Social Insurance

To me, it seems impossible to draw hard and fast lines between social insurance and public assistance.

—Arthur J. Altmeyer, aka “Mr. Social Security”¹

Temporary Assistance for Needy Familes (TANF) is conventionally thought of as a "welfare" program for parents with children—that is, a program that provides assistance to destitute non-workers who have no or very little work experience. This understanding of TANF—captured in the phrase "welfare to work"—keeps it from playing an effective role as part of America's larger system of social insurance. TANF should be thought of as a necessary part of America’s system of social insurance rather than as a welfare program that exists outside of, and has little in common with, that system.

Viewed at the federal level, TANF lacks many of the features that are common to most social insurance programs. It provides no federal guarantee of income support to beneficiaries. The block grant structure of the program, combined with its broad purposes and lack of requirements that funds be spent on any particular benefit, leave it having little in common even with other federally funded means-tested programs. However, the first purpose of the program, the provision of assistance so that “children can be cared for in their own homes” clearly establishes the social-insurance nature of the program.

The most common use of TANF funds is the provision of income supplements to “needy” families—families that lack adequate incomes due to unemployment or low earnings. Every state has a TANF program that provides such supplements, although there is great variation among states in the extent to which employed families with low earnings can receive income supplements, the amount provided to unemployed and employed families, and the requirements that families must meet in order to receive supplements.

If the core purpose of social insurance is to provide income security against common risks, there is little question that TANF income supplements are properly viewed as part of the American system of social insurance. Like other social-insurance programs, TANF income supplements provide families with protection against the risk of inadequate labor income inherent in a market economy. While it is true that TANF income supplements are generally conditioned on participating in various work, training, or social-services activities, such requirements are not inconsistent with social insurance. UI recipients, for example, may be required to register for work with the Employment Service and usually must be “able and available” to work. An injured worker receiving workers compensation benefits may be required to engaged in rehabilitation activities.

The conventional wisdom that TANF income supplements are “welfare” rather than social insurance is bolstered by the belief that most parents who receive TANF income supplements are able-bodied and have no work experience. This belief is incorrect. Most parents who receive (or should be receiving) help from TANF are workers. For example:

• More than two-thirds of families with below-poverty-level incomes included one or more workers in recent years.
• An HHS study of TANF beneficiaries in six states found that nearly 4 in 5 have worked at least half their adult years.
• Most families who receive TANF income supplements in a typical month are either working during that month, or are working at some point during the year in the year they receive TANF.

Of course, TANF income supplements are not limited to families with parents who are employed or have significant work experience. TANF is also an income supplement program for families with limited or no labor force experience, for example, very young parents who are still in school, recently divorced or separated parents (so-called “displaced homemakers”), and persons living in depressed areas with extremely limited job opportunities.

However, providing insurance for such individuals is not inconsistent with early historical understandings of social insurance. In an article in which he noted the conventional distinction between social insurance and public assistance, Arthur Altmeyer, went on to say: “Yet they are of the same kind, although people sometimes hold that those arising out of contributions paid by a person, or on his behalf, are the more valid. I do not believe that such a distinction can be made.” Similarly, in one of the first books written by an American on social insurance, Social Insurance: A Program of Social Reform (1910), Columbia Professor Henry Roger Seager explained that social insurance “consists in protecting wage-earning families which have developed standards of living from losing them, and in helping wage-earning families without standards to gain them.”

This is not to say that the TANF program as a whole should be thought of exclusively as a social insurance program. TANF funds are used for a variety of purposes other than supplementing the inadequate income of workers. Some of these purposes, such as the provision of child care, can be thought of as falling within the rubric of social insurance, while others, such as teen pregnancy prevention programs, clearly fall outside of the scope of social insurance since their primary purpose is not to provide protection against immediate economic insecurity.

One way to understand the importance of TANF income supplements in the American social insurance system is to consider how it fills gaps in “coverage” and benefit adequacy. In this sense, TANF is complementary to other social insurance programs that provide income support or supplements, including unemployment insurance (UI), the Earned Income Tax Credit (EITC), and SSI. Moreover, the existence of these programs does not eliminate the need for TANF supplements.

