What I mean by the title, is that I have a menial task for some plucky Free Advice fan. In this post from May 3, Krugman was arguing that there was no correlation between the size of government pre-crisis and how it did during/after the crisis. He posted this chart:
In reference to the above, he wrote, “So what is the correlation between the size of government and recent economic performance? None at all, as far as I can see.”
Now when I eyeball that chart, it seems that there is definitely a negative correlation. That doesn’t mean causation, of course, but my point is that Krugman isn’t able to “see” anything there, when I think I can see a pattern.
What struck me as funny about this is that very often Krugman will put up scatter plots that bolster his case (in his mind), even though there are a lot of outliers in those ones too. Because I don’t subscribe to the NYT, and also because I was busy, I didn’t bother going back and looking them up, but I remember Daniel Kuehn called Krugman out on one of them within the last few months.
Anyway, on May 21 Krugman put up a scatter plot that he thought showed how obvious the case for stimulus was:
So here are my requests:
(1) Can someone grab the data Krugman used to generate the above charts, and then plot the OLS regression line and give us the stats (like the R^2 value I guess)? I’ll PayPal you $15 if you do it right and I agree you accurately reproduced Krugman’s charts. (It will have to be the first person who emails it to me. If you do the work and someone beats you to the punch, that’s not my fault. So maybe you should coordinate in the comments.)
(2) Now I confess, I think the second chart looks like a way better correlation too. I’m just saying, Krugman is denying that the first one has any pattern, when I think I see one. But, I wonder if Krugman has ever posted another scatter plot in the past, where he also drew the line in himself to help the readers “see” it, that has a worse fit than the first plot above. If you can show me that, I’ll PayPal you $40. (If it’s really close, like you find him touting a chart that has only a slightly better fit than the one he says has no pattern at all, then I’ll probably give it to you.)
I remain final arbiter in all disputes over payment. I am the decider. But please talk to each other; this happened before when I wanted students at a school to dig up Paul Samuelson’s referee reports on my articles, and two guys did it simultaneously and I had to tell one of them I already paid somebody else. So be careful out there, kids.