A Quantum of Herring

Casey B. Mulligan, of whom I have written before, has a new post on the New York Times Economix blog, in which he attempts to school the less wise what policy impact assessment is all about. It is not about Red Herrings, for example. He references one of his recent posts that I opted to mostly let go at the time. Though I did make a comment not unlike the one he disparages.

In this post he says that the point of policy impact assessment is to compare what will happen if a policy is implemented to a baseline, without the policy. Fair enough, but is that enough? He says:

Policy impact quantifies how things are different as a consequence of the policy. [emphasis mine]

His analysis of the impact of the Affordable Care Act on the part-time labor market concludes that two of the things that keep people in full-time employment, access to health insurance coverage and higher pay, will be eroded by the ACA. The bit about the insurance coverage is obvious enough. Well done! The bit about the higher pay is not quite as obvious. The numbers Mulligan uses are telling, however.

The example he uses to make his case is of someone working at a job that pays $28 per hour, so if you assume as Mulligan does that they also get two weeks of unpaid vacation a year, the full-time worker gets $56,000, while someone working at the same job for only 30 hours per week would earn just $42,000. When you look at the differences between the full-time worker getting insurance coverage through her job and the part-time worker getting subsidized coverage through an ACA exchange, at the end of the year after taxes and health-insurance related costs, it’s a wash! They both get $34k take-home pay.

However, how likely is it that that worker will be able to get a job that pays $28/hour on a part-time basis? Not very. Using the 2012 Monthly Outgoing Rotations Group file (available here), we find that the 95th percentile of hourly wages last year earned $29 per hour. The average was $12.14 an hour and the median hourly wage for part-time workers was $9.50. Those are the likely kinds of jobs someone will get if they “choose” to work part-time (a choice Mr. Mulligan has never pondered, alas). Full-time workers getting paid hourly will be a little better-paid: the 90th percentile hourly wage for full-time works last year was $28! Their mean and median hourly wages were also higher ($16.67 and $14, respectively).

So, yes, it is good to not use Red Herrings when doing policy analysis. It is also good to use common sense. In the mythical world where more than just the top 10% of workers make the kind of money Mulligan imagines hourly workers do, perhaps workers all have the option of setting their own hours. In our economy, workers do not control the hours they work, for the most part, unless they are consultants. The monthly Current Population Survey, our country’s labor force survey, asks detailed questions of people working part-time: why are you working part-time; do you want to work full-time; etc. They do not ask full-time employees if they would rather be working part-time (though I am sure many of them would say yes). If we could imagine the survey then asking them why they would rather work part-time, I don’t think better pay will be on the list of options, even after the ACA. So, the “new” economics of part-time labor will apply to very, very few workers. Mulligan tells us that

As long as there are more than zero people whose full-time vs. part-time work decision depends on funds or insurance, there is the potential for policy impact.

Indeed. Another interesting thing to include in your analysis is how big the impact is liable to be. Remember the part about quantifying the impact? Mulligan’s case is that there will be an impact. But he does not quantify it. In this case, the quantity of the impact will be a quantum.