Commentary: Medical Savings Accounts in Singapore: What Can We Know?

The primary determinant of the cost, quality, and form of medical care premiums in my country is the country's own history, culture, social mores, and politics. Yet there is clearly a strong interest by theorists to try to use the variation in the way other countries choose to finance services, to yield generalizable conclusions on the effect of those financing systems. Both of these statements are represented in Michael Barr's interesting review of the health care financing system in Singapore. Beginning with the low share of gross domestic product (GDP) spent on health care as a fact to be explained, Barr is surely right in arguing that this relatively low proportion is not explained entirely by the design of Singapore's health care financing system. Other governmental health policies, public housing and social policies, and the unique ethnic and cultural mix also function to limit quantities, prices, and qualities of medical services. On the other hand, he seeks (as have many others) to understand whether the combination of medical savings accounts and catastrophic health insurance for the nonpoor has also made a difference. It is this second goal that I will primarily address.

Compared to What?

The main potentially generalizable claim made by advocates of a medical savings account/catastrophic health insurance (MSA/CHP) financing system in Singapore and in the United States is that this scheme leads to a significantly lower level of medical care spending than that which would prevail under insurance plans with lower (or zero) levels of patient cost sharing. No one, to my knowledge, has alleged that MSA/CHP leads to lower costs than all other alternatives. For instance, spending is probably higher under such a system than under one with neither insurance nor MSAs, or even a system with catastrophic insurance but no tax breaks or mandates for medical savings accounts. And though there has been some imprecise use of the term cost containment, I do not think that even the most ardent advocate would claim that MSA/CHP would (or should) stop dead the growth of medical care spending in an economy with rapidly increasing real income exposed to costly but beneficial new technology.

The essay by Barr is never explicit about the most fundamental analytic question intended to determine the effectiveness of some policy: What is the comparator? It is clear that his comparator is not the one suggested above--(nearly) full coverage insurance. As a result, he provides no evidence at all to refute the hypothesis, firmly grounded in economic theory and other empirical evidence, that, other things equal, Singapore's spending was significantly lower under MSA/CHP than it would have been if Singapore had instituted more generous insurance with no personal savings accounts. The considerably higher health care gross national product share of Australia and New Zealand, with more generous insurance systems, would seem to support this hypothesis, although it is virtually impossible (with a sample of one, unique, small country) to test it formally.

The benchmark Barr uses seems to be the inappropriate one of zero growth in spending; he is then shocked to find that spending grew in Singapore even with MSAs in place. I think few unbiased analysts would find this fact either surprising or interesting.

However, it is a legitimate point to argue that, while Singapore's MSA/CHP program effectively caused spending to be lower than it would otherwise have been, that program was not the only deviation in the Singapore system from a general system of universal comprehensive insurance. In particular, such low-income people as there are in Singapore generally do not have sizable MSA balances, do not choose to obtain the voluntary Medishield catastrophic insurance, and do generally use the lowest class of service in government-owned hospitals. Even these people receive some protection from the Medifund subsidy, but it remains true...