The Myth That Conservative Welfare Reform Worked--Part 3

by: Paul Rosenberg

Sat Mar 06, 2010 at 16:00

In the interests of digestibility, I've decided to split both the diaries planned for this weekend into two parts each, running a two-part diary on national data this weekend (splitting Section 3 into two sub-sections) , and two separate diaries on international comparisons (Section 4) and state-level comparisons (Section 5) next weekend.  In this diary, I will look at national data that helps explain the main false assumptions used in the conservative anti-welfare arguments made popular by Charles Murray in his 1985 book, Losing Ground.  In the next diary, I will concentrate on data that illustrates what the real problems actually were, and how welfare reform failed to perform as promised.

Section 3A: The Story of National Data-Debunking The Conservative Frame

In this part, we examine national level data showing that conservative welfare reform was not a success as it widely believed in Versailles.  To do this properly, however, we must take account of the arguments advanced to make the contrary point-no matter how misleading, and yes, downright dishonest they may be.  Thus, before any number-crunching can begin, we need to discuss how conservatives portrayed welfare as a failure in need of reform. At the end of 1970s, no one would claim that America's social welfare policies were a roaring success, but liberals would point to two major factors-first, a relative lack of effort compared to Europe, and second, the abysmal economic performance of the 1970s, which had stalled economic progress even for the middle class, much less for the poor. An obvious solution would be to take lessons from Europe-but that would be too logical. Instead, conservatives came up with a facade of research to "prove" their charge that not only had Lyndon Johnson's Great Society programs failed to reduce poverty, it had made things worse.

This was logic of Charles Murray's 1985 book, Losing Ground: American Social Policy, 1950-1980 (whose very subtitle was a lie--see below), and it was reflected in Ronald Reagan's claim in his 1988 State of the Union that "some years ago, the federal government declared war on poverty, and poverty won." But this had not always been the conservative position.  Indeed, in late December 1980, Martin Anderson, an Ayn Rand acolyte and a researcher at the Hoover Institute, had just been announced as Reagan's first chief domestic policy adviser, and an AP story quoted from his writings:

The 'dismal failure' of welfare is a myth.  There may be great inefficiencies in our welfare programs, the level of fraud my be very high, the quality of management may be terrible, the programs may overlap, inequities may abound and the financial incentive to work may be virtually nonexistence.

But if we step back and judge the vast array of welfare programs, on which we spend billions of dollars every year, by two basic criteria--the completeness of coverage for those who really need help, and the adequacy of help they do receive--the picture changes dramaticll.  Judged by these standards our welfare system has been a brilliant success.  The war on poverty is over for all practical purposes.

This was, of course, delusional.  But it was a very different delusion than the one the right would later embrace.

There were many flaws in Murray's book, as was revealed in a devastating review in The New Republic by Robert Greenstein, "Losing Faith in 'Losing Ground,"' The New Republic, 25 March 1985.  Unfortunately that review is no longer available online--or even in Lexis/Nexus.  But Daniel was able to find it for me in a Canadian database... not, however, before I had taken a different tack, drawing on a more exhaustive refutation in a multi-author booklet, Losing Ground: A Critique, Institute for Research on Poverty [IRP] Special Report Series #38, which was summarized in Focus, IRP's nesletter, Volume 8, Number 3, Fall and Winter 1985.  

Since Daniel was able to provide me with a copy of Greenstein's review, I'm going to focus briefly on the most central, devastating part of his critique, and then continue with the diary as I had prepared it before Daniel's late save.  

The crucial part of Greenstein's review starts off thus:

But before Murray s view that the programs have failed becomes the new conventional wisdom, careful examination of his use of the facts and figures is very much m order. Could it be that Losing Ground, with its endless recital of statistics, actually rests on deceptive numbers juggling?  That Murray has, in the service of a radical political agenda, consistently omitted or concealed critical facts and research findings that do not support his case? Close scrutiny of Murray's data suggests that this is precisely what has occurred. This is the untold story behind Losing Ground.

