In this recent post, I made the following claim:
The evidence for the Keynesian worldview is very mixed. Most economists come down in favor or against it because of their prior ideological beliefs. Krugman is a Keynesian because he wants bigger government. I’m an anti-Keynesian because I want smaller government. Both of us can find evidence for our worldviews. Whose evidence is better? I’m not sure it’s a meaningful question. My empirical points about Keynesianism won’t convince Krugman. His points don’t convince me. I am not saying that we will never get any kind of decisive evidence on the question. I’m saying it sure isn’t here now.
I have evidently disillusioned Daniel Kuehn:
This is a statement about Russ that I wouldn’t have even made before I read him say it himself. He holds his views on Keynesianism to conform to his ideology. That’s a really disheartening thing to read, even though I’ve never been particularly in agreement with Russ in the past. I would have figured he at least had other objectives. Scientific conclusions based on adherence to an ideology are worthless. This is how we get the persistence of ideas like creationism and geocentrism. That isn’t to say that in ruder stages of society creationism and geocentrism weren’t decent explanations – they were decent at one time. But when the evidence starts to stack against them, adherence to ideology is what impedes scientific advances. That Russ actually embraces this is dumbfounding to me.
Needless to say, I see no evidence at all that that’s why Krugman views Keynesianism favorably. Russ doesn’t appear interested in offering any reason for thinking that’s Krugman’s motivation. It doesn’t really make any sense. There are Keynesians who favor and oppose larger government (just as there was at one time a fairly active community of Austrian socialists). Your view on how the economy works doesn’t require a certain political philosophy.
I recently interviewed Valerie Ramey on the multiplier. In her work, the multiplier ranges from .8 to 1.2. A multiplier of 1 means there is no stimulus from the spending–GDP rises by the amount of the increase in G but by no more. Private spending doesn’t grow. When you include the taxes (either today or in the future to pay back debt from the increase in spending) that finance the government spending, it’s particularly costly. Ramey uses a very clever idea to generate her estimates, but she is respectful of what other people find and in her survey of the literature and in our interview she says that the multiplier may be as small as .5 or as large as 2.
That’s a four-fold range. That in itself is discouraging. (Ramey and I had a very interesting conversation about the implications of that range. You can listen to that podcast in about ten days.) Much more discouraging is the fact that most if not all of the people who think the multiplier is large are fans of larger government and most if not all of the people who think the multiplier is small are fans of smaller government.
What is one to make of this alignment of ideology and belief? Is it a coincidence? Or perhaps causation runs the other way. It is possible. It is also possible that the small estimates of the multiplier are the right ones. Or the large ones are. And the other side, whichever side that is, does econometrics poorly. But to me it suggests scientism rather than science. It suggests that we are unable to measure the impact of the government on the economy with any precision. This is a particularly persuasive idea when you consider that few (any?) proponents of one view or the other change their mind when confronting the findings of the other side. And each side would certainly concede that the other side’s proponents are exceptionally bright, well-trained economists.
My view is that we cannot accurately measure the effect of government spending on the multiplier. To think otherwise is the pretence of knowledge. I don’t view my view as anti-scientific but rather a view that recognizes the limits of knowledge and the tools we use to measure the impact of government on the economy. It is not scientific to use science for tasks it cannot achieve. That is scientism. Very dangerous.
Daniel is right that I provided no evidence for my claim about Krugman, that his views on stimulus are driven by ideology as I know that mine are. The evidence is implicit in the post but I should have made it clearer given the boldness of the claim. Daniel “sees no evidence” of the claim in his own reading of Krugman. Here is my evidence. I will be interested to see if Daniel finds it persuasive.
(And one more point before proceeding. I do not believe that ideologies are evidence-free. I hold my ideology for a wide range of reasons many of which are based on what I observe about the world and human behavior along with a set of beliefs about how the world would work if my ideology were more prominent in policy decisions. But I don’t pretend I’m against government spending because the multiplier is small. Now on to Krugman.)
1. Krugman demonizes those who oppose more government spending. He rarely or ever grants the possibility that they might be right and that he might be wrong. This is not the way a scientist thinks. It is the way an ideologue thinks.
2. Krugman cherry-picks data and stories that confirm his worldview. He doesn’t just dismiss data that challenges his worldview. He usually ignores it. He does not write about the Japanese malaise that persists after trillions of increased spending. He does not write about the growth in the US economy when World War II ended. He rarely if ever writes about the work of Higgs or Ramey or Barro who find that wartime spending during WW II hurt the US economy. If he does, he writes about their work dismissively. He does not concede the possibility that the failure of the 2009 stimulus might challenge his views. (I recognize the possibility that it may have failed because it was badly designed or because the problem is worse than we thought and it was too small. Krugman never to my knowledge writes that he might have to reconsider his views based on the evidence.)
I did find a post where Krugman conceded that John Taylor “actually has a pretty good point.” That point is that there was actually very little stimulus in the 2009 stimulus package. Krugman calls that point “pretty good” either because he really thinks it is or because it confirms his own world view. The evidence suggests the latter. In other posts on Taylor he refers to a “zombie claim” of Taylor’s or says Taylor has lost his mind or that his mind is corroded, that it’s a “good bet he [Taylor] doesn’t believe what he’s saying.” I mention Taylor because I know him a little bit. He’s a fine person and thoughtful and I’ve never seen him write or say anything like this about his intellectual opponents.
Krugman has written once or twice about the austerity of the World War II economy from a consumer’s viewpoint. He as explained that as being caused by rationing. He does not explain why rationing was necessary.
Certainly, Krugman has a story to tell that can explain the failure of the 2009 stimulus package, the on-going malaise of Japan, the World War II and post-World War II experiences. I’m sure he has a story about why stagflation was consistent with the Keynesian model of today (it’s been improved!). His stories would have facts. His stories could be right. What I find interesting is that I do not remember a time when he written for the New York Times where he has conceded uncertainty, or the possible virtue of the ideas of his intellectual opponents unless they comported with his own views. By the way, in his other writing, Krugman writes like a scientist. In his book, The Return of Depression Economics, he often says some issue is unsettled, we don’t know the full story and so on.
Of course I’m not a big fan of Krugman’s work in the Times. Maybe I’ve cherry-picked examples and failed to notice the times he was gracious or thoughtful about people who disagreed with him or more importantly where he conceded that his own views might be wrong or that further study was needed before reaching a definitive conclusion. Happy to learn about those writings from Daniel or others. But on the surface, he does not write like a scientist. He writes like an ideologue. That’s OK. He is an ideologue. Me, too. Nothing wrong with that. What is wrong is to be an ideologue while pretending that your ideology doesn’t affect your views on economics. I think it does. And pretending or claiming that economics has a great deal of certainty when you know that it doesn’t also doesn’t strike me as something a scientist should do. It’s what ideologues do.
Daniel is right that using ideology to judge science is bad science. I think the recent debates over the effects of the stimulus package are indeed bad science.
UPDATE: Matt Yglesias isn’t happy with me either.
UPDATE: Krugman responds to my first critique here.
