Good morning, Mr. Chairman and members of the Committee. My name is Maya MacGuineas and I am a Senior Fellow at the New America Foundation, a nonpartisan think tank here in Washington, where I work on fiscal policy. Thank you for inviting me to testify today. It is a privilege to appear before the Committee.
The Social Security Trustees' Report is the single most important and influential source of information about the financial health of the Old-Age and Survivors Insurance and Disability Insurance programs. Given the attention the report receives, it is certainly worthwhile to discuss not only the implications of the findings but also whether there are ways to improve either the content or presentation. Furthermore, given its unbiased analysis, the Trustees' Report can and should provide a framework for comparing various proposals as we move forward with the discussion about how best to reform Social Security.
In my comments today, I am going to discuss ways to provide more information about Social Security's effects on the unified budget, an analysis that extends beyond the actuarial window, a more detailed breakout of the sensitivity analysis of various economic and demographic assumptions, and some additional information with regard to benefits.
The main purpose of the annual report is to shed light on the overall financial health of the Old-Age and Survivors Insurance and Disability Insurance programs. The Trustees' Report is unquestionably the most comprehensive source of such information. In addition to evaluating short-range trust fund adequacy, the analysis relies primarily on five tools including: 1) Trust fund exhaustion dates; 2) Income and cost rates; 3) Trust fund ratios; 4) Actuarial balance; and 5) Social Security costs as a percentage of GDP.
In my opinion, some of these are more useful than others. Income and cost rates for instance, are reasonably straightforward and quite useful. They show that while taxes as a percentage of payroll currently exceed benefits by 1.88 percent of covered payroll, this relationship will deteriorate over time and that by 2025, benefits will exceed non-interest income by 2.90 percent. By 2080, the number will have grown to 6.68 percent. Likewise, viewing the Social Security program as a percentage of GDP is an extremely useful tool because it shows the level of resources the program will transfer across the entire economy. The Trustees report that Social Security will transfer 7.0 percent of the economy at the end of their 75-year valuation period as opposed to 4.5 percent today. Both of these sets of numbers are relatively easy to understand.
On the other hand, trust fund ratios, actuarial solvency, and exhaustion dates seem to cause at least as much confusion as clarity. While I do not take issue with the assumptions these numbers are based on, or the methodological approaches used to derive them, I am concerned they may divert attention away from more relevant issues.
The trust fund ratio is expected to peak at 471 percent in 2015, and decline thereafter. The trust fund exhaustion date, 2041, is the year when the trust funds' assets will be depleted. Actuarial balance measures Social Security's financial status over a 75-year time period, expressed as the difference between the expected income and costs as a percentage of taxable payroll in present value terms. Currently, the actuarial deficit is 1.87 percent.
But these numbers do not convey the full burden to the budget or the economy of meeting future obligations. Their inclusion of the trust funds, while appropriate in an accounting sense, masks the extent of the larger problem. One must also consider the burden the trust funds represent in order to view the funding problem in its entirety. To illustrate this point, you merely need to look at the idea that has been floated occasionally to increase the rate of interest paid on the bonds in the trust funds. Such a change would improve trust fund ratios, extend the exhaustion date, and decrease the actuarial deficit. Increase the interest rate by a fraction, and things would look a bit better. Increase the rate by enough, and all problems would appear on paper to evaporate. Would this change make any meaningful improvements to the situation we are facing? Of course not. The overflowing trust funds would not do a thing to make the task of paying benefits any easier. The money to pay the higher interest costs would have to come from somewhere, but neither trust fund ratios, actuarial solvency, or exhaustion dates reflect this.
One recommendation I would make, then, is to include a year-by-year cash flow analysis. Almost all of the information necessary for such an analysis is contained in the Trustees' Report but could be combined in a way that illuminates some important issues.
(See Table 1.)
A cash flow presentation would lay out the tax revenues that will flow to the program as well as their sources, and the annual costs. I show the numbers here in 2002 dollars; if this format were adopted, they would presumably be shown in both current and constant dollars. The surplus or deficit numbers are particularly helpful because they illustrate how much Social Security contributes to the rest of the budget in the short run and how much it will drain from it in the future. It is also useful to view how these deficits translate into payroll tax increases or benefit reductions necessary to keep the program balanced on an annual basis.
