47884 noTE no. 46 ­ Mar 2009GRIDLINES Sharing knowledge, experiences, and innovations in public-private partnerships in infrastructure What drives private sector exit from infrastructure? Economic crises and other factors in the cancellation of private infrastructure projects in developing countries Clive Harris and Kumar V. Pratap T he private sector exits only a fraction sold or transferred its economic interest in the of private infrastructure projects before project to the public sector; the private company the contract ends. Yet such cancella- physically abandoned the project (such as with- tions can have a sustained impact on a coun- drawing all staff); or the private company ceased try's program of public-private partnerships, operation or halted construction for 15 percent or reducing the private sector's confidence in more of the license or concession period, follow- the government's commitment as well as the ing the revocation of the license or repudiation of government's confidence in the robustness the contract. On the basis of this definition, they and "value for money" of these arrangements. found that only 1.9 percent of projects reaching Econometric analysis shows that macroeco- financial closure in 1990­2001, representing 3.2 nomic shocks nearly double the cancellation percent of total investment commitments, had rate. As today's global financial crisis greatly been canceled by the end of that period. About a increases the cost, and reduces the availability, third of the cancellations related to the Mexican of project financing, the number of cancella- toll road program. tions could grow. That would have implications for the role public-private partnerships can play This note, relying on the same definition of proj- in meeting the infrastructure needs of develop- ect cancellation, updates these findings using ing countries. data through 2006 from the PPI Project Data- base. Of the more than 3,800 projects reported Trends in investment in infrastructure projects, by the database as reaching financial closure in and renegotiation of the concessions and contracts 1990­2006, 179 were canceled by 2006. These underpinning them, have received considerable accounted for 4.7 percent of all projects and 4.9 attention. Less attention has been given to the percent of investment commitments. On average, premature exit of the private sector from public- projects were canceled 5.3 years after the date of private partnerships, including the factors that financial closure. influence this. In one study looking at this issue, Harris and others Recent trends (2003) reviewed the extent of the cancellation of private infrastructure projects using information PPIAF Approved Logo Usage Although the share of infrastructure projects in from the World Bank­PPIAF Private Participa- developing countries that are canceled has varied tion in Infrastructure (PPI) Project Database. This over time, recent years have seen an increasing database collects and disseminates information on Logo - Black trend in cancellations. While the rate of private infrastructure projects with private participation in low- and middle-income countries.1 A project Clive Harris is the infrastructure policy advisor in the East was deemed to have been canceled if, before the PUBLIC-PRIVATE INFRASTRUCTURE ADVISORY FACILITY Asia and Pacific Region of the World Bank. Kumar V. Pratap end of the contract period, the private company is a PhD candidate in the School of Public Policy at the University of Maryland. Logo - 1-color usage (PMS 2955) Helping to eliminate poverty and achieve sustainable development through public-private partnerships in infrastructure PUBLIC-PRIVATE INFRASTRUCTURE ADVISORY FACILITY 2 sector exit is low, that rate doubled between 2001 est rate of cancellation by number of projects, at and 2006. more than double the average for all energy proj- ects. The sub-sectors that have the lowest rates of Rates of project cancellation vary considerably cancellation are natural gas and port projects. among sectors (table 1). Projects in the water and sewerage sector are most prone to cancellation. About 9 percent of these projects--representing Factors driving cancellations more than a quarter of investment commitments in the sector--were canceled by 2006. This was These sectoral patterns are broadly intuitive. For significantly higher than the overall rate of cancel- example, the water sector has the lowest levels of lation. Distribution projects and bulk water cost recovery among infrastructure sectors, and treatment plants had similar rates of private sector the involvement of the private sector in provid- exit. But canceled distribution projects accounted ing water supply is sometimes viewed negatively for a larger share of investment commitments-- by stakeholders. Power distribution projects bring about a third of the total in the subsector--driven far greater political economy issues than do well- mainly by the large size of a few of the canceled structured generation projects. Sectors such as projects. ports and airports are typically highly commer- cialized and deal with large customers that are as Energy projects were the least likely to be canceled, interested in the quality of service as in the price. while canceled projects accounted for the smallest share of investment commitments in telecommu- Most projects in the more difficult sectors do not nications. Among energy subsectors, distribution get canceled, however; among water and sewer- and integrated power utility projects had the high- age projects, for example, fewer than 10 percent TAblE 1 Canceled infrastructure projects with private