Social insurance, like commercial insurance, is about protection against financial risk. In the United States, Medicare and the Social Security Administration's programs for retirement, disability, worker's compensation, and worker's life insurance have become dominant features of American public policy, amounting to more than 41 percent of the federal budget. Yet their fiscal centrality does not rest on anything like an understanding of what makes social insurance social—or why that is so important to American political life. This essay seeks to clarify the crucial differences between social and commercial insurance and elaborates on the conceptual justifications and distinctive operational features of America's social insurance programs.
How should health care be financed in the United States? At the origins of the American welfare state, in the 1930s, the advocates for more equal and adequate coverage would have answered, “through social insurance.” In the conditions of the Great Depression President Roosevelt acknowledged a need for immediate “relief” to families, with poverty as a precondition for eligibility. But his strongest defense of an increased role for American social policy was based on the principles of social insurance.
The concept had received growing attention in the United States in the decades leading up to the New Deal (Marmor, Mashaw, and Pakutka 2013; Rubinow 1913). In the mid-1960s two leading architects of Medicare—Wilbur Cohen, who served as undersecretary and then secretary of health, education, and welfare for President Johnson, and Robert M. Ball, then commissioner of Social Security—recalled that social insurance was part of their college educations. Cohen reported it was a topic in courses in sociology, politics, and economics. Other leading scholars of the time reported similar experiences.1 Roosevelt (1935) used the term in his message to Congress to accompany the report of his Committee on Economic Security. Social insurance per se was a theme that could make the headlines in major US newspapers. But that level of prominence declined sharply in later years (see figure 1).
This brings me to the first paradox of social insurance. Programs were created and developed to fit and extend President Roosevelt's vision: Medicare and the Social Security Administration's programs for retirement, disability, worker's compensation, and worker's life insurance. In fiscal terms, these programs have grown to dominate American social policy, amounting to more than 41 percent of the federal budget.2 Yet their fiscal centrality does not rest on anything like an understanding of what makes social insurance social—and why that is so important to American political life.
This matters because the principles and judgments incorporated in the concept of social insurance are central to the major policy controversies of our time. They relate to the backlash against the Affordable Care Act (ACA) and to the debate, rekindled recently, between the advocates of “Medicare for All” and those who advocate Medicaid expansion as the path toward universal health coverage. They are crucial to addressing the broader Republican critique of government's role in American social policy generally.
The “social insurance” category is still used by some analysts to define a template, an aspiration, and perhaps a fading dream (e.g., Morone and Fauquert 2015). But both the principles and distinctions denoted by the expression have been muddied, both by applying it very broadly, to almost all government programs to make people more secure, and by using other terms, with different meanings, to describe Medicare, Social Security, and other classic social insurance programs.
By the 1980s, leaders in the development of Social Security and Medicare were worried about the thin understanding of the programs' social insurance rationale among political elites and the general public. In 1988 Robert Ball and Wilbur Cohen addressed the reduced understanding of social insurance's fundamental premises. They solicited help from informed academics and knowledgeable retired officials and founded the National Academy of Social Insurance. Their hope was to counter the declining understanding of the programs that had in the 1970s become the targets of ideological and budgetary attack. Social Security retirement, Medicare, disability, and unemployment insurance were increasingly labeled as simply entitlements and charged with out-of-control spending and unaffordable benefits. Other critics advocated a much smaller social policy commitment, urging a less costly “safety net” for the deserving among America's poor citizens.
The National Academy of Social Insurance has helped preserve modest familiarity with the term in the policy sphere. Yet, in part because of Medicaid's programmatic importance in citizens' lives, much commentary conflates means-tested Medicaid with contributory Medicare and Social Security. Thus, Grabowski, Bruber, and Mor (2017) described Republican efforts to cut Medicaid under cover of repealing the ACA as “the largest single reduction in a social insurance program in our nation's history.” Paul Krugman (2017), Eduardo Porter (2017), and Thomas B. Edsall (2017) all deplored attacks on “social insurance” but misleadingly included means-tested programs like Medicaid within their definition.
How important is it that most contemporary reports on social insurance programs—and much social science scholarship—ignore their conceptual justifications and distinctive operational features? I would not be writing this essay if I did not believe, as one of three coauthors, that the title of our 2013 book identified an important problem: Social Insurance: America's Neglected Heritage and Contested Future (Marmor, Mashaw, and Pakutka 2013).
There are two issues that involve serious misunderstandings: the difference between social insurance and commercial insurance, and the difference between programs for which benefits are earned through contributions and programs with means-tested, often called “welfare,” benefits.
What Is Social about Social Insurance?
