Incidence of Distress Problems

The incidence of distress in PPP projects can be analyzed based on the information contained in the World Bank's PPI Database.140 This database has data on almost 3,600 PPI projects developed worldwide between 1991 and 2006.

The available variables for each project include the project's status, which can be: cancelled, concluded, under construction, distressed or operational.141 Out of these, cancelled and distressed projects may thus be viewed as projects in distress.

The PPI Database uses the following criteria to classify projects into these two categories:142

• canceled projects from which the private sector has exited in one of the following ways:

o selling or transferring its economic interest back to the government before fulfilling the contract terms.

o removing all management and personnel from the concern

o ceasing operation, service provision, or construction for 15 percent or more of the license or concession period, following the revocation of the license or repudiation of the contract

• distressed projects where the government or the operator has either requested contract termination or are in international arbitration.

As to the latter definition, it should be noted that the underlying criterion is quite restrictive, as it only accounts for projects that have reached a relatively advanced stage of conflict. Situations of financial, economic or operational distress that affect the service but have not triggered an open legal conflict with the government are not taken into consideration. Accordingly, the results of the analysis can be viewed as conservative with respect to the real incidence of distress situations in the universe of projects.

A first measure of the relevance of the distress phenomenon is related to the number of projects in distress relative to the total number of projects developed in the period. Table 21 illustrates this information, broken down by type of PPP.

Table 21: Incidence of Distress Number of Projects by PPP Type

Source: Prepared by us using PPI Database data

Out of the almost 3,600 surveyed projects, 179 have been cancelled and 58 are in distress. Their incidence is thus 6.6% (5% and 1.6%, respectively). Although not particularly high, this value indicates that the risk that a PPP project will encounter difficulties is not nonexistent. These values are relatively consistent across all types of PPPs other than divestitures, where the incidence of cancelled projects is relatively lower (3.2%), and that of distressed projects is relatively higher (4.1%).

A special remark should be made in connection with divestiture projects. In principle, these could be left out of the analysis as they were asset sales and, hence, not strictly PPPs. However, an analysis of the data in the PPP Database reveals that, because of the manner in which they were granted, many concessions have been classified as divestitures.143 Accordingly, we have found it more adequate to include them in the analysis.144

An alternative measure consists in considering the amounts invested instead of the number of projects. The investment figure is recorded in the PPI Database on the basis of disbursements and investments committed up to the execution of the PPP agreement. Table 22 provides this information, also broken down by type of PPP.

Table 22: Incidence of Distress Investment Commitments by Type of PPP

Source: Prepared by us using PPI Database data

Out of the slightly more than 1,000 billion dollars of investments disbursed and committed in the projects under consideration, 52 billion represent canceled projects, and 32 billion represent distressed projects. Both categories combined account for 8.1% of the total investment commitments (5% and 3.1%, respectively). The fact that measuring incidence in terms of investments leads to a higher incidence value than if measured based on the number of projects reveals that distress situations have higher incidence in larger projects.

As project size is frequently related to the project's technological, institutional and financial complexity, the results appear to match a priori expectations. The more complex the project, the higher the probability that, as a result of design and/or implementation problems, the project will encounter difficulties in the construction or operation stages. It is also worth noting that the greater difference in incidence relates to distress as compared to cancellations. This also matches a priori expectations: large projects are renegotiated, rather than abandoned.

The manner in which the PPI Database measures investment may also help explain these results. The values in the database represent investments committed throughout the life of the project. The higher failure rate for projects involving large investment commitments could then reflect a gap between the investment commitment and the economic conditions defined in the contract.

Another way to test the possible impact of the specific features of each project on the probability of distress consists in analyzing projects based on the economic sector in which they are inserted. Such breakdown by sector, considering the number of projects, is illustrated in Table 23.

Table 23: Incidence of Distress Number of Projects by Sector

Note: Distressed includes canceled and currently under stress.

Source: Prepared by us using PPI Database data

For this analysis, we added the cancelled and distressed categories together. As it is evident, the water sector behaves in a considerably different way than the rest. The incidence of distressed projects in the water sector is 10.1%, some 1.5 above the average for all sectors (6.6%). In the remaining sectors, the incidence ranking is, from lowest to highest, as follows: transport, energy and, lastly, telecom, all presenting relatively similar values (ranging from 5.7% to 6.6%).

