28 March 2008

About Social Security

What are the key political issues related to Social Security?

A trade group for certified public accountants has a nice report of issues related to Social Security reform three years ago, whose executive summary I quote liberally from below.

Poverty Prevention

Social security is arguably the most effective anti-poverty program in U.S. history:

Social Security provides more than half of the total income for almost 60 percent of beneficiaries. For almost 30 percent, it provides more than 90 percent of income. . . . The poverty rate among the elderly in 2000 was approximately 10 percent, down from a rate of 35.2 percent in 1959. Without Social Security, the poverty rate among the elderly would be 48 percent.

Note also that on top of their Social Security benefits that Medicare provides essentially everyone age 65 and up in the United States with health care under what amounts to a single payer program, and that Medicaid provides nursing home care to most asset poor senior citizens. The means tested SSI program provides further support to the elderly poor.


Social Security's operating transaction costs, both on the benefit paying side, and on the revenue generation side, are phenomenally low. Fraud in either part of the program is infrequent and small in dollar amount, in large part because there are few relevant facts necessary to produce a correct tax or benefit amount, and those facts are maintained by third party government and private sector professionals for the most part.


Social Security is not in a fiscal solvency crisis, although tweaks need to be made to match revenues to expenses in the long run. As of three of years ago, in an estimate that remains close to accurate now:

According to the Social Security Administration's "best guess" (intermediate) assumptions, the Social Security Trust Fund surplus will peak in 2028. Then it will decline steadily until 2042, at which time the Trust Fund will be exhausted. However, inadequate funds do not mean zero benefits. If no changes are made to Social Security, beneficiaries could receive full scheduled benefits through 2042. Thereafter, scheduled benefits would have to be reduced by 27 percent. In 2078, benefits would have to be reduced by 32 percent. This Social Security "deficit" could be funded by an immediate infusion of $3.54 trillion; by increasing the payroll tax rate from its current level of 12.4 percent to 14.3 percent; or by reducing current scheduled benefits 12.6 percent.

Other tweaks that can be made to improve the fiscal solvency of Social Security include changing the way it calculates cost of living adjustments, increasing the payroll tax wage base which is currently a little over $100,000 (quite a bit lower than the maximum income that can be taken into account in tax preferenced defined benefit plans in the private sector), and broadening the scope of the self-employment tax which substitutes for FICA for people who are self-employed rather than earning wages and salaries upon which a payroll tax is due.

Another, moderately more controversial way to improve the fiscal solvency of Social Security is to invest some part the Social Security trust fund, which currently consists of ledger entries representing U.S. Treasury bonds, in some other form of security, a strategy sometimes known as "pre-funding" the Social Security trust fund.

In addition to the fiscal solvency motivation, pre-funding would help inspire greater confidence that Congress will make good on its promise to provide Social Security benefits to future retirees.

Payroll taxes collect more than is currently needed to fund current Social Security benefits. But when there is no longer a "Social Security surplus" (i.e. excess of current payroll tax receipts over current benefit payments) which will happen much sooner than the 2040s, Congress will have to either raise general fund taxes to redeem the bonds in the Social Security trust fund in order to pay current benefits or publicly offer the Treasury bonds that make up the Social Security trust fund to members of the general public, first in an amount that would only be a trickle, but eventually, in an amount that would flood the Treasury bond markets and drive down Treasury bond interest rates. In contrast, the liquidation of a diversified low risk private securities portfolio in the Social Security trust fund to pay current benefits would not necessarily have the same overwhelming impact on the markets because the impact would be felt more widely.

For example, a gradual investment of part of the Social Security Trust fund in a mixed portfolio of municipal bonds and investment grade corporate bonds, in a manner designed to spread the investment more or less evenly over the entire market for each would be relatively low risk, give the Social Security trust fund psychological substance, and might increase the rate of return. Yet, it would give the federal government no additional say in how corporations or state and local governments are managed.

At the end of December 2004, the Trust Fund had an accumulated balance of $1.69 trillion, comprised largely of “special issue” securities available only to the Trust Fund. (The Trust Fund holds a very small amount of marketable Treasury securities.) The weighted average maturity of these bonds was 7.2 years, and the weighted average interest rate was 5.7 percent. . . . The Social Security Administration’s Office of the Actuary (SSA) estimates that a 1 percentage-point increase in the real interest rate [earned on the Social Security trust fund] would reduce the actuarial deficit from 1.89 percent to 1.38 percent of payroll and push back the estimated exhaustion date by approximately three years.