Social insurance, like commercial insurance, is about protection against financial risk. It is “insurance” in the sense that people contribute to a fund to protect themselves against unpredictable financial risks. These include outliving one's savings in old age, the early death of a breadwinner, the onset of disability that makes work difficult if not impossible, the high costs of illness, involuntary unemployment, and work-related injury. Contributions are not prices in a market and so, unlike commercial insurance, are not higher or lower depending on the customer's risk profile. Instead of a commercial insurance contract between enrollee and insurer, social insurance is for shared protection among participants in which they agree that each should pay for that protection according to their work income. The “insurer”—a government agency or, originally in Europe, a corporate body with a joint labor/management board—is the agent of the contributing enrollees. The social insurance contract, once created, cannot be voluntary and survive long. By law contributions are required, because otherwise adverse selection would be financially disabling.
Social insurance, then, spreads the costs of coverage according to a different logic than that of commercial insurers. In commercial insurance, price must reflect risk. Social insurance, by contrast, operates on the premise that contributions are calculated according to one's income and benefits are related to one's needs. But the central political feature of social insurance is that the contributors are also beneficiaries. This is not the case with social assistance programs with means-tested eligibility standards. As important as such programs are for those who experience poverty, taxpayers do not in general identify with welfare beneficiaries. And, finally, private insurers, unlike governments, cannot tax citizens to make up for losses. But private insurers do routinely try to select customers with an eye to signing up those least likely to experience the risk the insurance policy is to mitigate. That's why fire insurance is more expensive in poorer neighborhoods, not because insurance firms are racist.
The Vocabulary of Social Policy
Words make a difference to all thinking about public policy, but this is especially the case where conflicts of fundamental values are at stake. Consider, for example, the common use of safety net as a collective description of programs as diverse as Medicare and Medicaid, Social Security retirement pensions and food stamps, disability insurance and homeless shelters. This expression collapses the distinction between means-tested and social insurance programs. Further, the metaphor suggests that recipients have to “fall” into poverty to warrant the net's help. This is the opposite of social insurance “platforms,” programs on which one can count before economic risks arise. Note that safety nets can be high or low, porous or tightly knit. In American politics, generosity is not a label most analysts would place on welfare safety nets.
In recent years, much linguistic muddle has been created through the use of entitlements as the term of choice for discussing both social insurance and means-tested programs. One meaning of the term is technical and budgetary. Entitlements in American fiscal policy are simply those programs whose benefits and beneficiaries cannot be adjusted without changes in statute. Congress cannot appropriate less money without changing the rules that generate expenditure, and administrations cannot simply reduce a program's benefits to fit an appropriation. This meaning of entitlement became associated, in the budgetary world, with uncontrollable budgets (Tax Policy Center, n.d.). But not being controllable through the appropriation process is very different from being immune to change (White 1998).
What citizens believe about the appropriateness of a program differs sharply from the budgetary rules about changing its provisions. In ordinary discourse, entitlement can express the legitimate claim to social insurance benefits: that Social Security and Medicare beneficiaries have earned their benefits, and the government as their agent has a duty to provide them. Hence, there should be a stable governmental commitment to social insurance protection over long periods. The commitment is the basis on which people pay taxes to contribute to the programs.
Yet “entitlement” language is commonly used to express the opposite thought: that benefits are given, not earned. This view is an ironic result of the legal “new property” movement beginning in the 1960s, which expanded noncontributory programs through the judicial process (Reich 1964; Rosenbaum 2015). Then the benefit may be unjustified, and critics argue the giver—the government or the taxpayer—has every moral right to take it away.3
We see the power of the first view by default: few if any critics of Social Security or Medicare explicitly criticize their appropriateness. Instead, they concentrate on claims that the programs are unaffordable. As I describe below, Medicaid is more vulnerable to broader criticisms.
The Trust-Fund Paradox
Attacks on Medicare and Social Security combine budgetary diatribes—fearful projection of “uncontrollable” entitlements—with what I call solvency talk. When policy discussion turns to the fiscal projections of social insurance programs, critics and defenders alike turn to trust-fund language. If the old age retirement actuaries forecast a revenue projection of X in twenty-five years and the projected outlays of Y equal more than X, the Old-Age and Survivors Insurance or Disability Insurance “trust fund” is, according to this logic, in trouble. It will no longer have enough to meet its “bills” at that forecasted date. And if that shortfall were to continue, the necessary result would, in this frame of language, be insolvency.