Consider, for example, the "case of Harold and Phyllis." Harold and Phyllis are a fictional couple that Murray creates to illustrate one of his principal contentions--that the social programs have led to unemployment and illegitimacy among blacks by encouraging the poor to live off welfare rather than to work, Murray presents Harold and Phyllis as a young couple facing fundamental decisions about their lives. Phyllis is pregnant, and the couple must decide whether to get married. More important, they must also decide whether Harold should take a tedious minimum-wage job, or whether Phyllis should go on welfare and Harold should live off her benefits rather than work.

Naturally, Murray's analysis shows that the incentives in 1960 were clearly pro-work, whereas by 1970, they were anti-work.  Just one tiny problem, Greenfield explains--the analysis is "flatly wrong":

The Harold and Phyllis example is at the heart of Murray's case and is critical to the entire book. It is also flatly wrong, First, Murray's family budgets for l960 and 1970 are not based on welfare benefit levels in an average state. Instead, his data is for the state of Pennsylvania, a fact buried in footnotes at the back of the book. Welfare benefits grew twice as fast in Pennsylvania from 1960 to 1970 as in the nation as a whole. This allows Murray to portray the shift in incentives  over the decade as being twice as  great as they actually were. [Emphasis added]

If only Murray had had the integrity of Emily Litella, and just said, "Never mind," after Greenstein's review came out, he could have saved us all a great deal of grief.  But as Senator Al Franken is learning anew every day, national politics sorely lacks the integrity of Saturday Night Live, and so it was not to be.  There was another equally basic problem, however:

Murray makes a second error, In calculating the family's budget if Harold works, Murray incorrectly assumes Phyllis and her child cannot obtain food stamps, even though poor families who work are eligible for food stamps. In figuring the budget if they went on welfare, however. Murray does not make the same mistake and adds in food stamps when computing tho family s income. The error makes working, compared to welfare, appear less attractive than it actually was.

When the welfare vs. work comparisons are computed accurately, they show that taking a minimum-wage job was more profitable than going on welfare in most parts of the country in 1970. In some states with low welfare payments, such as southern status, minimum wage jobs paid almost twice as much.

These two problems were each enough, separately to have torpedoed Murray's thesis in the realm of serious research.   Together, they reveal Murray's thesis as a blatant fraud.  But there's much more wrong with his argument than that....

Three Main Criticisms From The Institute for Research on Poverty [IRP]

We now shift gears, turning to the IRP analysis, both in Losing Ground: A Critique and in the condensed report of its findings in Focus.

The critique explained:

"Dr. Murray argues that the growth of female-headed families during the past two decades was a response to increases in the generosity and availability of welfare programs.  His solution to the problem, which he offers at the end of his book, is to eliminate welfare benefits to single mothers and their children.  He claims this will 'drastically reduce births to single teenage girls...reverse the trendline in the breakup of poor families...[and] increase the upward socioeconomic mobility of poor families" (p. 227).

What's more, Murray also denied something that was obvious to anyone else at the time-the fact that the 1970s had been very rough, economically:

All agree, however, that progress against poverty was disappointing in the 1970s.... But why is poverty higher today than in the early 1970s?

Murray hypothesizes that the cause was the shift in social policy, not a lagging economy.  He argues that the period from 1970 to 1979 was one of strong economic growth: "Even after holding both holding both population change and inflation constant, per capita GNP increased only a little less rapidly in the seventies than it had in the booming sixties, and much faster than during the fifties.  Growth did not stop.  But, for some reason, the benefits of economic growth stopped trickling down to the poor" (p. 59).

It would be far too time-consuming to try to go through all of Murray's misleading or downright dishonest statistical tricks, but for our purposes, it's sufficient to just address three of them: The misrepresentation of the 1970s as a prosperous time when the poor ought to have been advancing, the misrepresentation of the timing of changes in black family structure, and the misrepresentation of the scope of his critique.  Together, exposing these three tricks is sufficient to discredit the main thrust of Murray's arguments, and to clear the way for a straightforward understanding of the actual statistics involved.

