The Chilean pension reform has given rise to significant debate in the last decade. It was the first pension system in which the core mechanism for producing pensions is individual capitalization and the first to hand over its management to the private sector. This pension concept is one of the rare institutions exported from Chile to the rest of the world, including countries in Latin America, central and eastern Europe, Africa, and Asia, as well as a few developed countries such as Sweden.
A quarter of a century after its inception, it seems natural to undertake a deep evaluation of the system's outcomes, strengths, and weaknesses. Corbo and Schmidt-Hebbel make a positive macroeconomic assessment, as well as good evaluation on financial grounds based on the high average return on assets.1 Several papers, however, argue that the level of competition is weak, which translates into high prices and high returns on equity.2 From a public finance perspective, Arenas suggests that the fiscal burden may be significant and increasing, while the average coverage ratio appears to be low.3 Finally, in an earlier paper, we undertake a preliminary evaluation of coverage based on affiliates' administrative data.4 These last two [End Page 227] studies reveal that one of the main weaknesses of the Chilean model is its low coverage.5
This paper studies the reasons behind the low coverage of the Chilean pension system by focusing on one benefit—namely, old-age pensions. It departs from previous studies in that it uses individual data from administrative records, as well as socioeconomic characterization of these same individuals provided by the 2002 Social Security Survey (EPS), to project the future benefits of the generations that are now contributing to the system.6 The use of individual data is relevant because it takes into account the large heterogeneity of individual experiences.
We use these data to simulate the distribution of old-age benefits, which is a key aspect of a defined-contributions environment. A defined-contributions system has no solvency problems, as there are few, if any, guarantees involved. Beyond the rate of return on investments, however, the final pension depends heavily on the amount, number, and timing of contributions. In a defined-benefits system, shocks affect the solvency of the system, whereas in a defined-contributions system, individuals bear much of the risk regarding the amount of the final pension.7 In fact, while the theoretical replacement rates are high for those who contribute throughout their working lives, they may be quite low for workers with infrequent contributions.8 The use of individual-level data allows us to identify some of the risks that generate such uncertainty in Chile's private pension system.
When the Chilean system was designed, it was expected to deliver replacement rates of around 75 percent.9 This is the case for around one-third of affiliates. However, the strong heterogeneity of individual income and contribution density profiles means that some people will be in a very difficult situation [End Page 228] when they retire.10 Self-employed workers (who are not mandated to contribute in Chile) and women (who have a relatively low participation rate) appear to make up the most vulnerable categories. Additionally, a considerable segment of the population shifts between salaried work and self-employment and between formal and informal employment. Even if they work most of their active life, such workers may end up with low contribution densities. Consequently, the average contribution density in the case of the Chilean pension system hides very different realities that should be considered when evaluating the system's performance.
The main results of this paper show that under current circumstances a considerable group of old-age retirees will receive pensions below the level of the minimum pension guaranteed by the state. At the same time, most of them will not be...