For example, TANF helps make what is sometimes called the “work support” system whole by ensuring that working families have supplemental protection against very low-earnings. For parents with low-wage or part-time jobs, TANF income supplements can make the difference between living above or below the poverty line. Moreover, because TANF income supplements are typically provided to working families on a monthly rather than an annual basis, they can play a more immediate role than the EITC in reducing hardships caused by low earnings or unemployment.

Similarly, TANF helps fill gaps in disability insurance coverage. Supplemental Security Income (SSI) provides income support to persons with disabilities who are not eligible for Social Security disability insurance or are eligible for very modest benefits. SSI is limited to individuals, however, and thus does not provide supplemental income for children of a parent with a disability. The SSI eligibility criteria are stringent.  SSI’s definition of “disability” is the same as that used in Social Security. A person must have a physical or mental impairment that will last at least 12 months or is expected to result in death. In addition, the person must prove that he or she is not able to engage in any “substantial gainful activity” as a result of the impairment. The definition is stricter than definitions commonly used in private disability insurance.  It also is stricter than definitions used in many public employee benefit systems for federal, state, or local employees.

In my next post in this series, I'll discuss how understanding TANF as a part of a larger system of social insurance can provide a better lens for thinking about the program's future.

Social insurance “consists in protecting wage-earning families which have developed standards of living from losing them, and in helping wage-earning families without standards to gain them.”

--Social Insurance: A Program of Social Reform (1910), Henry Roger Seager

I've been digging Matt's series of posts on the safety net and John's last one on whether a frame like "expand the middle class" is really all that different from one along the lines of "reduce poverty." I'm hoping to do some longer posts over the next few weeks on these matters. As a starting point, I think it's important for us to put to rest the distinction between welfare and social insurance. Conventional wisdom has it that programs like Social Security, Medicare and Unemployment Insurance are categorically different than programs like TANF, food stamps, and Medicaid. A common way of expressing this difference is that the first set of programs are “social insurance” and the second set are “means-tested public assistance” or, more pejoratively, “welfare.”

This distinction is both artificial and ill-conceived. Means-tested programs like TANF and food stamps should be thought of as a necessary part of America’s system of social insurance rather than as “welfare” programs that exist outside of, and have little in common with, that system. Like Social Security and other social insurance programs, means-tested public assistance programs protect Americans against various risks that can reduce their economic security. This essential similarity of purpose is more important that some of the design differences that exist among programs serving an income-security purpose.

In True Security: Rethinking American Social Insurance, Michael Graetz and Jerry Mashaw define social insurance as a set of programs and institutions that “cover common risks to income security across the life cycle of individuals” (45). In this conception, social insurance is defined by its core purpose—moderating the risks of income loss or inadequacy—and not by its funding mechanism or other design features. The Graetz/Mashaw project is best understood, not as a sweeping reconceptualization of social insurance, but as an attempt to develop a conception of social insurance that is more conceptually coherent and useful, and perhaps even more historically grounded, than the current conventional conception of social insurance in the United States.

In their view, the argument that social insurance programs cannot be means-tested is an “ahistorical” one that reflects a political strategy to distinguish Social Security and other programs from unpopular “welfare” programs like AFDC.

As a matter of history, protection against current low income because of defined personal or family circumstances, irrespective of past contributions or earnings, has long been a cornerstone of American social insurance arrangements. The original Social Security Act was a compromise between those who thought social insurance should be structured primarily as a protection against low income and those who saw it primarily as a protection against loss of prior economic status and wanted social insurance closely tied to workforce attachment. (62)

While the strategy of distinguishing social insurance from “welfare” may have been politically beneficial at times, Graetz and Mashaw view it as a “serious mistake.”

This artificial and ahistorical division of the social welfare world between contributory and non-contributory schemes strands crucial poverty reduction programs in political backwaters. It creates confusion in both public discourse and public perception whenever progressive benefit and contribution formulas for social insurance are proposed and discussed. This political separation poses political dangers for ‘contributory’ schemes as well. It highlights ‘individual equity’ or bank-account considerations in social insurance arrangements—represented recently by the ubiquitous calculations of each individual’s ‘money’s worth’ from Social Security—while submerging the social adequacy commitment that should be the fundamental norm in the design and defense of social insurance.