participation in developing countries, by sector, 1990­2006 Projects reaching Canceled projects as financial closure Projects canceled % of sector total While the rate Investment Investment by of private commitments commitments by investment sector exit Sector Number (US$ billions) Number (US$ billions) number commitments Energy 1,498 322.8 49 11.5 3.3** 3.6** is low, it has Electricity generation 836 190.2 23 9.7 2.8** 5.1 been rising Electricity distribution or integrated utilities 328 76.8 20 1.2 6.1 1.6** Natural gas 334 55.8 6 0.6 1.8** 1.1** Telecommunications 797 537.3 35 11.2 4.4 2.1** Transport 994 180.2 47 15.3 4.7 8.5** Airports 118 25.6 4 0.9 3.4 3.6 Ports 298 33.1 4 0.5 1.3** 1.5** Railways 101 36.8 7 4.6 6.9 12.6* Roads 477 84.7 32 9.3 6.7 10.9** Water and sewerage 546 53.9 48 15.3 8.8** 28.4** Treatment plants 257 11.0 23 1.1 8.9* 9.8** Utilities 289 42.9 25 14.2 8.7* 33.1** Total 3,835 1,094.2 179 53.4 4.7 4.9 Source: World Bank and PPIAF, PPI Project Database. Note: Data refer to projects reaching financial closure in 1990­2006. * Difference significant at the 5 percent level. ** Difference significant at the 1 percent level. What drives private sector exit from infrastructure? 3 database (2007 edition). These factors included TAblE 2 project-specific variables (including the type of Effect of statistically significant variables on project cancellation contract and the level of government granting the Percentage point contract), the sector and subsector in which the project was located, and country-specific variables, change in including country income level, measures of gover- probability of nance and institutional quality, and the occurrence Variable project cancellation of macroeconomic shocks. Today's global Macroeconomic shock 3.7 financial crisis Water and sewerage sector 8.4 The results from this analysis show that the occur- could increase Sub-Saharan Africa 8.8 rence of a macroeconomic shock (measured by depreciation in the exchange rate) increases the Project with foreign sponsor 1.9 the rate of exit likelihood of project cancellation from less than Contract granted by local 5 percent to more than 8 percent, controlling for government -2.2 other variables (table 2). Other things equal, a Project investment (log) 0.3 project's being in the water sector increases the likelihood of cancellation by more than 8 percent- Source: Harris and Pratap 2008. age points and its being in Sub-Saharan Africa Note: Model results with binary dependent variable taking a value of unity in case of project cancellation, zero by almost 9 percentage points. The presence of otherwise. Data relate to 3,835 projects reaching financial a foreign sponsor also increases the probability closure in 1990­2006 and are from the World Bank­ of cancellation, as does the size of the project. PPIAF PPI Project Database. Values in the table report statistically significant marginal effects calculated at the However, projects granted by local governments variable means using a probit model. were less likely to be canceled than those granted by other levels of government. were canceled. Thus other factors are clearly at Some of the results confirm intuition and earlier work, including bad project or concession design. work. Macroeconomic shocks can dramatically One important factor is the impact that economic increase the cost of project financing, through shocks can have on the sustainability of projects. exchange rate depreciation (where foreign currency Featuring prominently in the list of canceled proj- financing is used) or through increases in domes- ects are toll road projects in Mexico, water projects tic interest rates. These shocks may also affect in Argentina, and power generation projects in demand for services from the projects through Indonesia, all of which suffered from macroeco- lower economic growth. In the water and sewerage nomic shocks occurring at different times in these sector the starting point for many public-private countries. partnerships--low cost recovery, dilapidated assets, and the need for substantial investments--and Earlier experiences with private infrastructure political sensitivities around the very involvement also shed light on factors that might drive proj- of the private sector in water supply can make for ect cancellation. Gomez-Ibanez (1999), reviewing a challenging environment. the nationalization of electricity utilities in Latin The presence of a foreign sponsor arouses greater America in 1943­79, notes that the few that were political sensitivities than local sponsors would, not nationalized by the 1980s were domestically and foreign sponsors may feel more able to aban- owned. By contrast, all foreign-owned utilities had don a project in difficulty. Some have argued that been nationalized. He also notes that utilities in domestic firms are likely to be more accommodat- Argentina, which were regulated by municipali- ing in renegotiations than foreign multinational ties, were the first to be nationalized, which might firms. As a result, projects with domestic sponsors suggest that the level of government overseeing are less likely to be canceled even after a macroeco- the projects and the quality of institutions is nomic crisis, as with some of the Indonesian power important. projects after the 1997 East Asian financial crisis (Wells and Ahmed 2007). In addition, projects To examine the factors that might be driving project with foreign sponsors may be more likely to use cancellation, econometric analysis was undertaken foreign financing, which can lead to stresses when by combining data on individual projects from the revenues are earned in local