A replication of the analysis by sector, based on the investment commitment figures, leads to the values presented in Table 24.

Table 24: Incidence of Distress Investment Commitments by Sector

Source: Prepared by us using PPI Database data

In this case, differences between sectors are much more evident. The telecom sector has the lowest incidence, with only 4.5% of the value of investment commitments affected by distress situations. The energy and transport sectors present similar values around 9% (9.1% and 9.5%, respectively).

The water sector presents the greatest difference, with an incidence of 31.1% of the total investment commitments. The number of projects - investment commitment ratio also ranks at the top for this sector, which further supports the view that this sector has the more complex projects with a greater probability of facing situations of distress.

On average, these sectors have an 8.1% probability of distress (measured in terms of investment commitments). Leaving out the water sector, the average stands at 6.9%.

Overall, the aggregate results of the analysis based on the WB's PPI Database are consistent with other specific sectoral studies on the incidence of projects under stress in the energy sector and in the water and sewerage sector.

Using the same basis, Harris et al (2003) found 48 cancelled projects for the period 1990-2001, which accounts for only 1.9% of the total number of existing projects. In terms of investments, these projects added up to 24.2 billion dollars, some 3.2% of the total investment commitments. On average, projects were cancelled 4 to 5 years after their initial date, which is relatively early for projects estimated lives of 20 to 30 years.

For the energy sector, a study on electricity projects in distress developed by the PPIAF provides more detailed evidence of the causes and effects of situations of distress.

The PPI Database was the source of information for this study as well, supplemented by a World Bank and private investors' survey that provided a sample of 63 distressed projects in 18 countries.

This study found that "Overall, a stress situation in the electricity sector is a rare event, which has affected or is affecting only 4 percent of power projects in number and 10 percent in value. If projects that have been worked out are excluded, the risk of stress falls to 3.5 percent in number and about 7 percent in value. Considering that all projects that went through stress were either cancelled or worked out, it appears that in the past some 21 percent of the projects that incurred stress were ultimately worked out while some 33 percent ended up in cancellation."

For the water sector, there is a similar study of the IADB that uses a slightly different approach by focusing on the analysis of the "Salida de operadores privados internacionales de agua en América Latina" [Exit of international private water operations in Latin America]. Such work is not intended to quantify the incidence of this process in the universe of projects, but to "perform a count of cases and describe their main distinguishing features." The universe of analysis is different from ours in two dimensions. First, because such exit may be the result of business decisions that have nothing to do with distress situations, the study included cases that are irrelevant to our analysis. Second, there may be distressed projects that do not involve international operators or have not ended up in their exit.

The study analyzed a total of 14 cases across five countries.145 Of such cases, three relate to decisions made by the parent companies and have no connection with stress situations. The remaining 11 cases are classified into changes in national sectoral policy, social and political conflicts, alteration of the contract's economic and financial equilibrium.

This universe can be compared to the 24 cancelled and distressed projects (20 and 4, respectively) included in the PPI Database. The differences would, in principle, be attributable to the two reasons mentioned above: distressed projects that do not affect foreign investors and projects involving foreign investors but which did not end up in their exit.

To conclude, both the evidence from the analysis of the PPI Database and the sectoral studies referred to above reveal that, even if not leading to extremely high values, the incidence of distress in PPP projects is far from being immaterial. A relevant aspect is the strong differences between sectors, particularly the significant impact of distress problems on the water and sewerage sector.

____________________________________________________________________________________

140 http://ppi.worldbank.org/index.aspx

141 There is a sixth category, i.e. merged projects, which we have chosen to leave out of the analysis, as it covers projects merged into other projects and their inclusion might lead to a duplication of results. In any event, these are not significant in number.

142 PPI Database glossary http://ppi.worldbank.org/resources/ppi_glossary.aspx

143 This is the case, for instance, with electricity concessions in Argentina, where, because of the creation of new companies as concession holders, followed by the sale of their controlling interests, the transaction appears as a divestiture rather than a concession.

144 The greatest impact of such inclusion hits the telecom sector, where divestitures prevail. Still, overall, the inclusion of this category does not materially alter the results of the analysis.

145 Out of the cases under analysis, six represent Argentina whereas the remaining countries, namely Bolivia, Chile, Uruguay and Venezuela, present two cases each.