Reforming Social Security: A How-To Guide

In early October, the Mercatus Center published my guide to designing comprehensive Social Security reforms. The guide was written in response to interest expressed by federal lawmakers and their staffs about how to best avert Social Security insolvency. Time is rapidly running out to fix this problem. A solution enacted today would already need to generate savings roughly equal to a 27% across-the-board reduction in future benefit claims. By the time Social Security’s combined trust funds are nearing depletion in 2034, even suddenly stopping all new claims would not avert insolvency. Faced with this impending crisis, lawmakers have two choices:

1. Act responsibly, enacting comprehensive reforms to repair Social Security’s financing shortfall and address other problems facing the program.

2. Be irresponsible, papering over the financing shortfall with accounting gimmicks and further escalating federal debt.

What they can’t do is nothing, because then the program would become insolvent and the benefit checks would stop going out. This would be an intolerable situation not only for beneficiaries but for all the politicians who depend on their votes, so one way or the other, legislation will be enacted. My guide is meant to inform lawmakers who wish to take the responsible approach.

Social Security’s Unique Design

Social Security is unusual among federal income-support programs. In so-called welfare programs, the government doesn’t keep track of what each individual taxpayer has paid to support them. Those programs are financed from the general government fund, meaning they are supported largely by income taxes, which not all Americans pay. Eligibility for welfare program benefits is determined by perceived need, and thus many Americans will never receive such benefits.

Consequently, with a welfare program you have one set of people paying for it, and a different set of people collecting from it. This creates a competition of interests, which naturally results in frequent reconsideration of the terms of exchange, including eligibility rules, benefit levels, means tests, work requirements and other parameters. You would be very foolish to base a long-term personal income strategy around getting benefits from a welfare program, because you don’t know what the benefits of those programs will be years from now, or even if you will be eligible.

By contrast, people can and do make long-term income plans around the presumption of a certain level of Social Security benefits. They can do this because Social Security benefits are reliable and predictable in a way welfare benefits are not. This phenomenon is rooted in how Social Security is designed and financed. Workers make dedicated payroll tax contributions to Social Security that are separately quantified on their W-2 forms. The federal government represents to workers that these amounts are credited on their behalf to Social Security trust funds, which are permitted to spend only on Social Security obligations. In addition, the trust funds are not permitted under law to borrow or spend in excess of the financial resources they collect. Unlike with welfare, workers from the poorest to the richest both contribute to and collect from Social Security, and each individual’s benefits are a mathematical function of that individual’s earnings subject to the payroll tax. This is what enables voters to say to politicians at all times that the rules of the game can’t be suddenly and arbitrarily changed – because Social Security, unlike welfare, is a contributory insurance system whose financial resources were put there by the same individuals who are claiming benefits.

There is, however, a price to Social Security’s unique substantive reliability and political strength. This kind of system only works as long as lawmakers remain willing to do what they have always done in the past: namely, maintain a benefit schedule that aligns with what worker contributions can actually fund. If lawmakers are no longer willing to do that – if they pay out benefits without regard to the amount of revenue the program brings in – then Social Security can no longer operate as it does now. We could have a Social Security system, but it would need to be financed either wholly or in part from the federal government’s general fund, which means there would no longer be any causal connection between the amounts workers contribute and what they receive.

In my view, this would be a very bad change, for it would simultaneously eliminate the financing discipline historically imposed by the trust fund system (because then benefits could be paid without tax financing), while also sacrificing the predictability of the benefit structure (because benefits would simply be a function of lawmakers’ caprice, not of what workers’ contributions can pay). Further, the voting public has shown no indication that they wish to abandon Social Security’s current design and replace it with something more like welfare. Unless and until someone convinces the public to shift to a fundamentally different type of system, we must determine how to best finance the one that we have. That’s what my guide aims to help lawmakers do.

Benefits, Taxes and Time

Social Security policy basically boils down to three things:

1. The monthly benefits the program pays

2. The tax burdens it imposes on workers

3. The fraction of Americans’ lives spent drawing benefits

Most people intuitively understand that if we want to pay higher benefits, we must impose higher taxes. It is equally important, however, to understand the role played by the number of years that participants spend collecting benefits rather than making contributions. If Americans continue to claim benefits at the same ages despite living longer lives, then one of two things must happen: Either workers’ standards of living must drop (as they are hit with higher tax burdens to finance more years’ worth of benefits), or beneficiaries’ standards of living must drop (as the benefits incoming taxes can finance must be spread over more years, meaning less can be paid each month). A decision to leave current eligibility ages in place is effectively a decision to reduce Americans’ living standards.

