of private participation in the water sector both through the creation of mixed enterprises in the larger cities, and increasingly via public-private partnerships in smaller cities. However, while private participation has helped to bring about significant improvements in utilities management, it has never proved to be a major source of investment finance. Transport. Colombia's major program of toll road concessions i s now close to completion, having now covered just about all o f the high traffic corridors. The `first generation' contracts secured investments of U S $ l billion but triggered US$500 million of traffic guarantees. Subsequent contracts have substantially curtailed guarantees, limiting them to critical components of risk that cannot be readily transferred. A series o f successful concessions for sea ports leveraged private investment of US$600 million and resulted in major improvements in operational performance. 58. The government intends to regenerate private financing of infrastructure, although international investors will probably require some degree of credit enhancement. A list of US$6.5 billion of green field projects identified as suitable for private participation was recently published. In addition, the government announced its new policy of zero tolerance towards failing public enterprises, and its intention to divest from a number of major public enterprises. Given Colombia's loss o f investment grade rating, it seems likely that some degree of credit enhancement will be required to attract (particularly international) investors back into infrastructure projects. However, this does not imply a return to the generous guarantees of the early 1990s. A wide range of more sophisticated financial instruments i s now available for targeting such enhancements towards the most critical elements o f risk. 59. However, the exponential growth of pension deposits during the last decade offers a possible new source of long-term domestic financing for infrastructure projects. Colombian pension deposits reached US$9 billion in 2003, and are will grow at a rate o f US$lOO million per year into the medium term. However, harnessing such financing in a responsible manner will entail a number o f significant reforms in the pension industry, including the offering of multiple funds with a range o f risk-reward 35 ratios, the establishment of specific limits for infrastructure investments, and the establishment of a more sophisticated risk-rating capability. The infrastructure sectors themselves will also need to take measures to make their offerings more appealing to retirement investors, including financial repackaging of securities to separate out the safer post-construction income stream, and to pool together securities across infrastructure sectors. 60. The overall balance between funding sources and investment needs varies enormously across and within the infrastructure sectors. While the telecommunications sector shows a substantial surplus of resources relative to investment requirements, the water and energy sectors present modest but manageable deficits, and the transport sector i s woefully under-funded. Even for those utility sectors that are in overall balance, important differences exist across different market segments or groups of enterprises. In particular, generators affected by early 1990s Power Purchase Agreements, electricity distribution utilities outside the major urban areas, and some of the larger water utilities, present serious and persistent deficits preventing them from financing the investments that they need. In natural gas and mobile telephony, where investments are generally funded under a market paradigm, a significant step-up in capital expenditure will be required if plans to export natural gas and develop new mobile telephony networks are to be fulfilled. Moreover, the situation in the water and telecommunications sectors could rapidly deteriorate towards a deficit position, once the large and under-funded pension liabilities of public enterprises in both sectors start to result in significant retirement claims. 61. Overall, bringing together the productive and social investment needs identified above, leads to a desirable investment agenda of US$4,000 (or 5.1 percent of GDP). It i s important to note that 31 percent of the total (or 1.6 percent of GDP) corresponds purely to maintenance and rehabilitation expenditures for the water, electricity and transport sectors. The remainder corresponds to new infrastructure investment, of which 35 percent (or 1.8 percent of GDP) i s productive investment and 33 percent (or 1.7 percent o f GDP) i s social investment. The transport sector accounts for the largest share of the total with 46 percent (or 2.3 percent o f GDP), followed by electricity with 23 percent (or 1.2 percent of GDP) and telecommunications with 19 percent (or 1.0 percent of GDP). It i s estimated that about 33 percent o f these investment requirements (or 1.7 percent of GDP) would be met directly by the private sector. If the remainder of the publicly owned segments o f the electricity and telecommunications industry were privatized, this would transfer an additional US$508 million o f the investments into the private sphere, raising the share that could be financed by the private sector to 45 percent of the total (or 2.3 percent of GDP). 36 62. However, this agenda lies well above historicalinvestment levels, even during the peak period of the late 1990s. It i s necessary to confront this bottom-up desirable investment agenda, with the country's overall budget constraint for infrastructure investment. As noted above, Colombia has managed to sustain one of the higher and more stable levels of infrastructure investment in the region. Notwithstanding, Colombia has typically managed to invest no more than 3 percent of GDP in infrastructure, of which around 2 percent of GDP from the public sector and 1 percent of GDP from the private sector. Moreover, even at the historic peak during the late 1990s, public finance for infrastructure amounted to no 4 percent of GDP, of which 2.5 percent of GDP from the public sector and 1.5 percent of GDP from the private sector. The implication i s that the desirable investment agenda falls 1-2 percentage points of GDP short of the resources likely to be available, even under a best case scenario. Given that Colombia will continue to face a tight fiscal balance over the next few years, it will not be possible to fulfill this investment agenda inits entirety. 