David R. Henderson  

Does Drawdown of Savings Explain the Postwar Miracle?

The Deal Delusion... Economists' Self-Conception...

Yesterday, co-blogger Arnold Kling referenced my work on the U.S. post-World War II austerity. I had pointed out that Keynesian wunderkind Paul Samuelson had blown it with his prediction of a postwar slump.

In the comments, wd40 writes:

During WW2, there was forced saving (war bonds) and investment went into the war effort. Hence, the robust economy of 1946 when consumers could draw drawn their savings and consume items that were not not available during the war. Naive regressions did not sufficiently account for this pent up demand.

This has become the standard response and, in fact, you can find it in textbooks, to the extent textbooks talk about this event.

The problem is that the part about the drawdown of savings is wrong. Here's what I wrote in my Mercatus study, "The U.S. Postwar Miracle."

Keynesian economists also explained why their glum postwar predictions hadn't come true by arguing that people drew down their savings to finance their "pent-up demand" for the various goods they could not have during the war: cars, tires, refrigerators, stoves, and so on. In 1943, Paul Samuelson, in the article quoted at the beginning of this paper, laid out the idea that pent-up demand for consumer goods would cushion the blow of demobilization. Cited in almost every textbook on U.S. economic history, this explanation has become the orthodox one. There's a problem with this explanation, though: it doesn't fit the evidence.

There are two parts of this explanation. The first, which is plausible, is that there was pent-up demand due to the heavy rationing that the government imposed during the war. People were ready to buy cars, for example, after having not been able to do so for over three years. But Samuelson pointed out that this would be a short-term cushion at best. Of course, one could argue that the two years from 1945 to 1947 were short term. But then, after this pent-up demand was satisfied, there should have been a major drop in economic activity and a major increase in unemployment in the medium term. That didn't happen. The unemployment rate was 3.8 percent in 1948 and kicked up to only 5.9 percent in 1949.

The second part of the explanation is that people drew down their savings that they had accumulated during the war. But the term "savings" is what economists call a stock, whereas "saving" is a flow. If I draw down my savings this year, not only do I not save anything this year, but I also spend some of my stock of savings. So, if people were
drawing down their savings, they would have a negative rate of saving. They didn't. While the personal saving rate did fall substantially from a wartime peak of 25.5 percent in 1944 to 9.5 percent in 1946 and 4.3 percent in 1947, it remained positive.

Comments and Sharing

CATEGORIES: Fiscal Policy

COMMENTS (6 to date)
Ken B writes:

David: I am not familiar with the numbers but while private citizens were saving the govt was spending, and spending on stuff to destroy. Now I think those were wise choices considering the options (you demur I believe) but can you say anything about the net overall investment in productive facility in the economy in 39-45 vs 32-38 say? I find it inherently implausible that blowing stuff up makes us richer, but the increased application of man hours during the war could. And how would a Keynesian view that?

david writes:

The signs are consistent with the Keynesian explanation - reduced saving over dissaving, say - this seems like an argument about the magnitudes.

Mark Little writes:


Thank you for your work on this important issue.

What about the history of the post-WWI period? As I understand it, the Harding administration responded to the first post war recession with the opposite of the Keynesian prescription. Reduced government spending, balanced budgets, and a cautious monetary policy was followed by the "Roaring Twenties", with rapid growth, low unemployment, high rates of innovation, combined with a slightly deflationary trend in prices.

Unless my understanding of that history is faulty, the post-WWII Keynesians were not just wrong; they were ignoring what was then very recent history contradicting their predictions.

Are you aware of any studies of the post-WWI episode comparable to your work on the post-WWII history?

wd40 writes:

David Henderson writes: “… if people were 
drawing down their savings, they would have a negative rate of saving. They didn't. While the personal saving rate did fall substantially from a wartime peak of 25.5 percent in 1944 to 9.5 percent in 1946 and 4.3 percent in 1947, it remained positive.”

He is correct that people did not draw down their savings. What people did is increase their marginal propensity to consume because of their high (but forced) savings during WW2. Given the basic Keynesian model (C = A + BY; Y = C + I) with unemployment, an increase in B will lower the savings rate, but not savings for a fixed I. And if investment increases, savings will increase along with income (holding G constant). So post WW2 events did not contradict the basic Keynesian model

David R. Henderson writes:

@Mark Little,
Thanks. No, I don’t know of a study that looks at the post-WWI data carefully. A big part of the reason, I think, is that we don’t have as good data from before the Great Depression.
I’m glad we’re agreed that people didn’t draw down their savings. That was the point of my post.

jc writes:

@Mark Little

It's not a study. Just some basic numbers at a glance, really. But David Friedman has written a little about this before, at least w/ regard to spending and employment. (Perhaps he knows of a study?)

From 1920 (roughly a year after WWI ended) to 1923, federal spending (nominal, real, and as a percentage of national income) decreased every year and was ultimately cut in half. Unemployment during that time period: 5.2, 11.7, 6.7, 2.4 percent.

From 1929-1932, spending increased every year, nearly doubling (real spending) or tripling (% of nat'l income) over 4 years. Unemployment during that time period went from roughly 3% to 24%.


Comments for this entry have been closed
Return to top