Click here for a printer-friendly version of this article.
This is the second of two pieces explaining why I have a more favorable view of the Simpson-Bowles Social Security plan than that expressed in a recent paper from the Center on Budget and Policy Priorities (CBPP). In the first installment, I listed instances where I mostly agreed with their analysis – sometimes agreeing with the policy criticism, sometimes not. In this piece, I give instances where I have disagreements with both the critical analysis and the policy conclusions.
To see the first part of this analysis, click here.
Criticism #4: The Plan “Relies Excessively on Benefit Cuts.”
I have a fundamental disagreement with some characterizations of Simpson-Bowles’s relative reliance on benefit cuts to achieve solvency. The statement was made that the lion’s share of the plan’s savings “comes from benefit cuts. The benefit provisions account for nearly two-thirds of the savings over the 75-year period and four-fifths of the savings in the 75th year.”
These figures are arrived at by dividing the plan’s provisions into revenue and benefit pieces, and then adding up the relative size of provisions in each category. I do not believe this method provides for a complete, accurate assessment of the plan’s total effects. The total reliance of Simpson-Bowles on benefit changes is much less.
The reason is that some plan provisions affect both revenues and benefit obligations. Raising the cap on taxable wages, for example, increases total system revenues, but it also increases total benefit obligations. So too does the plan’s provision to expand coverage to state and local workers. Simply listing these provisions as being on the revenue side does not capture their full effects.
Fortunately, a memorandum from the SSA Actuary has done most of this work for us. According to the memo (see Table 1d), roughly 46% of the plan’s net improvement in solvency over 75 years would be achieved through increased revenues, the remaining 54% from cost containment. This is merely an average over 75 years; it varies along the way. By the 2080s, roughly 65% of the annual improvement would be from cost containment. But in the program’s peak cost years in the 2030s, roughly 56% of the improvement would come from increased revenues. This evolution over time can be more readily seen on the chart below.
This particular allocation may not be to everyone’s liking (again, I personally would prefer a plan that does more to slow cost growth). It is, however, a significantly smaller reliance on benefit changes than has been suggested. Simpson-Bowles charts a resolutely middle ground on the benefit-revenue spectrum.
In this respect, Simpson-Bowles contrasts significantly with the imbalanced proposal of the Bipartisan Policy Center, which would raise revenues by enough to resolve the entire Social Security shortfall (and would actually increase total cost obligations).
A General Concern about Benefit Analyses
CBPP’s paper asserts that current Social Security benefits are modest and uses “wage-indexed dollars” and “scheduled benefits” as reference points to conclude that the proposal would “cut” benefits. I am not arguing with their characterizations of today’s benefit levels; my perspective instead focuses on future benefit levels. I’m concerned that scheduled benefits significantly exceed the amount that can be financed from tax rates Americans have been willing to pay. I also believe that a fuller portrayal of Simpson-Bowles would show that future per-capita benefits would rise substantially (not fall) under the plan.
Whether Social Security benefits are modest is an inherently subjective value judgment. Gauging the appropriate level of Social Security benefits would be much easier if Americans funded their own Social Security benefits in advance; in that instance, the adequacy of retirement benefits could be weighed against the contribution requirements that Americans were willing to bear.
But this is not how Social Security works. The first generations of Social Security beneficiaries received benefits that far exceeded the value of their own contributions. Their benefits were financed by taxing the next generation of workers. Similarly, benefits for today’s workers will be financed primarily by taxing future workers.
Accordingly, we don’t know how much each generation is willing to pay for its own benefits. The current benefit formula only reflects what each generation is willing to ask the next generation to pay for.
The case for Social Security’s current benefit formula would be far stronger if Americans were demonstrably willing to carry the tax burdens necessary to fund it. Thus far, this has conspicuously not been the case. Funding the current formula would require a total program cost rate of 16.7% facing workers in the mid-2030s – fully one-sixth of their total taxable wages. We do not yet know whether our children will be content to carry this burden. We do know, however, that current and previous generations have repeatedly declined to impose comparable tax burdens on themselves to meaningfully advance-fund their own benefits.
The 1976 Consultant Panel on Social Security declared about the current benefit formula just then being adopted, “This Panel gravely doubts the fairness and wisdom of now promising benefits at such a level that we must commit our sons and daughters to a higher tax rate than we ourselves are willing to pay.” But this is exactly what the current benefit formula would do.