63. As a result there is an evident need to prioritize, while at the same time exploring options for leveraging additional resources. As far as prioritization i s concerned, maintenance and rehabilitation expenditures to preserve and restore existing assets should generally take precedence, since they typically bring significantly higher rates of return than investments in new assets. On the public sector side, the estimates suggest that 1.4 percent of GDP i s needed to maintain and rehabilitate assets, which would leave only 0.6 percent of GDP for financing o f new investments by the public sector. However, increased privatization of commercially viable enterprises currently operated by the state, could potentially open up an additional 0.6 percent of GDP doubling the public resources available for the creation of new infrastructure assets. 37 F.RECOMMENDATIONS 64. Based on the above analysis, the main report develops some forty recommendations, most of which are fiscally neutral. A summary overview of these recommendations classifiedby issue andby sector i s provided in Table 3 below. The vast majority o f these recommendations are fiscally neutral, or suggest ways o f makingbetter use o f resources within the current envelope, as well as leveraging additional resources from the private sector. The number given in parenthesis after each recommendation enables the reader to cross-reference it to the main report, where a fuller description can be found. Although many o f the recommendations relate to very specific sectors and issues, it i s important to understand how they inter-relate with each other in support of the two over-arching objectives of promoting economic growth and competitiveness as well as improving equity and social welfare. 65. The bulk of the recommendations are ultimately linked to the objective of promoting economic growth and competitiveness (Figure 8). In order to meet this objective, Colombia needs to work consistently across three broad areas: the refocusing of public investment; the reactivation o f private finance; and the revision of a number of regulatory aspects with direct impact on competitiveness. Figure 8: Integrating recommendations for growth and competitiveness I GROWTH& COMPETlTIVENESS Source: Own elaboration 66. The first issue is to refocus public investment so that it concentrates on those areas that the private sector cannot finance, and make effective use of existing public resources. On the one hand, this means reducing the role of the state in the telecommunications and electricity sectors, which have been making excessive fiscal demands, and where there i s significant scope for increased private participation. On the other hand, it implies improving the efficiency and efficacy o f the existing resource envelope in more traditional public finance areas such as transport and water. This could 38 be achieved, for example, by providing more stable financing and agile contractual mechanisms for road maintenance to ensure that the roads agency i s able to make use of its full budgetary allowance. Another key issue i s to improve the framework of fiscal transfers to sub-national governments, so as to improve accountability for transfers earmarked to the water sector, and to strengthen incentives and obligations to allocate adequate sub-national resources to maintenance of secondary and tertiary roads. Finally, a review o f the parameters in the national cross-subsidy framework could help to reduce the recurring deficit associated with this policy. 67. The resources thereby released should be focused on addressing strategic `public good' infrastructure bottlenecks for the productive sector, as well as improving road maintenance. Key examples include road and sea access to major ports, development of logistics zones and airport freight facilities, and creation of multi-modal facilities to permit integrated use of road, rail, river and sea transport channels. Such investments need not necessarily be very large, but have important strategic value, and in some cases could potentially be co-financed with the private sector. In addition, additional resources need to be allocated to road maintenance to prevent the accumulation of major future liabilities for rehabilitation. 68. The second issue is to reactivate private finance for infrastructure to take on investments in commercially viable sectors that are beyond the scope of the current restricted fiscal envelope. These include areas such as electricity generation, telecommunications, and port facilities. However, in order to revive private capital flows, particular attention will need to be paid to addressing deficiencies in the sectoral regulatory frameworks. There are three particularly critical areas. First, the need to provide a more stable regulatory environment for electricity generation, without which the private investment-n which the future expansion of thermal generating capacity i s premised-is unlikely to be forthcoming. Second, the need to address a number o f more general deficiencies in the regulatory framework including strengthening of the relevant agencies, an improved anti-trust framework, clearer boundaries between policy and regulation, as well as a more effective appeals channel. Third, the need to modernize the legal framework for telecommunications to provide a more level playing field between technologically converging services, such as fixed line, mobile telephony, and internet. 