Awarding winning journalism, terrible social science

I avoid using this blog for commentary on ephemeral current events, but sometimes they lend themselves well to big picture issues. Case in point: Pulitzer winner dug beyond politics to explore impact of food stamps on American families. Reading this interview, we can look at some important themes.

1. The interview is bad, really bad, awful social science. This is not surprising given that Saslow is a reporter and not a social scientist, but the accolades given to his work make this a relevant fact, and high-level theory does have real effects. Specifically, Saslow appears to take as exogenous things that are actually endogenous. It’s true that many people rely on food stamps* in the US. However, this behavior is not predetermined. If no government food assistance programs had ever been created, nobody would rely on them. Yes, some people would be worse off, but the knowledge that soup kitchens, churches, and food banks were the only institutional help available would be a motivating factor in peoples’ work, spending, and saving decisions. [In the long run, it would also likely lead people to create still other means of assistance.] By changing the set of options available to people, you can change their behavior. That people behave in some ways with a given opportunity set does not mean they will behave in the same ways with a different opportunity set.

One of the classic examples in the economics of risk that we can use to illustrate this point is the idea that seat belts allow people to drive more recklessly than they would if every car had a sharp spike pointing out of the steering wheel. In the latter scenario, driving behavior would be very different than it is today.

This does not necessarily entail cutting off all assistance programs. It is not a political point. It does mean treating recipients like us—like rational human beings—who can and do adjust behaviors in different opportunity sets.**

2. Seen and unseen! It’s true that food stamps are a major revenue source for some locations and businesses. That is seen. It’s also true that the same revenues are by definition not spent in other ways. That is unseen. It isn’t as if the government simply creates wealth to give out; what it has was taken from others who no longer have it. This point does not rule out transfer payments entirely, but it does mean that it’s bad social science to focus only on the places/firms that benefit now and ignore the places/firms that are worse off now relative to alternative arrangements.

3. Coming closer to the contingent world, it’s lamentable how quickly new political conditions are seen as timeless. As Friedman put it, “Nothing is so permanent as a temporary government program.” Though Saslow admits “it’s a program that’s grown four times in size in the last 10 years,” there’s no reflection on the fact that 11 years ago there wasn’t the social problem of a massive die-off by people who would qualify for assistance under the new guidelines.

I can already hear the objection: but there was a huge recession within the last ten years and many people were worse off than they had been previously! Setting aside peoples’ endogenous reaction to their worsening conditions, this still demonstrates my point. What is the expiration date for the expansions? If they were made necessary by a lot of genuine hardship in the Great Recession, at what point will they be made unnecessary by the return of good times? (I expect few people who supported these expansions when they were made to argue for contractions in the next few years.)

The final point is how readily people will read this as a political post with an axe to grind rather than a plea for better social science thinking. Value-free social science and political partisanship are not the same, and you do yourself and the intellectual climate you operate in a great disservice to confound the two.


* I know there are more official terms for this, but this is the term the interview uses and the term is more widely used and understood than a more neutral term like SNAP.
** There is a very small percentage of people with mental illnesses for whom a separate set of policies would be more appropriate, though even most of them are not completely random thinkers who can’t adjust their behavior at all.

Quotation of the Day: the Hayekian worldview

“If we are to understand how society works, we must attempt to define the general nature and range of our ignorance concerning it. … The misleading effect of the usual approach stands out clearly if we examine the significance of the assertion that man has created his civilization and that he therefore can also change its institutions as he pleases. … In a sense it is true, of course, that man has made his civilization. It is the product of his actions or, rather, of the action of a few hundred generations. This does not mean, however, that civilization is the product of human design, or even that man knows what its functioning or continued existence depends upon. … If we are to advance, we must leave room for a continuous revision of our present conceptions and ideals which will be necessitated by further experience. We are as little able to conceive what civilizations will be, or can be, five hundred or even fifty years hence as our medieval forefathers or even our grandparents were able to foresee our manner of life today.” (F. A. Hayek, The Constitution of Liberty [The Definitive Edition edited by Ronald Hamowy], Chicago, IL, University of Chicago Press, 2011, p. 74)