Saturday, April 7, 2012

Budget Committee Inequality Report: Welfare Reform

Ryan states in his report that:

According to the CBO’s study, one contributing factor to the declining share of transfer payments going to lower-income households was the decline in spending on Aid to Families with Dependent Children and its successor, Temporary Assistance for Needy Families, relative to market income. These relative declines occurred as a result of bipartisan efforts in the late 1990s to transform cash welfare by encouraging work, limiting the duration of benefits, and giving states more control over the money being spent.

At the time, opponents of these reforms charged that the reform effort would cause large increases in poverty and despair. Instead, the exact opposite occurred. These reforms cut welfare caseloads in half against a backdrop of falling poverty rates, falling child poverty rates, low unemployment, and rising income for the lowest quintile. In fact, over the period studied by the CBO, the lowest quintile’s income gains were most pronounced after the passage of the 1996 welfare reform law (see Figure 1 on page 2).
As a result of methodology limitations in the CBO’s study, this successful welfare reform effort appears to have reduced government’s role in making the distribution of income more equal, but it does not follow that welfare reform caused income inequality to get worse. In many cases, welfare checks were replaced by paychecks. The CBO report states that its analysis does not account for the incentive effects of changes to taxes or transfer programs. In the case of welfare, that means that the analysis does not account for the sea change in the incentive structure of the program that led large numbers of welfare recipients to escape the grip of government dependence in favor of gainful employment and independence.
 He goes on to state:
Creation of new means-tested entitlements/tax credits: Also, as the CBO’s study notes, the decline in transfer payments to low-income families with children was largely offset by increased spending on other means-tested programs and tax credits. In the late 1990s and early 2000s, Congress created a number of new means-tested entitlements (such as the State Children’s Health Insurance Program) and tax credits (such as the Earned Income Tax Credit, or EITC, and the Child Tax Credit, or CTC) that raised the after-tax income of lower-income households.

The EITC and the CTC are refundable credits, meaning that if a taxpayer’s credit exceeds his or her tax liability, he or she receives the balance of the credit in the form of a payment. The CBO notes that an increase in the refundable portion of these credits largely took the place of cash welfare spending, as government income-support programs migrated to the tax code and became tied to work status.

The increase in these credits was also partially responsible for the observed increase in the progressivity of the federal income tax over the period studied. The CBO study noted that “The lowest income quintile has a negative average federal tax rate because, as a group, households in that quintile qualify for more in refundable tax credits than they owe in income taxes before the credits are applied.”

I've quoted these two section at length because I think it is necessary to point out what I find deceptive about these passages.

The first thing to note here is that what has happened is that American anti-poverty measures have shifted from being assigned to transfer payments to instead being assigned to the tax code. A large portion of the reduction in the progressivity of transfer payments and the increase in progressivity of income taxes is due to this shift, anti-poverty spending has simply been shifted to different columns of the government's balance sheet. This provides an illusion of lower transfer payments and anti-poverty spending which can deceive those not well acquainted with the accounting used for these programs.

The second issue I want to get into is Ryan's claims in the bolded sections. Ryan claims that the government's role in fighting poverty has been reduced and that the result has been less poverty and more employment.

However, the facts don't support this conclusion. The sharp decline in the welfare caseload and decline in poverty rates began in 1994, not in 1996 when welfare reform occurred. In 1994 the EITC was introduced, most analysts contend that EITC expansion was the single largest factor in both poverty reduction and in increased workforce participation. While Ryan briefly does mention the role of other Federal spending, he fails to mention that these programs had less strict means and asset testing than the AFDC program had and that their net cost was significantly greater than what had been spent on AFDC. His causal story centers on the reduction in AFDC benefits that occurred in the 1996 reform, however, neither the sequencing of poverty reduction and employment or the actual net changes in Federal spending support his assertions.

It is also notable that Federal anti-poverty spending had a net increase, not decrease, as a result of the 1990s reforms. While Federal spending on cash welfare support for non-disabled, non-elderly adults fell from $24 billion in 1988 to $13 billion in 1999, Federal spending (including categories other than cash welfare supports which were the centerpiece of the 1996 reform) on working low-income families increased from $11 billion in 1988 to $66.7 billion in 1999 (all in 2000 dollars).* It is also notable that the minimum wage increased substantially during the 1990s relative to earlier periods. Most analysts feel this significantly increased the labor supply and reduced poverty among low income families (while not the place to get into this in detail, most studies have shown that increases in the minimum wage do little to reduce the supply of jobs and tend to reduce poverty and increase labor supply, quality, and retention by exceeding the reservation wage of marginal individuals, the supply effect on marginal laborers tends to dominate the demand effect through increased costs to employers at minimum wage rates seen in most areas of the US, though I wouldn't extend this to all situations either in the US or abroad).  The numbers don't match with Ryan's narrative, government intervention and spending increased, not decreased, in the time period under consideration. Ryan is able to conceal this by defining the declining spending as transfers while concealing the increased spending as taxes, but this sleight of hand is easily revealed once concrete numbers are used in place of vague assertions. Ryan attempts to deceive his readers by assigning all of the causality for poverty reduction to the numbers that support his story, while ignoring that once all categories of Federal anti-poverty spending are taken into account the government role is shown to have increased in the period of poverty reduction.

