I haven’t read the papers Barro refers to in his column in the Wall Street Journal. I assume they are well researched and well founded, but the column itself didn’t sit well with me.

Barro seems to be suggesting that WWII represents a good test case for the effect of government stimulus. I can see the motivation. Indeed, its common to point to WWII as the force that ended the Depression.

However, there seem to be some serious issues here, perhaps they are addressed in the paper but skipped over in the column. First, doesn’t war, especially WWII radically alter expectations? Indeed, the larger the war the more radical the effect on expectations.

Second, wasn’t it the case that WWII was also accompanied by an excess profits tax. The idea being that business was going to be directly asked to sacrifice for the war effort.

Third, aren’t war expenditures hard to value. Some of the expenditure is simply hiring soldiers, who are valued at their typically low salaries. Indeed, there was a draft suggesting that the military is paying well below what soldiers consider their reservation wage. Some of those soldier would have gotten higher paying jobs had they not be drafted.

Fourth, it seems to me that multipliers should be inversely proportional to the amount of stimulus. First, you employ  the people who really have no options. Then the people who have options but don’t like any of them. Eventually, you’re eating into the people who are happy with what they’re doing.


Some of this seems to be captured by the unemployment rate, but aren’t we interested in some measure of cyclical unemployment. Which would probably have hit zero long before 1944’s 1% unemployment rate.

All in all it would seem that a massive increase in defense spending, much of which is on drafted soldiers and part of which is funded by an excess profits tax is a going to have a much lower multiplier than a more moderate stimulus that involves mainly cutting taxes, forgoing increases in state taxes, and providing funds for the short term unemployed.