Over at Vox, economist Andrew Rose argues that "maybe protectionism isn't countercyclical after all," that is, protectionism does not correlate with economic downturns. He looks at a number of factors, but two in particular stood out to me. One factor is whether economic growth (or lack thereof) correlates with the number of WTO disputes. This seems like a strange approach to me. There are around 20 or so WTO disputes initiated each year, but there must be thousands of instances of protectionism. I wouldn't think that such a small sample size of WTO disputes could tell us a lot about how much protectionism is taking place at any given time. He also looks at the number of antidumping cases initiated. This one seems more promising, but to make it more useful I might look at the number of trade remedy cases in which duties or other measures were actually imposed.
Coincidentally, I think, also at Vox, economists Chad Bown and Meredith Crowley have a piece that looks at how US and EU trade policy reacted to the great recession. If I understand it correctly, they look at the full scope of trade remedies, and find that temporary trade measures did rise, although not by as much as was expected. While this wasn't necessarily the goal of their piece, their findings may undermine Rose's argument.
Not being an economist myself, I have no great insights here. Any other thoughts from readers?