Tweet [1]
Krugman has responded to my claims about empirical evidence, confirmation bias, and the lack of science in macro policy discussions. Here’s the argument so far.
I suggested on Twitter that economists see what they want to see. We tend to latch on to evidence that supports our worldview and ideology and ignore evidence to the contrary.
Krugman responded [2] by saying that might be true of me, but not him. He is a Keynesian because of the evidence. He is mainly unbiased. Nor does his desire for a larger government role in the economy pre-dispose him to interpret the effects of fiscal policy in a particular way.
I responded that the evidence in support of Keynesian fiscal policy is very mixed. There is too much evidence. People on each side of this debate can easily find lots of evidence–some historical, some econometric to support there view. As an example of the challenge of using evidence vs. bias to reach a policy conclusion, I pointed out that in 2013, Krugman predicted that the fiscal contraction from increased taxes and reduced spending would hurt the economy. When economic growth turned out to be healthy in 2013 and greater than in 2012, Krugman responded by arguing that “other stuff” explained the failure of his prediction. This is a common response by people on both side of these kinds of debates. Very rarely does one side concede that their worldview may need revision. There is always “other stuff.”
Simon Wren-Lewis responded [3] by arguing essentially that the evidence for the Keynesian story about fiscal policy is very strong and that those of us who are skeptical of the evidence have no evidence of our own. (Scott Sumner responded to Wren-Lewis and Krugman here [4] and here [5].)
Krugman then cited Wren-Lewis and argued [6] that the 2013 story really was consistent with Keynesian analysis after all. He had been wrong back in 2013. Not about fiscal policy. Like most of us, he never ever cites evidence that might shake his worldview. What was he wrong about? The amount of austerity in 2013. There was very little of it turns out. That was his error. His prior views on fiscal policy can now be safely maintained.
Krugman’s other point is that we who disagree with him have no evidence of our own. All we can do is play “gotcha.”
Here is my response.
When a theory or model can always invoke “other stuff,” it is more of a religion than a science. Ex ante predictions that can always be salvaged ex post are not really predictions. They’re just rhetoric.
I am less interested in gotcha and more interested in whether there is any evidence that might get any of us to change our fundamental worldviews about deep and important policy issues. I will have more to say about this another time, but my point here is that Krugman and others rarely if ever concede that their views have been shaken by empirical evidence. Evidence that might shake our worldview is easily dismissed as inconclusive, muddied by “other stuff,” or the result of a poorly executed study. Krugman is not unique. He is human just like me and therefore prone to self-deception
Both Krugman and Wren-Lewis argue that economists who are skeptical of their confidence in the power of debt-based government spending have no empirical arguments and must therefore resort to other strategies. So let me provide some evidence. I’ll ignore the 2013 story. There’s plenty elsewhere.
The two greatest changes in fiscal policy in the United States were the beginning of World War II and its end. The massive increase in government spending in the 1940’s is often cited as the most powerful evidence that Keynes was right–a natural experiment large enough to overwhelm the “other stuff.” The expansion of government during the war did result in higher GDP and lower unemployment. But as Robert Higgs [7] and Valerie Ramey have pointed out, it is easy to lower unemployment when you force millions of people into the army. The real test is whether the massive increase in debt-financed government spending stimulated the private sector through the Keynesian multiplier. It did not. The 1940s were not prosperous in America unless you were a weapons manufacturer. Nor did private employment expand. The end of World War II, a 60% decrease in government spending did not, despite the predictions of the Keynesians of the day [8], lead to a massive Depression. The economy in 1946, thrived.
You want econometric evidence? Look at Valerie Ramey’s work [9]. No evidence of the effectiveness of fiscal stimulus even at the zero bound. Look at Cogan and Taylor [10]. Look at Barro and Redlick [11]. Look at the evidence before the Great Recession, before people’s views were so politicized. Here is an overview of the empirical literature [12] published pre-crisis in the spring of 2008 that concluded that the empirical literature is inconclusive on the effects of government spending on consumption and wages:
The empirical literature has been unable to resolve this controversy. Depending on the nature of the identifying assumptions, the empirical literature finds disparate stylized facts regarding the responses of some variables to government spending shocks. The responses of consumption and wages, in particular, can take on different signs depending on the assumptions used to identify fiscal policy shocks.
In other words, each side can point to evidence that supports their position. The evidence is inconclusive. We should not be disdainful of the other side. We should be honest about our disagreements. We are prone to self-deception and cherry-picking. We are not simply scientists dispassionately surveying the data and coming to a conclusion. We are prone to exaggerate the empirical evidence that supports our views. Those views are a complicated mix of philosophy and ideology and past positions. They are not unaffected by evidence or data but the nature of evidence is so murky that it is never conclusive. We can always live to argue another day.
Krugman and Wren-Lewis are offended that I suspect their view of government stimulus is something other than a straightforward scientific question and that it instead reflects an underlying philosophical or ideological bent that all of us have. Matt Yglesias made the same point in 2013 after watching Krugman and Ramey discuss the efficacy of fiscal stimulus at the AEA meetings. Calling the discussion a “dialogue of the deaf” Yglesias concluded:
I think if we tried to jog researchers out of basic left-right political alignments we might get a little bit more clarity on what everyone thinks the real issues are. I get the overwhelming impression that with a few exceptions the issue is basically that right-of-center economists (like right-of-center people) generally think the existing level of government spending is too high and that additional government spending is likely to be wasteful—the equivalent of conscripting soldiers to fight in an unnecessary war—while left-of-center economists have the reverse view.
We are all human. As economists we need to be more humble about the certainty we profess about how the world works. I don’t pretend that the evidence for my view is ironclad. I know there are people on the other side who can argue against it.
I have invited both Krugman and Wren-Lewis to be guests on EconTalk to discuss these issues. I hope that one or both will agree.