A Targeted Welfare State
For centuries, governments have provided aid to the poor and needy. But over the past 100 years, an alternative ideal has come to the fore: extending publicly funded benefits to all citizens, regardless of their ability to provide for themselves. The shift in approach has yielded an enormous expansion of government, leaving public finances increasingly dominated by the cost of health and retirement benefits for the middle class.
This shift has been the subject of fierce controversy, most pointedly illustrated in an acrimonious debate between English scholars Richard Titmuss and Arthur Seldon during the mid-20th century.
Richard Titmuss advocated expanding publicly funded social benefits to all citizens. He viewed public benevolence for those in need as stingy and unreliable, and believed that "separate...services for poor people have always tended to be poor quality services." He resented means tests as inherently unfair to those above eligibility cutoffs, and argued that they would inevitably exclude many who required help. He also believed that restricting eligibility would deprive programs of durable cross-class popular support and compel value judgments that would impose a "sense of inferiority" on beneficiaries. Underlying Titmuss's ideal of uniform collective provision was a passion for socialism, which he argued was "about community as well as equality."
Arthur Seldon, by contrast, believed that publicly funded benefits should be reserved for those who couldn't provide for themselves. He criticized Titmuss's "jejune sentimentality" for distracting from two essential considerations: scarcity and cost. No benefits given, he argued, are actually free, and he deemed it "foolishly wasteful and inhumane to hand state funds to 100% of the populace in order to make sure that the 10%, or 15%, or 20%, who need it get it."
Pointing to survey data showing that voters prefer to reserve benefits for the needy, Seldon wrote: "[T]o erect a vast State machine for taking from some who need it in order to give to many who do not deprives the act of grace and makes it a cause of cynicism, bitterness, and social disruption." Universalism, he noted, forces each part of society to engage in political conflict as the only way to improve the services they receive. That struggle is inevitably dominated by those "more literate and moneyed than the average," rendering redistribution "inegalitarian as well as arbitrary and 'unjust.'"
Did Titmuss or Seldon have the better argument? As the cost of federal entitlements becomes increasingly unsustainable, determining the answer is more important than ever.
THE GROWTH OF UNIVERSAL ENTITLEMENTS
Debates about the welfare state often focus on the merits of redistributing resources to the poor. But most public entitlement spending in America now goes to the middle class.
English-speaking colonies in North America always provided public assistance to the poor, initially administered through local government. In the pre-industrial era, when most Americans struggled from sun-up to sun-down to grow food for their families, this aid was meager and barely covered subsistence. Policymakers then, as now, feared that excessively generous benefits would lead people to give up work to claim assistance. As industrialization and economic growth caused household incomes to surge, aid was increased. But as economist Stanley Lebergott and demographer Tony Wrigley each documented, over the past 200 years, aid to the poor has remained remarkably constant in relation to basic wage levels.
By contrast, the 20th century saw the establishment of enormous public entitlements for the middle class. Impressed by Bismarckian Germany, progressive-era reformers argued that much poverty could be prevented by requiring households to purchase publicly administered "social insurance" against common risks to their livelihoods — namely the family breadwinner's unemployment, sickness, disability, or death, as well as the danger of outliving household savings in old age.
Social-insurance benefits, paid in proportion to prior payroll-tax contributions, appeared to offer a method of increasing assistance for those in need without deterring employment or fomenting taxpayer backlash. In the United States, where the Great Depression had suddenly impoverished millions, the idea formed the basis of the Social Security Act of 1935. In Europe, as continental war and hyperinflation destroyed private savings, investment, and insurance schemes, pay-as-you-go pensions allowed governments to fulfil the middle-class retirement expectations that had existed prior to the war.
World War II excused a great increase in taxes. From 1939 to 1945, the proportion of U.S. households subject to income taxes leapt from 7% to 72%, while government revenue surged from 7% to 20% of GDP — a plateau from which it has not greatly shifted. Following the return of peace, defense costs declined and the economy grew. Politicians soon realized that they could win votes by shifting expenditures to increases in social-insurance benefits.