This doesn’t mean that public assistance programs are solely social insurance, or even that all public assistance programs are social insurance. While Graetz and Mashaw believe social insurance is not limited to Social Security and Medicare, their conception of social insurance as a protection against income insecurity is “considerably narrower than all the public activities that might be said to support American family income.” As an example, they note that education programs are not social insurance, since they don’t provide insurance against “a current loss of economic well-being.” Instead, such programs are more appropriately viewed as an investment in future economic opportunity. (58)

This distinction isn’t completely clear-cut. Education obtained in one’s youth, after all, does enhance income security over the life cycle. But Graetz and Mashaw argue that such a narrowing of the definition is necessary for pragmatic reasons.

If the definition is too broad, ‘insurance’ becomes a useless metaphor.... Important ideas about good program design—identifiable risks, moral hazard, adverse selection, and so forth—lose their salience. If the criminal justice system qualifies as “social insurance” (protection against loss of income or assets through theft, embezzlement, and the like), the concept fails to define a distinctive area of public policy. (57)

Social insurance also is distinct from other types of insurance, including insurance provided in markets where there is considerable government involvement or regulation. The adjective “social” is important in making this distinction. Social insurance is different from other forms of insurance because it is a “social rather than an individual (or group) contract” and is made for “the purpose of collective provision, subsidy, or regulation.”

In part II, I'll discuss a specific case—why that quintessential "welfare" program, TANF, is best thought of in social-insurance terms.

Aussie John Quiggin is on to something:

....

The growing social salience of risk and uncertainty suggests the possibility of a more general reorientation of views about the welfare state, its role and significance. One way of approaching this reorientation is in terms of Nicholas Barr’s distinction between the ‘Piggy Bank’ and ‘Robin Hood’ functions of the welfare state. In Barr’s terminology, the Robin Hood function refers to the redistribution of wealth (either in lump sums or as flows of transfer payments) from the ‘lifetime rich’ to the ‘lifetime poor’. In a society where endowments of physical wealth and earning capacity were equal, the Robin Hood function would be unnecessary.

By contrast, the ‘Piggy Bank’ function of the state involves the smoothing of individual consumption over time and over a range of risky outcomes. This function would be relevant even in a society where lifetime incomes were equal. It is just as relevant, (perhaps more so) for those on middle incomes as for those lower down the scale.

In traditional presentations of the case for social democracy, these ‘smoothing’ functions were commonly seen as peripheral. Advocates of a targeted welfare system saw the provision of services to households that could afford to provide for themselves as an undesirable side-effect of provision for the poor - in other words, ‘middle class welfare’. On the other hand, advocates of universal provision saw it as politically necessary to build support for redistribution.

Traditionally, more attention has been focused on the Robin Hood function of the welfare state than its role as a Piggy Bank. Opponents of universalism have argued that ‘middle-class welfare’ constitutes wasteful churning and leads to an excessively large state that nevertheless does a poor job in equalising income. Supporters argue that universal programs build social solidarity and cement support for the Robin Hood function, even among those who are net contributors.

But Nicholas Barr persuasively argues that both sides miss the point. Consumption smoothing and risk-pooling are valuable in themselves, and the role of the state in these activities needs to be assessed independently of distributional issues.

In fact, there is a strong case that redistribution plays a vital role because it pools risks that arise within individual lifetimes. In other words, redistribution deals with the risk of being born into a poor family instead of a rich one, possessing the wrong type of job skills for a particular labour market, or living through a sustained economic downturn. On this analysis, the primary role of the welfare state is managing risk, not redistributing income.

See also, Quiggen's The Risk Society: Social Democracy in an Uncertain World.

What linguists call conceptual metaphor refers to the "understanding of one idea, or conceptual domain [the "target domain"] in terms of another" [the "source domain"]. Here's a striking recent example in which John Wagner, Director of the California Social Services Department, manages to use two metaphors derived from animal husbandry—"wean" and "carrot and stick"—to describe Gov. Schwarzenegger's proposal to cut Temporary Assistance income supplements:

The assistance boost would apply to people whose incomes are low enough to qualify for food stamps but have already been weaned from the CalWorks program, which provides temporary financial assistance to needy families, Wagner said.

....

Along with the carrot of more assistance for some working families, the stick could be strengthened as well. California currently cannot cease all assistance from families who fail to comply with work or other activity requirements, one of only a handful of states with such lenient policies, Wagner said.

