The Behavioral Justification for Public Pensions : A Survey

Unfunded public pension systems are primarily justified on grounds that many individuals lack sufficient capacity to appropriately save for retirement. We begin with a review of the known principle that a standard life-cycle/permanent-income consumer who discounts the future at an exponential rate can benefit from an unfunded public pension system only if the internal rate of return exceeds the private rate of return. However, a pay-as-you go program with a below market internal rate can in fact improve lifetime utility if the consumer misestimates social security benefits, uses a hyperbolic discount function rather than the exponential function, uses a short planning horizon, behaves impulsively, or if a fraction of the population do no saving at all. A literature has consequently arisen to study how severe these behavioral defects need to be in order to justify a pay-as-you-go program. We survey this literature, and we conclude that the results are highly mixed as to whether an unfunded public pension that earns a below-market internal rate of return can be justified on grounds of shortsightedness in model economies. The challenge for this literature is that the conclusions crucially depend on the particular values of the preference parameters that are used in the simulation experiments, and these preference parameters are not observable, nor is there much consensus concerning the values that should be used in simulations. In fact, even when the analysis is confined to a small and reasonable space of the unobservable preference parameters, it is possible to reach nearly any policy conclusion. We offer some guidance for future work in this area