The remarkable decline in the number of people on welfare is one of the most dramatic social and economic developments of the past decade. At the end of 2000, there were fewer than half as many people receiving welfare benefits as there were in 1993. Experts agree that this decline - from more than 14 million persons to under 6 million - cannot be explained as a byproduct of the strong economic expansion or the generally low level of unemployment in recent years. Instead, a radical policy shift really ended the option of lifetime welfare and motivated welfare recipients to seek work. Reinforcing that policy change, additional programs have increased the reward to work for low skilled people. As a result, most former welfare families are better off now than they were before the reform.

We recently attended an interesting conference at Harvard's Kennedy School of Government that focused on the economic policies of the Clinton years. Among the wide range of subjects covered was welfare reform, reviewed in an excellent paper by economists Rebecca Blank and David Ellwood. We learned a good deal from their paper as well as from the comments of the conference participants that included Clinton Administration alumni, Republican economists, and other experts.

Everyone seemed to agree that low income families are better off now than they were a decade ago, despite the ending of AFDC (Aid to Families with Dependent Children), the primary welfare program . A new combination of temporary assistance, income from work, and expanded tax credits for low income working people (the Earned Income Tax Credit or EITC, as well as tax credits for employers hiring low income workers) has helped move individuals off welfare and made it possible for them to earn enough to more than offset the decline in welfare payments.

In 1996, President Clinton signed The Welfare Reform Act despite his opposition to critical components of the bill. The new program replaced AFDC, which had provided long term assistance to those who met eligibility requirements, with a work based approach founded on the idea of temporary assistance for people who are expected to seek employment. A key feature of the legislation gives states flexibility to design their own programs, setting their own eligibility and work requirements, subject to a 5 year lifetime limit on federally financed benefits. This lifetime maximum was particularly controversial, though states are allowed to use their own funds to support people beyond this lifetime limit.

The amount of federal dollars that go to each state now depends on that state's success in reducing the number of people on welfare. Although there is no separate federal requirement that former welfare recipients actually find work, the experience has been a substantial increase in work among the former welfare population. About half of those single mothers who were previously receiving welfare and not working are now employed. The average woman who has moved off welfare has increased her cash income through work by more than enough to offset her decline in welfare payments.

Key factors in this success story have been the expansion of the EITC and greater child support. The EITC is very different from traditional public assistance. Under the traditional system, welfare beneficiaries received monetary payments that were reduced as incomes increased, often a dollar for dollar reduction. In contrast, the EITC actually reinforces the earnings of low wage workers by giving them more money for every dollar they earn until they reach a threshold income level. Benefits are tied both to the level of earnings and the number of children in the family. In that way the EITC increases the incentive to work along with raising the standard of living of low wage people. At a certain level, the EITC ceases to provide any incremental transfer and then if earnings go above $13,000, there is a reduction in the EITC at a rate of 21 cents for each dollar of earnings, a significant marginal tax on top of the payroll tax and eventually the income tax that causes the total marginal tax rate to exceed 50 percent for individuals with incomes of about $30,000.

But despite these high marginal tax rates at lower-middle income levels, the new system substantially increases the reward for work among the very poor. Blank and Ellwood estimate how much the payoff for work increased for a single mother with two children who works full time at the minimum wage. In 1988, such a woman would have had a disposable income by working of $10,937 compared to her disposable income of $8,612 if she relied solely on welfare programs, a gain of only $2,325 for full time work. And this small difference would be diminished by a loss of Medicaid benefits, virtually eliminating the net gain. In 1999, this mother would have a disposable income$15,018 if she worked, a net gain of over $7,000 compared to her potential benefits of $7,967 under the current temporary assistance welfare plan if she did not work. And her children under 16 remain eligible for Medicaid. It's clear that the payoff from work has increased dramatically over the last decade.

The increased payoff from working explains why the labor force participation of low skilled women has increased significantly over the last decade, substantially decreasing their poverty rates. Those decreases would be even greater if the changes in welfare procedures didn't discourage some individuals from claiming the food stamp benefits for which they are still eligible.

The willingness of Congress to enact the new welfare system reflects in part the success of similar welfare reforms in individual states like Wisconsin that had been given "waivers" by the federal government under the old system, allowing them to experiment with work incentives and welfare limits. The success of the welfare reforms is therefore also a reflection of the success of the American federal system in providing an opportunity for state level experiments with alternative approaches to social problems.