While the "safety net" metaphor has limitations, it's at least a better and more respectful metaphor to use in discussing programs for the adult people Wagner is referring to. Parents trying to get by on a low-wage job are walking a tightrope. Programs like Temporary Assistance provide both a "balancing pole" that makes it possible to continue moving forward and a safety net that catches them if they become unemployed. Even better, if less metaphorical, is the concept of social insurance, which has the added benefit of suggesting a collective endeavor to provide protection against labor market risks.

This useful new report, put out by Pew's Economic Mobility Project, attempts to determine how much the federal government spends on economic mobility. What I particularly like is that it looks at the whole range of federal spending going to individuals and families, including tax preferences and subsidy program that provide the greatest benefits to upper income families. (These subsidies include the mortgage interest deduction, and tax subsidies for employer-provided health care and retirement plans).

Among the notable findings:

• The federal government spends substantially more of its "mobility budget" on tax preferences/tax subsidies than on direct spending programs.
• The federal government spends some $242.4 billion a year, or 1.9 percent of GDP, on employer-related work subsidies, mostly tax subsidies for employer-provided health care and retirement plans.
• Most of the benefits from the largest subsidy for homeownership, the mortgage interest deduction, go to families in the top fifth of the income distribution.

I do have, however, a quibble with the report's definition of economic mobility spending.

... we focus on mobility in the acquisition of private income and assets, including human capital and education, but not on improvements in consumption levels, either absolute or relative. In effect, we look at whether programs aim to enhance the ability of households to increase their pre-transfer, pre-tax household income or assets, including human capital and educational attainment.

This definition means, for example, that the mortgage interest deduction for a wealthy family with two homes is treated as a mobility expenditure, while rental housing assistance for a working class family is treated as "income maintenance" and not as a mobility expenditure. This strikes me as being exactly backward. For the wealthy family, the mortgage interest deduction is really just a consumption subsidy that has little or no effect on their economic mobility. By comparison, for the working-class family, a rental housing subsidy is best thought of as a mobility subsidy since it provides them with the stability needed to improve their standard of living by working. (And, if it the subsidy is a rental housing voucher it provides them with an addition form of mobility, the ability to move somewhere else in search of better work, something that homeownership limits to some extent.)

Similarly, the report defines Pell Grants and the EITC as mobility expenditures, but Temporary Assistance income supplements as non-mobility income maintenance. Yet, Temporary Assistance supports work and education in much the same way as Pell Grants and the EITC—in most places participation in such activities is essentially a condition of receipt of Temporary Assistance—and Pell Grants and the EITC can be used to maintain consumption in much the same way as Temporary Assistance.

As a result, the report's definition of economic mobility spending seems both under- and over-inclusive. If I have the time, I might try my hand at a remix of the data ...

In his important 2005 book Growing Public, economic historian Peter Lindert drew on over a century of economic experience in the West to demolish the conventional wisdom that public social spending inevitably reduces economic growth, and that there is a trade-off between efficiency and equity. According to Lindert's research, it's possible to equalize incomes and improve life expectancy at zero cost in terms of reduced economic growth.

Last year, the National Academy of Social Insurance held a seminar on Lindert's work. This useful research brief summarizes the proceedings. Of particular note, is Lindert's emphasis on the importance of broadly universal social policies:

In examining countries with high social spending, Lindert finds that part of the reason their policies have not harmed their economies is that the policies cover almost everyone. For instance, northern European countries are much better than the United States at providing support for all working mothers. “Mothers’ human capital, their ability to become productive over the lifecycle, is very much affected by society’s willingness to give them help with childcare and early-parental leave.”

The brief also discusses research by GW political scientist Kimberly Morgan on why European social insurance programs have maintained support:

... these social insurance programs are universal, and thus are likely to have more political strength than means-tested programs targeting a much narrower, and less politically powerful, constituency. The second critical piece of this model is how these programs are financed, with a particularly heavy reliance on consumption taxes and payroll taxes. Morgan stated that consumption taxes paid at the time of purchase are less visible than income taxes, which are reported annually to the government. And consumption and payroll taxes may be more acceptable to groups that might otherwise lead an anti-tax movement, namely the wealthy or business groups.

....

“Some people are starting to argue that the financing of European programs really rests on a sort of cross-class compromise, one in which seemingly regressive forms of finance, which are less likely to antagonize powerful groups in society and are simply less visible to the public, basically fund progressive redistribution through social benefits,” Morgan said.

