Correlation, causation, and big changes
Scott Alexander recently argued that building more housing in a given city causes housing prices to go up in that city. He acknowledged that previous studies had found the opposite relationship, but suggested that he was more impressed by the strong positive correlation between population and prices when there are big changes:
Matt Yglesias tries to debunk the claim that building more houses raises local house prices. He presents several studies showing that, at least on the marginal street-by-street level, this isn’t true.
I’m nervous disagreeing with him, and his studies seem good. But I find looking for tiny effects on the margin less convincing than looking for gigantic effects at the tails. When you do that, he has to be wrong, right?
I also believe that looking at the tails (big changes) is often more revealing than looking at lots of small changes. But only if you’ve got causality right. And in this case, Alexander hasn’t necessarily done so.
Here’s an analogy: Suppose you wanted to compare the mainstream view of monetary policy (raising interest rates is deflationary) with the NeoFisherian view (raising interest rates is inflationary.) So you focused your attention on cases where there were truly massive increases in interest rates—say to 20%, 30% or 50%. In virtually all of those cases, inflation will be very high when nominal interest rates are very high. That seems to support the NeoFisherian position. (I wonder how Alexander feels about this debate.)
But this sort of correlation doesn’t address the issue of causation, and thus most economists reject the notion that a high interest rate policy is inflationary, despite the clear correlation. They see this as an example of “reasoning from a price change.”
In my research on market reactions to policy news during the 1930s, I argued that big changes were especially revealing. But in those sorts of “event studies” the direction of causation is clear—policy news leads to immediate changes in asset prices.
Big changes don’t help if the house price model you are criticizing is also consistent with the stylized fact that bigger cities tend to be more expensive. That stylized fact would be true even if building more houses reduced house prices at the margin.