Over at Marginal Revolution, Tyler Cowen cites an abstract of a recent paper by Derek A. Epp and Enrico Borghetto:

This article investigates the effects of economic inequality on legislative agendas. It considers two competing hypotheses: (1) that policymakers will act to counter rising inequality by renewing their focus on redistributive social policies, and (2) that rising inequality makes legislative agendas especially vulnerable to the influence of economic elites, and that these elites will attempt to keep redistributive social policies off the agenda. Empirical tests, which are designed to arbitrate between these hypotheses, use data on public laws and parliamentary bills introduced in the legislatures of nine European countries between 1941 and 2014. The evidence is supportive of the second hypothesis: as inequality becomes more acute, European legislative agendas become systematically less diverse and this narrowing of attention is driven by a migration away from social safety-net issues toward issues relating to law enforcement, immigration, and national defense.

Tyler introduces the abstract by saying that “most of you won’t like it.”

I’m one of those who doesn’t like it, but my not liking it has nothing do with the bottom line and everything to do with the sketchy methodology that got them there.

There are two problems. The first, and possibly smaller, problem is that they look at legislation introduced between 1941 and 2014. Was there anything else going on in Europe in 1941? Oh, yes: World War II. And Europe wasn’t just in World War II the way the United States was. It was being fought on much of their territory. The nine countries they study are: Belgium, Denmark, France, Hungary, Italy, the Netherlands, Portugal, Spain, and the United Kingdom. Let’s consider them in order.

Belgium: Although the government declared its plan to stay neutral, the German government took over and occupied Belgium.

Denmark: Denmark’s government was officially neutral but Germany’s government took over and the U.S. government occupied Greenland.

France: France lost to Germany early on and was split into occupied France and Vichy France.

Hungary: Hungary was allied with Germany during the war.

Italy: Italy was part of the Axis with Germany and Japan.

The Netherlands: The Netherlands was invaded by the German government.

Portugal: Portugal was neutral during the war but its policies were strongly influenced by Germany and Britain.

Spain: Spain was neutral, in varying degrees, during the war.

The United Kingdom: No need to report their role; it’s obvious.

I would expect that policies in these countries during the war were strongly influenced by the war. Granted that this accounts for only 5 years of their 54 years of data, but wouldn’t you expect them to run their data with and without those 5 years? But they don’t appear to. At least I couldn’t find it. You might think that they are counterexamples to Henderson’s First Law of Econometrics. But they’re not. Both authors are political scientists, not economists. [In my original law, I should have explicitly said, rather than left it implicit, that the econometric studies I’m referring to are done by economists.]

The second problem with the study is possibly more important than the first. The authors don’t give enough detail about the content of the legislation for the reader to judge whether their conclusion is correct.

Take immigration, the one that stood out to me. Does more focus on immigration indicate more influence by the relatively affluent? That’s what they’re saying. But then it’s crucial to know the content of the legislation, not just that the topic was immigration. If, for example, there was a push to slow immigration, that may well be due to concerns about income inequality because it could be due to the less-affluent citizens’ worries about immigrants competing for their jobs. If the legislation was about allowing more immigrants, that certainly could be due to owners of capital wanting more, and somewhat cheaper, labor to work with. Maybe that’s their story. But that simply points to the fact that knowing the content of the immigration legislation is crucial.

Even in this latter case of more immigrants, though, overall world income inequality is likely to be lower as people immigrate from l0wer-wage to higher-wage countries. So the authors’ concern, expressed throughout the paper [see especially page 6, where they mention “efforts to solve inequality”], that in societies in which the rich dominate, not enough attention is paid to income inequality, may be exactly wrong.