Turns out that Dennis Kucinich might not be that far off when he says "joblessness is a weapon of mass destruction. In a new NBER working paper, Daniel Sullivan and Till von Wachter find that mass-layoff-caused "job displacement leads to a 15-20% increase in death rates during the following 20 years."

If such increases were sustained beyond this period, they would imply a loss in life expectancy of about 1.5 years for a worker displaced at age 40. These results are robust to extensive controls for sorting and selection, and are consistent with estimates of the effects of job loss on mortality pooling displaced workers and stayers that are not affected by selective job displacement. To examine the channels through which mass layoffs raise mortality, we exploit the panel nature of our data -- covering over 15 years of earnings -- to analyze the correlation of long-run career outcomes, such as the mean and standard deviation of earnings, with mortality at the individual and group level, something not possible with typical data sets. Our findings suggest that factors correlated with a decrease in mean earnings and a rise in standard deviation of earnings have the potential to explain an important fraction of the effect of a job displacement on mortality.

Although burdened with an idiotic title, Indicators of Welfare Dependence, an annual report put out by HHS, has a fair amount of useful information in it. Some interesting tidbits include:

• More than half of the people who received Temporary Assistance Supplemental Income¹ in 2004 lived in families with at least one worker.
• A lot more adults receive Supplemental Security Income (4.083 million between the ages of 18-64, and almost another 2 million age 65 and over) than receive Temporary Assistance Supplemental Income (1.277 million adults).
• During 2001-2003, the typical individual who started receiving Temporary Assistance Supplemental Income received it for 4 months or less.
• The typical person who started receiving Temporary Assistance Supplemental Income in the 2001-2003 period, applied for it because their earnings from work fell.
• As the chart below shows, Temporary Assistance has been fairly non-responsive to economic changes. More individuals who have seen their incomes drop have become eligible in recent years, but fewer are actually get help.

There are a number of improvements that could be made to Temporary Assistance Supplemental Income to ensure that more workers who are eligible for help actually get it. But the big barrier to making these kinds of improvements is mostly attitudinal. Until the program is seen as one that provides a form of social insurance to workers, it will be hard to get much traction on positive changes.

Part of the conclusion from an interesting recent IRP discussion paper from Neil Bania and Laura Leete:

....

In a logistic regression model of food insufficiency estimated for lower income (nonelderly) households, we find that income volatility—not just the mean level of household income—is a statistically significant determinant of food insufficiency. These effects are robust across different model specifications, with more variable income being associated with higher rates of food insufficiency. Furthermore, the relative importance of income volatility versus income level in determining food insufficiency increases as household income falls. For lower income households (below 150 percent of 35 poverty), income variability is a statistically significant determinant of food insufficiency while average (mean) income is not. These findings are consistent with models in which such households face either greater liquidity constraints or more binding constraints in spending that are associated with contractual expenditures.

....

A second important set of findings relates to the degree and source of income volatility itself. We find that income volatility in this sample is considerably larger for lower than for higher income households and that it increased between 1992 and 2003. Furthermore, this increase in volatility was largest for households in deep poverty and those at risk for welfare usage (with 64 and 78 percent increases, respectively). In simulating the probability of food insufficiency for households in deep poverty, we show that the increase in income volatility between the two time periods would double the impact of a one standard deviation negative income shock on predicted food insufficiency.

In order to examine the sources of increased household income volatility, we decomposed income into three components: earned income, AFDC/TANF income, and all other income. Our analysis indicates that the increase in volatility is largely attributable to the shift in the composition of income from welfare payments to earnings and other income sources among the poorest households and among welfare at-risk households.

....

Large recent increases in income volatility, taken together with the importance of volatility in determining food insufficiency for low-income households, suggests that the nexus of income level, income volatility, liquidity constraints, and access to a social safety net may be more generally important to the well-being of lower income households than previously recognized in this and related literatures. Various aspects of these relationships deserve continued attention.

I think a big part of the problem is that both Temporary Assistance income supplements and Unemployment Insurance aren't doing enough to smooth the income volatility of low-wage workers. People need to stop thinking of Temporary Assistance as primarily "welfare" for people without labor market experience and start thinking of it as a social insurance program for low-wage workers.