Because Social Security costs automatically grow faster than its tax base of worker earnings, the problem can’t be solved just by raising taxes. Unless cost growth is moderated, the best a tax increase can do is buy time before taxes must be raised again. Similarly, it’s too late in the game to realistically close the shortfall solely through benefit changes. A couple of decades ago the problem could have been fixed easily and essentially painlessly without a tax increase, simply by changing the indexation of benefit growth from the Average Wage Index to the Consumer Price Index. Unfortunately, lawmakers have procrastinated for so long that’s no longer the case. Today, maintaining solvency without tapping any additional tax revenues would require a real decline in future benefit levels relative to inflation.

Thus, despite ongoing partisan posturing, any realistic solution will require a combination of eligibility age changes, moderation of benefit growth and increased tax revenues. Since there is no objectively correct way to balance these three options, my guide assumes a starting point of doing roughly one-third of the job through each mechanism. Lawmakers can dial any of these options up or down according to their own value judgments.

Choosing Among Competing Values

To develop a solution that usefully serves future Americans, lawmakers must be mindful of why the shortfall exists, as well as who is being treated better and worse under current law. The shortfall exists because people who have already entered the system (either as beneficiaries or current workers) are scheduled to receive benefits far exceeding what their own tax payments can finance. Social Security can’t create income out of nothing, so if it gives windfalls to one group, it must extract the money from another group.

Social Security generally treats those born earlier better than those born later. The excess of benefits over taxes for those who have already entered the system equals 3.8% (in present value) of all projected future earnings of American workers. Translated into English, this means that unless older Americans contribute something more to restoring solvency, enormous burdens will befall younger workers, to the extent of making them nearly 4% poorer over their lifetimes. And if Social Security makes the average worker nearly 4% poorer, it would make many below-average-income workers poorer as well. A solution that relies solely on raising taxes on future earners, while older generations continue to collect large windfalls, cannot result in a future Social Security system that serves as an effective bulwark against poverty.

Another value judgment lawmakers must make is how progressive or redistributive Social Security benefits should be. Two long-standing features of Social Security exist in some tension with one another. On the one hand, the benefit formula is progressive, giving lower-income workers a greater rate of return on their contributions than higher-income workers receive. On the other hand, it’s not welfare: Rich and poor alike pay into Social Security, and both accrue benefits whenever they do. Social Security’s progressive benefit structure serves its objective of providing income insurance against poverty, while its contribution-benefit link underpins its political support and renders its benefits more secure.

There is a consensus across the U.S. ideological spectrum that Social Security should continue to be progressive, and specifically that higher-income workers should bear the financial burdens associated with maintaining solvency. But critical details of implementing these consensus judgments still need to be decided. Specifically, how much of a worker’s contributions should go toward accruing benefits for herself, as opposed to being redistributed to protect lower-income households? In addition, how much should the better-off contribute through higher taxes, and how much through moderated benefit growth?

Regardless of how those questions are answered, one thing is clear: Lawmakers must moderate benefit growth for high-income earners. This is true, first, because current cost growth exceeds sustainable rates and must therefore be slowed, and no part of the U.S. political spectrum would support cutting benefit growth for low earners but not high earners. It’s also true because recent U.S. income growth has been concentrated among older Americans. We cannot have the Bill Gateses of the world contribute meaningfully to solvency by raising the payroll tax rate or the payroll tax base, because they will not be taxpaying wage-earners in the decades ahead, but rather beneficiaries. Moreover, a solution that focuses exclusively on taxing higher earners will concentrate the burdens on younger workers just coming into the labor force – the very group getting hit the hardest under current law. No plan can enhance intergenerational equity or optimize progressivity unless it slows benefit growth on the high-income end.

Options for Reform

My guide reviews options for adjustments in three categories: eligibility ages, benefit growth and additional revenues. Social Security’s current eligibility age requirements are unrealistic in the 21st century because they result in Americans spending a continually rising proportion of their lives as beneficiaries. Back in 1940, when life expectancy at age 65 was seven years shorter than it is today, one could not claim benefits until age 65. Today the full retirement age has barely moved (to 67) and the most common age of claim is actually 62, an early eligibility age established in the midst of the baby boom.

Beyond the obvious financial strains it places on the system, the current position of the early eligibility age (62) at a full five years younger than full retirement age (67) harms beneficiaries. Social Security benefits are automatically adjusted downward the earlier they’re claimed relative to the full retirement age, Thus, beneficiaries who claim five years early, at age 62, are locked into low benefit levels for the rest of their lives, with a greater risk of ending up in poverty after outliving their savings. The earliest eligibility age should be adjusted back to 65 as soon as possible, and the full retirement age gradually adjusted to reflect rising life expectancy; the guide offers illustrative schedules for doing so.

The guide also offers multiple options for moderating benefit growth. Benefit growth changes are an especially powerful engine for financial corrections, because even a slight change in the rate of growth compounds to enormous savings over time. Lawmakers could actually get most of the way toward a permanently sustainable system simply by reindexing the growth of the benefit accrual brackets from wage inflation to price inflation, merely a ~1% difference in the annual growth rate. The guide illustrates a variation on this option: reindexing the brackets to the midpoint of wage growth and price growth, which would still allow for real benefit growth over time – that is, benefits that grow above and beyond the cost of living.

Speaking of the cost of living, correcting the Consumer Price Index from its current, inaccurate CPI-W to the chained C-CPI-U index would by itself close about a sixth of the shortfall. Current CPI-W overstates price inflation on average by about 0.3 percentage points a year by not adequately accounting for changes in consumer purchasing patterns. Fixing CPI is also one of the few ways that boomers and older Gen Xers can contribute a little something to the solution, given that lawmakers will not want to cut benefits for current retirees. By contrast, continuing to inflate the COLAs paid to current beneficiaries – the generations already coming out farthest ahead – unnecessarily worsens the financial shortfall and should be stopped.

Raising taxes is the least attractive category of reform options for several reasons, but unfortunately lawmakers have procrastinated so long that by this point additional tax revenues will be needed. Without additional revenues, Social Security will not be able to maintain current real benefit levels without becoming insolvent. To avoid a future benefit decline, lawmakers could either raise the payroll tax rate, broaden the compensation base subject to the payroll tax, or increase the maximum amount of earnings subject to the tax.

This last option of raising the “tax max” is by far the most commonly endorsed reform option, especially among left-of-center advocates. Unfortunately, it would do far less to resolve the shortfall than many assume. Even if every penny of future U.S. wage earnings were subject to the full weight of the payroll tax, it would close barely more than one-quarter of the program’s long-term annual cash deficits. The guide illustrates a less extreme option of taxing 90% of all national earnings, roughly matching the high point over Social Security’s history. This would close between one-fifth and one-quarter of the actuarial shortfall, though closing less in the later decades of the projection period. As a general rule, the financial efficacy of tax increases fades over time unless cost growth also slows, whereas the financial benefits of slowing benefit growth increase over time.

The Road to Reform

Finalizing a reform plan involves a lot more than adding up provisions on a table and aiming to close 100% of the shortfall. The plan only works if it serves people in Social Security, and there are two main criteria advocates use to assess reforms: equity and adequacy. Equity addresses whether the program is fairly treating differently situated participants, and adequacy addresses whether benefits are sufficient to serve as useful income insurance upon departure from the workforce.

The Social Security chief actuary’s office provides analyses that speak to both questions. With respect to equity, for example, they calculate “money’s worth ratios” for different participant prototypes. These ratios compare the present value of an individual’s lifetime benefits to the present value of their lifetime contributions to Social Security. A value of 1 means the individual has come out even. If it’s higher, then they have come out ahead. If it’s lower, they have lost money (in each case, the ratio accounts for what they could have generated by saving those contributions and receiving interest earnings). Under current law, low-wage workers get better returns than high-wage workers, which is reflected in their higher money’s worth ratios. Most advocates wish to preserve this feature of Social Security. Those born later get much lower returns than those born earlier, however, a trend that must be corrected if Social Security is to serve future generations.

Lawmakers and their staffs need to know how to accurately read the actuaries’ money’s worth memoranda to craft good policy. Some of the tables in these memoranda don’t really mean anything – for example, table 1 assumes that scheduled benefits are paid without anyone being taxed sufficiently to pay them, an assumption that naturally generates higher apparent returns than would occur in real life. An example of a meaningful table is 3a, which calculates returns based on what Social Security could actually pay. A good plan would raise the money’s worth for a medium-wage two-earner couple born in 2004 from its current 0.85 to closer to 1, while lowering the returns for those born in 1964 and 1973 from their current levels, which are much higher than 1 (meaning that under current law they take out much more than they contribute).

Equally important is understanding how to fairly measure benefit adequacy. A trick often played by partisan opponents is to compare benefits under reform plans – not to today’s benefit levels or to what Social Security could actually pay in the future – but to current-law benefit schedules at a distant date decades into the future. This comparison is grossly misleading because those benefit levels would be much higher than today’s in real terms, would require an enormous tax increase to fund, and wouldn’t be paid under current law anyway. A better analysis is to show the growth of benefits at the full retirement age under the plan, adjusted for CPI, as in one column of these tables generated by the Social Security chief actuary. A good plan will maintain current benefit levels or increase future benefits relative to inflation, although the future rate of increase would need to be much slower than the unsustainable growth rate to date. The bottom line is that benefit adequacy is properly measured by quantifying the purchasing power of beneficiaries, not by comparing to an inflated baseline that cannot realistically be funded.

It is my hope that with the aid of these materials and other similar resources, federal lawmakers can craft and enact comprehensive reforms to maintain Social Security’s solvency and to enable the program to usefully serve future generations. If they can do so, the hundreds of millions of Americans who participate in Social Security will benefit enormously.