currency. PPI Project Database with data from the World Bank's World Development Indicators database Larger projects are politically more visible, and (2008 edition) and World Governance Indicators they may also impose a large financial burden on 4 the government, making their cancellation more and financing becomes most costly and less avail- likely. able--and as demand for services from the projects perhaps falls. While difficulties in public-private Other results are less easy to explain. Projects partnership projects can sometimes be self-inflicted in Sub-Saharan Africa might have significantly by the private sector--for example, through aggres- higher cancellation rates because of weak insti- sive bidding that leads to unsustainable project tutional capacity. It may also be that in countries terms--macroeconomic shocks are clearly exoge- that are smaller or that have fewer repeat projects, nous impacts. Dealing with these shocks effectively the private sector might expect fewer penalties will be important in ensuring that, once financing from walking away from a project. The finding that conditions improve, governments' private infra- projects granted by local governments were less structure programs will recover. likely to be canceled could reflect a higher level of local support. It may also reflect the private sector Notes being given an easier ride through less effective A second note by the authors will examine what happens to canceled regulation. projects: do they stay in public sector hands, or are they placed back with the private sector? The analysis found no significant impact of insti- 1. A joint product of the World Bank's Infrastructure Economics and Finance Department and PPIAF, the PPI Project Database tutional quality on cancellation. When canceled collects information from publicly available sources and includes and distressed projects were used as the dependent transactions worth $1 million or more or, in water and electricity variable, however, control of corruption was found distribution, transactions that would provide 5,000 or more new connections. See http://ppi.worldbank.org. to significantly reduce the probability of cancella- tion, other things remaining the same.2 2. Distressed projects are those in which the government or the operator has requested contract termination or that are in international arbitration. Corruption, as defined in the World Governance Indicators database, measures the extent to which Conclusion public power is exercised for private gain, including both petty and grand forms of corruption as well as "capture" of the state by elites and private interests (Kaufmann, Kraay, and Mastruzzi 2007). Although relatively few projects are canceled, these References cancellations can have considerable impact. They can reduce the credibility of the government's Gomez-Ibanez, Jose A. 1999. "The Future of Private Infrastructure: Lessons from the Nationalization of Electric Utilities in Latin commitment to contracts in the eyes of the private America, 1943­1979." Discussion Paper, Taubman Center for State sector, and they can leave the public sector feeling, and Local Government, Kennedy School of Government, Harvard often legitimately, that public-private partnerships University, Cambridge, MA. can lead to adverse outcomes for both govern- Harris, Clive, and Kumar V. Pratap. 2008. "What Drives Private Sector Exit from Infrastructure? An Analysis of the Cancellation of ments and consumers. Also true, however, is that Private Infrastructure Projects." World Bank, Washington, DC. commercial discipline and the "freedom to fail" are Harris, Clive, John Hodges, Michael Schur, and Padmesh Shukla. a big part of the rationale for turning to the private 2003. "Infrastructure Projects: A Review of Canceled Private sector, and project cancellations should therefore Projects." Viewpoint series, no. 252, Private Sector and Infrastructure be expected, since some projects or concessionaires Network, World Bank, Washington, DC. will underperform. Kaufmann, Daniel, Aart Kraay, and Massimo Mastruzzi. 2007. "Governance Matters VI: Aggregate and Individual Governance Indicators 1996­2006." Policy Research Working Paper 4280, The issue takes on particular significance at the World Bank, Washington, DC. present time. Private infrastructure projects will Wells, Louis T., and Rafiq Ahmed. 2007. Making Foreign Investment come under pressure as liquidity in inter- Safe: Property Rights and National Sovereignty. New York: Oxford University Press. GRIDLINES national and domestic markets shrinks PPIAF Approved Logo Usage Gridlines share emerging knowledge on public-private partnership and give an Logo - Black overview of a wide selection of projects from various regions of the world. Past notes can be found at www.ppiaf.org/gridlines. Gridlines are a publication of PPIAF (Public-Private Infrastructure PUBLIC-PRIVATE INFRASTRUCTURE ADVISORY FACILITY Advisory Facility), a multidonor technical assistance facility. Through technical assistance and knowledge dissemination PPIAF supports the efforts of policy makers, nongovernmental organizations, research institutions, and others in designing and implementing Logo - 1-color usage (PMS 2955) strategies to tap the full potential of private involvement in c/o The World Bank, 1818 H St., N.W., Washington, DC 20433, USA infrastructure. The views are those of the authors and do PhoNE (+1) 202 458 5588 FAX (+1) 202 522 7466 not necessarily reflect the views or the policy of PPIAF, PUBLIC-PRIVATE INFRASTRUCTURE ADVISORY FACILITY gENERAl EMAIlppiaf@ppiaf.org WEb www.ppiaf.org the World Bank, or any other affiliated organization.