[An updated version of this article can be found at Redistribution of Income in the 2nd edition.]
Since the Great Depression most Americans have agreed that a principal responsibility of government is to redistribute income from the well-to-do to the impoverished and to those who are temporarily disadvantaged, most notably the unemployed. While many people complain about waste, fraud, and abuse in government income-transfer programs, or about the extent of income redistribution, few dispute the proposition that some level of redistribution is needed. Over the last twenty years, however, many economists—including some on the political left—have raised serious questions about the effectiveness of current transfer programs in helping the poor. While government policies do redistribute enormous amounts of money each year, the actual benefits to the poor may be much smaller than people presume.
Most people, of course, are certain that the government helps the poor by transferring income to them. Almost without exception academic studies and journalistic accounts of government's effect on the well-being of the poor focus exclusively on the effectiveness of transfer programs designed to redistribute income only to those in need. The fact that some government programs do indeed help the poor is taken as sufficient evidence that government helps the poor. But to know whether the net impact of all government transfer policies is really to help the poor, we need to examine government's many other transfer programs.
Such an examination quickly yields a striking fact: most redistribution by government is not from the rich to the poor. Instead, government takes from the relatively unorganized (e.g., the general taxpayer) and transfers to the relatively organized (lobbying groups with common interests or characteristics, such as the elderly and farmers). Moreover, the most important factor in determining the pattern of redistribution appears to be political power, not need. Of the more than $500 billion a year spent on public assistance and social insurance programs, only about 25 percent is allocated through means-tested programs. The other 75 percent—more than $400 billion a year—gets distributed regardless of need. Social Security payments shift approximately $270 billion of income a year to the elderly regardless of their wealth, and on average the elderly possess about twice the net worth per family as does the general population. And because qualifying for Medicare requires only that one be 65 or older, most of the more than $100 billion in annual Medicare benefits go to the nonpoor.
What's more, the direct transfer of cash and services is only one way that government transfers income. For example, government also transfers income by restricting competition among producers. The inevitable consequence—indeed, the intended consequence—of these restrictions is to enrich organized groups of producers at the expense of consumers. Here the transfers are more perverse than with Medicare and Social Security. They help relatively wealthy producers at the expense of relatively poor (and in some cases absolutely poor) consumers. Many government restrictions on agricultural production, for example, transfer billions of tax dollars to farmers annually and also allow farmers to capture billions of consumer dollars through higher food prices (see Agricultural Price Supports). Most of these transfers go to a relatively few large farms, whose owners are far wealthier than the average taxpayer and consumer (or the average farmer).
Restrictions on imports also transfer wealth from consumers to domestic producers of these products. Again, those who receive these transfers are typically wealthier, on average, than those who pay for them. Consider, for example, the distributional effect of the restrictions that were imposed on steel in 1984. Economist Arthur Denzau estimated that these restrictions saved 16,900 steel production jobs in the United States. Unfortunately, Denzau found, the resulting higher-priced steel raised the cost of production in U.S. industries using steel and caused a loss of 52,400 U.S. jobs. This represents a net loss of over 35,000 domestic jobs. Furthermore, according to Denzau, the workers who lost their jobs earned, on average, about 40 percent less than the steelworkers whose jobs were saved. In other words, government made many lower-paid workers poor (at least until they found new jobs) in order to help higher-paid workers.
Although transfer programs that come in the form of higher prices do not show up in the government's budget, they are just as real, and could be as large, as those that do. And the crazy-quilt pattern of subsidies, import restrictions, and the like indicates that regardless of whether or not a transfer shows up in the government budget, the size and distribution of that transfer has far less to do with the relative income of the recipients than with their relative political influence.
Not only do the poor receive a smaller percentage of income transfers than most people realize, but the transfers they do get are worth less to them, dollar for dollar, than transfers going to the nonpoor. That is because only about 30 percent of the value transferred in means-tested programs in recent years has been in cash. In contrast, well over half of the transfers to the nonpoor are cash. The remaining 70 percent of the means-tested transfers come in the form of in-kind transfers such as food stamps, housing, and medical care. Economists who study poverty point out that the value to the poor of these in-kind transfers is well below their cost to taxpayers. The reason is that the poor—like the rest of us—value cash more than in-kind transfers because if they have cash, they can choose what to buy. Therefore, a dollar taken from taxpayers and given to a poor person in the form of, say, medical care, is often worth much less than a dollar to the poor person. And it is worth much less than each dollar going to Social Security recipients.