The original use of trust-fund language in social insurance had more to do with trust than with funds. President Roosevelt rightly felt in the 1930s that the contributory ethos of social insurance would come to be central to its secure political status. A population believing that contributing workers had earned their social insurance benefit would not politically tolerate substantial budgetary cutbacks. The idea of a trust fund, then, was to emphasize the special status of a program whose benefits would be paid decades after a contributor's payments.4 It was language meant to highlight reliability, to suggest a governmental appreciation of an especially protected program. The sad and second paradox is that this language has been turned upside down, bringing needless fear that the funds will “run out.” Roosevelt's protective rhetoric backfired as the original understandings of social insurance weakened while the popularity of the programs remained substantial.
Trust funds as a mechanism and a metaphor create a view of old-age social insurance programs that might sound absurd for other programs. Will the retirement benefits Social Security has scheduled for the future be funded? Can one count on it? Anyone who asked whether the Defense Department will “be there” in 2040 would be considered at the very least odd. Nobody writes about the Defense Department going broke or becoming insolvent. Indeed, no sensible analyst would make twenty-, thirty-, or forty-year forecasts for defense expenditures. Uncertainties about the medical world may be as great as those about national security, but the trust-fund logic leads to health care expenditure forecasts long into the future. The very preoccupation with solvency generates anxiety.
In a further irony, promoters of solvency fears are inconsistent in applying the trust-fund logic. In budgetary terms, the fact that Part A of Medicare has a trust-fund device and Part B has not means talk of insolvency is doubly misleading. Part B funding is from general revenues and beneficiary premiums; it cannot go broke, but it can be reduced. Part A funding is based on dedicated payroll taxes, so its revenues can be seen as limited. Because American social insurance for medical care is predominantly limited to the elderly, demographic projections of a growing population over age sixty-five prompt concern. As a speaker I face questions about dire predictions of “insolvency” regularly. I urge such questioners to dwell for a moment on how a growing proportion of senior citizens can be politically compatible with large reductions in future Social Security benefits. Put another way, how could the “sacred cow” of Social Security—the language of its critics—face such a fate?
Medicaid, Medicare, and Social Insurance
Some health scholars specializing in the evolution of the Medicaid program have argued that Medicaid should be the model for future expansion of public health insurance (e.g., Sparer 2017). Medicaid is now a very large public program, with 74 million beneficiaries in July of 2017 (MACPAC 2017: exhibit 11) and legions of medical industry supporters for whom Medicaid outlays are income. Its enrollment has grown in part through a series of legislative expansions, including raising income eligibility cutoffs for benefits for children in the 1980s and expanding eligibility for adults in the ACA. Enrollment has also risen due to the growing disabled population and somewhat less rapid increases in the number of elderly who are poor enough to qualify for extra help with Medicare's cost sharing and premiums. An increasing number of formerly middle-class beneficiaries have accessed Medicaid's long-term care benefits after their assets have been depleted. Because Medicaid's beneficiaries are less thoroughly impoverished than they were before the expansions, and some were middle class for most of their lives, the difference between Medicaid and Medicare may now seem less important. So why not promote Medicaid for all instead of Medicare for all?
The answers to that question are, from a social insurance viewpoint, straightforward. Medicaid's very structure is ill designed for broad expansion. Its means-tested mode of eligibility creates a fiscal cliff. That is the term for when an additional dollar of a beneficiary's income means the loss of a program's benefit.5 No wealthy democracy has found a way to set the cliff at a point where some other arrangement and a means-tested program can combine to create universal health care. The design problems are particularly evident in the supposedly more middle-class part of the program, long-term care. The means testing and asset testing create cliffs and incentives for fraud. Quality problems can be severe, and yet many people go without needed services. For some scholars, this experience demonstrates a clear need “to socially insure long-term care” (Feder 2015: 267).
Medicaid's recent expansions not only still have the technical flaws of means testing but also have earned at best ambivalent public support. As crucial as the ACA's measures have been to millions of poor American families, the political response has been fraught. Even as of November 2017, eighteen states had refused to accept the substantial fiscal incentives the ACA offered. Work requirements have become a major theme of Republican proposals at both state and federal levels. Then followed talk about drug testing as a condition of program protection. These are exactly the responses one would expect on the basis of historic conservative conceptions of welfare programs: the “deserving” poor might be helped, but only when they behave properly. These views of and restrictions on behavior are what social insurance programs avoid. At the very least, there is serious question about what Medicaid advocates can say in response to these objections.
There is little to say about Medicare in the context of conflict over the ACA. Democrats did not rely on expanding Medicare, save for a brief dalliance with voluntary coverage for Americans fifty-five to sixty-five years of age. The “public option” relied on Medicare's ability to limit prices but not on explicit social insurance principles.6 Indeed, there was no serious discussion of how social insurance ideas were relevant, even though they were important in fact. The regulatory innovations of Obamacare represent earnest efforts to regulate commercial health insurance to become more like social insurance. Requiring insurers to guarantee issue at a fixed price regardless of preexisting conditions would reduce risk selection that social insurance eliminates directly. Requiring commercial insurers to offer a basic benefit plan also moves toward the common coverage that social health insurance offers automatically. The partial measures in the ACA appealed to the values underlying social insurance, but almost no connection was made between these measures and a principled vision of social insurance.
