A few weeks ago Bill Dickens and I argued about poverty: see here and here for previous rounds.  Now Bill’s written a lengthy response to my last post.  Italics indicate that Bill’s quoting me.  Enjoy!


“I’m not merely saying that “bad behavior is bad for you.”  I’m saying that bad behavior is a major cause of poverty.  If I’m right about this, there is a great, neglected remedy for poverty: Poor people should stop engaging in bad behavior

If this seems flippant, that’s not my intention.  Poverty: Who To Blame will largely be a work of economic philosophy.  Part of my project is to provide intellectual foundations for what I perceive as Americans’ justified frustration with welfare recipients.”

So this is the crux of it. You subscribe to two central right-wing memes: government coddles the poor and won’t make them face the tough choices everyone else does, and welfare recipients are overwhelmingly lazy and undeserving. Anyone with firsthand experience dealing with a wide range of the poor or those receiving government assistant (with the later being only a small subset of the former) knows these two things to be false. Further, I think your view on the causes of poverty and your view of the culpability of the impoverished for their circumstances are internally contradictory. Even Charles Murray acknowledged that beliefs in both the importance of cognitive ability and its genetic basis is grounds for progressive taxation. It is no leap at all to the notion that the poor are not culpable for their circumstances.

Let’s start with the absurd notion that better behavior is a “…neglected remedy…” Neglected by whom? Certainly not the poor themselves. When I worked on Clinton’s welfare reform taskforce I was struck by a number of things, but one thought that would never have occurred to me was that instructing people in the consequences of bad behavior would have the slightest impact on their condition. I must have spoken with, in the neighborhood of, 100 welfare recipients when I was working on the reform (and lots more people with precarious lives when interviewing people for a number of labor market projects). Overwhelmingly those on public assistance were full of regret and/or a sense of hopelessness that they are fated to their condition. They know they should have worked harder in school, they know they should be working to support their family, they know it would be better if their children’s father was there to help support their kids. There is no shortage of hectoring from society, welfare caseworkers, family members, and the media. Consider that even before the passage of TANF most women on welfare worked at least some during every year (on or off the books). Most welfare mothers are not drug abusers or alcoholics (when they have been tested only a tiny fraction fail). A lot had their children with a husband or boyfriend they had hoped to marry. A lot of the AFDC caseload cycled on and off welfare as people made repeated attempts to return to work (attempts that were often stymied by lack of adequate child care – one of the most common reasons for returning to welfare was being fired by a low wage employer for missing work when child care arrangements fell through).

Bryan, there is one thing I really think you ought to do before you write this book and that is to spend some time with people who are receiving state benefits and the wider group of poor who do not receive benefits but remain poor. If you do I think you will find it very hard to believe that these people are unaware of the consequences of their actions and that they aren’t, overwhelmingly, trying to make their lives work.

So now let’s turn to the notion that welfare recipients are undeserving. This meme, again, is based on a huge misperception/lie. Over and over when I talk to people about government income support programs I’m told that they have no objection to giving money to the truly needy, but that they don’t like supporting lazy bums who don’t like to work. When I tell them that overwhelmingly government support goes to families (usually single women) with children they don’t believe me. You, I think, know better so I wonder how you can hold the view that people’s frustration with welfare recipients is justified. Overwhelmingly, government money goes to families with children. Whatever you may think of the parents’ behavior the main recipients of government money are children and the disabled. You want to take money away from them? If not then you aren’t going to have much impact on the income support budget (at least not without going after SS or Medicare).

Go through the list of government programs and look at the criteria for receiving money or the formulas for determining the size of benefits: TANF — temporary and exclusively for families with children with benefits depending on the number of children, Food stamps – much more generous for families with children than for those without, EITC – depends on family size, only goes to those who work and the amount initially grows the more you work, WIC – nutrition support for pregnant women and infants, SCHIPS – health care for low income children who don’t qualify for Medicaid, Disability Insurance programs – for the disabled. In terms of major government support programs that leaves four. The typical American does not consider recipients of Medicare and Social Security to be undeserving. That leaves Medicaid and unemployment insurance as the only major transfer programs that are conditional only on income and assets (and in the case of UI previous work experience) with no bias towards families or the disabled.