In additional to showing the cash flow numbers in dollars, I would suggest showing them as a share of total government revenue. Since the government's actual tax base is only a little more than half the value of GDP, showing cash flow numbers as a share of total government revenue gives a more realistic picture of the tax rates required to achieve balance. By incorporating Congressional Budget Office assumptions, one can see that spending on Social Security will rise from 23 percent of the budget today to 33 percent by 2025 and continues to rise thereafter. The cash flow deficit will grow to 6 percent as a share of total revenues over that time period and then double over the next fifty years. While in my mind viewing these numbers as a share of the budget is most helpful, they could also be calculated as a share of GDP, covered payroll, or any other denominator deemed appropriate.
This format would not only be helpful in viewing trends in a way that is relevant to Social Security and the unified budget, it could also serve as a useful benchmark for comparing the effects that changes in assumptions or policies would have on the program. For instance, the Trustees include in their report high, low and intermediary cost assumptions for their underlying economic and demographic assumptions. In Appendix D of the report they perform a sensitivity analysis by altering one variable at a time. The findings are reported in terms of summarized income and cost rates and actuarial balance over 25, 50, and 75-year periods. Reporting these results using the annual cash flow framework would be more useful in conveying the timing and magnitude of these effects. This analysis would be helpful in clarifying some misconceptions, such as the notion that we can grow our way out of the problem without other changes to the program, or that the financing challenges result solely from a demographic bubble that we can weather with a few minor changes.
Furthermore, cash flow tables would be extremely helpful when comparing and contrasting specific policy recommendations. In evaluating the effect of increasing the payroll tax cap, for instance, one could see by how much the top line -- income from payroll tax -- would increase. Similarly, this analysis would convey the extent to which shifting from wage indexing to price indexing would affect benefits over time. Proposals to create private accounts could be evaluated in the same manner. If a private account plan specified revenue or benefit changes, they would be reflected above the line and to the extent a plan depended on general revenue transfers, that would be reflected below the line and shown as a share of the budget. Over time, the money available from the accounts would provide another source of income to be added to tax revenue.
In addition to adding cash flow tables, there are a few other changes that would be helpful. First, actuarial solvency is somewhat confusing not only because it takes into account the Social Security trust funds while ignoring where those funds will come from, but also because it is evaluated over a 75-year period. This focuses attention on policy changes necessary to keep the program balanced over that period and that period only. This approach to reform suffers from the problem of the "cliff effect" where when the evaluation period is lengthened by a single year, the program promptly falls out of balance, thus making further changes necessary. It would arguably be better to evaluate the program's well being in perpetuity, shifting attention away from actuarial solvency over a limited time period to sustainability -- a far more important objective. Calculations reported in recent work by Kent Smetters and Kevin Brennan show that the actuarial shortfall is twice as large when evaluated in perpetuity. In particular, the shortfall increases from around $3.3 trillion over the next 75 years to over $6 trillion when evaluated in perpetuity.
Finally, the Trustees should consider including tables that show both lifetime benefits and net transfers on a generational basis. Lifetime benefits rather than average annual benefits would be helpful in reflecting how costs rise along with increases in life expectancies.
Net transfers -- the present value of a generation's benefits less the taxes they pay -- would be useful in evaluating generational equity. Smetters and Brennan show that this measure gives a more objective view of liabilities than standard trust fund accounting. We could, for instance, make Social Security appear to be healthy by all evaluation techniques including cash flow by simply passing a law that the payroll tax would be increased as necessary to cover promised benefits. The program would be actuarially solvent, cash flow deficits would be zero, and the trust fund would never dip below zero. But younger workers and future generations would suffer huge losses, which would be captured in a net transfer evaluation while missed in other assessments.
To conclude, the integrity with which the Trustees' Report is constructed and its unbiased content play a crucial role in providing the information needed to evaluate the financial health of Social Security. My suggestions here should in no way be taken as a criticism of the work that is currently done, but rather as suggestions about other information that might be useful for the purposes of analysis and comparison. I look forward to your questions, and once again, thank you for holding this hearing and inviting me to testify.
