Time for a Guaranteed Income?
The pros and cons of a welfare idea championed by liberals and libertarians alike
Switzerland will soon hold a nationwide referendum on granting a guaranteed and unconditional minimum monthly income of $2,800 for each Swiss adult. In America, where Lyndon Johnson's War on Poverty just celebrated its 50th anniversary of failing to achieve victory, liberals jumped on the Swiss news to reconsider the un-American-sounding idea of a universal basic income.
Surprisingly to some, they were joined by many libertarians. The list of intellectuals who have made cases for a guaranteed minimum income over the years includes such laissez-faire luminaries as Milton Friedman, Friedrich Hayek, and Charles Murray.
Friedman favored a negative income tax (NIT), in which taxpayers who earn less than the established minimum taxable income level would receive a subsidy equal to some fraction of that difference. (A watered-down version of this became the Earned Income Tax Credit.) Hayek defended a minimum income floor, in which the government provides a conditional income to each adult. Murray's 2006 book In Our Hands argued for an unconditional $10,000 annual cash payment to all adult Americans, coupled with a repeal of all other welfare transfer programs.
Their proposals aim to fully replace the current welfare state with a less-bad alternative. In a world where government already redistributes income, with all of the inefficiency that comes with overlapping bureaucracies, the idea of direct cash payments has an intuitive appeal because of its comparative simplicity and fairness.
Any alternative might seem preferable to the welfare system we currently have. Federal welfare in the U.S. today consists of a highly complex maze of 126 separate anti-poverty programs, many of which are redundant. (There are, for instance, seven different housing programs.) While the system benefits the many government employees who manage these duplicative programs, it is neither easy for poor Americans to navigate nor is it an effective way to deliver anti-poverty services.
According to Cato Institute analyst Michael Tanner, the federal government spends close to $1 trillion each year at the federal, state, and local levels on anti-poverty programs-everything from Medicaid to job training to food stamps. After adding in the bureaucracy that attends to applying for food stamps, rent subsidies, and everything else, it isn't hard to imagine how moving to a cash transfer system would make the entire process far less time-consuming and humiliating for the poor. In addition, getting rid of the bureaucrats who administer these programs would save between 10 and 15 cents on every welfare dollar, a significant amount.
Welfare programs are demeaning by design, because they dictate to poor people what they must spend on food, housing, or health care, rather than letting them make those trade-offs themselves. The government even dictates what food poor people may or may not buy with food stamps. The libertarian interest in a guaranteed income scheme proceeds not simply-or even mostly-from the desire to make government smaller and more cost-efficient. It stems from a belief that all individuals have the capacity to promote their own interests, and in fact are better able to make decisions about their lives than anyone else.
However, the abstract idea loses some of its appeal when one starts looking at its realistic cost. The details vary from one version to another, but even in the best of possible worlds, none are likely to save much money, if any. Consider a plan to provide a $12,000 annual subsidy to every adult above 18.
Giving $12,000 a year to the 237 million adults in the U.S. above the age of 18 would cost $2.8 trillion a year. If we add this amount to the other big-ticket budget items, such as the $550 billion we spend on the Pentagon and the $200 billion devoted to misguided corporate welfare and other wasteful programs, this plan would break even with the current system, if and only if we get rid of all other anti-poverty programs and tax breaks, unemployment insurance, Obamacare subsidies, Social Security, Medicare, Medicaid, and so on.
Another possibility would be to limit transfers to the estimated 106 million individuals who currently qualify for welfare programs by earning less than 200 percent of the federal poverty level (i.e., $23,440 a year). That would cost $1.3 trillion a year, an amount that exceeds the current $668 billion in welfare spending at the federal level and the additional $284 billion at the state and local levels (not all of which would disappear if Washington replaced the old federal welfare state with a basic income). Restrict payments to the 100 million people currently receiving means-tested public assistance, and it will still cost $1.2 trillion.
Even if we assume that all other anti-poverty programs will disappear (which is assuming a lot), it's conceivable that taxpayers would not save money compared to the status quo. Don't forget, they would still be on the hook for the military, for corporate welfare, and-if middle-class entitlements stay in place-for part of Social Security and Medicare. In 2012, the government spent roughly a trillion dollars on Medicare and Social Security recipients (excluding the Social Security benefits that go to Medicaid recipients). That amount will grow as more baby boomers retire in the next decade.
A cash transfer program to adults at the poverty line level ($11,720 a year) would still cost around $600 billion a year. And, of course, if the transfer extends to those above the 200 percent level, that cost would go way up.
The appeal of a guaranteed income also diminishes when judged against its ability to move people away from government dependency. There is some evidence that a guaranteed minimum scheme would undermine incentives to seek employment. Four landmark experiments in the 1960s and '70s examined the Negative Income Tax's impact on labor supply. The recipients of NIT grants tended to work fewer hours compared to control groups that did not receive the grants.
Making the NIT more progressive in order to placate these disincentives to work, as Milton Friedman suggested, does not seem to help. Pointing to a series of 30 welfare experiments conducted in the 1990s, National Review's Jim Manzi argued in 2011 that of all the policy options tested, only welfare policies that included work requirements pushed people off welfare and back to self-sufficiency. Manzi concluded that taxpayers' moral aversion to subsidizing sloth will ultimately undermine any move to a guaranteed income or negative income tax scheme that lacks work requirements. People, he demurs, seem to prefer the paternalism.
But my main objection to a guaranteed minimum income is rooted in the wisdom of public choice: The poor structure of government incentives ensures that good intentions and elegant theories rarely equal expected results in public policy. The biggest risk in implementing a guaranteed income is that it won't completely-or even partly-replace existing welfare programs, but instead simply add a new layer of spending on top of the old. Friedman learned this the hard way: After years of promoting the NIT, he wound up opposing Richard Nixon's NIT-inspired Family Assistance Plan precisely because it would not displace the preexisting welfare state.
So what are libertarians to support? If nothing else, more research: We could use a new series of voluntary, dispersed trials aimed at finding ways to avoid work disincentives while delivering payouts more efficiently and tying the hands of special interests and politicians.
But more importantly, as economists Peter Boettke of George Mason University and Adam G. Martin of Kings College in London remind us in a recent paper, libertarians shouldn't forget that "the most robust protection against poverty comes from institutions that generate a harmony of interests rather than those that foment distributional conflicts." A guaranteed income may or may not be an improvement over the current state of affairs, but a massive transfer and regulatory state harms the poor either way.