I suspect that the Trust Fund's weighted average interest rate now is lower than it was in three years ago.

The biggest problems with pre-funding the Social Security trust fund with private sector assets are that (1) even a diversified portfolio of private securities invested for the term could be invested in a way that carries more investment risk than Treasury securities or produces a lower return, if not optimally managed, and (2) the trust fund would be such an 800 pound in gorilla that this species of sovereign wealth fund would permit political agendas to inappropriately manipulate the private securities market place and thereby harm the economy.

A far more controversial proposal is to great private accounts invested in publicly traded securities in lieu of some of the existing core social security benefits, at least for young Social Security system participants. This minimizes, although it does not eliminate (since someone would have to decide what investment choices were available to participants), the sovereign wealth fund investment issues. But, this raises a host of additional difficult questions involving transition arrangements, increased administrative costs, how to handle cases where individuals make poorly performing investments for themselves that cause a hardship in retirement, and how this changes the relative benefit that individuals receive from the system. Suffice it to say that I oppose private accounts that would reduce the core of benefits currently provided by Social Security, that Congress and public opinion agree.

The modest fiscal solvency issues that the Social Security system can be solved in ways that don't create their own problems in the way that private accounts do. For example, if the investment meddling by the federal government was the primary concern associated with pre-funding the Social Security trust fund collectively in private securities the federal government could entrust state treasurers and/or state pension fund managers with the investment of a share proportionate to social security benefits paid to state residents to invest on a pooled basis with state funds.

Also, if tweaks were made to a combination of payroll tax rates, the Social Security payroll tax base amount, the self-employment tax base scope, and a factor that impacted scheduled benefits like the manner in which inflation adjustments are calculated, no one adjustment would have to be as large in magnitude as the ones suggested in the blockquote above.

Also, it is worth noting that solvency solutions are well suited to being put in place on a piecemeal basis. Even if only half of the long term solvency gap is fixed by a proposal, or the trust fund insolvency date is only pushed out by ten or twenty years, an early half measure now makes the job easier for those who have to address it in the future.


Individual returns on investment from Social Security vary a great deal. In the worse case scenario, someone who works at a high income for his or her entire life, without ever suffering from a disability, and then dies without a surviving spouse (or ex-spouse to whom the person was married for at least ten years) and without minor children, who dies right before reaching retirement age, has made large contributions to the system and gotten almost nothing in return. Others can receive significant benefits based upon contributions made by someone else like a spouse or parent without making any contributions personally.

But most of those difference are the capricious happenstance of any defined benefit system. While two men born on the same day with the same marital history and having the same work history might end up receiving very different benefits, no one could predict that in advance. There is no systemic bias towards one man or the other.

While Social Security's tax base (a combination of FICA payroll taxes and self-employment taxes), is regressive by design, because it does not tax investment income and does not tax income over a certain dollar threshold, when the tax cost and benefit returns to individuals are taken together, the Social Security system is actually moderately progressive in effect, an element of the program which is fairly uncontroversial because it is intended to be a social safety net rather than make complete provision for seniors in their retirement.

Also, while high income people, on average, get something of a raw deal in terms of returns on their contributions, they also get the lion's share of the tax benefits associated with tax favored non-Social Security retirement accounts and defined benefit plans.

The benefits that older recipients receive relative to younger ones, call it the Ponzi scheme effect, while real are also not horribly controversial. This distinction is gradually losing its currency as the most advantaged oldest beneficiaries pass away. To some extent this is also an inevitable consequence of the fact that people are having, on average, fewer children, and are living longer, reducing the number of people of working age per retiree.

The more controversial fairness issues in Social Security reflect the fact that its formulas were designed with a traditional breadwinner husband, homemaker wife, with common children family model in mind. AICPA explains who wins and who loses in the Social Security benefit formula:

Most of today's Social Security recipients are receiving – and will continue to receive—more in benefits than their actuarial "fair share" based on their contributions. Even if all promised benefits were paid, future retirees, particularly singles, two-earner couples and those with high incomes, will earn below-market rate returns on their contributions.