Murray's Misleading Claim About Scope
We take the last of these first, as it requires very little more than quoting a passage from article in Focus:

Poverty as officially measured among the aged has been reduced by between 30 and 50 percent since 1967 (see Table 1).  Public speneding on this group and the totally disabled, primarily through Social Security, Medicare, and Medicaid, accounted for over 75 percent of all 1980 expenditures for major income transfer programs.  Another approximately 18 percent was spend on programs for those who were neither elderly nor totally disabled--chiefly Unemployment Insurance and Workers' Compensation.  Thus, though the subtitle of Murray;s book is broad: "American Social Policy, 1950-1980," he is in fact emphasizing one the 7.3 percernt of the 1980 income transfers that go to nondisabled,nonelderly recipients of AFDC and food stamps.

The upshot of this is quite simple:  The Great Society/War on Poverty involved quite a bit more than just assistance to poor women, and once this is admitted, one must not only limit any conclusions drawn after the fact from Murray's arguments to only those policies he discusses, one must also .question ahead of time why his arguments about the supposed destructiveness of the Great Society/War on Poverty don't also apply to all other programs that came under that umbrella, such as, for example, Medicare.

Murray's False Claim About 1970s Prosperity
You'll recall from above that:

Murray hypothesizes that the cause [of stagnation in poverty rates during the 1970s was the shift in social policy, not a lagging economy.  He argues that the period from 1970 to 1979 was one of strong economic growth: "Even after holding both holding both population change and inflation constant, per capita GNP increased only a little less rapidly in the seventies than it had in the booming sixties, and much faster than during the fifties.  Growth did not stop.  But, for some reason, the benefits of economic growth stopped trickling down to the poor" (p. 59).

But everyone knows the 1970s were a tough time, economically.  The was stagflation, two unprecedented oil price shocks, and two debilitating recessions.  The (inflation-adjusted) GDP expanded 36.85% during the decade, but this was dramatically below the 50.77% expansion during the 1960s, and about 10% behind the 40.76% expansion during the 1950s.  Murray tries to get around this by fiddling with basis for comparison-choosing  per capita GNP (roughly equivalent to GDP), and pretending that this chose of measurement disadvantages the 1970s.  But he's actually doing is fairly easy to see through. During the 1950s and 1960s, the population grew more rapidly, due to the baby boom.  But during the 1970s, the potential workforce grew much more rapidly as the boomers entered the workforce in large numbers.  Making no distinction in age and ignoring the difference between GNP/GDP and total household income, Murray has fastened on a statistic that's convenient for his argument, but entirely divorced from the reality he's pretending to analyze. If one looks at the changes by decade for GDP, population and households, then things start to get much clearer:

The GDP per population measure makes the 1970s look better than the 1950s-by more than 20%, whereas just by GDP growth, the 1970s were about 10% worse.  But if we look at GDP per household, then the 1970s are suddenly more than 40% worse than the 1950s, which turns Murray's argument upside-down, and returns the analysis to the realm of reality, as can readily be seen by looking at the changes in median income (a 3-year average is used to smooth out annual fluctuations):

As we can see, there's no real mystery why "for some reason, the benefits of economic growth stopped trickling down to the poor"   They were no longer even trickling down to the median wage-earner--whose incomes were falling more often than not throughout the 1970s, whereas they had never fallen during the 1950s and 1960s. In times that tough, what chance did the poor have?

Murray's Convenient Confusion Over Illegitimacy Rates vs. Ratios And Timing

The third way in which Murray misleads is by confusing the birth rates outside of marriage with percentage of births outside of marriage. As we'll see below, black birth rates outside of marriage started growing  significantly long before the Great Society programs were enacted, and were actually headed downward at that time--a trend that continued for a decade after the programs were put into place.  But because fewer black women were getting married--and they were having fewer children when married--the percentage of black births outside of marriage rose at an increasing rate after 1965.  Murray's arguments logically depended on welfare causing an increase in illegitimate birth rates, but the data show the opposite.  So he used illegitimate birth ratios instead.