Social-insurance schemes initially justified larger payments to middle-class enrollees by reference to the fact that these enrollees had previously paid more in taxes. But after wartime inflation, Congress in 1950 voted to increase benefits beyond the value of prior contributions. In the subsequent two decades, politicians competed to claim credit for arbitrarily distributed "inflation adjustments" before enacting legislation to automatically increase all benefits according to wage growth in perpetuity — distributing the largest windfalls to the more affluent.
Disability benefits had also been added, again paying more to higher earners. In 1965, Congress established Medicare to fund comprehensive medical benefits for seniors for the rest of their lives without regard to financial need or any pretense of limiting benefits to the extent of prior contributions. Medicare's benefits package has since been expanded with every advance in medical capabilities.
The cost of Medicare and Social Security climbed from 3.8% of GDP in 1970 to 8.9% in 2023, and now accounts for the majority of federal entitlement spending. Seniors at all income levels are the largest net beneficiaries of redistribution every year, while the poorest working-age households receive only slightly more in benefits than they pay in taxes. Affluent retirees are paid higher pensions and consume more medical services than the poorest every year. Now that the richest retirees live twice as long on average past the age of 65 as the poorest, this disparity in federal support has expanded significantly.
THE LEAKY BUCKET
The primary constraint on any welfare state is the need to pay for it. If expenditures are not reserved for those who cannot provide for themselves, soaring costs will increase the net tax burden on those who are less well off.
Economist Arthur Okun once suggested that redistribution through taxation and spending should be thought of as a leaky bucket. This "leakage" — funds that are lost in the transfer between taxpayers and intended beneficiaries — occurs not only due to administrative costs and imprecision in identifying need, but also through incentives to claim public aid rather than to work, to consume rather than to invest, to deploy resources less productively, and to employ costly schemes to avoid taxation.
In democracies, the rich are the first to be taxed. But there are limits to how much money can be drawn from them. The highest incomes typically come from capital, which is highly mobile and can easily evade taxation. Only so much revenue can be captured from a wealthy minority of the population before tax-avoidance measures — and the resulting reduction in economic output or diversion of investment — dissipate much of the potential revenue gain.
The cost of basic government responsibilities, such as national defense and infrastructure, already exceeds tax revenues from the rich. Broadening eligibility for benefits beyond the poor, therefore, requires increasing net taxes on those who are less affluent.
Such can be seen in European countries with the largest welfare states, where the poor pay a higher share of their incomes in taxes than the rich. Denmark, frequently endorsed as the model universalized welfare state, raises less tax revenue per capita from its richest income decile than the United States, but it also takes substantially more from each of the eight poorest deciles. A 2021 study by the European Commission found that, because pension expenditures dominate the welfare state, the more European countries redistributed to affluent seniors, the less net aid they provided to poor residents of working age.
The more a welfare state is expanded, the greater the rate of leakage. There are, at first, many easy cases (such as the elderly poor and infirm) who can be aided without greatly displacing self-reliance. But the more eligibility for entitlements is broadened, the more benefits will serve to crowd out private provision rather than to fill unmet needs.
For instance, countries with higher levels of public spending on pensions typically promise to maintain the standards of living of both the poor and the middle class into old age. But this does not make their seniors more prosperous; it just reduces the amount they set aside for themselves in retirement. Similarly, extending publicly funded health-care entitlements to the middle class crowds out private health-insurance funding rather than increasing the total resources available to fund medical care. As a result, health-care systems more dependent on a single public payer offer less access to cutting-edge medical care while also burdening those with modest incomes with higher taxes to pay for care that the most affluent receive.
Leakage of redistributive expenditures might be acceptable as collateral damage from necessarily imprecise public-assistance policies to help those who cannot work, or as a deliberate attempt to allow the elderly or mothers of young children to drop out of the workforce. But it cannot simply be justified in terms of improving on the supposed inefficiency of private insurance methods. Economists at the University of Chicago estimated that a universal basic income of just $12,000 per year would reduce average earnings by slightly more than $6,000, while also requiring a tax increase of 27 percentage points across the board. Similarly, Medicare for All would require increased taxes on the millions of elderly, disabled, or low-income Americans currently enrolled in Medicare and Medicaid to pick up health-care costs for more affluent workers who do not fit these descriptions.