What most people seem to have in mind when they talk about welfare is General Assistance – that is aid to able-bodied individuals with no children. Such programs became prevalent during the 1930s but are almost non-existent today. As of 2005 only two states still had any GA program. Some counties and cities run such programs but the benefits are extremely limited. In general, Medicaid and food stamps are the only programs available to able-bodied adults with no children and they do not provide enough income for food cloths and shelter.

So where are the undeserving recipients of aid programs? It is certainly true that any attempt to provide a safety net is going to lead some people at the margin to rely on the safety net and work less than they would if it wasn’t there. You cite my estimate that extended benefits increased the equilibrium unemployment rate by about 1 percentage point (as evidenced by an outward shift in the Beveridge curve consistent with a change of that magnitude that took place right around the time extended UI benefits were phasing in). There are several things I would like to point out here. Art Okun used to talk about transfer programs as leaky buckets. They do not create first best solutions. The question is whether the amount that reaches the deserving recipients, and the good that does, isn’t justification for what is lost through the “leaks.” That is a question to consider very carefully. Better buckets are good. Very leaky buckets need to be dropped or replaced. But leaks are not a reason by themselves not to try to use a bucket.

Another thing to consider is what people would be doing in the absence of a transfer program. In the case of the increase in the unemployment rate induced by unemployment insurance there are two ways UI could cause that. You are thinking that it causes people not to take jobs they otherwise would, but the unemployment rate also goes up if you keep people in the labor force who would otherwise cease looking for work. Recent evidence suggests that the impact of UI is much more the latter than the former (also I’m moving away from my original view on this as a considerable decline in the number of people claiming long-term unemployment insurance has done nothing to shift the Beveridge curve back in so we are back to the drawing board in trying to explain this phenomena).

Now let’s consider the case of a bucket that was probably
too leaky and needed to be replaced. As you know I was converted by my
experience with Clinton’s welfare reform task force to the belief that AFDC
needed to be time limited. Over and over I heard young women tell me that they
didn’t think much about having a baby because that is what people in their
world did. “You get to be 16, you get yourself a baby and you get yourself a
check and an apartment.” AFDC as a career choice was a serious problem back
then. But even as we went around preparing the welfare reform we heard
over-and-over again that the word was out that welfare was going away and you
were going to have to do something else now. Starting in the early 90s – long
before TANF actually limited benefits to 2 years – AFDC caseloads started
dropping and ultimately dropped enormously.

I strongly suspect that today the career welfare moms are gone leaving two main types; what we called the “easy cases” and the “hard cases.” The easy cases (which substantially outnumber the hard ones) are the ones where a working mother gets sick and can’t support her kids, or one where a housewife suddenly finds herself divorced with no support from the ex-husband, or where a woman has been driven into a shelter by a battering husband or boyfriend. Those cases are relatively easy because those women typically have the resources to support themselves and only need fairly short term assistance (at least during normal times – in today’s economy all such bets would be off). The hard cases (we figured about 20% of the peak AFDC caseloads in 1992) were those with multiple problems. Drug and alcohol addiction often went together with a history of abuse. Many of the women were mentally ill, and/or had serious chronic physical health problems or both. Some you felt sorry for, but in a number of cases you might justifiably feel that the woman’s poor choices had brought the problems on herself. But again, it’s the woman’s children who benefit most from the aid. Even in the worst cases of an undeserving mother, the children are more deserving than ever. Those who want to push the case of the undeserving welfare recipients never want to discuss the children.

In your post you list low IQ as one of the main causes of poverty and dependency. I believe you also accept the notion that differences in IQ are largely driven by genetic differences and that those differences are at least very hard to change if not immutable. How do you square that with the belief that American’s frustration with welfare recipients is justifiable? To the extent that poverty is caused by people not having the mental resourced to take care of themselves and their families what is the basis for blaming them for their bad choices and the subsequent results? (Of course cognitive ability plays only a very small roll in explaining the income distribution – only about 16% of the variance in a cross section regression and only about 25% if you bend every assumption in favor of there being measurement error when interpreting current income as a measure of permanent income.)

