Social Security

I frequently Ž nd economists who express a view of the system that is very far from mine. For example, many young economists and economics students say that they expect to get no beneŽ ts at all from Social Security. This expectation does not seem sensible to me. If there is no legislation changing Social Security, trust fund assets and payroll tax revenue (and revenue from the taxation of beneŽ ts) are projected to be sufŽ cient to pay all the beneŽ ts scheduled under current law until 2042 (Board of Trustees of Social Security and Medicare, 2003). After the trust fund assets are exhausted the payroll tax revenue would continue to be available to pay beneŽ ts, with the  ow of revenues at that time sufŽ cient to pay roughly three-quarters of the beneŽ ts scheduled in current law. The estimate for the end of the 75-year projection period shows enough revenue to pay roughly two-thirds of scheduled beneŽ ts. With initial beneŽ ts indexed to earnings, average real beneŽ ts would be higher than today, although replacement rates would only be roughly 60 percent of current levels for the medium worker. This projection is a far cry from no beneŽ ts. Moreover, I anticipate that Congress will act before the trust fund is exhausted, both lowering beneŽ ts relative to those scheduled under current law and providing additional revenues to Ž nance higher beneŽ ts than are payable after 2042. After all, the Ž nancial problem of Social Security is not so very large (unlike the larger and more complex set of Ž nancial problems of Medicare and Medicaid). An increase in tax revenue of just over 15 percent of currently projected payroll tax revenues would handle the projected cash  ow problem for 75 years on a present value basis. On an annual cash  ow basis, the share of GDP needed to provide all of the beneŽ ts scheduled in current law would increase from 4.4 percent of GDP today to 7.0 percent in 2077. Like almost everyone else, I do not favor addressing the projected deŽ cit by simply adding more revenues with no other changes. Nor do I picture that solution as having any political prospects. But solving the problem with a mix of beneŽ t reductions and revenue increases does not require large changes, nor does it require a fundamental restructuring of the program. It is not just in the perception of the projections and the forecast of politics that I Ž nd myself in disagreement with opinions that I often hear. More generally, I think the system works better than many economists think. I hope to convince you that the approach inherent in the current U.S. system broadly makes good sense. In particular, I will argue that it makes sense to mandate taxes to Ž nance a reasonable replacement of earnings after retirement; that it makes sense to mandate that retirement beneŽ ts be paid as an annuity; that it makes sense to mandate protection for family members, both young children and surviving spouses; that it † Presidential Address delivered at the one hundred Ž fteenth meeting of the American Economic Association, January 4, 2004, San Diego, CA.