69. Any reactivation of private investment needs to pay attention to financing mechanisms. Moreover, given Colombia's loss o f investment grade, some carefully designed limited credit enhancements may also be needed to attract international investors. These should not repeat the major contingent liabilities incurred in first generation pubic-private partnerships, but should rather focus on isolating and attenuating key dimensions of risk. In addition, Colombia's burgeoning pension fund deposits provide an unprecedented opportunity to harness long term domestic savings to finance infrastructure projects. However, doing so will require significant modifications to the regulatory framework for pensions, as well as suitable structuring of infrastructure investment opportunities to provide the necessary low risk ratings required by these types of investors. 39 70. The third issue is to address a number of specific regulatory issues that have a direct impact on the competitiveness of the infrastructure sectors. Most salient among these is the development of a policy to modernize the trucking industry that carries 80 percent of Colombia's internal freight, and presents significant problems in terms of the efficiency and quality of services provided. However, there are also important considerations relating to the regulatory framework for self-generation by large consumers that currently limit the scope for taking full advantage of co-generation possibilities. 71. Given Colombia's relatively strong performance on infrastructure social policy, fewer recommendations are offered in this area (Figure 9). Most o f these are concerned with regulatory issues that restrict the effectiveness o f current policies. A few examples can be given, such as: the need to ensure greater stability in the flow of existing financial resources to rural access programs in the electricity and telecommunications sector; the fact that tariff regulations for electricity distribution in the interconnected system do not provide adequate incentives for distributors to invest in coverage expansion; and the existence of restrictive technical norms in the water sector that , prevent innovation into lower cost technologies. Figure 9: Integratingrecommendations on social welfare and equity I /, SOCIAL WELFARE & EOUITY 1 FINANCE FOR SOCIAL POLICY REGULATION Source: Own elaboration 72. Moreover, notwithstanding extensive policies to promote affordability of public services these are found to be very poorly targeted with respect to low- income households. Thus, as much as 60 percent o f utility cross-subsidies and 80 percent of urban transport subsidies are being captured by upper income groups. In the case of utility cross-subsidies, targeting improvement i s not straightforward, and i s likely to require a major review o f the stratification system. Inthe case of urban public transport 40 subsidies, a significant rethink will be needed to identify better ways o f channeling these resources towards lower income households. 73. Finally, although financing of social programs is generally good, two areas of vulnerability are identified. The first i s the volume of financing for rural electrification, which appears to be small relative to the size of the unserved population. Simulations suggest that historic rates of expansion o f rural electricity are not high enough to meet universal access objectives within a reasonable time frame. The second i s the issue of rural roads, where substantial progress has been made under the `Roads for Peace' initiative that i s now drawing to a close, and there i s need to find ways of sustaining maintenance efforts inthe future. Table 3: Overview of recommendations by sector and by theme I Growth Social Finance Regulation Utilities Reform design o f cross- Improve commercial Reducegovernment's subsidy framework [2e, orientation of public conflicts of interest [4a] 3dl enterprises [Sc] Clarify boundaries between Reduce regulatory risk for policy-makingandregulation privateparticipation; [4bl especially inelectricity generation [4i, Sd] Strengthen human resources for 3 RCs andSSPD [4c] Increaseprivate participationin Improve functioning of commercially viable SSPD [4d] activities; especially electricity and Reduceneed for enterprise telecommunications [Sg] intervention[4e] Consolidate anti-trust in a single institution [4fl Create an effective appeals mechanism[4g] Strengthen application of anti-trust regulation; especially inelectricity and telecommunications [4h] Electricity Improve security of Revisetariff regulations to Reform failing distribution electricity supply [Zc] strengthen incentives for utilities [4j] coverage expansion on the Review regulations on SIN [3i] self-generation [2fl Natural gas Replace cross-subsidies Create incentives to expand with connection subsidies gas transportation for new thermal plant [2d] Telecom Consider needs of Phase out cross-subsidies Harmonize financial Developa legal framework productive sector in for local telephony [3fl contributionsfor rural consistent to technological rural telecom programs telecom [3j] integrationo f the sector [4k] [2nl Water Revisetechnical norms to Differentiateregulatory permit use o f low cost regime for different groups o f altemative technologies utilities [41] [3cl Develop regulatory Develop explicit policy to framework for quality of promote access inrural service [4m] areas [3h] 41 Table 3: Overview of recommendations by sector and by theme (continued) Growth Social Finance Regulation Transport Increasefinancingto Rethinksocialpolicy Developa mechanism Developcapabilityfor roadsector [Za] for urbantransport [3g] for stable funding of integratedmulti-modal roads sector [Sa] planning[40] Developpolicy for trucking industry [2b, Adjust regulatory 4 4 framework for ports to preparefor FTA [4p] Investin removing strategictransport bottlenecks for I intemationaltrade [2hI Multiple I Increasevolume of Develop multi-annual sectors funding for rural access budgeting programs;especially electricity,water and instruments [5b] roads [3a] Developmore Improvereliability of sophisticatedcredit funding for rural access enhancements [Se] programs;especially electricityandtelecom Reformregulatory [3bl frameworkof pension > funds [Sfl Improveaccountability of sub-national govemmentsfor fiscal transfers [Sh] Source: Own elaboration 42