But a caveat to the reader: I say that the book “could have been” one of the most important economics books because, although Mulligan occasionally writes well and clearly, much of the book is too technical and hard to follow. Ph.D. economists who are up to speed on highly technical issues might be able to zip through the book; however, even though I’m a Ph.D. economist, I found it hard slogging. There are parts of the book that appear to be important, but the way it is written obscures—rather than reveals—Mulligan’s meaning. Typical readers of Regulation are probably better off reading my review of Mulligan’s book than reading the book itself.

So, first, I’ll tell his story and some of his important conclusions, and then I’ll discuss how he deals with some Keynesian criticisms of his conclusions.

Expanded welfare state | Consider the food stamp program, which is now called the Supplemental Nutrition Assistance Program (SNAP). Between 2007 and 2010, notes Mulligan, the number of families with income below 125 percent of the federal poverty guideline increased by about 16 percent. But during that time, he writes, “the number of households receiving SNAP benefits increased 58 percent.” Why? Because the program was changed to allow more people to qualify for it. With the Farm Bill of 2008, which President Bush signed, the government increased the maximum benefit and relaxed the asset and income tests that determine who qualifies for benefits. President Obama’s 2009 American Reinvestment and Recovery Act (ARRA) granted state governments relief from the work requirement that had previously existed in the program and also increased the maximum benefit.

Another major expansion of the welfare state was the legislated increases in unemployment insurance (UI). In the regular state-funded UI system, people who are unemployed can get benefits for up to 26 weeks, and there is an automatic benefit trigger that adds 13 weeks of eligibility for federally financed extended benefits when the unemployment rate in a state goes above a certain level. In 2008, Congress and Bush added an automatic 13 weeks to states that qualified for extended benefits. That meant that an unemployed person in one of those states could get UI benefits for 52 weeks. Later in 2008, still under Bush, the Unemployment Compensation Extension Act added 20 weeks of eligibility. With various extensions and additions, unemployed people in some states could get benefits for up to 99 weeks. That’s positively French!

In a related development for unemployed people, Obama’s ARRA paid 65 percent of the health insurance premium for people who were laid off and who wanted to keep their employer’s health plan. Previously, under a law called COBRA, laid-off employees could keep the employer’s plan, but the employer could charge them up to 102 percent of the premium that would have been charged for their policy had they stayed employed. A 65 percent cut in this premium was a big discount.

Mulligan goes through, in detail, four of the six “safety net” programs: food stamps, UI, Medicaid, and the COBRA subsidy. (The other two are a break on federal income taxes and a break on payroll taxes.) Then, for people in various income and family categories, Mulligan derives the disincentives to work that the expansions of these programs have caused and are causing. Among six hypothetical households chosen to represent the most common households affected, Mulligan finds that the “policy impact” on a household “with an unemployed primary earner” ranged from $227 to as much as $2,190 per month in added disposable income for not working. He notes that these are the amounts by which the change in rules, alone, added to the disposable income of various unemployed people. Those numbers at the bottom end are substantial and at the top end are huge.

Mulligan devotes a large part of the book to estimating the portion of the drop in hours worked by U.S. workers since 2007, when the recession began, that is due to these expansions of the safety net. He writes, “I conclude that at least half, and probably more, of the drop in aggregate hours worked since 2007 would not have occurred, or at worst would have been short-lived, if the safety net had been constant.” He spends much of the book making that case, and this is where, frankly, he lost me in the technical weeds. So it is hard for me to fully evaluate his case.

Is there demand? | Mulligan realizes, though, that he must contend with an alternate claim that other economists, primarily Keynesians, have made: that the problem during the recession and subsequent weak recovery was not that people became less willing to work, but rather that employers became less eager to hire because of weak demand for their goods. Mulligan counters this claim with three sets of evidence, all of which confirm his view that labor supply is key and cast doubt on the Keynesian view that the key constraint has been the demand for labor.

First, Mulligan notes that, to the extent that the safety net changes are responsible for the drop in hours worked, the following would result:

• Labor hours for the elderly would increase because they were not much affected by the specific safety net increases.
• Labor hours would decrease for single people more than for married people because many of the changes made it easier for single people to get government aid.
• Labor hours would decrease less in regions that had more-stable housing prices because one government program gave people who had suffered a drop in housing price an incentive to cut their income.
• Labor hours would decrease less for high-income people because they got virtually none of the added government benefits.

Mulligan presents evidence that affirms all four expectations. The Keynesian model has no such implications.

Second, notes Mulligan, if the driver of the decline in work hours was a decline in aggregate demand, one would expect a similar decrease in the use of other factors of production. But, Mulligan writes, “Output declined sufficiently less than work hours to make it appear that other production inputs (aside from work hours) tended to increase during the recession” (emphasis his).

Third, Mulligan looks at data on summer employment. If, as Keynesians claim, the demand for labor was the constraint, then the seasonal increase in the number of hours worked by young people in 2009, a recession year, should not have been as high as it was in normal years. But Mulligan shows that the seasonal pattern in 2009 was similar to that in prior years.

Mulligan also shows that after housing investment collapsed in 2007, investment in non-residential structures increased. This is evidence that factors of production, including labor, left one sector and went to the other. I’m not sure why he presents this evidence, though. It demonstrates labor mobility, but he seems to see it as something more—specifically, as evidence against the Keynesian view. But most Keynesians don’t seem to deny that even in a recession, labor can move between sectors. That could happen even if the Keynesians are right that wages are inflexible.

Of course, the most inflexible wage is the minimum wage because it is set by law. Mulligan points out that the minimum wage increased from $5.15 to $5.85 in July 2007, to $6.55 in July 2008, and finally to $7.25 in July 2009—a 41-percent increase. Bush had signed that phased increase into law. Because the inflation rate for those two years was very low, the real increase in the minimum wage was a whopping 36 percent over those two years. Mulligan estimates that this increase destroyed over one million jobs.

Conclusion | The good news is that some of the changes in the welfare laws were temporary; unfortunately, others were permanent. In the food stamp program, for example, the relaxation of income and asset tests under Bush is permanent, as is the increase in the maximum benefit. The easing of work requirements under Obama’s ARRA, by contrast, was temporary, expiring in October 2010, and the further increase in the maximum benefit expires in November 2013. In UI, the added 13 weeks of eligibility for extended benefits that Bush and Congress passed is permanent. The Obama-granted 65 percent cut in COBRA noted above was temporary, lasting only from April 2009 to May 2010. Although Mulligan never comes out and says it—because his is not a partisan book—it appears to me that Bush’s actions in this area will cause more long-run damage to employment than Obama’s.

I have one remaining gnawing problem with the book’s argument, though. If the decrease in work hours is due mainly to the drop in labor supply caused by an expansion of the welfare state that discourages work effort, why do so many people show up when jobs are advertised, even at Walmart? It’s hard for me to believe that Mulligan’s is the whole story. He has persuaded me, though, that his labor supply explanation is a significant, and largely unreported, part of the story.