Subject: Social Insurance or Mandatory Individual Saving?
Contributor: PANELIST: Robert D. Reischauer
Social Insurance or Mandatory Individual Saving?
Mr. Davis raises a fundamental issue in his May 23 comment¯it is whether Social Security should be a social insurance program or simply a system of mandatory saving for retirement. If it¢s nothing more than the latter, the case for personal accounts with a good deal of worker control over investment decisions may make more sense that if we are trying to structure a social insurance program.
First, let¢s dispose of the unfunded liability issue which led him to raise his question. No matter what assets the contributions (taxes) of early cohorts of retirees were invested in, the account balances of these workers (those retiring between say 1940 when the first benefit payments were made and about 1965) would not have been sufficient to provide an adequate retirement income and the vast majority would have been poor. The nation¢s policy makers decided to provide these early cohorts more adequate benefits than their tax payments justified. This was a welfare function that properly should have been supported through general revenues, not through payroll taxes. It saved federal and state and local taxpayers a lot of income taxes that would have been needed to support the Old Age Assistance (OAA) program, the predecessor to Supplementary Security Income (SSI) whose rolls were smaller because Social Security was providing benefits to many. (Some of the participants in this debate would do well to ask their parents and grandparents how they felt about receiving *excessive* Social Security benefits and what their lives would have been like if they had relied on the very low benefits offered by OAA welfare.) Because a big chunk of the payroll tax contributions of workers during these first decades was used to provide these immediate benefits, they were not available to invest and that loss has been passed on from the shoulders one cohort to shoulders of the next.
To compensate for this, a general fund contribution could be made to the trust fund, which is what the president has proposed to do by crediting the trust fund with a portion of the projected budget surpluses. Alternatively, the transition costs of moving to a privatized system could be supported by general revenues. Mr. Hart and others have suggested this approach and the Archer/Shaw reform plan, the plan suggested by Prof. Martin Feldstein of Harvard, and other proposals that use the budget surplus to fund private accounts follow this approach.
So having taken care of (or pushed aside) the unfunded liability problem, let¢s consider what type of fully funded system we would like to establish forthose entering the workforce¯say those turning 20. Should we require them to participate in a fully funded social insurance program or force them to save for their own retirement needs in a privatized system? There is no right or wrong answer to this question. It is a matter of what one values most.
Social insurance spreads risks broadly across the society, a program of mandatory retirement saving concentrates risk (and responsibility and opportunity) on the individual. In social insurance, workers contribute a portion of their earnings in return for a promise that they will receive retirement benefits that will be calculated according to some agreed-upon set of rules. Some of the participants in the debate have argued that these promises are uncertain and I would agree that they are if the system is partially funded or unfunded. Under a fully funded defined benefit system, however, it would close to impossible to change the benefit rules retroactively to reduce benefits.
Of course, benefits may not be proportional to contributions in a social insurance system. Relative to contributions, those with the highest earnings usually receive the lowest payback. Special benefits may be provided to survivors, to spouses who stayed home to raise children, and to divorcees. This is the *social* aspect to Social Security. A rational person, even one with graduate degrees and great earning potential, might want to participate in such a program because, at age 20, he or she realizes that life is uncertain. Those great career prospects may unexpectedly go up in smoke because technological change makes one¢s skills obsolete (stockbrokers after e-trading becomes pervasive), soaring imports devastate the industry (cars and steel in the 1980s) in which one works, or personal demands and problems rule out a time consuming career (taking care of an ailing parent or the birth of a child with a disability).
In a fully funded system, collectively investing the reserves in a diversified portfolio of private and public assets makes sense because it would produce a higher rate of return on the reserves than investing solely in relatively low yielding government securities. (To the debate participant, like Mr. Dermant, who bridled at my use of the term *relatively low yielding government securities,* I respond that this return is low relative to the return on private bonds and equities. The special non marketable Treasury securities in which Social Security surpluses are invested pay the average yield on Treasury issues with maturities of 4 or more years that are outstanding at the time the special securities are issues. In 1998, the yield on Social Security¢s reserves was about 6.9 percent.) Mr. Jones is correct that with a defined benefit system, there is nothing that automatically translates a higher return on investments into a better return for the participants. However, unless policymakers decided to deliberately overfund the system, higher yields would mean that payroll taxes could be cut or benefits raised. In a representative democracy, I have a hard time imagining that deliberate over funding would be a problem.
The alternative to social insurance is a system of mandatory retirement savings. In such a system, workers would be required to salt away some portion of their earning in a personal retirement account which they would *own.* The amount of control workers would have over their accounts could be a lot or a little. The pensions supported by such a system would depend on the earnings of the workers (which would determine the contributions to their accounts), the timing of these contributions (a dollar contributed when one is 20 will be worth a lot more when one turns 65 than a dollar contributed when one is 55), and the performance of the assets in which the contributions are invested (its a lot better to have invested in Microsoft than in the Penn Central Rail Road).
In such a mandatory retirement savings system, there would be no additional benefits for those whose spouse stayed home to raise children, care for a sick relative, or contribute to the community through volunteer work. Divorcees would have to get along on the pension provided by their own personal accounts (unless the accounts of the husband and wife were combined and then divided in half at divorce). Retired 65 year olds with small children from a late first or second marriage would not get dependents benefits. (I am not arguing here that the amounts of such social assistance the current system provides are appropriate; rather, I am just pointing out the differences in the two approaches.)
Pensions would depend on investment performance and, therefore, would vary from individual to individual within a cohort depending on which assets each invested in and from cohort to cohort depending on market performance over time. Carolyn Weaver and many others have pointed out that this system will allow workers to decide how much risk they want to take. Many of modest means are relatively risk adverse¯at least that is what their allocation decisions for 401(k) plans show. They will invest conservatively and earn relatively low returns. Those who have higher earnings and the security provided by other assets will invest more aggressively and, if history is any guide, will on average earn greater returns. Therefore, the pensions provided by the mandatory retirement saving system are likely not just to be proportional to earnings but rather higher relative to earnings for those with high earnings than those with low earnings. Under such a system, the income disparities of the retired are likely to be greater than those of the working population.
What happens to those who end up with account balances that are insufficient to support an adequate pension? Mr. Davis suggests that they should be supported by welfare (SSI). Unlike many other countries, we are pretty miserly when it comes to welfare and welfare is viewed as demeaning. SSI provides less than $4,000 a year to an elderly individual with no other income and there is a reluctance of many elderly who have led independent lives to apply for the cash and in-kind means tested assistance for which they are eligible (SSI, Food Stamps, and Medicaid) when they are retired. Under a mandatory retirement savings system, an increased number of elderly would end up dependent on welfare living pretty tough lives. Some will have invested their account balances unwisely, others will have suffered some unforseen economic or personal catastrophe during their prime earning years.
I think a fully funded social insurance system whose reserves are invested in a diversified portfolio of private and government assets is the best way to assure basic retirement income for all workers and their dependents. This should be supplemented with personal retirement saving (IRAs) and employer-sponsored retirement plans (401(k)s) in which workers can exercise their taste for risk without jeopardizing their basic retirement incomes. Others advocate a system of mandatory retirement saving back up with a welfare safety net.
