David R. Henderson  

Great Moments in Economic Estimation

The power of wishful thinking?... A Public Transportation Patch:...

Even the great Irving Fisher was, at times, a lousy microeconomist.

"In an attempt to memorize poetry," Irving Fisher wrote in 1926, "Professor Vogt of the University of Christiana found that on days when he drank one and one-half to three glasses of beer it took him 18 per cent longer to learn the lines." To an extreme dry like Fisher--which is to say, to an extreme dry who also happened to be one of America's foremost economists--this was a "fact" to fall in love with. Fisher wasn't going to miss the chance to attribute the strength of the U.S. economy to computations made by Professor Vogt during his attempt to master the Odyssey, in Norwegian, while a little bit in the tank. He had made a similar calculation in 1919, when he extrapolated from a study showing, he said, that "two to four glasses of beer will reduce the output of typesetters by 8 per cent." From this he determined that withholding those beers from those typesetters, and from everyone else in the American labor force, "will add to the national output of the U.S. between 7 [and] 1/2 to 15 billion dollars' worth of product, every year."

Fisher seemed to find evidence for the wonders of Prohibition behind every statistic he encountered--for instance, the discovery, in 1924, that arrests in New York City for the use of "foul language" had dropped 20 percent since pre-Volstead levels. He was not inclined to consider other causes--like, possibly, a wider acceptance of profanity in the Jazz Age city--when he could grab a number and pronounce it evidence. When Fisher determined that a single drink reduced efficiency by 2 percent, he said that this translated to more than a billion dollars of GNP.

These two passages are from Last Call: The Rise and Fall of Prohibition by Daniel Okrent.

To get serious for a few moments, these passages above demonstrate that even the great Irving Fisher, whose fantastic book The Theory of Interest the late Jack Hirshleifer had us read my first year in graduate school at UCLA [I read 250 pages in one sitting, the only time I've done that with an economics book], could be a poor economist when he wanted to impose his own values on others.

Consider the first paragraph above. I trust that I don't need to explain why extrapolating from a group of typesetters to the whole labor force is, well, problematic. That's a simple statistical mistake.

But he also makes a fundamental microeconomic error that shows his inability or unwillingness to apply my Seventh Pillar of Economic Wisdom: the value of a good or service is subjective. Let's say it had been true that when people in the labor force drank, their productivity fell. So what? Most of them enjoyed it. When people played tennis, they weren't being productive either. But does that mean tennis should have been banned? Fisher fell victim to what I've called "GDP [GNP] fetishism."

Now consider the second quote above. Fisher fails to take account of substitution in enforcement. Throughout the book, Okrent lays out how enforcing Prohibition sucked resources away from other enforcement activities by local police. The most likely explanation for the decline in arrests for foul language is that police were busy enforcing Prohibition.

Postscript: I just realized that Okrent is the former ombudsman for the New York Times who, in a parting shot, was critical (and justifiably so) of Paul Krugman, Maureen Dowd, and the late William Safire.

Comments and Sharing

COMMENTS (8 to date)
Daniel Kuehn writes:

I'd be really interested in a separate post on your postscript, David. My impression was always that Okrent had egg on his face after his encounter with Krugman. The employment data, the retirement income issue, the long-term unemployment data, the tax cut data - they all came down on Krugman's side. The one area Okrent mentioned that I remember where Krugman had goofed, Krugman issued a correction even before the Okrent "parting shot", although I gather Okrent was upset with how the correction was issued.

I can already anticipate how a lot of commenters will respond to this comment, but I know you're not like that David and I am genuinely interested in what part of the exchange you found especially justifiable.

Daniel Kuehn writes:

In a way that sort of thing should not be surprising. It's like the Giles/Piketty thing. Journalists that think they can double as professional economists often wander into trouble (and vice versa - it's just the division of labor).

David R. Henderson writes:

@Daniel Kuehn,
I’m on the road and don’t have a lot of time to do searches. Because sometimes comments with lots of links get held up in the spam filter, please send me, by e-mail, links that you think establish that Krugman was right. Then, when I’m sitting in an airport, I’ll check them out.
Here’s my guess, but it’s only a guess at this time: Okrent caught Krugman in what Bob Murphy calls a Krugman Kontradiction. But, as I say, I will look at the evidence.

Daniel Kuehn writes:

Lots of links are provided in the exchange that the NY Times hosted (http://publiceditor.blogs.nytimes.com/2005/05/31/new-public-editor-hosts-paul-krugman-daniel-okrent-debate/), so this one should get through and I'll email it to you too. It's worth clicking through some of those links and looking at the original material, I'll add. There are some real oddities, like Okrent claiming Krugman says something about UI and the federal deficit but if you click through he never even mentions the deficit - he talks about government spending.

Also notable is the big difference in tone between the two.

Daniel Kuehn writes:

On Fisher, the really dangerous thing here is that although the numbers seem big it's not an implausible claim. Drinking reduces productivity seems reasonable (and we do care about GDP even if we're not being fetishistic about it!). It's precisely the reasonable claims that people sometimes aren't careful about thinking through.

Jason Braswell writes:

Hmmm. I appreciate the urge to be charitable, but I don't find the thinking attributed to Fisher in the above to be very plausible, especially for a person who had thought as much about "things" as Fisher undoubtedly had. The numbers don't just seem big; they are big, ridiculously giant, in fact.

Hopaulius writes:

Wouldn't beer consumption (and thus production) also add to GDP?

Daniel Kuehn writes:

Jason - is that to me? I agree, they're very big. I wasn't thinking my comment was charitable. But the "drinking reduces productivity" relation is a reasonable one, so he was able to see what he wanted to see and rationalize the big numbers.

Comments for this entry have been closed
Return to top