Amity, Meet Eric

Amity Shlaes clearly doesn't like [1] that my recent post called out that her revisionist history of the New Deal [2] is based on misleading numbers. She responded on the National Review blog The Corner: [1] (emphasis added)

Scher says that The Forgotten Man has been “found to have used misleading numbers.” As Scher, quoting some or other, points out, there are differing data on unemployment for the period. The data in The Forgotten Man show unemployment averaging well above ten percent for the decade. Those data are not from some obscure source. They are based off the work of Stanley Lebergott, the authority for the field. The data that Scher praises are the obscurer data. They include make work jobs of short duration. Scher is correct that economy grew off its tiny base in the 1930s. But by the two measures we most frequently use now, unemployment and the Dow, we never got back to the 1920s levels.

Interesting that Shlaes does not want to mention who I quote and the specific criticism of her numbers. [2]

But the "some or other" is Eric Rauchway, the UC-Davis history professor who reviewed Shlaes anti-New Deal book for Slate. [3] I'll quote at length again, because it ably rebuts most of Shlaes rebuttal :

But if the New Deal did not end the Great Depression, was it doing some good? Historical Statistics of the United States says yes: Except in the 1937-38 recession, unemployment fell every year of the New Deal. Also, real GDP grew at an annual rate of around 9 percent during Roosevelt's first term and, after the 1937-38 dip, around 11 percent.

So on the numbers, the U.S. economy improved briskly during the New Deal. Things that are moving quickly and in the right direction, but still haven't reached their destination after a while, are things that have a long way to go—which is true of the U.S. economy recovering from 1932. Historians disagree on which part of the New Deal most encouraged economic growth, but at the least the New Deal did not prevent this recovery.

Shlaes makes a different argument about numbers, because she uses different numbers. She starts each chapter with a rat-a-tat of just-the-facts, but instead of GDP, which represents the overall economy, she quotes the Dow Jones Industrial Average, which represents the maybe 10 percent of Americans who owned stock. And though she quotes an unemployment number, she doesn't quote the figures I've just mentioned. Instead she chooses different estimates of unemployment that (she acknowledges) show a much larger share of Americans out of work during the New Deal.

If you want to know how the New Deal treated ordinary Americans, this choice really matters. Let's look at a figure Shlaes gives twice in her book and again in her Wall Street Journal editorial: She has unemployment at 20 percent in the 1937-38 recession. That's appalling—almost as bad as 23 percent in 1932. Based on such a statistic, you could think the New Deal wasn't alleviating the Great Depression. But that number hides something: A third of the people Shlaes counts as unemployed had a job that the New Deal gave them through its relief programs.

Now, you may say, wait: Those people really shouldn't count as employed—we're not interested in government make-work, we're interested in the real economy. Fair enough—and if you look again at Historical Statistics of the United States, you'll see another measure of unemployment—private, nonfarm unemployment—measuring the real, industrial economy. And on that measure, unemployment again runs markedly lower under Roosevelt than under Hoover. John Maynard Keynes might have explained that the New Deal wasn't just offering make-work, it was stimulating the economy—and Shlaes in fact at one point says the same: "[I]t functioned as Keynes ... hoped it would." Yet of all the possible ways to measure unemployment, Shlaes chooses the only way that hides the effect of New Deal relief programs and makes it look as though the economy performed as poorly under Roosevelt as under Hoover.

Shlaes cavalierly dismisses Historical Statistics of the United States as the "obscurer data." But it's the official historical data sanctioned by our federal government. [4]

But what of this Stanley Lebergott, "the authority of the field" of historical economic stats according to Shlaes?

Rauchway has him covered too. Lebergott may be a respected economic historian, but that does not make his work above critique.

For the blog The Edge of the American West [5], Rauchway wrote (footnotes and graphs can be found in the original post [5]):

Here’s what Lebergott had to say about counting unemployment in the 1930s:

These estimates for the years prior to 1940 are intended to measure the number of persons who are totally unemployed, having no work at all. For the 1930’s this concept, however, does include one large group of persons who had both work and income from work—those on emergency work. In the United States we are concerned with measuring lack of regular work and do not minimize the total by excluding persons with made work or emergency jobs. This contrasts sharply, for example, with the German practice during the 1930’s when persons in the labor-force camps were classed as employed, and Soviet practice which includes employment in labor camps, if it includes it at all, as employment.

Did you catch that? People who painted murals for the WPA fall into the same category as internees in Mauthausen or the gulag. So they count as unemployed!

One could say a few things about that.

(1) Wow, that’s a lot of ideology to cram into a single data series;

(2) if you’re using the unemployment data to answer the question, “did the New Deal help people,” then this data set is going to give you the wrong answer, because it’s going to show people suffering unemployment who in real life had a job, as Lebergott says;

(3) but what if people in emergency work acted like the unemployed—i.e., they were looking for a job...

...Now, as it happens it looks like the answer to (3) is, mainly they didn’t—people who had an emergency job acted like they had a job (perhaps because they had a job) and probably shouldn’t count as unemployed.

Furthermore, Boston University Economist Robert Margo noted two years ago on EH.net Economic History Services [6] that Lebergott's treatment of 1930's economic data was his "first high profile controversy" as it "generated a lot of theoretical work for macroeconomists who thought they had to explain how unemployment rates could remain above 10 percent while real wages were rising (after 1933)."

Darn that government "make work" and its dastardly increase in wages!

So for Shlaes to hold up Lebergott as an untouchable "authority" on economic data, unequivocally superior to the Historical Statistics of the United States, is news to other economists.

Lesson: even when government action demonstrably and quantifiably helps people, conservatives will just use data that pretends it never happened.

Remember that as the battle for robust public investment in our weakening economic foundation heats up after Election Day.