[Aside]: Before proceeding, it's also worth noting that during the 1970s teen birth rates in general were declining in the US, but not as rapidly as in other countries such as Britain, Canada, France and Sweden, which had far more generous programs of support.  This will be taken up more fully in my international comparisons section next weekend, but is worth noting here simply to underscore that even if Murray were not being deeply disingenuous with this particular argument, more recent data would have undermined his argument--particularly when compared with the outcomes of more generous welfare policies in other countries.

From here on, to explain the foregoing in detail, I quote extensively from one section of the Losing Gound: A Critique:

Sarah McLanahan: Charles Murray and the Family

Dr. Murray argues that the growth of female-headed families during the past two decades was a response to increases in the generosity and availability of welfare programs.  His solution to the problem, which he offers at the end of his book, is to eliminate welfare benefits to single mothers and their children.  He claims this will "drastically reduce births to single teenage girls...reverse the trendline in the breakup of poor families...[and]increase the upward socioeconomic mobility of poor families." (p227).

Murray's claim for the importance of welfare programs in producing the growth of female-headed families is based primarily on his analysis of the trend in illegitimate births.  In Chapter 9 he presents a figure showing the trend in the illegitimacy ratio for blacks and whites between 1950 and 1980.  Both trendlines increase substantially during the sixties and most of the seventies.  However, the black line jumps sharply in the mid-1960s, coinciding almost prefectly with the initiation of the Great Sociaety programs.  Murray places a good deal of emphasis onthe jumps in the two trends and argues that the increase in welfare has caused the increase in illegitimacy.

All of the evidence presented by Murray is based on the illegitimacy ratio, which is the ratio of nonmarital births to all live births.  If, however, we look at the trend in the illegitimacy rate, which is the ratio of nonmarital births to the total number of women at risk for such an event (single women between the ages of 15and44), a very different picture emerges.  The trends in illegitimacy rates fro blacks and whites are presented in Figure 1 below.

After the charts (there is one--less dramatic--for whites as well), McLanahan notes:

For black women, the illegitimacy rate rose sharply between 1945 and 1960,levelled off between 1960 and 1965, and began to decline after 1965.

This, of course, directly contradicts Murray's thesis.  Which is arguably why he ignores the illegitimacy rate and concentrates on the illegitimacy ratio for the rest of his analysis, as McLanahan goes on to note:

Murray, is aware of the discrepancy between the two trends and says so at the beginning of his discussion.  Having acknowledged the difference, however, he focuses exclusively on the illegtimacy ratio which, of course, is consistent with his explanation.

Next, she explains what's so important here:

Why, one might ask, is the illegitimacy rate any more valid than the ratio?  And how does one decide which statistic to use for examining a particular trend?  The answer is as follows: If one is interested in describing the experience of a particular group, e.g. children, the ratio is the appropriate statistic. It tells us what proportion of children are born to nonmarried women and presumably what proportion are at high risk for being in poverty.  If, however, one wants to talk about behaviors or propensities, the rate, not the ratio, is the appropriate statistic.  The illegitimacy rate tells us what proportion of nonmarried women are having children out of wedlock.  If the propensity of nonmarried women to have children is increasing, it should be reflected by an increase in the illegitimacy rate.

In short, the rate tells us about the choices and behavior of women as a group-what proportion of unmarried women are having children. The ratio tells us about the experience of children as a group-what proportion of children are being born out of wedlock.  It is Murray's core assumption that the former by and of itself drives the later. To a certain extent he's right.  If all other things are equal, that's exactly what it will do.  But the chart above shows that all other things aren't equal at precisely the point in time that he identifies as crucial-the point in time when the Great Society programs began.  Rather than trying to make his argument-which the statistics are overwhelmingly against-he simply assumes it to be true.  That's what he does when he uses the ratio and ignores the rate.

The following chart indicates how misleading and oversimplified Murray's argument is:

As the chart makes clear, not only can the illegitimacy ratio be increased by a decrease in the "legitimacy rate", the number of children raised by single mothers can also be affected by the divorce rate for women with children as well as the marriage rate for single mothers.  The obsessive focus on single motherhood--particularly among teens--is indicative of a pathological narrowing of attention.

