Evan Soltas
Jun 27, 2012

Uncertain on Policy Uncertainty

Does policy uncertainty meaningfully reduce economic growth?

Paul Krugman, Brad deLong and Dean Baker, three left-leaning economists, have argued that it doesn't and that policy uncertainty isn't a compelling explanation of the anemic recovery in the United States. Krugman has argued again and again -- here's one example -- that policy uncertainty is basically irrelevant. DeLong put a big, fat "no" to answer this question on his blog. Baker referred to a "regulation monster," a fictional creature akin to the "confidence fairy" speciously invoked by the political and economic Right.

On the Right, economists such as Edward Lazear, Robert Barro, and John Taylor have argued that policy uncertainty really does matter, both in general and as an explanation of current economic conditions. Lazear, in a recent op-ed in the Wall Street Journal, writes that "[i]t would be difficult to argue that government polices over the past three years have enhanced confidence in the U.S. business environment. Threats of higher taxes, the constantly increasing regulatory burden...and the enormous increase in government spending all are growth impediments." Barro blames the lack of "clarity about future policies." And Taylor? "In my view, unpredictable economic policy...is the main cause of persistent high unemployment and our feeble recovery from the recession."

There is also a recent study by three economists which creates a "policy uncertainty index" which is then used to explain changes in investment, industrial production, and unemployment during the most recent recession. The authors of this study even wrote a recent piece for VoxEU arguing that the rise-and-fall behavior of the recovery corresponded with movement in their measure of policy uncertainty.

Now it's my turn: I have found everyone's arguments unconvincing, as well as correct and incorrect in varying respects.

On the Left, the rejection of policy uncertainty as a meaningful explanation seems to me partisan and made out-of-hand without any serious consideration of the evidence -- or, given the shortage of formal evidence, an examination of what they see as the strongest arguments against their position.

For example: I think all three economists would agree that future government policy can be more or less certain; that is, there exists a "policy uncertainty" variable. And I think all three economists would agree that marginal increases in this variable would, in almost any economic model, correspond to marginal decreases in investment and other real variables. (You can insert any "story" of causes here: loss-averse behavioral responses by firms to policy uncertainty strikes me as particularly compelling.) And I think all three economists would agree that there have been marginal increases in policy uncertainty -- more tax provisions expiring, the debt ceiling debacle, etc. Therefore it follows that all three economists must think that policy uncertainty can reduce economic growth and is, on the margin, doing so today.

I agree with them, of course, that a lack of nominal spending or aggregate demand remains the largest factor in explaining the recession and the weakness of the recovery. But I don't think that their mockery of policy uncertainty -- or, as I have argued in the past, confidence -- reflects well on them as thinkers.

For that matter, I think that the Right is too quick to embrace the narrative which blames policy uncertainty. There is almost no formal evidence behind the contention -- it isn't cited, and it doesn't exist.

And the one major study which merely begins to make the academic argument is not that strong. The authors argue that the measured increase in policy uncertainty causes economic contraction. But the authors' methodology is not particularly rigorous. They cannot eliminate the most obvious alternative story of causation: (1) there is an economic shock, (2) the shock induces or even necessitates a policy response, (3) that policy response is discussed beforehand, during, and after in the media, which shows up as an increase in measured uncertainty.

The study does not show that policy uncertainty causes economic contraction. Instead, the study shows merely that policy uncertainty and economic contraction happen at the same time. There is a relationship, though the direction of causation is unclear. It could run in the direction they hypothesize, the opposite direction, or even both. The authors, to be fair, seem to concede as much at one point in their study -- though not, pointedly, in any of their media discussions of the study. There is also, I imagine, a strong relationship between uncertain policy and bad policy -- that is, policy changes which receive a lot of discussion in media also tend to be policies with potentially large impacts on economic growth.

I think their method of research -- aggregate counts of news stories -- may be a dead end. To better examine the effects of policy uncertainty, I would try a method used in work which is similar but on different topics: looking at lagged changes in employment, output, and other variables among counties which straddle state borders in the United States, or regions within member nations of the EU, when policies are under uncertainty in one state but not the other.