Casey B. Mulligan

Side Effects: The Economic Consequences of the Health Reform
I'm writing this review as an economist who looks at policy issues regarding Medicaid in a state that was among the first to adopt the ACA and push for expansion. I, like others, followed the health care debate and the claims by various health care economists like Jonathan Gruber very closely. This book is the most technical and helpful I have seen thus far in analyzing the ACA, but it is too technical for the average reader with plenty of equations and explanations of the convexivity of his curves and such. Dr. Mulligan ought to have someone write an executive summary and put it out for free, this text can be bought by those interested in the mathematics behind his conclusions. The book is an applied microeconomics text, I would recommend it for a health care economics course.

This book cannot be read without some understanding of Mulligan's previous work (a book and various articles in the NY Times and on his blogs) about labor economics and the effects extending unemployment insurance had on disincentivizing people to seek employment during the Great Recession. Mulligan calculates that the "jobless recovery" through 2013 was caused by the generous expansion of the social safety net. Unlike most economic pundits, however, Mulligan has carefully put together a case by analyzing data and doing the math. Like other Chicago-school economists, he is focusing on supply rather than demand. It's his position that U.S. recovery begins when the generous UI extensions expire, and then the ACA permanently pushes back on the recovery by providing even greater incentives for people to reduce their hours or stop working altogether.

"In other words, I expect the ACA to stop the recovery as measured by hours per adult, and likely significantly reverse it" (loc. 4227).


Mulligan calculates that the greatest damage (a little more than half) is done by the employer mandate, which has not been fully implemented (and Mulligan projects will be repealed by 2017). The Obama Administration has claimed that the delay is to help businesses with the increased administration costs, but the reality is that it creates a strong incentive for firms to reduce the working hours of employees.

This book sets out to establish a framework by which to measure the ACA's effect on employment and economic growth going forward. I'll note that Mulligan has been mostly silent on his blog since this book was published and in much of 2014. I would say the great drop in unemployment and the fastest growth of employment since 1999 has caught him by surprise, he does not predict this either in his previous work or this book. We'll see what he says in 2015.

This book was quite helpful and I came close to maxing out the allowable highlights on my Kindle version. The noteworthy conclusions:

"while Medicare and Medicaid help their target populations and give them a bigger slice of the economic pie, the programs also diminish the pie itself" (loc. 168).

"(The ACA adds) about six percentage points to the marginal tax rate faced, on average, by workers in the economy" (loc. 203).

"The ACA will have the nation working fewer hours, and working those hours less productively, so that its nonhealth spending will be twice diminished: once to pay for more health care and a second time because the economy is smaller and less productive" (loc. 320).

"Overall, weekly employment rates (not to be confused with the fraction of persons who work sometime during a calendar year) will be depressed about 3 percent...I assume that aggregate hours worked fall, in the long run, about 0.36 percent for every 1 percent that taxes reduce the economywide average reward to working...the average 2016 worker under the ACA keeps 68.7 percent of what he earns at the margin, as compared to the 75.0 percent that he would have kept if the ACA had not been passed...ACA-induced misallocations will reduce both wages and GDP per quality-adjusted hour worked by about 0.6 percent in the long run and reduce total factor productivity by 0.4 percent...the productivity loss is about $100 billion per year...the ACA increases by 4 percent (14 percent of 24.7 percent) the fraction of workers who can get assistance from, and would participate in, Medicaid...Without its employer penalty, the ACA would depress log aggregate work hours “only” 0.023, or about 2.3 percent, as compared to 0.031 with the penalty...the long-term impact of both ACA and non-ACA events on the log of real GDP per capita relative to its trend is 0.009 plus the range of impacts for log labor hours per capita: the combination is the range of −0.019 to 0.019 by 2016–17, which means that by then the log of real GDP per capita will have deviated from its trend by −0.014 to 0.014...the ACA—will include about 3 percent less employment, 3 percent fewer aggregate work hours, 2 percent less GDP, and 2 percent less labor income...sixty million people (workers and dependents) will be experiencing significant labor market disruption, not to mention those experiencing lesser changes in their wages or work schedules." (2233, 2259, 2269, 3776, 3779, 3985, 4091, 4294, 4696-4698).


The employer mandate ends up creating a class of "29-ers," people working around 28-29 hours and no more. It also creates a huge marginal tax on the 50th worker of medium-sized firms. Unlike the health insurance it could have offered employees, the penalties paid under the ACA are not tax-deductible. When factoring in the tax deductions it could have earned through higher employee wages, the opportunity cost to the employer is greater than the nominal amount of the penalty:

"In effect, the fiftieth employee-year costs $40,000 more with the ACA than it would without it. As long as the employer restrains hiring to remain below the threshold of fifty, the $40,000 hiring disincentive does not appear as government revenue even though it affects the labor market...For every employee for whom the employer’s coverage is not affordable and who receives exchange subsidies, the employer owes $3,000 (plus health cost inflation after 2014) per year...I estimate that salary equivalent of the employer penalty (for the purpose of comparing to 2014 salaries) will be $3,113 in 2015...half of the full-time workers at penalized employers earn less than $15 per hour and would each have to work at least four hours per week for free in order to compensate his employer for the penalty owed because of his employment...Large employers not offering coverage and having more than thirty full-time employees in 2016 will, as a consequence of the employer penalty, owe the salary-equivalent of an additional $3,163 per year for every full-time employee they add to their payroll, and save an additional $3,163 per year for every full-time employee they remove from their payroll...that fiftieth employee costs her $62,265 plus the employee’s normal salary and benefit" (loc. 498, 507, 931, 955, 973, 977).


