Drum on Sanders on Welfare Reform

Kevin Drum wrote “Here’s Why Bernie Sanders Doesn’t Say Much About Welfare Reform”
He presents a graph of total social welfare spending in the USA divided by the number of people in households under 150% of the poverty line.

He concludes

“There are two obvious takeaways from this. First, overall spending on social welfare programs has increased by 3x since 1980. That’s pretty substantial. Second, if the 1996 welfare reform act had any effect on this steady rise in spending, you’d need a chart the size of my house to make it out. Perhaps Bernie Sanders knows this, and understands that in the great scheme of things, welfare reform just isn’t worth fighting over anymore.”

I am trying to control my outrage which is all the more bitter, because I have such a high opinion of Kevin Drum. Briefly, I think his post has no merit whatsoever and is not defensible.

The title discusses the presidential campaign. On this point, I agree with commenter CAinDC that Sanders doesn’t discuss welfare reform, because he knows his position is extremely unpopular. It is a matter of politics (appropriate to a campaign) not policy. Politicians who hate and detest the 1996 welfare reform law don’t (and shouldn’t) mention it, because they know that, if forced to choose, many voters will choose welfare reform over them. I entirely agree with Sanders’s and Clinton’s decision to discuss the issue as little as possible. I think it is best to try to undo the damage of welfare reform by stealth, by proposing a new program and calling it anything but AFDC or welfare *and* by making sure that, even if it is less efficient policy, it doesn’t look at all like AFDC.

In fact, Kevin Drum knows this perfectly well. In 2014 he wrote

why does the WWC [white working class] continue to loathe Democrats so badly? I think the answer is as old as the discussion itself: They hate welfare. There was a hope among some Democrats that Bill Clinton’s 1996 welfare reform would remove this millstone from around Democrats’ necks,

With, at most, a moment of thought, he must understand why Sanders doesn’t talk about welfare reform — he answered the question in 2014.

But the post isn’t just a answer to the question about Sanders’s political strategy. It includes a discussion of welfare reform, and the assertion that welfare reform didn’t have a big impact on average US economic well being (confusingly the technical term used by economists for this is “welfare” which I will reserve for AFDC/TANF). I disagree with this assertion. Oddly I first saw the very very convincing evidence against the assertion in the current Kevin Drum post in an earlier Kevin Drum post which noted the dramatic increase in severe poverty (income below half the poverty line)*.

The post as written just will not do. Drum looks at a ratio of total inflation adjusted dollars to a number of people when discussing social welfare spending. If one thinks that the effect on well being of a dollar is about the same whether it goes to one person or another — that the total not the distribution matters — then one can conclude that welfare reform wasn’t very bad. One must also conclude that optimal social welfare spending is zero. The argument that AFDC disbursed a small number of dollars so it didn’t have a big effect on average economic well being can only be justified by the assumption that a dollar to the extremely poor is just like a dollar to someone at 1.5 times the poverty line. If this assumption, which Drum must make to argue that his graph is relevant, were valid, then there would be no justification for AFDC, TANF, the EITC, SNAP, or Social security. It is not OK to ignore inequality when discussing social welfare policy. This is, in fact, exactly what Drum did. Partly because of my great respect for Drum, I am putting this as politely as I possibly can after prolonged efforts to calm myself.

The point is that there has been an increase in inequality among people who receive means tested benefits. If one cares about the distribution of income at all, one should not ignore this. Consider a family with no cash income. They are not helped at all by the EITC or the minimum wage. They will get food stamps. They have a small chance of getting a rent voucher (one in 10 poor people live in a family which gets one). At least the children will get medicaid (not the parents in some states).

People can’t live on food and medical care alone. The increase in the cost of medical care (which includes both inflation and genuine increase in care provided as new treatments are discovered) can’t substitute for clothing and shelter.

Now one might ask if we can tell, for sure, whether welfare reform had large effects. The answer is yes, because there was a genuine experiment in Florida. It is now known, as well as anything can be known in the social sciences — that welfare reform killed people.


I stress that it is hard to generalize from the case of Florida, because the Florida reform included an extraordinarily high level of support for former AFDC recipients. It is almost certain that simple extrapolation of the solid experimental results would lead to an underestimate of the number of people who have been killed by welfare reform. It is very important that the deadly effects of welfare reform were reported by supporters of welfare reform who were surprised by the results — this makes the paper much more convincing.

Now one might argue that those deaths are no big deal. I suspect that, nationwide since 1996, they are not many orders of magnitude greater than 3000. If you think 9/11 was a minor matter, then you might consider welfare reform to be a minor matter too.

In any case, even if you decide you don’t believe in experimental results, you should understand that Drum’s post is indefensible. He bases his discussion of social welfare policy on the assumption that the distribution of social welfare spending doesn’t matter. If his graph is relevant, we shouldn’t mind if all social welfare spending were converted to a block grant to Bill Gates. The post is one huge category error and unworthy of Kevin Drum. Hell it would even be unworthy of me.

*google is giving me trouble — there is too much evidence that the well being costs of welfare reform are large. OK I give up. I can’t find the original Drum post. I don’t know if it was here or at washington monthly or at calpundit. The best I can do is
this follow up to Drum’s post


Here are some relevant links I found when googling.

CHART: Welfare Reform Is Leaving More In Deep Poverty

New Research Documents Growth of Extreme Poverty in US


This pdf is very thorough