TARGETING TOOLS
The displacement of self-reliance associated with publicly funded entitlements can be minimized by narrowing eligibility to beneficiaries who cannot provide for themselves.
The establishment of compulsory old-age, unemployment, and survivors' benefits for the middle class did little to reduce the need for gratuitous aid to the poor. Long-term poverty in America is due more to low earnings histories than the failure of middle-class workers to provide for old age or other risks to their incomes. Nor are the causes of temporary poverty entirely beyond individuals' control. According to economist Rebecca Blank, family breakdown and reductions in working hours are the "two main reasons why people become poor."
Noting the adverse incentives on earnings, marriage, and out-of-wedlock births arising from benefits to low-income single mothers, political scientist Charles Murray observed: "Any social transfer increases the net value of being in the condition that prompted the transfer." This concern has led some to endorse a shift from targeted to universal benefits that are not contingent on specific circumstances and needs.
Yet the capacity for self-provision is highly unequal between the disabled and able bodied, the elderly and working age, and parents and childless adults. Uniform assistance that is too generous to the able bodied would be inadequate for the disabled.
Governments do not need to distribute aid indiscriminately. Just as the combination of claims reviews, cost sharing, and risk aversion can protect private insurers from succumbing to moral hazard, so the bad incentives resulting from publicly funded entitlements can be greatly reduced by targeting or conditioning assistance.
Entitlement programs are typically caught between three conflicting goals: making assistance generous, controlling program costs, and avoiding work disincentives. The trade-offs between these can be alleviated by limiting eligibility to causes of need beyond individual control — such as old age, disability, and dismissal from employment without cause.
And while targeting tools each have their drawbacks, these can be mitigated if multiple targeting tools are used in concert, or with sensitivity to specific circumstances.
Disincentives for the temporarily unemployed to find a job, for instance, can be checked by excluding those who voluntarily quit jobs, by imposing time limits on the receipt of benefits, or by conditioning aid based on prior earnings levels. Asset tests might be an appropriate method of screening for need among those who are disabled and permanently unable to work. Alternatively, more nuanced evaluations of needs, though potentially costly, could be made on a case-by-case basis (as with disability determinations or work-support programs).
Means testing benefits for the elderly may be less advantageous than for other cohorts. Even wealthy retirees often have relatively low incomes, and asset tests may lead people to forego savings, spend down investments, and fail to purchase insurance plans. Yet there are other ways to ensure that affluent retirees do not inflate the tax burden on workers by needlessly claiming public aid. We see this with long-term care: Those who deliberately spent down assets can be excluded from eligibility, while exemptions of assets from means tests may be coupled with subsequent collection of repayment from estates.
There is a trade-off between providing aid to those who don't need it and ensuring that those with the greatest needs are not excluded. Restrictions on eligibility can lead to gaps in coverage, whether they be from the rules' failure to identify those in need, the difficulty of satisfying eligibility determinations, or the challenge of navigating complex program rules. Adverse circumstances for which individuals might be at least partially responsible (such as drug addiction, family breakdown, or poor employment prospects) can leave them in acute need without triggering any particular form of targeted assistance.
Political theorist Robert Goodin has therefore argued that policymakers should "err on the side of kindness" in setting eligibility. But mercy in aiding those who have behaved irresponsibly need not imply universality and public provision for those Americans who can provide for themselves. Responding to the deficiencies of specific targeted programs by hurriedly embracing the opposite extreme of universality is misguided and ignores associated costs. When targeting methods leave out very needy people, it is always possible to respond to gaps in coverage by establishing additional targeted assistance.
