Since alteration of the saving behavior of the present working generation may have such a significant impact on capital stock growth, some concern has arisen about the effect of the Social Security system on saving behavior. In response to this concern numerous empirical and theoretical analyses have been forthcoming. Most of the theoretical analyses, either in the form of overlapping generations models [9; 11] or simpler life cycle models [6; 15] have indicated that unless specific parameter values, utility functions, and, in some cases, production functions are chosen the effect of a social security system on saving is indeterminate. For example, Kurz [11] has shown that the introduction of a simple pay-as-you-go social security system in an overlapping generations model in which individuals do not have a leisure choice will not have a predictable effect on the steady state level of the capital stock unless parameter values and the form of the production function are known. Similarly, in life cycle models that include a leisure choice the effect of social security is difficult to determine, although Feldstein [6] shows that when the leisure choice is suppressed in such a model a social security system will have a negative impact on saving. This indeterminacy has not been completely resolved by empirical work, although the bulk of evidence [7; 8; 10; 12] tends to support the view that Social Security has depressed saving in the past.' Unfortunately, with the exception of the work of Sheshinski and Weiss [16], little of the theoretical analysis deals with the effect on saving behavior of participants who are faced with uncertainty about future prices, wages, and social security benefits. Sheshinski and Weiss have shown that in an overlapping generations model with uncertain lifetimes the effect of increasing the role of social security (through balanced increases in benefits and contributions) will actually lead to an increase in saving. Their analysis, however, does not include a leisure choice nor does it deal with other forms of uncertainty facing participants. Consequently, we still lack more complete answers to the following basic questions: Will the introduction of a social security program in the presence of uncertainty