The most important question, of course, is whether the poor have benefited from the large increase in the percentage of the national income that has been channeled through government in the name of reducing poverty. The answer, surprising as it may seem, is that we really do not know. To determine the effect of government transfer programs on the poor, we would have to know how the poor would have fared had these programs never existed, and we have no way of estimating that with any degree of confidence.
Most studies that have attempted to measure the benefit to the poor from government transfers compare the income of the recipients with what their incomes would be if all transfer income were eliminated. The assumption is that the entire transfer is an increase in the income of the recipients. Such studies conclude that government programs have significantly reduced the poverty rate.
But such studies overstate the benefits the poor receive from government transfers. For one thing, means-tested transfers that diminish as the recipient's income from working rises reduce the incentive to work. Although there is controversy over the magnitude, all economists agree that these disincentive effects exist. The late Arthur Okun, President Johnson's chief economist and a strong advocate of government transfers to the poor, compared such transfers to a leaky bucket to illustrate the fact that the increase in recipient income is less than the amount transferred.
At least to some degree, government transfers to the poor have substituted for income the poor would have earned, rather than adding to that income. Also, along with all other taxpayers and consumers, the poor have to pay for the large government transfers that go primarily to the nonpoor. What the poor receive from some transfer programs may be largely taken away from them in the form of higher income taxes, higher Social Security taxes, and higher consumer prices to support other transfer programs.
Probably the best information on what the poor have received from the growth in transfer programs comes from examining what has happened to the distribution of income over time. If government transfers have helped the poor, as advertised, by redistributing income to them, then this should show up in a more equal distribution of income. However, studies that have investigated the trend in income distribution, after adjusting for taxes and adding in the value of all transfers, including in-kind transfers, find little if any change over the last forty years. One of the first such studies, by economists Morgan Reynolds and Eugene Smolensky, covered the period 1950 through 1970. Reynolds and Smolensky found that the households with incomes in the bottom 20 percent of the income distribution received 6.4 percent of the national net income in 1950 and 6.7 percent in 1970. Those households with incomes in the top 20 percent of the income distribution received 39.9 percent of the net national income in 1950 and 39.1 percent in 1970.
More recent studies on the trend in the after-tax/after-transfer distribution of income continue to show little evidence of redistribution from the rich to the poor. Based on these studies, economist Robert Haveman concluded that "in spite of massive increases in federal government taxes and spending, we are about as unequal in 1988 as we were in 1950."
Still, we can never know what would have happened if government transfers had not increased. It is possible that the distribution of income would have become more unequal. The slowdown in the growth of wages since 1973, the increase in the number of female-headed households, and the aging of the population have been cited as reasons why the income distribution would have become more unequal without increased government transfers. Yet some of these very changes may have been accentuated by increases in government transfers. A partial explanation for the slowdown in the growth of wages is that governments have required private firms to increase nonwage compensation and to pay higher payroll taxes.
In sum, many economists would agree that the transfer programs designed specifically to help the poor have in fact improved their well-being. The magnitude of the improvement, however, is much less than the dollar amounts transferred—because so many of the transfers are in-kind instead of cash, and because transfers cause at least some of the poor to work less than they otherwise would. More important, looking only at transfers to the poor gives an incomplete answer—and possibly a wrong answer—to the question of whether the net impact of all government policies that affect income distribution is to help, or harm, the poor.
Dwight R. Lee is the Ramsey Professor of Economics at the University of Georgia. In 1997, he was president of the Southern Economics Association.
Denzau, Arthur T. How Import Restrictions Reduce Employment. Formal publication no. 80. Center for the Study of American Business, Washington University (St. Louis), June 1987.
Haveman, Robert. Starting Even: An Equal Opportunity Program to Combat the Nation's New Poverty. 1988.
Okun, Arthur M. Equality and Efficiency: The Big Tradeoff. 1975.
Reynolds, Morgan, and Eugene Smolensky. Public Expenditures, Taxes, and the Distribution of Income: The United States, 1950-1970. 1977.
Sawhill, Isabel V. "Poverty in the U.S.: Why Is It So Persistent?" Journal of Economic Literature 26 (September 1988): 1073-1119.
Tullock, Gordon. Economics of Income Redistribution. 1983.