Summer 2012 Adam Wagstaff and Wanwiphang Manachotphong
In Thailand universal coverage reduced the probability of people being too ill to work by 20 percent
The developing world—and some of the developed world too—is in the midst of a major push toward universal health coverage. A key objective is to narrow gaps in coverage and thus improve the population’s health. Improvements in health would be of value in their own right. But they may also have an economic benefit—by enabling people to work who would otherwise be too ill to do so, allowing people to take less time off work because of sickness, and increasing labor productivity.
That the expansion of health insurance might yield health benefits is certainly plausible: cost sharing at the point of use may lead people to delay seeking care, discontinue treatment prematurely, or forgo it altogether. Yet two recent reviews conclude that there is surprisingly little hard evidence on whether health insurance does indeed lead to better health and, if so, how large the effects are. The problem is that insurance status is rarely random, and in practice—because of selfselection or deliberate targeting—it is likely to be correlated with unobserved variables that are correlated with health status.
Randomization would be enough to eliminate concerns over selection bias. But randomized control trials of insurance are few in number. In any case, randomization is unnecessary; what is needed is a setting where people’s insurance status is exogenous to their health. One such setting is where a program is rolled out in a staggered fashion so that the number of people who are entitled to or have health insurance increases over time, but the individuals affected do not decide their eligibility and may not even have a say in whether they are covered. This approach is used in a new paper by Wagstaff and Manachotphong, who exploit the staggered rollout of Thailand’s universal coverage program to estimate its effects on one measure of health status—whether people report themselves in the Labor Force Survey as being too ill to work.
By using general revenues to cover everyone not covered by the civil servant or social security scheme, Thailand achieved universal coverage within a period of 12 months. Four clusters of provinces implemented the new program in waves, enabling the authors to credibly estimate the effects of the reform on health status by linking a person’s self-reported health limitations at a particular date to the person’s length of exposure to the universal coverage “regime”; exposure depends on the survey date and the person’s province of residence.
Statistical power comes from the fact that the quarterly labor force surveys in 1997–2005 include an average of 160,000 respondents. Since the month in which universal coverage was implemented is also available from 2001, the authors can precisely identify the timing of implementation in each province. Seventeen of the 32 surveys they use predate the launch of universal coverage, enhancing the accuracy of their estimates.
The authors find that universal coverage reduced the probability of people reporting themselves in the Labor Force Survey as being too ill to work. One year after its launch, the program reduced the probability by 0.4 percentage points (or 20 percent). This effect rises to 0.7 percentage points (35 percent) after two years and stays at the same level in the third year since implementation. The authors find much larger effects among those age 65 and over.
These effects are mostly statistically significant at conventional levels. But are they large in economic terms? To answer this question, the authors also estimate the effects on the probability of being too ill to work of another major public program introduced at around the same time as universal coverage: the Village Fund. This provided Thai villages with seed money (around $22,500 per village) to set up a microcredit scheme; villagers used the money mostly to fund agricultural activities, though some used it to finance daily expenses and nonfarm businesses.
While not a health intervention, the Village Fund might have been expected to have had some beneficial effects on health status, for example, by increasing food consumption and giving households extra income to purchase medicines and other medical goods not covered by the universal coverage program. The authors find that the fund’s effect on the probability of someone reporting themselves as too ill to work is zero in the year of implementation and in the next two years. Only in the third year after implementation is there a nonzero effect, but it is about a quarter of the effect of universal coverage.
Adam Wagstaff and Wanwiphang Manachotphong. 2012. “The Health Effects of Universal Health Care: Evidence from Thailand.” Policy Research Working Paper 6119, World Bank,Washington, DC.