Lesson 7 of 7
Two contractions
The 1920 to 1921 contraction is the episode most used to argue for this account. Here is the argument, and here is why two episodes prove very little.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- Published NBER dating: the contraction beginning January 1920 lasted 18 months to a trough in July 1921. The one beginning August 1929 lasted 43 months to a trough in March 1933.
- The argument built on that contrast is that the shorter contraction saw much less intervention, and that the longer one was prolonged by attempts to prevent adjustment.
- The comparison is weak evidence. Two episodes that differ in a dozen ways at once cannot isolate the effect of any one of them, and this lesson says so before it says anything else.
This is the empirical exhibit most often put forward for the account in this course. It is worth understanding, and it is worth knowing exactly how much weight it can carry, which is not much.
Duration of the two contractions, as dated by the NBER
- January 1920 to July 192118 months
- August 1929 to March 193343 months
The argument as its proponents make it
Both episodes began with a sharp contraction. The first ended in eighteen months. The second ran for forty-three and was followed by years of high unemployment.
Rothbard's version, in America's Great Depression, is that the government pursued a largely laissez-faire policy in the depression of 1920 to 1921, and that the 1929 contraction was met by the opposite. He also reports that the National Industrial Conference Board recorded wage rates falling by 19 percent in one year during the earlier depression, and that from then on a high-wage doctrine took hold, with more and more people adopting the view that wage cutting would dry up purchasing power and prolong a depression.
The inference offered is that prices and wages adjusted quickly in the first case and were held up in the second.
Why the comparison cannot carry the argument
Start with the obvious. Two episodes is a sample of two, and the two differ in almost everything: the size and character of the preceding expansion, the state of the banking system, the international monetary arrangements, the scale of the shock, the condition of agriculture, and the policy response. Attributing the difference in duration to any one of those requires holding the others fixed, which a sample of two cannot do.
Then the measurement. Statistics for the interwar period and earlier were largely constructed after the fact by researchers rather than collected by statistical agencies at the time, and different reconstructions of unemployment, output and prices for these years disagree with one another. A comparison resting on those series inherits their uncertainty.
And finally, the counterfactual is unobservable. Nobody can run 1929 again without the policy response, so the claim that the response caused the difference cannot be checked directly.
What the comparison is good for
Framing a question, not answering one. It identifies a genuine puzzle: why did one severe contraction resolve quickly and another not.
Every account of the cycle owes an answer to that puzzle, including this one. The next course puts four answers beside each other on the later episode, which is the honest way to use a case that no single theory can claim.
Check yourself
How long were the two contractions, on published NBER dating?
- Both about two years
- 18 months from January 1920 to July 1921, and 43 months from August 1929 to March 1933
- 43 months and 18 months respectively
Show the answer
Answer: B.
These are the reference dates published by the National Bureau of Economic Research, with its business cycle data last updated in March 2023.What does Rothbard report about wage rates in the earlier depression?
- That they were held steady by agreement
- That the National Industrial Conference Board recorded a 19 percent fall in one year, and that a high-wage doctrine took hold afterwards
- That they rose throughout
Show the answer
Answer: B.
The inference he draws is that prices and wages adjusted quickly in the first case and were held up in the second.Why is the comparison weak evidence?
- Because the episodes are too far apart in time
- Because two episodes differing in many ways at once cannot isolate the effect of any one difference, the statistics are reconstructions, and the counterfactual is unobservable
- Because the NBER dates are disputed
Show the answer
Answer: B.
It frames a real puzzle that every account of the cycle owes an answer to, which is what the next course is for.Sources
- National Bureau of Economic Research, US business cycle expansions and contractions, the published table of reference dates, for the peak of January 1920 and trough of July 1921 with a contraction of 18 months, and the peak of August 1929 and trough of March 1933 with a contraction of 43 months. Business cycle data last updated 14 March 2023.
- Rothbard, America's Great Depression, 1963, part three, chapters 7 and 8, for the statement that the government pursued a largely laissez-faire policy in the depression of 1920 to 1921, for the National Industrial Conference Board's figure of a 19 percent fall in wage rates in one year during that depression, and for the spread of the high-wage doctrine afterwards.
What this leaves out
- This is a heterodox tradition. The mainstream account of the same phenomenon differs, and nothing here is a forecast or a view on any security.
- The NBER durations are published reference dates. The 19 percent wage figure is reported by Rothbard citing the National Industrial Conference Board and was not checked against that board's own publication for this lesson.
- Statistics for these years were largely reconstructed after the fact and different reconstructions disagree. Any comparison resting on them inherits that uncertainty.
- The comparison in this lesson is not evidence for the account in this course. It is described because it is widely used, and its weakness is stated in the lesson rather than left to a footnote.