Ryan also contends that the experts were wrong regarding the impacts of welfare reform. This is partially true, they did believe the 1996 reforms would lead to mass poverty and desperation. What they were wrong about, however, isn't that government programs created dependency and that reducing them would decrease poverty. What they were wrong about was the adaptiveness of the bureaucracy to welfare reforms; AFDC had been the gateway program and they didn't realize that other state and federal programs would grow to take up the slack, the impact of the economy on the ability of the poor to find work; and the behavioral response to work incentive programs, experts seemed to share the idea that many of the poor preferred not to work (though for more defensible reasons than most conservatives) and failed to appreciate the large response to work incentive programs (despite state level studies that existed and suggested this at the time).

Much of this is revealed in Blank's study when she examines studies looking at the particular impacts of various components of welfare reforms. Most notably, she finds in her literature review that programs designed to reduce government exposure through greater work requirements and means testing have only a small impact on employment and significant increases in poverty, though they do save the government money. Programs that instead use work incentives (such as reduced benefit reduction rates, work subsidies, child care subsidies, and job training) are found to increase employment by the largest amount, reduce poverty by the largest amount, and cost the government the most (as well as show other positive benefits like increased school performance in children). She reviews mostly studies looking at state level welfare reforms, though she also includes Canadian reforms. The numbers point in the opposite direction Ryan contends, the programs with the best outcomes involve higher government expenditures with the broadest base and weakest means testing. This is consistent with international findings, broadly based welfare programs that have the least discrimination based on means show the largest increases in employment, reduction in poverty, and increased government expenditure (alongside increases in other measures like educational attainment and income mobility).

These conclusions are further reinforced by more recent studies. To quote Bitler, Hoynes, "The State of the Social Safety Net in the Post-Welfare Reform Era."**
We find that TANF provides less protection, or at least no more protection, in an economic downturn than the AFDC program that preceded it, but that the noncash welfare safety net (and especially food stamps) is providing significantly more protection. In our analyses using both the official measure and our own alternative measure of poverty, the point estimates imply that the increase in poverty in an economic downturn is greater after welfare reform than it would have been before reform...
Overall, we find no evidence that the prevalence of negative family or household well-being in an economic downturn has improved after welfare reform, and some weak evidence that it has worsened. Further, it appears that food stamp benefits are playing an important role in mitigating adverse impacts on income in post-welfare reform recession.
On the whole, Ryan seems to get the actual impacts of welfare reform completely wrong. This is a well studied area, there is no excuse for his misrepresentations. He tries to conceal increases in government involvement and increased spending by separating out tax effects from traditional transfer payment, he isolates the impacts of cash-welfare benefits from the substantial increases in other programs that happened concurrently, and he neglects the substantial body of evidence regarding state level studies that would have revealed the problems with his statements. I would strongly encourage everyone to read the two studies that I cite below, they have a wealth of charts and graphs that provide more compelling evidence regarding welfare reform than I am able to produce here. These studies surely would have been available to the House Budget Committee in producing their report, it is notable that instead they rely off of the Wall Street Journal, the Cato Institute, and the CBO instead (I really like the CBO, but they act under rules set by Congress so they tend to maintain polite fictions like the difference between transfer and tax programs in reducing poverty that researchers not bound by political considerations would find specious). There are a number of other studies on this topic as well, but I think these two studies provide the best starting point on the subject by covering such a broad array of issues and perspectives.

*Blank, Rebecca M. "Evaluating Welfare Reform in the United States." Journal of Economic Literature.  Vol. 40, No. 4. 2002.  pp. 1105 - 1166.

**  Bitler, Marianne P, Hoynes, Hilary W., Jencks, Christopher, and Meyer, Bruce D. . "The State of the Social Safety Net in the Post-Welfare Reform Era [with Comments and Discussion]." Brookings Papers on Economic Activity. 2010.  pp. 71-147

No comments:

Post a Comment