John Schmitt, Senior Economist of the CEPR, left a comment on my last post on the CEPR report (which Schmitt co-authored) on entrepreneurship in the OECD and the implications for American health care reform. I would like to thank him for responding, but also address some of his specific claims. His argument, essentially, is that his speculation in the CEPR report -- that the lack of universal health insurance is a major reason why the U.S. has lower rates of small business employment -- is plausible and fits with anecdotal evidence. It certainly is plausible, as I wrote in both of my posts on the report. But it isn't the only plausible explanation, and in my first post I sketched out a few other possible causes.
Schmitt raises some issues in his comment that I'd like to discuss. First, this part:
Scott Shane (who's blog post you quote) believes, instead, that what is causing low small-business employment in the U.S. is our high per-capita income. To support his view, he notes that there is a strong negative correlation between national per-capita income and the self-employment rate.
Shane's view is also plausible, but odd. Why would we expect higher incomes to cause lower small-business employment? If people are on average richer, wouldn't they have more money to start their own small business? Wouldn't they have more money to buy the products and services of small businesses? (And, how is it that more small business, which is supposed to be making our country more prosperous is, on average, strongly correlated with being poorer?) A correlation is not an explanation, and neither Shane's post nor your's helps readers to understand why the U.S. lags.
Schmitt doesn't bother disputing the correlation, but instead tries to redirect attention by pointing out that Shane's result is counterintuitive. That may be so, but it doesn't refute the evidence. What it does do is indicate that there may be complex institutional or structural factors at play that make simplistic explanations (i.e. "the U.S. has less small business employment because it lacks universal health care") fall short no matter how intuitive they may be. For example, the regulatory environment for small businesses is much different in the U.S. than in most other OECD countries. This is independent from health care policy, but must have some effect on small business employment. So is the general business environment. National attitudes towards Big Business and preferences for mom-and-pop shops are also different in the U.S. compared to, say, France.
Further, there is no intuitive reason to think that small business owners should be richer than employees of large firms. Many small businesses struggle mightily to stay afloat even if the owners work much more than 40 hours per week and refuse to take vacations (as a former part-owner of a failed small business, and an employee of several others, I have some experience in this regard). If we don't limit our thinking to the OECD, this may be even more clear. In poor developing countries, almost everyone is self-employed as a subsistence farmer or a small-time trader. And yet they are very poor. An economist should quickly recall the efficiency gains that often come from scale, and the ability of larger firms to develop and use new technologies that enhance productivity and make them wealthier. In other words, we should not be at all surprised that poorer economies have more small business employment.
(For a quick case study, consider the differences between wine production in the U.S. and France as described in Mondovino. France's mode of production involves many small, family-run businesses; the U.S. relies more on several major corporations. Which do you think is more profitable? It's not the small businesses, which are going out of business (or selling out to a conglomerate) at ever-increasing rates. Saying this has anything at all to do with the health insurance systems in the U.S. or France is absurd.)
More from Schmitt's comment:
Fairlie, Kapur, and Gates (RAND Working Paper, November 2008) extended and updated Wellington's work. They also found that 65 year-old men (who were thus eligible for Medicare) were significantly more likely to own a business than men just a little younger (and thus not eligible for Medicare). (Fairlie, Kapur, and Gates's work was funded by the Kauffman-RAND Institute for Entrepreneurship Public Policy, hardly a source to participate in "baseless speculation motivated by partisanship.")
(Last part first: the "baseless speculation motivated by partisanship" bit was a jab at Krugman, who will seemingly say anything that reinforces his priors and disparages those who disagree with him, not the CEPR authors. I have plenty of respect for the CEPR and Dean Baker, and had no intention of casting general aspersions on their work. And I enjoyed this report a lot, although I would have preferred that they stick with the data rather than make speculative claims.)
Now, then. I do not dispute the findings of the RAND study, but I do not accept on its face that this effect is explained by the Medicare cut-off age. Why? Well, what else happens after the age of 65? For most workers, the answer is retirement. Many workers do not need to work full-time after the age of 65, but they also are not ready to stop working entirely. So what do they do? Start a part-time business like a consulting firm, or a "hobby business" to give them some enjoyable work to do a few days a week. Perhaps they purchase a summer home and rent it out when they aren't using it. Doing so pushes them into the ranks of the self-employed, but health care has nothing to do with it.
Shane notes in his post that the U.S. does not rank last in the OECD in small business employment in any category defined in the CEPR report, and all the countries below the U.S. have universal health care systems. If the driving force for low rates of entrepreneurship in the U.S. is lack of universal health care, then this fact must be explained. Schmitt may be right that neither Shane's post nor mine helps readers understand why the U.S. has lower rates of small business employment, but neither does his report for the CEPR because the data aren't clear. It's certainly possible that health care plays a major role; it's also possible that variance in regulatory structures and market pressures are the driving factors. Most likely (in my mind), it's some combination of the three.
My whole point in all of this is not to say that differences in health insurance systems have no effect on differences in rates of entrepreneurship. In fact, I've said in all three of my posts on the topic that that is a plausible explanation. But there are other plausible explanations as well, and we should seek out better evidence and more rigorous testing before accepting one hypothesis over the others. After all, isn't that what scientists do?