Those are my main objections to your stated objectives. Now for some point-by-point on your reply.

If you do across-country comparison, I suspect you’re right: Low-income Swedes do seem less self-destructive than low-income Americans.  If you do an over-time comparison within countries, however, the opposite seems to be true.  For starters, the American poor had very high rates of employment and high marital stability when the welfare state was smaller.

I disagree that the time-series correlation goes the other way. Look at European countries that introduced more generous social welfare benefits. Did they see an increase in working class pathology? I don’t think so. But take Steve Sailors example of Britain. One of the few other countries that has experienced the same gutting of good work for working class people. I suspect you will find (in both the US and Britain) that the timing of the deterioration of white working class circumstances is more consistent with the economic change explanation than the growth of welfare.

In any case, have you considered the possibility that countries with relatively prudent low-income populations can afford to have more generous social welfare systems?

The US white working class population was indistinguishable from any European population in the same class position (or if anything was more like the middle class than those populations in other countries). I doubt you could show this to be true.

It’s still easy for today’s white working class to get jobs that would support a family at the income level of the 1950s. U.S. per-capita income in 1950 was under $10,000 (1990 dollars).  That actually understates the ease of maintaining earlier living standards because modern household technology and family size make it a lot easier for mothers to work.

You can’t be serious. First, I don’t know where you got that, but here is my comparable calculation. In 1959 the average non-supervisory nonagricultural worker made $78.78 a week. Multiply by 52 weeks gives you $4,096. Inflation since then has been about 740% giving us an inflation adjusted annual income of more than $30k. Take the low skilled individual who drops from a job at a steel mill or an auto plant to working for $10 an hour and that guy now makes $20k. That person has to overcome two huge disadvantages. The inflation adjustment takes into account the fact that a lot of manufactured goods are much cheaper in terms of work hours now than they used to be as is food. However, the cost of housing has gone up so that someone making $20k a year isn’t going to be able to live in the same sort of neighborhood that a working class person in the 50s and 60s could afford. The other big item is that if that person wants his kids to have any chance to do better than him (something the people in the 50s and 60s could basically take for granted) he’ll have to send them to college at huge costs. Mothers can work for sure, but first before and after school child care (let alone full day care for children who aren’t in school yet) costs about $6k per year per child – there goes mom’s income. Also, working mothers end up having to spend a lot more on prepared foods and dinning out.

That said, I also happen to think that reducing the generosity of the welfare state and making assistance conditional on good behavior will (eventually) reduce bad behavior

That is the nature of TANF now. You have to seriously look for work or make progress in job training to receive the benefits. Ultimately, if your benefits run out, you’re out of the program. There is a reasonable argument that could be made for giving caseworkers more discretion in the granting of benefits. This seems to work well in some European countries and in some tenant run housing projects in the US. In fact, this is the way the system functioned prior to some court decisions in the late 1960s. The problem is that the courts found that the caseworkers were arbitrary and discriminated on the basis of race. There is also the problem of what do you do with the children when you cut off a mothers’ benefits.  I would be all in favor of moves back in this direction and would like to see more experimentation with more extensive caseworker involvement and discretion. Of course in the short run this will cost more money and for the states that run these programs such money is in very short supply right now.

If the American working class had responded to long-term economic changes by forming stable two-earner marriages, they’d still be doing fine.

I expect you will demonstrate this in the book? You will provide a budget for such a stable family that includes child care and will take into account the volatility of income for such people and the likelihood of extreme stress (what fraction of marriages break up over money problems)?

Not quite.  Some people make very very bad trade-offs; others make very very good trade-offs. 