The rate of return earned on an individual's Social Security contributions is affected by gender, marital status, and income level. Social policy considerations weaken the direct link between contributions made and benefits received. The Social Security benefit formula includes a declining fraction of income in the calculation. As a result, low income beneficiaries benefit from the formula, high income beneficiaries do not. Married couples benefit from spousal and survivor benefits.

The magnitude of that differences is significant.

Rates of return in the 1960s and 1970s were still much higher than one could expect from a private pension plan or from a mature pay-as-you-go system. The rate of return for a single male was 10.97 percent in 1960, declined to 1.82 percent in 1995, and is projected to be 1 percent by 2030.

I display below numbers from the Report's Exhibit 3.2A showing average annualized real rates of return on contributions expected for beneficiaries of each type based upon retirement age. I omit the entries for beneficiaries who retired in 1960, who are almost all dead now, and for beneficiaries who retired in 1980, who are now age 92 or 93, and are likely to be irrelevant to political issues involved before any change could be enacted and take effect.

Single male
1995 - 1.82%
2010 - 1.16%
2030 - 1.00%

Single female
1995 - 2.89%
2010 - 2.09%
2030 - 1.90%

One-earner couple
1995 - 4.79%
2010 - 3.64%
2030 - 3.37%

Two-earner couple
1995 - 3.54%
2010 - 2.52%
2030 - 2.29%

It is worth recalling that the greater return on contribution that single females receive on average at any given income is largely due to women's longer life expectencies, and hence to their greater need for benefits.

Also, it is worth maintaining a sense of perspective. None of these types of beneficiary categories is making less than a 1% rate of return on their contributions, and for persons retiring in 2030 (i.e. people who are currently age 45), no one is making more than a 3.37% return on their contributions. Everybody make a return in the range of what one might earn in a savings account, certificate of deposit or a government bond fund, which is appropriate at an order of magnitude level, for a system that is providing emergency fund like support in the event that there are no other retirement savings.

About 95% of the population gets married at some point and some people who remain single all their lives die before reacing retirement age. When you are married and how long you are married to particular people matters.

Most of this couples’ advantage comes from the availability of spousal and survivor benefits. Spouses are eligible for the greater of (1) the benefit they earn on their own; and (2) the amount they are entitled to as spouse or a surviving spouse. Contributions to the system by a one-earner couple can result in benefits generating a greater rate of return because the surviving, non-worker spouse can continue to receive benefits. A single individual’s rate of return on investment would be lower because benefits stop at his or her own death.

One-earner couples also receive a greater return than two-earner couples based on their lower contributions, even with comparable benefits. The availability of spousal benefits diminishes the value of Social Security benefits for two earner couples relative to one-earner couples – particularly when there is a significant disparity between spouses’ incomes. If the spousal benefit is larger than the benefit based on the surviving spouse’s own earnings, the survivor’s contributions have earned no return. If the survivor’s earnings generate a greater benefit, the additional contributions made by the deceased spouse decreases the couple’s rate of return as does the “net” benefit – which is the increase over the benefits the survivor would have received as a spousal benefit. . . .

One-earner couples versus single individuals and two-earner couples. The spousal benefit provides tremendous value at little or no cost to beneficiaries. Even as Social Security rates of return are projected to decline rapidly, one-earner couples are projected to receive a very beneficial deal in the future.

The income effect is significant, but still keeps most people within the return range associated with safe preservation of principle oriented investments. On average, only man who is single all his life and has a high income will not manage to at least achieve a return of all his contributions, and men in this situation will still come very close.

Although contributions are made at a fixed rate based on earnings, benefits are calculated using only 15 percent of average indexed monthly earnings over the second bend point ($3,689, or $44,268 annually, in 2004). This creates a lower rate of return as income increases, as illustrated in Exhibit 3.4. Income taxes paid on Social Security benefits received by higher income retirees further reduces rates of return.

Exhibit 3.4

Impact of Social Security for Low-, Middle-, and
High-Income Beneficiaries Turning 65 in 2000

A. Real Rates of Return

Single male
Low Income 2.02%
Middle Income 1.15%
High Income -0.07%

Single female
Low Income 2.77%
Middle Income 2.05&
High Income 0.81%

One-earner couple
Low Income 4.39%
Middle Income 3.54%
High Income 2.39%

Also, were it not for the preferrential treatment that Social Security provides low income workers, many elderly individuals would be reduced to poverty and expensive to administer welfare programs that would mitigate any costs savings from providing more modest Social Security benefits would have to be established to support them.

No comments: