In the Nation's Interest
“Why Aren’t Social Security and Medicare Means Tested?”
By Joe Minarik
A few weeks ago, I had the pleasure of speaking to a twice-yearly luncheon of the BOB (Bureau of the Budget) and OMB (Office of Management and Budget) Alumni Association. These meetings are always a pleasure. My former colleagues were among the most dedicated public servants you can find. They were willing to go above and beyond for every president of either party, and to provide impartial, fact-based advice from a wealth of institution knowledge that they carefully preserved across generations of career staff. And most of them are older than I am, which is an encouraging change from just about every other setting in which I find myself these days.
The subject that the group asked me to address was economic inequality. (We have just discussed that issue in a CED monthly Trustee conference call as well.) This BOB/OMB conversation provoked a question from one attendee, a former political appointee in a Republican Administration. He asked: Why are you Democrats so opposed to means testing Social Security and Medicare?
The premise of the question, which I recall the questioner articulated at least in clear shorthand, was fairly straightforward: The overall federal budget is in trouble, and the contributory and putatively self-financing Social Security and Medicare programs are not living up to their financial billing either. Although many of the elderly are poor in terms of both income and assets, some have comfortable (or even more than that) wealth, which throws off comfortable (or even more than that) income, and yet they still receive benefits from these universal programs. So why would anyone resist means testing those programs to close their financing gaps – through reducing the benefits for those who need the benefits the least, without affecting those who need the benefits the most?
You (like me) probably have heard this question often, and have seen it in print as well. Accordingly, it would seem to be worth discussing in some detail.
And the answer, in a nutshell, is that Social Security and Medicare are already means tested. This answer comes as a surprise to many people, so an explanation is called for. But that answer will not preclude a follow-up question as to why Social Security and Medicare should not be even more aggressively means tested than they already are, and that follow-up question deserves an answer, too; we will come to that in due course.
But to begin, some backup for the (to some) surprising assertion. The notion that Social Security (Medicare discussion to follow) is not means tested flows from a fairly simple view of the program’s cash benefits. Retirees who earned comparatively higher wages receive higher benefits than those who earned lower wages, the reasoning would go, therefore the program cannot possibly be means tested. Furthermore, the amount of wages subject to the payroll tax that funds the program is capped – and therefore the tax is regressive, and so the overall program must be pro-rich and anti-poor.
But as some understand, that simple reasoning misses two key aspects of the program’s operations. First, the program’s benefit formula favors lower-wage workers. Benefit amounts are based on the lifetime earnings history. The first dollars of a worker’s wages are replaced at a 90 percent rate. Earnings in a second bracket (mechanically like an income-tax-rate bracket) are replaced at a 32 percent rate. And any earnings above that level are replaced at only a 15 percent rate. Thus, although higher-wage workers receive more dollars in absolute terms, they receive less back in Social Security benefits per dollar of tax paid over their lifetimes. At the extremes, the difference in the implicit rate of return on those contributions is enormous.
The second key program feature is that a fraction of Social Security benefits can be subject to income taxation, on a progressive basis – and then the income tax that applies to those included benefits itself is progressive. Once a beneficiary’s total income (including half of Social Security benefits) exceeds $25,000 ($32,000 for a married couple), the first dollar of benefits begins to be taxable, up to inclusion of one-half of benefits. And once income (including half of Social Security benefits) exceeds $34,000 ($44,000 for a married couple) the portion of benefits included in taxable income begins to rise further, up to a maximum inclusion of 85 percent of benefits. This provision is designed to have no effect on the low-income elderly, while gradually increasing its impact as total incomes rise.
The Social Security Administration estimates that the average beneficiary in 2013 received an annual benefit of $15,528. If that person had no other income, he or she would owe no income tax – on Social Security benefits or anything else (although that person certainly would not enjoy a luxurious lifestyle based on that benefit). As incomes from whatever source rise above the $25,000 threshold noted above, benefits progressively become subject to tax, at income tax rates that start at 10 percent and rise according to the total amount of taxable income. Tax is calculated after a $3,900 personal exemption and a $7,600 standard deduction (for single persons of age 65 or older). Thus, a beneficiary with a taxable income of $11,500 (from taxable Social Security benefits or any other source) would owe no income tax, with tax beginning to accumulate above that level only at the bottom-bracket 10 percent rate.
The theoretical maximum annual benefit in 2013, for a lifelong maximum wage earner retiring at the full eligibility age, was $30,396. If that person also received other income of roughly $40,000, the full 85 percent of that person’s benefit would be subject to income taxation. At the absolute upper end of the scale, a beneficiary in the highest income tax rate bracket with 85 percent of benefits subject to tax would pay tax on benefits at the margin at a rate equal to about 34 percent (that is, 85 percent of the 39.6 percent top-bracket tax rate). And to get into that top income tax rate bracket, that taxpayer would need to have a taxable income, after the personal exemption and any standard or itemized deductions, of over $400,000.
In other words, the income tax due on Social Security benefits is quite progressive, with benefits for probably most of the elderly not subject to tax at all, but with more than a third of benefits paid back in income tax by the most well-off. Taking into account the lower benefit-formula conversion rates that apply to high-wage retirees, their after-tax returns on their lifetimes of payroll tax contributions, measured in investment terms, would be quite modest – if not in some instances negative (further reflections on that fact later). So Social Security is means tested in reality, if not in name.
Then there is Medicare. The conclusion on Medicare is the same, but more so. For starters, consider that the nature of the Medicare benefit is different from Social Security. The high-wage Social Security beneficiary might get a significantly lower rate of return on his or her past contributions, but still gets more dollars in absolute terms. However, the high-wage Medicare beneficiary get exactly the same health-insurance policy as the lower-wage retiree.
But to get that identical insurance policy, the higher-wage Medicare beneficiary paid much more in taxes. This was true even some years ago, when the Medicare payroll tax was collected on the same capped wage base as the Social Security payroll tax. But effective in 1993, the maximum earnings cap for the Medicare payroll tax was eliminated. Then in 2013, an “Additional Medicare Tax” of 0.9 percent was imposed on earnings from labor in excess of $200,000 for a single person ($250,000 for a married couple). In the same year, an additional “Net Investment Income Tax” of 3.8 percent was imposed on such incomes of persons above the same income thresholds.
So the original, basic Medicare payroll tax could have been said to constitute a means testing of Medicare, but all three of the additional provisions – the lifting of the earnings cap for the regular Medicare payroll tax, the “Additional Medicare Tax” on labor earnings and the “Net Investment Income Tax” on income from property – all add to the effective means testing of Medicare. Because of all of these features, persons with higher incomes pay more for the same Medicare health insurance coverage than do persons with lower incomes.
Perhaps the most common first-blush view of what would constitute means testing of Medicare would be in some way to have more-affluent beneficiaries pay a higher percentage of their medical bills, such as through higher deductibles or co-pays. But that would have the unattractive properties of burdening people when they are sick, and of collecting more from sick beneficiaries than from healthy beneficiaries who are precisely equally well off. That probably would strike people, upon careful reflection, as unfair. It also would raise administrative issues, particularly if Medicare beneficiaries ran up large bills in a final illness or injury. How do you collect at that point? And if you don’t collect, how do you justify the unequal treatment relative to someone else who is identically situated but who survives, and then has to pay the higher medical bills? The real benefit that people get from Medicare is insurance coverage against risk, not medical services; and the coverage against risk is the same whether the risk materializes or not.
Thus, people who call for “means testing” of Social Security and Medicare are just a little late; means testing is already there. In fact, it arguably was embodied in the original features of both programs, and it has been made only more pronounced over time.
As noted at the outset, one might argue that even the current means testing is not enough. And truth be told, as both programs are altered to make their financing sustainable over the long haul, the kinds of features described above are likely to be tweaked still further (although many people probably still will not recognize them as “means testing”). However, there is an arguable case that we should employ some caution if we proceed down the road toward greater effective means testing.
First of all, the kinds of conditional additional taxes now imposed at upper-income levels add complexity to the tax system. They are far from transparent, and might catch out especially persons who just exceed the thresholds and might not have expected the extra charge that appears only when they fill out their tax returns well after the end of the income year in question.
Second, those additions to marginal tax rates do reduce incentives for productive activity. And perhaps to greater effect, they increase incentives to engage in manipulation and create tax shelters to avoid the higher burden. The result is an endless game of cat and mouse between the taxpayer and the revenue authorities, in which successive and ever-more-intricate manipulation schemes are answered with successive and ever-more-complex regulations that in the end erode taxpayer morale and benefit no one.
And finally, although Social Security has always aspired to provide at least a fair return on taxes paid by all income groups of workers (understanding that because of the annuity nature of the program, longevity will play a major role in determining the return for any particular worker), that aspiration is being challenged – notably for high-wage retirees. The ideal of an investment return for all that is at least fair helps to maintain a societal consensus in support of the program. Conversely, the pressure that we sometimes feel from people who perceive that their own interest would be better served if they could opt out of the program threatens to undermine its political viability in the long term. The returns received by particular workers are driven by circumstance. Social Security is relatively more generous to the traditional “Ozzie and Harriet” family (where Ozzie is a lifetime high-wage worker, and Harriet stays at home, keeps up the TV program filming stage, and raises the two future rock-and-roll stars) than it would be to Ozzie alone (if he chose to live the life of a free spirit, unburdened by the wife and kids), for example. But generally, the higher the income, the lower the investment return to Social Security contributions.
There are any number of sides to the inevitable debate about the generosity of Social Security to different income groups. A vigorous line of argument comes from some who observe the shorter average life spans of low-wage workers, and who therefore claim that Social Security provides a bad deal to the worst-off. But that claim is answered by others who point out that for the very same reason, Social Security’s disability and survivor benefit programs provide correspondingly greater protection to those same low-wage workers. There is a continuing administrative challenge from that same disability program, where there are mutually conflicting needs to provide prompt rulings on behalf of suffering applicants, but also to investigate fully the inherently complex claims that those applicants submit. Disability can be very hard to verify.
Some people argue that any pairing of the payroll tax with Social Security or Medicare benefits to assess the fairness of the entire programs is artificial; these people want a progressive retirement benefit financed by a progressive income tax. Others counter that the American people want an earned benefit, and argue that the contributory nature of the programs is responsible for their generally fiscally responsible handling up until this point – benefit increases always have been paid for, at least in keeping with the best estimates at the time – and for the programs’ extraordinary longevity and perceived success. I could go on with this back and forth, but you would be late for dinner. Tomorrow.
The central underlying issue probably boils down to the nature of social insurance itself. “Social insurance” contains two words. The “social” part suggests some measure of assistance from the more well-off to those less fortunate. “Insurance” denotes a program where individuals are willing to pay in the event of a good outcome to lessen the blow of a bad outcome – like fire insurance on a home, or accident insurance on a car. Those terms paint an innocuous picture – until some people contemplate a future in which they believe they will receive less from Social Security or Medicare than they have hitherto paid in. We all would like to enjoy a heads-I-win-tails-you-lose system, from which we get both protection from a bad outcome and a fair investment return in the event of a good outcome. In the real world, that is decreasingly possible as demography has its inexorable effect.
In its early years, the real-world performance of the Social Security program hid an inherent long-term tension in the concept of the program. The early Social Security program has been characterized as “immature” in the sense that upon its inception it could pay immediate benefits to an elderly population that had not contributed at all. For those persons, every benefit dollar constituted a windfall – and therefore a much better-than-fair return. Social Security is now fully mature in that respect, however, with virtually all of today’s beneficiaries having contributed over their entire working lives. The financial comparison of today’s benefits against the actual contributions paid in is therefore obviously far less favorable than it was in the program’s early decades. Add to that the demographic challenge from steadily increasing longevity and falling birthrates, plus the population irregularity of the baby boom that could not have been forecast at the creation of the program, and what had been an easy-to-achieve better-than-fair deal for all has become a close calculation for some, and therefore a much more difficult political lift.
So against the easy early favorable outcomes from an immature pay-as-you-go system now are weighed the hard realities of a mature program in challenging times. We need to make policy decisions going forward on the basis of the current mature system. And that raises questions about the core values behind the program.
Looking at the concept of social insurance very broadly and considering big-picture options, some might prefer to abolish Social Security (or make it voluntary, to the same effect) and strike a firm bargain with all workers at the beginning of their careers: Save for your own retirement or deal with the consequences on your own when the time comes. In semi-practical terms, that raises the unaffordable problem of how to finance the benefits of today’s retirees who already have contributed to Social Security over their entire working lives. But such show-stopper issues aside, even the strongest opponents of Social Security might want to give the issue a re-think. Even taking such a free-market approach on its own terms, Social Security may look like the best possible protection for society. There can be circumstances beyond the control of even the conscientious. In the absence of Social Security and Medicare, society might find it impossible to refuse the claims of the worst-off of the elderly, whatever explicit bargain was struck at the beginning of the working life. Dealing with such situations on a piecemeal basis with outright welfare programs probably would be the worst of all possible worlds. Social Security and Medicare might avoid such an unpleasant outcome for society.
There is a value to today’s defined-benefit Social Security and Medicare programs. They fill a void that has been left by the demise of private defined-benefit pensions and retiree health-insurance programs, particularly for working people with modest incomes. But given economic and demographic realities, that void will become gradually ever more difficult to fill.
We face hard choices going forward. And although Social Security and Medicare already are means tested in a realistic sense of the term, we may need to revisit those decisions, and perhaps to push them further. We will need to proceed with care, weighing the risks on both sides.