An editorial in today's NYT notes that the United States devotes relatively little, as a share of its economy, to public social expenditures:

You would think that we were living in the lap of the Nanny State. One of the most puzzling facts of the political debate is how much traction Republicans still get from their calls to cut taxes and public spending, and how timorous Democrats are in arguing against them.

The United States has long had one of the most meager tax takes in the industrial world. America’s social spending — on programs ranging from Medicare and Social Security to food stamps — is almost the stingiest among industrial nations. Among the 30 industrialized countries grouped in the Organization for Economic Cooperation and Development, only four — Turkey, Mexico, South Korea and Ireland — spend less on social programs as a share of their economy.

The data the NYT is relying on is from the OECD Factbook 2007. Here's more detail:

PS: I made the chart using the new Numbers spreadsheet program in Apple's latest version of iWork. I've only started testing it out, but so far I like what I see.

The WSJ on the most recent research on income volatilty:

Weighing in on an intensifying debate on income insecurity, three economists—including two from the Federal Reserve—have found that American families today are more likely to experience big drops in their income than three decades ago.

....

The study found that the volatility of household income rose 23% between the early 1970s and early 2000s. While small changes in family income are no more frequent, large changes in income—more than 50%—are.

The probability that a family will experience a decline in annual income of 50% or more, compared with their average income in the previous three years, rose to 1.8% in 1995 from 0.6% in 1973. After 1995, the probability dipped, and has risen back to 1.7%.

"The increase in volatility we document is not trivial," Mr. Elmendorf said in an interview. "Our work is quite consistent with being concerned about the level and increase in volatility of household income."

....

The new study finds that volatility in all the components of income -- wages, capital income, and government payments -- has risen in the past few decades and income volatility has risen for both high school and college graduates. It finds that for individuals, income isn't any more volatile than 30 years ago, because an increase in volatility for men has been offset by a decrease in volatility for women. But because men typically earn more than twice as much as their wives, their experience dominates the trend for households.

Thus, the study reinforces the view among academics that household volatility has risen, although the magnitude is in dispute. On the other hand, it is at odds with the CBO study, which used extensive Social Security record and found no increase in volatility for either male or female wage earners since 1980.

Mr. Elmendorf and his co-authors emphasize they don't yet know why volatility has risen, in particular whether it reflects voluntary reductions in work, such as to care for children, or forced reductions, such as layoffs, or whether it reflects a shift in the makeup of households, such as more single-parent families.

....

Today's Progress Report from the Center for American Progress notes in its State Watch section that NYC Mayor Michael Bloomberg annouced a plan "to improve poor New Yorkers' access to healthful food and exercise." There are some interesting things in the plan, particularly the piloting of paperless applications for food stamps. But it should be noted that around this time last year Mayor Bloomberg refused to implement a policy change recommended by his top social services officials (and similar to one implemented by then-Governor Pataki in other parts of the state) that would have allowed unemployed adults not caring for children to receive food stamps for longer than three months. That policy change would have done much more to increase food security in New York City than the proposals annouced by the Mayor today.

The conclusion of Elizabeth Warren's testimony at today's Senate Finance Committee hearing Economic Issues for America's Working Families:

America was once a world of three economic groups that shaded each unto the other—a bottom, a middle, and a top—and economic security was the birthright of all those who could make it to the middle. Today the lines dividing Americans are changing. No longer is the division on economic security between the poor and everyone else. The division is between those who are prospering and those who are struggling, and much of the middle class is now on the struggling side.

The economy has changed, and middle class families are struggling to change with it. Laws like social security, Medicare, FHA, consumer product safety, fair credit reporting and a host of other statutes were designed to help middle class Americans cope with the risks in the economy of the mid-Twentieth Century. With a strong safety net to back them up, Americans innovated at a rate unparalleled in world history. Today’s families face new costs and new risks, and they need help so that they too can achieve security and prosperity for themselves and a stronger, healthier economy for everyone.

The correct answer is nothing, but for reasons I can't figure out, the usually excellent American Prospect has published a piece on welfare reform in its latest issue with the unfortunate title Using Carrots and Sticks. The piece includes this line:

It was the stick of welfare reform that induced mothers to leave welfare for work; it was the carrot of work-support benefits that supplemented the mothers' earnings and led to substantial reductions in poverty.

This is an inaccurate assertion, but before I say more about the actual facts of the matter, I need to say my piece about why progressive media should never use the phrase "carrot-and-stick" in an article about lone mothers, or pretty much any one who isn't a CEO or serving on a corporate board.