Universal programs must still ration scarce public resources, but they have a smaller and less-precise set of tools with which to do so. When entitlement programs maintain a pretense of providing everything to everyone, rationing tends to occur not through the transparent deliberation of legislators, but through the obscure, furtive, and ad hoc actions of administrators behind closed doors. When health-care spending is regulated for aggregate purposes, the decision not to purchase an extra piece of equipment becomes a product of the same cost-benefit calculus as the decision not to help a patient with lesser prospects of being restored to health. Such determinations are likely to come systematically at the expense of the disadvantaged.
Targeted benefits have been criticized as administratively arbitrary and stigmatizing, but such features are not inherent to them. The largest means-tested benefits now confer legally enforceable entitlements that do not require appropriations, are typically administered automatically and discretely through the income-tax system, and are subject to federal rules against discrimination. Indeed, under the Poor Laws of the Victorian era, it was often the absence of targeting that, by lumping the elderly and children together with drunks and vagrants, made claiming assistance particularly unpleasant.
THE POLITICS OF ENTITLEMENTS
Framing entitlements as a form of "insurance" has skewed public expenditure to the affluent beyond what is actuarially justified. As columnist Polly Toynbee has noted: "All governments cling to vestiges of national insurance, by now mostly a sham, because residual belief in it makes it a more acceptable tax." The principal effect of this illusion is to support taxes on the poor and benefits for the rich exceeding what would otherwise be politically feasible.
Political scientists generally conclude that the claim that benefits targeted at the poor are inherently politically vulnerable does not withstand scrutiny. An analysis by Robert Greenstein, founder of the Center on Budget and Policy Priorities, discovered that while some entitlements have grown more than others, this depended mostly on whether they undermined work, delivered benefits in kind, were fully federally funded, or were seen as effective. In fact, targeted programs tend to have an advantage in budget fights because they do more with every dollar to fill unmet needs. In recent years, bipartisan budget agreements have repeatedly cut Medicare spending while altogether exempting means-tested benefits from sequestration.
The task of determining who is sufficiently incapable of work to deserve public aid is unavoidably judgmental and controversial. The question of how much mothers who bear children out of wedlock should be supported by taxpayers or expected to work is even more so. Effective targeting allays concerns that funds will merely support indolence, profligacy, and other vices, thereby increasing public confidence in and support for programs to aid the needy.
Indeed, polls suggest the unpopularity of welfare programs in the 1990s was in large part due to the perception that benefits were going to those who "could get along without it" rather than to people "genuinely in need of help." This concern remains, and also applies to non-means-tested benefits: In June 2021, 52% of Americans surveyed wanted a $300-per-week Pandemic Unemployment Compensation supplement eliminated immediately, while only 16% wanted it to continue indefinitely.
Arthur Seldon argued that "ordinary people will pay more for a service if they can see their families will benefit than they will in taxes for a service in which they can see no benefit for higher tax-payment." And in fact, the middle class is eager to escape being trapped in the same fiscally constrained programs as the poor. A 2020 assessment of public opinion across 20 countries by political scientists Marius Busemeyer and Torben Iversen concluded that, although support for universal entitlements declines where private alternatives are available, the means-tested aspects of the welfare state "appear to be politically secure."
AN EFFECTIVE, AFFORDABLE WELFARE STATE
Arthur Seldon was right. Public funds are scarce, and raising revenue is costly. The extent to which the welfare state improves social well-being thus depends on how well spending is targeted at filling needs that would otherwise go unmet.
By the mid-1960s, Richard Titmuss himself acknowledged that the universalization of the welfare state had fallen short of his hopes. "[W]e put too much faith in the [1940s] on the concept of universality as applied to social security," he lamented. "Mistakenly it was linked with economic egalitarianism. Those who have benefited most are those who needed it least."
Aid for those who are less fortunate is typically popular with American voters, due both to compassion and a fear of falling into similar need. Only 13% of Americans surveyed in 2018 responded that the nation spends too much on the poor. But this only holds if assistance is not so indiscriminate and open ended that taxpayers feel they are being taken advantage of. The more government expands eligibility for entitlements to those who can provide for themselves, the less faith voters have in the system.
America's entitlement programs should be fiscally sustainable, politically accountable, and effective at aiding those who need it most. A targeted welfare state is the best means of ensuring all three.