What do I mean when I call Murray's focus "obsessive" and "pathological"?  Am I just being hyperbolic?  Not at all.  He is focusing attention on a fixed object in the midst of a complex dynamic.  And his fixation--as well as the response his "analysis" gained--is clearly informed by centuries of white male obsession with black female sexuality--particularly among young girls. This obsession was clearly attended by a good deal of guilt and shame, giving rise to negative projections onto young black women and girls. Murray--and those who championed his work--were so deeply in the grip of this dynamic, that were utterly oblivious to the conceptual confusion their arguments wallowed in.

Returning to McLanahan's discussion of the two statistics, and the significance of how they differ, she explains:

Why are the trends in the two statistics so different?  First, it is clear that the illegtimacy ratio jumped during the mid-sixties, not because of an increase in the rate of births to single women, but because of a dramatic decrease in marriage rates and a decline in the fertility of married couples.  Now Murray might argue that his theory is about the decvline in marriage--teh decision of Phyllis and Harold [fictional characters Murray uses to illustrate his arguements] NOT to marry. But in truth, his explanation speaks only to the decline in marriage among persons eligible for welfare, i.e., poor pregnant women. It says nothing about the behavior of middle-class women or nonpregnant women who in fact experienced an even greater decline in marriage during the late sixties and early seventies than the women describd by Murray, even though they had nothing to gain from the expanding welfare programs.

But, of course, Murray's argument has nothing to say about "the behavior of middle-class women or nonpregnant women who in fact experienced an even greater decline in marriage" because they aren't the objects of his lust/guilt/projection-based fixation.  They might as well not exist for him, because they are irrelevant to the underlying psycho-sexual fantasy which has permeated race relations in America since the first slaveholder raped his first teenaged female slave, thus giving rise to entire literatures of lies.  (For example, in one of his essays in Ronald Reagan: The Movie Michael Paul Rogin makes note of the fact that in Birth of a Nation, a leading "carpetbagger" is given a mulatto mistress, thus transferring the secret shame of Southern aristocrats onto their most immediate white enemies.)

And liberals, of course, are condescending (according to the narrative promoted by Gerard Alexander)--not because they bring up this psycho-sexual subtext--which Murray's social science criticis at the time did not--but because they dare to write the sort of painstaking critical analysis that McLanahan did, rather than just taking Murray's word for it that "welfare policies... discouraged work, marriage and the development of skills... with devastating effects," and they were the principal cause of poverty in Ronald Reagan's America.

Conclusion

In this diary, we've seen how Charles Murray made at least five major errors in interpreting--or constructing data--that when corrected utterly destroy his argument.  (This vastly understates the number of errors pointed out--it only refers to the ones I've presented.) Greenstein points out two of them: First that Murray usese data from an unrepresentative state--Pennsylvania--to construct an argument about the country as a whole.  Second, that he fails to include food stamp benefits in his calculations of how well off a working poor family would be.

The IRP authors pointed out three major errors that I discussed: First of all, the programs Murray discusses are only a small fraction of those he targets in his argument.  He not only makes no effort to extend his argument to programs not covered, he basically pretends they aren't even there.  Second, Murray pretends that the 1970s were a rosy time economically, so only the poor themselves (or those who have lead them astray!)  can be at fault for not getting out of poverty.  Third, Murray pretends that the Great Society unleashed a tsunami of illegitimate children when the illegitimate birth rate was actually declining at the time those programs came into effect.  With these three lies cleared out of the way, Murray's entire thesis is swept away as well.

This is not to say that there's nothing at all to Murray's arguments.  There's certainly some tendency for some people in any population getting "something for nothing" to become dependent and irresponsible.  We're certainly seen that throughout the financial sector, for example.  And we can see it in the monopoly sport leagues, where George W. Bush made his millions off of the taxpayers of Arlington, Texas. But once we realize that Murray's story is a fantasy, and we turn our attention to the national data presented in the next diary in this series, we will clearly see that things generally tend in the opposite direction, and that the negative incentives that Murray dwells upon are relatively trivial in the grand scheme of things.