The Federal Reserve is at it again, publishing bad climate research.  Last year in Econ Journal Watch, a peer reviewed academic publication, I criticized a Fed working paper about climate. The author, Fed economist Michael Kiley, never responded, though he was promised space in the journal for a reply. He published his paper in Economic Inquiry. Kiley made a few changes, but did not address my criticisms. I have written a new critique of Kiley’s updated work.

Nobel Prize-winning economist William Nordhaus has made it clear that, compared to future expected economic growth, any effect of warming temperatures would be small. Hence, climate fearmongers try hard to show that warming would reduce the rate of growth of GDP.  Kiley’s paper claims that an extra degree of temperature will lower median world GDP growth by 84 percent through the end of the century, with an even larger reduction of growth during bad times.

Kiley’s analysis gives equal weight to 124 countries, including Rwanda, where a genocide in 1994 caused GDP to fall by 64 percent in one year. Blaming the genocide on a slightly higher than average temperature in that year is ridiculous, particularly since the year’s warm weather happened after the genocide took place. Kiley’s data also includes Equatorial Guinea, where, following an oil discovery, GDP rose by 88 percent in one year!

Other statistical properties of the data are classic causes of spurious results, as described by Nobel winning economist Clive Granger. All economic data are imperfect, but the problems with Kiley’s data are extreme.

I find that eliminating countries that experienced the worst wars, coups and genocides during the sample period also eliminates Kiley’s main result.

Weighting observations by the inverse of growth volatility, a common practice in econometrics, reversed the sign of the estimated effect, meaning the data tell us that warming will increase economic growth. I also find that a number of alternative model specifications cause the statistical significance of Kiley’s result to disappear, and in some cases to reverse sign.

Examined individually, not a single country illustrates the pattern that Kiley is trying to prove, suggesting that variation over time does not support his hypothesis. The fact that cross sectional analysis suggests that warm countries tend to be poorer has been noticed for centuries, and so Kiley is not adding anything new.  Economists have not reached consensus on why poor countries tend to be warm, but examples like Singapore and Florida suggest that the relationship is not causal.

I also constructed simple simulations of countries over time to show how common time trends in growth and temperature combined with some unusual observations can trick Kiley’s statistical methods into finding an effect of temperature on growth.

Much of the research published in economics journals has little relevance to consequential policy decisions.  But a finding that CO₂ emissions might cause an 84 percent reduction in growth has huge implications.  It is absolutely necessary to get research like this right, particularly when it comes from the Federal Reserve.

Fed economists always include a disclaimer that their research does not represent the views of the Federal Reserve. But the paper was first published as a Federal Reserve working paper on the Fed website, and Kiley is a Fed employee. The Fed is not a public university, required by the First Amendment to provide professors with academic freedom. The Fed’s legal mandate is to maximize employment and price stability. Why is it pouring money into climate research? And, given the low quality of its climate research, how can we trust its judgment about the soundness of its research on banking and monetary policy?