Social Security Reform Can Make Things Worse

Summary of working paper 7362
Featured in print Digest

Postwar Americans lose 5 cents out of every dollar they have earned or will earn over their lifetimes in the form of payroll taxes paid into OASI in excess of benefits received.

Social Security gives people born in the second half of the 20th century a bad deal, according to Jagadeesh Gokhale and Laurence Kotlikoff, and the $8 trillion funding shortfall facing the system in the 21st century means that things are going to get worse. In Social Security's Treatment of Postwar Americans: How Bad Can It Get? (NBER Working Paper No. 7362), the two economists show that choosing to keep Social Security in the black while maintaining the current pay-as-you-go approach has important consequences for which postwar generations, and individual members of every age group, will pay.

Gokhale and Kotlikoff find that a straight cut in benefits distributes the burden more equally across generations than a straight tax increase. The other policy proposals - all variations on tax hike and benefit reduction themes - come somewhere in between the two benchmark cases in terms of intergenerational burden-sharing. One clear result of this research is that the youngest postwar generations have the most to worry about. Tax increases will affect them over their entire working lifetime. Graduated benefit cuts will be fully phased-in by the time they retire.

According to the Social Security Trustees, an immediate and permanent 38 percent increase in the Old Age Survivor Insurance (OASI) payroll tax rate would restore present value financial balance to the system. So would an immediate and permanent 25 percent reduction in all OASI benefits. These are the two benchmarks that Gokhale and Kotlikoff use, although they stress that the Trustees have tended to understate future funding shortfalls.

Because the Social Security benefits an individual receives depend on longevity, lifetime earnings, marital arrangements, and fertility, working out precisely how bad a deal postwar Americans receive requires an actuarial approach. The necessary data, following individuals from their first encounters with payroll taxes through to the end of their lives, has to be simulated. Gokhale and Kotlikoff do this by using a model developed by Cornell sociologist Steve Caldwell to generate a sample of 68,688 individuals which they divide into 11 five-year birth cohorts starting with 1945, and into lifetime earning quintiles. The model's starting point is a representative sample from the 1960 U.S. census, extended for demographic and economic changes each year through 2100.

As currently legislated, the authors write, postwar Americans lose 5 cents out of every dollar they have earned or will earn over their lifetimes in the form of payroll taxes paid into OASI in excess of benefits received. This can also be described in terms of an internal rate of return on contributions of 1.86 percent, which is less than half the rate now paid on inflation-indexed long-term government bonds.

An immediate payroll tax hike, by 38 percent, would be paid not just by Americans born between 1945 and 1999 but also by those born this year and in future years. However, the postwar cohorts would bear almost half the burden of filling the long-term budget gap of $8.1 trillion in present value terms. This is also the case with an immediate 25 percent benefit cut.

Both policies would raise the lifetime net tax rate for all postwar generations. But they have very different intergenerational impacts. The tax hike has a much greater impact on later generations, with more years left to work and so pay the tax, than earlier generations. In the case of the benefit cut, all generations are similarly affected because none has yet started to receive benefits. The tax hike raises the net lifetime tax rate for the generation born in the five years from 1945 through 1949 from 5.3 to 5.7 percent. It raises the lifetime net tax rate for the generation born in the six years between 1995 and 2000 from 5.4 percent to 8.4 percent. In contrast, cutting benefits leaves both cohort's lifetime tax rates at 6.0 and 6.1 percent respectively.

Eliminating the earnings ceiling, with or without a change in the ceiling on benefits, hurts younger generations much more than older ones. This also goes for linking benefits to inflation (cutting out the link with wages) and eliminating the real growth in benefits. Accelerating the already legislated increase in the normal retirement age would hurt older cohorts - those close to retirement - more than younger cohorts.

Both the benchmark tax increase and the benefit cut are harder on the lifetime poor than the lifetime rich. In the lowest income quintile of the 1995-2000 cohort, a 38 percent tax hike means losing 4 cents more per dollar earned. Those in the highest quintile only lose an additional 2.6 cents per dollar earned under the policy. In terms of the benefit cut policy, the difference is even more striking. The poorest quintile in the 1995 cohort lose 3.3 cents on the dollar, the richest lose less than half a cent.

All of these proposals are in the framework of a pay-as-you-go system. Alternative reforms that include an investment based element can have very different effects.

-- Andrew Balls