Casey Mulligan, Loren Baker, Michael Bordo, Kate Bundorf, Elizabeth Caucutt, John Cochrane, Mark Duggan, Richard Epstein, Bob Hall, Lori Hodrick, Dan Kessler, Peter Klenow, Lance Lochner, Lee Lockwood, Paul Peterson, Josh Rauh, Gopi Shah Goda, Kathryn Shaw, Richard Sousa, Peidong Sun, John Taylor.
Casey Mulligan, Professor of Economics at the University of Chicago, discussed “The Employer Penalty, Voluntary Compliance, and the Size Distribution of Firms: Evidence from a Survey of Small Businesses.”
Casey Mulligan presented some of his ongoing work on how social programs give disincentives to work. He focused on the ACA or Obamacare in this talk. Obamacare is gradual, so it is hard to see the results. But some parts came on abruptly so we can see their effects. He studied one in particular, the employer penalty.
(slide 2) In 2015 “large” firms with 50 full time equivalent (FTE) employees were designated, and then in 2016 they either had to provide insurance or pay a penalty. So there was an incentive to not be large starting in 2015. About 30-35% of these companies did not offer health insurance beforehand. The penalty is based only on full time employees, and the definitions are complex.
(slide 3) The penalty is supposedly $2,000 per employee but in reality it is bigger. It is indexed, so it is $2323 this year. Penalties are not deductible from business income tax, so they are equivalent to the loss of $3449 revenue. Crossing the threshold of 50 employees, and paying for all 50 is about $69,000 per year.
(slide 4) (orange) presorts the distribution of the penalty in terms of lost hours of work per week — how much employees must work more to pay for $3,500 per year.
(slide 5) Measuring the effects on employment, and how many firms stay inefficiently below 50, is hard. People who evade the penalty don’t want to be measured. The regulatory definition of employment is different from the economic definition so it is hard to know who is on the margin.
(slides 6,7) These give the fraction of firms under 50 in France, which has similar laws. You can see the bunching — companies are careful not to get over 50.
(slide 8, Figure 1) We can see the spike in 2015 and 2016 of companies under 50 employees.
(slide 9, “Hanover survey”) Casey did his own survey, with the help of the Mercatus Center. He surveyed managers of companies doing compensation and hiring. They were surveyed at home, and paid about $100 per hour to respond. They all answered.
(Slide 10 Non-ESI) The brown line gives the number of fringe benefits offered by companies of different sizes, other than health insurance. These include maternity/paternity leave, 401k plan, etc. The blue bars show who offers health insurance. You see a huge hold below 50 employees. This measures firms who are deliberately staying below 50 to avoid paying for insurance or a penalty. Alternately, it may be efficient for these companies and for the workers to get their insurance on the exchanges. (The vertical scale for blue is 60 to 95 roughly.)
(Slide 11 non-ESI). This graph reports which company managers report directly that they are reducing employment to avoid the Obamacare mandate.
Mark Duggan suggested that companies make a jump to very large, where economies of scale outweigh Obamacare costs.
Further discussion pointed out that health insurance is important not so much for small or large businesses but for different kinds of people — young, people with family member who have insurance, etc. don’t care much about health insurance.
(slide “Growth in businesses by size”) Again we see big growth in the just under 50 bucket, and some growth in very large firms.
Back to slide 1, employment. It could be that these costs just affect productivity lowers wages. Presumably these companies were at an optimal size of, say, 75, and reducing scale makes them less inefficient. But slide 1 shows a suggestive peak in employment at the moment of the regulation, suggesting that firms fired people rather than lower wages.
Mulligan presented some calculations of how much inefficient size lowers overall productivity. $69,000 times the number of firms that changed size divided by two (the usual Harberger triangle area) adds up to about $2 billion lost output.
(Last slide) Mulligan discussed some results in Duggan, Goda, and Jackson that finds no labor decline due to the ACA. Gopi Goda gave a spirited response, and much discussion followed. The bottom line seemed to be that Duggan, Goda, and Jackson’s evidence is largely that the crosssectional (one state vs. another) differences in Medicaid did not produce big results. Mulligan’s evidence about aggregate effects of the employer penalty is a different issue.
- John Cochrane