One needs to recognize that employers pass the cost of the ACA penalties onto their employees in the form of lower salaries. Firms will increase the hours of already full-time employees while reducing the hours of those close to 30 hours in order to avoid the mandate, where possible. While estimates are that this is a relatively small portion of the U.S. workforce, it is still a few million workers who will be affected.

Aside from the employer mandate, Medicaid's availability to 138% of the poverty line and the subsidized insurance offered to those qualified also create disincentives to earn additional income-- implicit marginal income taxes.

"Many families below 400 percent of the poverty line but above about 220 percent of it would have a marginal income tax bracket of about 25 percent: 7.65 percent for employee payroll taxes, 15 percent for federal individual income taxes, and roughly 3 percent for state individual income taxes.22 The federal rate would commonly be between 21 and 36 percent (instead of 15) for families between roughly 100 and 220 percent of the poverty line as the federal earned income tax credit is phased out (Congressional Budget Office 2012b), putting the combined rate over 30 percent and perhaps higher than 45 percent" (loc. 1411).


There are always some who are worse off if their employer pays them more, either making them ineligible for Medicaid or causing them to lose potential subsidies.

"(I) suspect it will be increasingly rare for employers to offer coverage to part-time employees, because the offer will usually prevent them from getting an exchange subsidy...a household with a 25 percent marginal tax rate has to earn another $1,333 in order to make up for loss of $1,000 of its subsidies because it would pay $333 worth of tax on the additional $1,333 it earned...the ACA puts some workers in the position where a part-time schedule gives them more disposable income than a full-time schedule would" (loc. 1394, 1407, 1492).


He is not certain of the exact number, but states that some workers will face a marginal tax rate on an additional dollar earned of over 100. Mulligan also notes which demographics would be affected, the distribution of effects is uneven:

"Unmarried people are more likely than married people to face the one-hundred-plus rate because the former do not have a spouse with opportunities for family coverage. Unmarried women are more likely to do so than unmarried men because the former are more likely to be working thirty to thirty-nine hours even without the ACA...more than ten million nonelderly household heads and spouses—about 9 percent of the workforce—will face those new taxes in 2016, and among them marginal income tax rates will be about twenty percentage points higher than they would be without the ACA" (loc. 1574, 2082).


This aspect of the ACA is forgotten or ignored by those advocating an increase in the minimum wage. Until the poverty line adjusts, increasing the minimum wage will lift some households out of eligibility for Medicaid, making them worse off by having to purchase health insurance (until prices rise such that the poverty line eventually adjusts). Mulligan mentions the exacerbation of the effects of an increase in minimum wage but its combination with the ACA is beyond the scope of this book.
Advocates are disingenuous if they argue that these wage increases will both help the poor and save the government money by reducing their transfers or subsidies.

The last chapter deals with those who argue against Mulligan's propositions by citing the lack of adverse effects seen in Massachussets under "Romneycare." Mulligan contrasts the two plans, showing that the employer mandate was much smaller ($295/employee), and was deductible from federal income tax. The marginal implicit tax was about five times lower than that of the ACA. Massachussets already had one of the most generous Medicaid plans in the U.S., offering it to much higher incomes than the rest of the country, so little change occurred under the ACA.

Some possible criticisms: If I were looking to make a counter-argument, I would point out that while Mulligan rightly predicts that the number of uninsured falls dramatically under the ACA, he fails to account for the gains due to increased productivity of a healthier workforce. I'm not saying that would completely offset the productivity losses he calculates, but it is worth noting that there is a benefit to the supply of laborers of being healthier (there are positive externalities to my co-worker being more healthy as well, I'm less likely to be unhealthy). ACA proponents argue a demand-side effect that Mulligan mentions but does not address. He focuses his criticism on those studies' lack of any analysis on the increases in implicit marginal tax rates due to the ACA. In a state that adopts ACA expansion, the increase in federally-subsidized health care means more disposable income for its low-income population. This has been estimated for Kentucky and chronicled anecdotally in various outlets like the New York Times.

There is also some confusion about productivity, in one section he writes that productivity (as the BLS measures it) will actually increase as the low-income, low-skilled workers reduce hours or leave the workforce to obtain subsidies, leaving the high-skilled, higher-income workers in the workforce. But in his conclusions he writes that productivity overall decreases, perhaps he means skill or wage-adjusted productivity.

This book serves as a great reference for information on the ACA. It's a must-read and will be worth coming back to as more data is available. So far, 2014 saw a continued decrease in labor force participation, about half of which is explained by demographics (and most of that is due to aging population) but the other half is not as easily explained. I can give local anecdotes of hours and income intentionally reduced in order to maintain subsidies. But the growth in RGDP and overall employment in 2014 is not something predicted by Mulligan's analysis, even without the employer mandate. Time is needed to judge Dr. Mulligan's analysis. 3 stars out of 5 (4 stars on Amazon... "I liked it" is the best description).… (mais)

Mulligan examines the multitude of government programs that create incentives for workers to leave the work force and for employers to not hire. His conclusion is the aggregation of these programs have resulted in lower employment and longer periods of unemployment.

It is a difficult book for non-economists, which is unfortunate since that makes it less likely it will be read by policy makers, who seem more wiling to engage in policies based on what they want to happen rather than policies based on the evidence of what is likely to happen.

Mulligan could be this generation's Milton Friedman if he had that ambition.… (mais)