Scott Sumner employs a little bit of Wittgensteinian logic to show us why the monetarist (Friedman’s and & Schwartz’, that is) explanation of the Great Depression is more accurate than the Keynesian one (Sumner’s summation: “inherent instability of capitalism”). Writes Sumner,
The Great Depression was originally thought to be due to the inherent instability of capitalism. Later Friedman and Schwartz blamed it on a big drop in M2. Their view is now more popular, because it has more appealing policy implications. It’s a lot easier to prevent M2 from falling, than to repair the inherent instability of capitalism. Where there are simple policy implications, a failure to do those policies eventually becomes seen as the “cause” of the problem, even if at a deeper philosophical level “cause” is one of those slippery terms that can never be pinned down.
I think that to some degree Sumner probably chose his wording wrong, so maybe his post here does not warrant as much criticism as I am about to give it. What he seems to be saying here is that the principle reason Friedman’s and Schwartz’ explanation of the Great Depression is accepted is because its policy implications are more appealing and easier to implement. This seems to me as the wrong way to go about explaining history. It is akin to saying, “This theory must be right, because it is the only one that justifies the easiest form of interventionism.”
Of course, this is not the methodology used by Friedman and Schwartz in A Monetary History of the United States. Friedman and Schwartz conducted an empirical study and recognized a dramatic drop in the supply of money between 1929 and 1933. Given the correlation between the worsening of the Great Depression and changes in the money supply, they concluded that it must have been this monetary deflation which made the 1929 stock market crisis into a great depression. While I do not agree with Friedman’s and Schwartz’ empirical approach — I think it leaves a lot of the causes of the Great Depression out —, Sumner’s explanation for the popularity of the theory strikes me as ridiculous.
If a theory does not properly explain an event, then how can we be assure that the policy implications of the theory still apply? This goes well beyond acknowledging the fall in the quantity of money. For instance, if the Austrian explanation of the Great Depression is the correct one (humor me), then we would realize that a massive indiscriminate increase in the supply of money may not be the correct one. This is because we are not sure that increases in the supply of money would help with the post-boom structural readjustment. No less, it is impossible to tell if it was the failure of the Federal Reserve to maintain the supply of money that made the depression “great.” In fact, there are reasons (reasons that Sumner may not agree with) to believe the opposite.
For what it is worth, the policy implications of Keynesian theory are not as complicated as Sumner leads his readers to believe. Keynes called both for monetary stimulus and fiscal stimulus; Keynes was simply not privy to monetary tools available when an economy hit a liquidity trap. Keynes did not necessarily support the regulative legislature passed into law during the 1930s (for example, Keynes did not support the National Industrial Recovery Act [NRA]), meaning that his purpose was not to justify a massive encroachment of government into the capitalist system. While to some degree I think it is difficult to predict what exactly Keynes would have supported, I think it is safe to assume that to some degree he would have opposed a large regulatory web. His only purpose, in The General Theory, was to explain why there existed such a large gap in industrial production (disconnect between savings and investment) and how the government could fix the issue (socialization of investment and monetary stimulus).
Keynes’ explanation of the crisis, I think, is much deeper and more comprehensive than Friedman’s and Schwartz’. It seems to me that Keynes provided us with a causal explanation of why an increase in savings failed to materialize into an equal increase in the rate of investment, while the latter did nothing of the sort. They noted an empirical truth and prized it with causality (this is sort of simplistic, given that Friedman and Schwartz borrowed from economists such as Hawtrey and Warburton). In any case, if the policy implications of Keynesian economics are more complicated than those of monetarist economics, it seems to me that there is more reason to believe that Keynesian policies will be more effective than just monetarist policies.
My point is that Sumner’s justification for his policies must be more nuanced than “they are easiest,” because the ease of implementation says absolutely nothing with regards to their accuracy in pinpointing the problems at hand. I do not think that the theoretical debate ever finished. So rather than starting from the prior that some interventionist policy must be the right one, so better use the ones which are easiest to implement, I think it would be more fruitful to return to the theoretical debate. That way we can judge the merits of a theory before it is applied to any incomplete set of data.
Late edit: I just read Cafe Hayek’s 15 December “Quotation of the Day.” Don Boudreaux quotes from William H. Hutt’s The Keynesian Episode: A Reassessment. Here is the most relevant part,
Already in 1936, although I had been bewildered by it, I had seen clearly and predicted that [Keynes's] The General Theory would have a quite unparalleled influence by reason of what I judged to be its demerits as a contribution to thought. Its policy implications appeared to have been chosen for their political attractiveness.
Hutt, as I interpret it, is claiming that how the economics profession chose its preferred framework is by looking at the policy implications first, then at the theory. I cannot comment on the accuracy of Hutt’s accusation — it is one I have tried to shy away from —, but after reading Sumner’s post I can see more merit in it.