According to Merriam-Webster's Online Dictionary, this is the etymology of "carrot-and-stick":

from the traditional alternatives of driving a donkey on by either holding out a carrot or whipping it with a stick

Now if I was, say, the editor of the National Review or the Weekly Standard, I certainly wouldn't mind running a piece that uses a metaphorical concept that effectively compares lone mothers to donkeys that need to be driven to work. But, if I was the editor of the American Prospect—which is, after all, "the magazine for liberal intelligence"—I'd very quickly red-pencil it.

One can support conditionality in social programs—and I often do, depending on the program and the degree—without resorting to dignity-denying metaphors like carrot and stick.

I've heard carrot and stick used before by liberals in regards to labor force participation by lone mothers. Who knows, it's probably even passed across my lips once or twice when I've been talking to conservatives and trying to speak their language. But it's not a term I've ever heard used by liberals when it comes to talking about unemployment compensation, dislocated factory workers, or middle-class families. And I know the Prospect would red-pencil it if used in those contexts.

Ok, enough with my William Safire bit, on to the substance of the assertions in the Prospect piece. Frankly, I'm quite puzzled by why the magazine of liberal intelligence published this piece in the first place, just as I'd be puzzled if the Weekly Standard accepted my latest on the need for carrots-and-sticks to rein in excessive CEO compensation, get Wal-Mart to be nice to its low-wage workers, or encourage small-business owners to pay more than the current minimum wage.

The piece tells a neo-liberal story about the decline in poverty in the 1990s that starts with the conservative 1996 welfare law, instead of where it really belongs, with the massive expansion of wage supplements (the Earned Income Tax Credit). The EITC was basically tripled under legislation passed in 1993 with only Democratic votes. Moreover, the piece fails to even mention the increase in the minimum wage in 1997.

Consider this part of the carrot-and-stick story as told in the Prospect piece:

Adding tax benefits, primarily the EITC, does not reduce poverty at all in 1990, but reduces it another 5 percentage points in 1999. The combination of work and work supports reduces poverty a full 12 percentage points—or by about 4.5 million people—more in 1999 than in 1990. It was the stick of welfare reform that induced mothers to leave welfare for work; it was the carrot of work-support benefits that supplemented the mothers' earnings and led to substantial reductions in poverty.

This last sentence simply isn't evidence-based. Research suggests that the ETIC and other "carrots" played as much, and quite possibly, much more, of a role in increases in work and reductions in poverty than the "sticks" in the 1996 welfare law. Remember, almost half the decline in families receiving AFDC/Temporary Assistance happened before the 1996 law was actually implemented by most states (in 1997 and 1998) and after the massive EITC expansion. UCLA’s Jeffrey Grogger, one of the leading welfare reform researchers, concluded in a study published in 2003 that the EITC may be the “single most important factor” in the increases in work and earnings, and the decline in AFDC/Temporary Assistance receipt that occurred in the 1990s.

Moreover, among welfare reform programs that have been rigorously evaluated, the single most effective one—the Minnesota Family Investment demonstration that began in 1994 and was evaluted by MDRC—was pretty much all "carrots" including a higher basic grant level for all beneficiaries, and wage supplements for low-wage workers. The program had a kind of mild conditionality—requirements to participate in work and various other activities—but failure to comply with these conditions meant only that you got 90 percent of the higher basic grant instead of 100 percent, and there were no time limits on assistance. Still, the demonstration boosted employment, earnings, children's school performance, and even marriage rates (with no federally funded marriage-promotion counseling, btw!), while it reduced domestic violence and poverty. You hardly hear about it anymore, however, because the facts of the demonstration don't fit the dominant "sticks matter more than carrots" frame about welfare reform.

The paramount role that wage supplements (either in the ETIC form or like those of the MFIP demo) played in producing good outcomes in the 1990s shouldn't come as a surprise. The EITC expansion happened on a uniform basis throughout the United States before the state-by-state implementation of the welfare law in most states. In fact, EITC payments to low-wage workers reached $25 billion in 1995 (I'll need to double check but that's probably as much, if not more, than was spent on AFDC that year). And, we know from research conducted by Chris Jencks, Kathy Edin, and Laura Lein that low-wage (and mostly unreported) work was actually quite common among beneficiaries of AFDC. In fact, as Jencks noted in his introduction to Edin and Lein's 1997 book Making Ends Meet, in one of their surveys of AFDC beneficiaries, conducted in Chicago in 1985:

None of them lived on AFDC alone, and none of them reported all their income to the welfare department. The average mother got only half her cash from AFDC alone, and none of them reported all their income to the welfare department. The average mother got only half her cash from AFDC. The remainder came primarily from off-the-books employment, family members, boyfriends, and absent fathers.