Because of this, I believe we should be very circumspect about making “scheduled benefits” and “wage-indexed dollars” the sole reference points for benefit analyses. Embedded in these figures are very subjective value judgments that the appropriate policy goal for Social Security’s benefit formula should be “wage replacement” – that is, that benefits should rise faster than inflation in proportion to wages – even though this would cause tax burdens to rise markedly over time. That subjective view is held by many respected advocates, but it is not the only view. It is also a view that leads to outcomes that many would find highly undesirable.
I also don’t believe that most people know what “wage-indexed dollars” are. The use of such figures, I believe, again reflects a particular policy viewpoint rather than how most people understand a presentation given in dollar terms. Many people are likely to look at such figures and conclude (wrongly) that the future purchasing power of seniors would decline, although what they actually show is a comparative point: that benefits simply wouldn’t rise as fast as wages would.
For this reason, I believe that it is more useful to show future benefit levels in real (inflation-adjusted) dollars. Presented this way, it becomes clear that the purchasing power of seniors would continue to increase over time under Simpson-Bowles, not be reduced from current levels.
I also have a concern about all presentations that appear to show benefits declining, not only by using “wage-indexed dollars,” but because they also show benefits only at the claim age of 65 – when in fact the normal retirement age (NRA) is higher under current law. Individuals cannot receive their full retirement benefit until NRA – now 66, soon to head to 67. The apparent reduction we see at age 65 on such graphs is really just the normal actuarial adjustment for early retirement – not an actual decline in full benefit levels.
Elected officials in 1983 made a deliberate policy decision to increase the retirement age; if we instead present our benefit projections only for age 65, we are implicitly making a contrary value judgment that we should frame our benefit targets around what individuals can claim at that age. Not only does this run counter to legislators’ apparent intent in raising the NRA, but it would establish an even more expensive standard for benefits – one resulting in even higher projected tax burdens than the current formula.
A General Concern about Retirement Age Alternatives
The CBPP paper lists – without endorsing – a number of proposals floated to shield physical laborers from the effects of raising the retirement age. Some of these would establish differing individual ages of eligibility based on such factors as lifetime earnings, years of work, or occupation – all with the aim of allowing physical laborers to retire earlier than others.
Unrelated to any of CBPP’s analysis, I believe these are all very bad ideas. Social Security should continue to offer the same retirement eligibility ages for everybody.
An individual’s benefit at a given age of claim is a function of two factors: the benefit formula (now based on one’s wage history; it could if desired be based on years of work as well), and the NRA. One can choose to retire early or late, and to accept a proportional benefit adjustment for making that choice.
Suppose we decide that we want to provide a worker with a given wage and work history a 10% higher benefit at a particular age of claim. This can be done directly through the benefit formula without also giving that worker a different eligibility age. Social Security is complex enough without having different eligibility ages for everyone. We simply don’t need to go down that road.
In a previous piece, I explained Simpson-Bowles’s proposed retirement age change. It is hardly draconian. It would phase up the retirement age in the future more slowly than is already being done under current law. Even toward the end of the 21st century, workers would still be able to claim benefits earlier than they could during FDR’s time.
Final Concerns and Conclusion
It is fair to critique individual provisions of Simpson-Bowles. Such criticisms are most useful if they are accompanied by clear guidance to elected officials as to a preferred policy alternative. One concern I have with some recent criticisms is that they can point two ways at once. It is reasonable, for example, to express the concern that the substantial progressivity of Simpson-Bowles would reduce work incentives and weaken the contribution-benefit connection. But if we instead leave the retirement age where it is, if we instead bolster SSI, and if we instead provide additional benefits for sporadic earners beyond those in Simpson-Bowles, the result will be a package that actually has far worse rewards for work.
Most Americans feel positively about Social Security despite various flaws and inefficiencies, and despite its uncertain future finances. Were Simpson-Bowles implemented, it would not be free of quirks and would undoubtedly require subsequent tinkering (though, I believe, substantially less than the 1972, 1977 and 1983 Social Security amendments are already requiring). I believe that Simpson-Bowles would ameliorate many of the flaws of current law, would more equitably target system resources, and would chart a middle ground between revenues and cost containment. Most importantly, by repairing system finances it would give Americans sound reason for added confidence in Social Security.
Charles Blahous is a research fellow with the Hoover Institution and the author of Social Security: The Unfinished Work.