There are at least two plausible criticisms of this essay's arguments about relearning the appeal of social insurance principles. One is that the world has changed dramatically since the birth of social insurance in the late nineteenth century, let alone since the 1934–35 Committee on Economic Security provided a blueprint for expanding social insurance in American public life. Changes in long-standing European social insurance programs suggest that major adjustments in the American programs are required as well. The second is that this article's neglect of employment-based health insurance imbalances the discussion of the development of health reform ideas since World War II. What follows are some brief considerations of these claims.
The world has changed in many ways, but every one of the risks noted in this article—outliving one's savings, involuntary unemployment, medical costs, and disability—remains. Nor have the moral and practical differences between contributions and means testing disappeared. I doubt, in other words, whether social insurance is in conceptual trouble. But economic restructuring does require adaptation of the details of how social insurance operates.
The range of these adjustments in Europe—such as consolidation of occupational risk pools and diversification beyond the payroll tax as a funding mechanism—is beyond the scope of this article. But clearly all social insurance programs have to take into account the realities of workers not in regular employment. Workers in the “gig” economy face the same risks that prompted social insurance in the first place. This is much on the mind of EU officials worried about the spread of contract employment. Yet these and related developments are if anything a greater challenge to the voluntary, employment-based health insurance in the United States, which becomes ever less adequate (Altman 2017) and less available (Long et al. 2016). In both cases, reduced employment in regular jobs requires a search for other sources of funding—maintaining the contributory principle while diversifying beyond payroll.
Similarly, the assumption of a traditional family consisting of male worker, female housewife, and children, we now know, is less accurate in the United States and any other industrial democracy. Coverage based on relationship to a worker, the traditional mode, has become less viable. Yet other countries have managed to make adjustments without losing the sense of legitimate benefits. Financing American social insurance from sources in addition to earmarked payroll contributions is no threat as long as the common benefit idea remains crucial and the identification between contributor and recipient continues to be central.7
The second objection would be that employment-based health insurance has grown to play a role that changed the financing choice as posed in the 1930s. As Tim Jost (personal communication, 2010) has emphasized, European social policy did expand from nonprofit social health insurance to universal programs. The United States has not experienced this development, leaving the country with a patchwork of programs. There are good reasons to doubt a transition will occur anytime soon.8 Yet the weakening of that system over the past decades shows the need to rearticulate the premises of social insurance.
I return in closing to the central claim of this essay. Social insurance programs dominate American domestic policy, but what that means for the country's politics or the programs' operations and legitimacy is too little noted or explained. That criticism extends not only to harried reporters but also to a significant amount of the policy-analytic community. The coming assault on Medicare and Social Security pensions makes understanding their social insurance roots and structure all the more important.
I thank Joe White for his remarkable helpfulness in the writing of this article and Kieke Okma for her careful review of its language. Obviously, I take full responsibility for the published version.
Both described their education in personal conversations with the author, as did scholars at the University of Wisconsin and University of Minnesota.
Calculation from CBO 2017: tables 1.1 and 1.2, for fiscal years 2016 and 2017.
See the comments about “respecting the taxpayer” by Office of Management and Budget director Mick Mulvaney, quoted in Edsall 2017.
Note that this time frame is peculiar to pension programs. It has no necessary relevance to social insurance for health care except in the United States, where the social insurance program for health care was linked directly to pension finance. The idea that people contribute at age eighteen for benefits that may be paid when they are in their nineties explains why the annual reports project solvency seventy-five years into the future.
If a Medicaid benefit were worth, for example, $14,000 for a family of four, losing that benefit because of a dollar of increased income would produce an almost infinite marginal tax rate. The ACA creates a similar cliff when its subsidies disappear at 400 percent of the federal poverty level.
This is not meant as a criticism of the idea's sponsors, who were making quite reasonable judgments of the political conditions of the time.
In a sense, the United States pioneered a mix of earmarked contributions with general revenue within the Medicare program. The very confusions about “solvency” reflect the mistaken view that contributions to Medicare's Part A program also create a claim to benefits from the program's Parts B and D.
Jost agreed that the “original Social Security Act” drew a “distinction between social insurance and public welfare programs that . . . this article” traces. He noted that we may be witnessing the “end of employer-sponsored coverage as an alternative to both commercial insurance and public programs.” But not soon.