My own guess is that the EITC expansion not only led to direct increases in labor force participation by lone mothers, but also led to a big increase in the number of working mothers who stopped working off the books and started working on them because their work was now rewarded, rather than penalized.

Letting neo-liberals tell the story their way isn't without consequences. I believe intelligent liberals can wholeheartedly support reasonable conditionality in social programs, and I certainly do. But the most important and problematic new element added to the welfare reform mix by the 1996 law wasn't conditionality, it was Reagan-style block-granting and the complete abdication of any direct federal responsibility, conditional or not, to low-wage working families in the Temporary-Assistance program.

Most intelligent liberals would agree it's a mistake to endorse more block-granting of federal social programs like Medicaid, Food Stamps, or Unemployment Compensation. But enshrining the 1996 law as a great leap forward effectively does just that. Again, that's fine for the Weekly Standard or National Review, but I expect different from the Prospect.

Admittedly, a big part of the problem is that liberals have failed to tell their own coherent and positive story about why some good things did happen in the late 1990s, despite the passage of the 1996 welfare law and contrary to their own dire predictions about it. To the extent that liberals do have such a story, it tends to be overly complex, with the emphasis typically placed on a laundry list of carrots, err, programs, and an overreliance on what American University's Matthew Nisbet calls "data dumps" to tell the story.

We need to start telling a much simpler story, one that gives policies that aim to make low-wage jobs better ones by boosting wages—in the form of income supplements (EITC) and minimum wage increases—their proper place as the leading driver of what should be called progressive welfare reform. And we need to stop with the data dumps.

The experience of rising poverty during the economic slowdown under the Bush Administration provide further support for this thesis. Unlike Temporary Assistance, which has provided income supplements to fewer and fewer low-wage workers even as poverty increased over the last several years, the EITC works like it should, helping more people when more people need it. The EITC now does more to ease earnings volatility and provide insurance for low-wage workers against downturns and temporary unemployment than Temporary Assistance.

Why has the EITC turned out to work so much better than Temporary Assistance? I think it's because the EITC is a focused federal program with an entitlement-funding structure (like nearly all tax benefits) based solidly on the principle of reasonable reciprocity and a core federal responsibility to low-wage workers. Temporary Assistance lacks all of these elements.

For some additional thinking on a way to tell this story, see this post from last August.

Lots of interesting findings in Elizabeth Kneebone's new Brookings Metro Center report on recent trends in the Earned Income Tax Credit is the extent to which the EITC—and over time the Child Tax Credit. Kneebone finds that the EITC helped to buffer the impact of the economic downturn and slow recovery in the first half of this decade. From tax year 2000 to 2004, the number of tax filers receiving the EITC increased by amost 3 million—from 18.8 million to 21.7 million. Kneebone notes:

Many central cities showing the largest increases in the share of filers receiving the EITC coincide with those experiencing increased overall poverty during that same time period. Job loss and slow economy recovery hit the Detroit, MI, Dayton, OH, Cleveland, OH, and Greensboro, NC metro areas particularly hard over the first half of the decade. Increases in EITC usage rates in their cities reflect that economic dislocation and sluggish wage growth did not cause all affected families to drop out of the labor market completely, but rather contributed to a rise in lower-wage and part-time work that qualified families for the credit.

This confirms my own view that the EITC is less a "work incentive" and more a social insurance program that protects workers who end up in poor-paying jobs, or part-time work involuntarily. This is one of the many reasons why it would be a bad idea to adopt a proposal made by Ron Haskins and Isabelle Sawhill of Brookings to tie EITC expansions to full-time work.

Other interesting factoids: "Red America" is also EITC America—Kneebone has a great map showing that the highest rates of EITC receipt are found in the South; as shown below, most low-wage workers pay to have their returns prepared and the share of workers hiring paid preparers is on the increase.