You know better than this. Impatience is the norm. For most people impulse control is a major problem. It isn’t a matter of cool calculation. I have always been amazed by how you seem to have no sense of this. I suspect that there is something very very right with you which makes you a very productive scholar but has the side effect of making you unsympathetic with the vast majority of us who, in the words of the great comedic vocalist Allen Sherman, “Let the diet start tomorrow while today I drown my sorrow in a double chocolate milk.” You know Khanaman’s two systems theory. There is the rational calculating mind and there is the intuitive system. The vast majority of our decisions are made intuitively and that intuitive mind is liable to all sorts of well-known decision errors. You want to blame people for not making more use of their rational facilities, but you want to put them in situations where they find it harder to do so. We know from experimental evidence that the more people feel threatened/at risk the LESS likely they are to make reasoned decisions and the more likely they are to act emotionally. This gets us back to the discussion before of whether the notion that bad behavior causes bad outcomes is “neglected.” People know they make bad decisions. They often know when they are making them that they are bad. Telling them that they are being stupid isn’t news to them. Find ways to change the system to help them make better decisions and I’m all with you. Take money away from children because their mothers and fathers made bad choices I’m very disappointed. Overlook all the people who are receiving aid not because of bad choices, but bad luck and I’m more than disappointed – I’m angry.

Last point: I realize that my perspective on poverty may seem fundamentally immoral to you, Bill. 

I doubt this. Unless you really are going to start punishing children for their parents’ sins I think our views mainly diverge on the extent to which we are willing to assign blame for mistakes to the people who make them. I suspect this may be due, at least to some extent, to the degree to which we find controlling our own behavior difficult. It may also have to do with our personal experiences with people who have mental illnesses or are otherwise less than completely competent. I think our views also diverge on the extent to which those receiving government support are there because of their bad decisions as opposed to their bad luck.

Let me close by asking you a question I publicly asked Paul Krugman:

Why are you so forgiving of people with irresponsible lifestyles, but so outraged by people who don’t want to pay taxes to help people with irresponsible lifestyles? This seems morally perverse.  If you’re going to single anyone out for condemnation, it should be the person who behaves irresponsibly in the first place, not the complete stranger who asks, “How is this my fault?”

This is an entirely fair question and yet another reason why I doubt that I would ever conclude that your views are immoral. So let me answer your question. First, I’m not “outraged” by people who don’t want to pay taxes to support the government transfer system. A few of them may be selfish and/or racist jerks. There are few enough of them that I could care less. I believe that most people with that view are misinformed about who gets government transfers, how the programs are administered, the amount of the benefits, and how much of their taxes go to such programs. I think the vast majority of people, if they knew the facts, would not object to paying taxes for the system. They don’t know the facts in large part because cynical right-wing politicians and journalists miss-portray the system and its participants.

Now I suspect that you and I have a fundamental disagreement on the legitimacy of government coercion. I feel fairly certain that the state mainly makes us much better off than we would be without it and that justifies its existence and the reason why the overwhelming majority of people in democracies all over the world consent to be governed. I suspect that in your view the exact opposite is true – government makes things worse and we would be better off without it. To some extent this difference is driven by a difference in values – we both value individual liberty a lot, but I don’t weight it as strongly as you do.

So my bottom line is this:

  1. Overwhelmingly government transfer programs help children who are innocent of any mistakes made by their parents and overwhelmingly I think American’s would want to see them taken care of by the state, as would I. In the overwhelming majority of cases the best (and cheapest) way to accomplish this is to provide resources to their parents.
  2. Beyond the children, the vast majority of recipients of government support are in a position where they need it due to bad luck. No doubt you could come up with ways they could have avoided being in that situation (they could have gone to school longer, they could have saved more money, they could have looked more aggressively for a job when they lost their job), but my response would be that many many other people in similar circumstances did not suffer as they did and that a system where the government provides insurance against such risks is superior to one in which people are forced to provide for their own security no matter what the cost.
  3. Bad behavior certainly plays some roll, but its roll in putting people on government assistance is limited. To the extent it is a problem we should be constantly looking for ways to improve the system. To the extent the problem can’t be solved we should accept that the bucket sometimes leaks and that that is part of the price for providing for those in need.