According to Casey Mulligan, the increase in the minimum wage which took effect in July 2009 caused the loss of 800,000 jobs. Writing in the New York Times Economix blog, University of Chicago Professor Mulligan explains that if we lowered the minimum wage, we would likely see an increase of 800,000 jobs. It appears that Mulligan is talking about part-time jobs, not full time jobs, based on this chart.

I can see that an increase in the minimum wage might cause a drop in employment, but that seemed too big to me. The minimum wage increased $.70, from $6.55 to $7.25 per hour. Adding in employer taxes and something for benefits, we get about $.80, using a cost of 15%. The difference over a 20-hour week is about $16. Employers expect to make money off their employees. It’s hard to believe that there are 800,000 jobs where the employer would make money at the lower pay but not at the higher pay. That is particularly true where there are so many unemployed people looking for work that the employer should be able to find better people who will be more productive, even in a part-time position.

So I find a copy of the paper, Simple Analytics and Empirics of the Government Spending Multiplier and Other “Keynesian” Paradoxes, (love the scare quotes) in the B.E. Journal of Macroeconomics, which you can get from your public library. Mulligan explains the paper as follows:

The purpose of this paper is to exploit the ready availability of obvious factor supply shifts during this recession to test the Keynesian pass-through hypothesis that is at the heart of the paradox of toil and many of the government spending multiplier results. The empirical analysis can be interpreted as tests of whether government spending stimulates private spending that are admittedly indirect, but not reliant on the historical data.

I wade into a pile of second year college math (differential equations! Lagrangians!) and basic economics, which eventually leads to a discussion of flexible-price and sticky-price models, and what they mean for figuring out whether government stimulus spending produces an increase in consumer spending, and what that multiplier might be, and whether government spending crowds out consumer spending, and whether any of this is different in a horrible recession. We can skip that part.

Part V lays out Mulligan’s argument about the minimum wage, along with a discussion of labor supply shifts associated with the seasons (for example, teen hiring in the summer), and the collapse of residential construction spending. Mulligan uses these three events “…to test the hypothesis that factor supply expands output, both at the industry and aggregate levels.”

For the minimum wage discussion, factor supply means an increase in people looking for work. The question Mulligan asks is if the number of people looking for work goes up, does that lead to more production of goods and services. That is a novel idea: that employers create more jobs because more people want to work. If only. This analysis turns on the above chart.

… [Mulligan’s] estimates of the pre-hike regression model … predict that part-time employment would have continued to increase during the second half of 2009 because, prior to the hike, part-time employment tended to increase with full-time job losses.

I am not a native speaker of economics, so I’m not quite sure how to reconcile this with Mulligan’s explanation that his paper tests hypotheses without reliance on historic data. But, pressing on, it turns out that this chart is only facts, and doesn’t explain causes. Mulligan offers explanations for why the differences should be attributed to the minimum wage hike that don’t relate to anything in the paper or the charts.

One reason to attribute much of the gap between actual and predicted part-time employment to the July 24, 2009 minimum wage hike is that, as noted above, the real federal minimum wage was substantially different after July 2009 than it was before, but not expected to significantly affect the full-time employment of prime-aged males that are the basis for the forecasting model.

Let’s see if we can think of other reasons for the massive decline in part-time work. Here’s one. Employers use part-time workers to pick up the pace without committing to the full costs of hiring full-time workers. If business is down, letting the part-timers go means they can keep their core workers.

Here’s another. The economy was in free-fall for much of 2009 and into 2010, so fewer workers were needed to meet demand. Full-time workers were being cut at a rapid rate. The remaining workers were the more productive workers, and were able to do the work employers needed with fewer man-hours. Employers didn’t need part-time workers.

Here’s another. Businesses were failing at an increased pace as the Little Depression (Paul Krugman’s phrase) deepened. Partly that was caused by tighter lending standards by frightened banks. I saw several businesses fail because their lenders refused to roll over their debt, putting all kinds of people out of work, both full and part-time. I don’t know of any business that failed because they couldn’t manage to pay the increase in the minimum wage.

It turns out that for all the math, this paper turns on opinions about the way the economy works in the real world. Maybe it’s all those years I spent with people struggling through financial problems that lead me to disagree with Mulligan. It was fun to read the paper, though.