It's a perennial favorite for op-ed screeds and cocktail party chatter: are American doctors paid too much? While many might demur from weighing in on seemingly more complicated matters such as how quickly the Fed should inflate the money supply, there likely are few Americans unwilling to express at least some opinion on doctor pay. Moreover, this is a debate that assuredly will grow more heated as a growing share of the burden of health spending gets shifted onto the backs of taxpayers. Answering this question accurately very much matters since it affects important public policy questions such as whether the Medicare "doc fix" needs to be made permanent or whether the U.S. should, like most other countries, subsidize much more of the medical education costs for doctors in general, as opposed to just those who elect to practice in medically underserved areas.
Pundits at Slate and The Daily Beast have assured us America's doctors are overpaid. Moreover, 21.5% of physicians are in households whose income places them in the top 1% of income (versus only 12.8% for lawyers, as one example). Yet half of doctors feel they are not fairly compensated and my own book shows that the rate of return on a medical education is far below that of students who pursue law degrees or MBAs. Uwe Reinhardt last week weighed in with ample evidence that both answers might be right. It all depends on our reference point of comparison. While it might not be very gratifying to fans of Harry S. Truman ("Give me a one-handed economist! All my economists say, On the one hand on the other") I'm here to argue that Prof. Reinhardt is correct, making use of data recently compiled by yet another health economist, Harvard's David Cutler:
Compared to annual pay for doctors in other major industrialized countries, average physician earnings in the U.S. are considerably higher-about 78 percent higher, on average. Unfortunately, this is not quite an apples-to-apples comparison since average earnings includes both specialists and primary care docs. Consequently, comparing average doctor pay in the U.S. (where more than 70% of doctors are specialists) to that in nations such as Canada and France (where less than half of doctors are specialists) is not very illuminating. Moreover, even two-handed health economists generally recognize that pay for primary care physicians in the U.S. lags behind that of specialists, which helps explains why we have a growing shortage of the former, but not the latter. Thus a comparison of average pay may mask the possibility that American specialists are overpaid while generalists are underpaid relative to their international peers.
The next few columns in Prof. Cutler's table help sort out some of these data problems. When specialty earnings are compared to average GDP per capita, it turns out that physicians everywhere are paid at least several multiples of GDP per capita. This should surprise no one, although it's worth noting that things were not always this way. In fact, 150 years ago, "medicine rarely offered a path to wealth;" Princeton sociologist Paul Starr estimates that prior to the Civil War, most doctors had annual incomes that put them at the lower end of the middle class. It probably will surprise few readers to see that this ratio is lowest in Scandinavian countries such as Norway (1.9) and Sweden (2.3) and is highest among the countries most reliant on private health insurance/market competition, such as Australia (5.3), Germany (5.4), the U.S. (5.8) and Netherlands (8.7). But even taking that general pattern into account, specialists in the U.S. are paid 35 percent more than their counterparts in other countries-that is, the specialty pay/GDP per capita ratio is on average, more than one third higher here than in the other countries listed.
But as Prof. Reinhardt astutely points out, the more relevant comparison is not to the average worker but to members of the talent pool from which American physicians are recruited. When specialty earnings are compared to those of high earners-i.e., those in the 95-99% of the earnings distribution-it turns out that physicians in nearly all other countries, not just the U.S., are paid reasonably well. Once again, in the Scandinavian countries, the ratio of specialty earnings to annual compensation of higher earners is lowest (e.g., Norwegian doctors are paid 22% less than high earners), with ratios being highest in countries such as Netherlands (2.56) and Australia (2.54). American specialists, in contrast, earn only 37 percent more than high earners. So if we think high earners is the more appropriate reference standard, American specialists actually are slightly underpaid compared to their counterparts elsewhere: on average, the specialty to higher-earner pay ratio is 6 percent lower here than elsewhere.
What's especially interesting is the story for general practitioners. As you might expect, in the U.S. general practitioners are paid about 8% less than high earners even as U.S. specialists are paid 37% more-re-affirming the general story of generalists being less well-compensated than those who invest extra time to train as specialists. This is consistent with the figure from my book showing that on average, the annual rate of return on a medical education is higher for procedure-based specialists (17.9%) than for primary care doctors (15.8%).
However, the "new news" in Prof. Cutler's table-at least from all the evidence I've pored over during the past few years-is that American general practitioners are actually slightly less underpaid than their specialist counterparts. That is, the phenomenon of paying generalists far less than specialists is by no means unique to the U.S., as a quick inspection of the last two columns reveals. However, the average comparison country pays generalists 6% less than its high earners whereas the U.S. pays its generalists only 8% less than its high earners-a trivial 2 percentage point differential. In contrast, the average comparison country pays its specialists 45% more than its high earners whereas U.S. specialists are paid only 37% more. Thus, when high earners are used as the reference of comparison, Americans "underpay" their doctors more in other nations. If we believe-based on comparisons to other nations where "social justice" is accorded a higher value than it is here-that specialists "deserve" to be paid 45% more than the highest earners, then the U.S. underpays its specialists by 6%. By the same metric, the U.S. underpays its generalists by only 2%. In short, if we were to align our physician compensation patterns to match those in other countries, we would increase physician pay, and would do so by increasing the pay of specialists slightly more than that of generalists.
I'm not arguing that increasing physician pay is necessarily a good idea. We probably ought to pay physicians whatever amount is needed to obtain an "adequate" supply of physicians-which simply opens another large can of worms that will have to be addressed in another post. But by that standard, primary care pay in the U.S. is in greater need of upward adjustment than that of specialists. My point is that these new cross-national comparisons provide a new and different perspective on the age-old question of whether American doctors are overpaid. As with so many health policy matters, the answers too often are never as simple as we might like.
 I have calculated these figures from IPUMS (Integrated Public Use Microdata Series) data reported in this interactive chart at New York Times. While 29.3% of lawyers working for security, commodity brokerage & investment companies fall into the top 1%, there is no broad occupational group overall except physicians (as best I can tell from that the NYT tabulation) that has greater than 20% of its members living in the top 1% of households.
 See p. 84 of The Social Transformation of American Medicine. Prof. Starr concedes there's a paucity of data on this issue, but does a commendable job of assembling what little information is available from that period. He reports that by 1904, average physician earnings had climbed to $750 (an AMA survey estimate that he concedes may have been self-servingly underestimated) versus $540 for the average occupation (and $1,000 for federal employees!). In 2012, the Bureau of Labor Statistics reports that physicians and surgeons earned an average of $190,060 compared to $45,790 for the average worker. Thus, the differential in physician pay compared to the average worker has grown quite dramatically over the past 150 years.
 This is not exactly a pure apples-to-apples comparison either given that 21.5% of doctors live in families with incomes in the top 1 percent. However, the rationale for excluding the top 1 percent of earners is that it is an open-ended category whose highest income families (with income in excess of $100 million) considerably distort the estimated average income for the group).
 The comparative data from Prof. Reinhardt imply that on average, U.S. specialists earn nearly 50 percent more than U.S. general practitioners (1.37/.92=1.489). There's 2 reasons the differential in annualized rates of return between specialists and primary care doctors is not nearly so large. First, my rate of return figures for primary care doctors include some "specialists" such as general internal medicine and pediatrics, whereas the general practitioner category used in the cross-national data does not. Second, specialists generally spend more time in training than do primary care physicians, thereby substantially increasing the size of the investment that has to be paid off in the form of higher earnings. That is, all other things equal, 1 added year in residency translates into 1 additional year of lost income which both increases the dollar amount of the investment but also reduces by 1 year the number of remaining career working years available to pay it back.