The Productivity Growth Paradox
Looking at the Great Depression and the Solow residual
So one of the beautiful things about economics to the new student is the apparent depth of inquiry in so many different areas--everywhere I go in this subject, it feels like someone has already broken a path. Occasionally that's frustrating to the wannabe trailblazer inside of us, or at least me, but such is the way of things. I am reminded, too, of Isaac Newton's famous words "If I have seen further it is by standing on ye sholders of Giants."
Anyway, I've been doing more thinking about what I called the "paradox of productivity growth," which hypothesizes that recessions can be caused by sudden coordinations of firms seeking to increase productivity, but as they all try to do this at the same time, labor income and aggregate demand both fall sharply enough for firms to have to cut back on output, which prevents the realization of the attempted productivity gains in the short run. In the long run, this sudden coordination of productivity gains can actually undermine productivity growth and potential output growth. This is because the long-term unemployed lose their skills, and so even if "employed productivity" is rising, "productivity at potential" can be falling, reducing the growth rate of the economy in the long term. (See my original post for a more detailed overview, and links to the empirical evidence behind my argument.)
All of this runs substantially counter to the main line of thinking that productivity gains, since they are almost everything in the long run--a point I do not contest--are always good, no matter when and how such gains occur. They are wrong, just as the classical economists were when Keynes argued that the sudden bursts of coordinated saving were dangerous.
In the meantime I came across a few items over the past couple of days which I realized appear to be further evidence for the contended paradox.
First is the research of Alexander Field on total factor productivity during the Great Depression. Field finds that innovations and investments made during the Depression--such as advances in chemistry and manufacturing, and transportation infrastructure--were responsible for the soaring productivity of the postwar era. Simultaneously, my follow-up research shows that productivity fell sharply at the start of the Depression, and that despite the labor-saving innovations, employers still had to cut back on the hours of their remaining employees due to falling demand. If we look at the Great Depression as a massive positive productivity shock--as both Arnold Kling and Joseph Stiglitz seem to (how often does that happen?)--then the empirical evidence adduced appears very strongly in favor of the "paradox of productivity growth" story.
Second, is the Solow residual, which is the fraction of productivity gains which cannot be explained by capital accumulation. Turns out it's highly procyclical. Now, I do not pretend to be knowledgeable on the larger model yet, but in exploring my theory I feel like I've been wandering around a dark house, and pushing past a curtain, have stumbled into a room full of people. Solow treated the residual as exogenous, and Romer as endogenous, and although Solow's treatment of it might have been convenient at the time, I think the paradox mechanism I've just written about makes it very clear that Romer's story is closer to the truth.
[By the way, my correlation above is robust to p=0.01.]