Health Insurance and Consumption:
Evidence from China’s New Cooperative Medical Scheme
Chong-En Bai  Binzhen Wu*
Tsinghua University
Abstract
The precautionary saving motive is a popular explanation for the Chinese Saving Puzzle, but few empirical studies quantify the importance of the precautionary savings in China. We exploit the quasi-natural experiment provided by the introduction of a health insurance program in rural China to examine how the insurance coverage affects household consumption.
The results show that the health insurance coverage on average stimulates non-healthcare related consumption by more than 5 percent. The effect exists even for households that hav
e no health-care expenditure. In addition, the effect is stronger for poorer households and households with worse self-reported health status, who both have higher risk of relatively large health expenditure.
We also find that the insurance effect varies with households’ experience with the program. Particularly, the effect is only significant in villages in which households witness some reimbursement resulting from the insurance coverage. In addition, the program stimulates more consumption for experienced participants than for new participants of the program in these villages.
Keywords: New Cooperative Medical Scheme; Consumption; Health Insurance; Precautionary Savings; Chinese Saving Puzzle
JEL Classification Nos.: D12, E21, I18
1 Introduction
Chinese high and rising saving rate has attracted a lot international attention. Household
saving rate has risen by about ten percentage points between 1995 and 2008, reaching 28 percent of the disposable income in 2008, which is higher than most of other countries including East Asian countries (Prasad, 2009). The literature has proposed many explanations for this Chinese Saving Puzzle. A popular one is that the dissolution of the traditional social safety net has created more precautionary savings (Chamon and Prasad, 2008; Meng, 2003). Chinese government has made a lot effort to improve its safety net. The social insurance programs disbursed 1.2 trillion RMB in 2009, with an annual growth rate of 19.4% since 2000. However, there have been very few empirical studies that quantify the size of precautionary savings in China. Given that Chinese saving rate has important global impacts, it is crucial to learn how much the public insurance programs affect consumption and savings in China.
The existing empirical literature for the developed countries delivers quite mixed results on the role of the precautionary savings. The results range from being very small (Dynan, 1993; Guiso, Jappelli, and Terlizzese, 1992; Hurst, et al., 2010; cooperativeStarr-Mccluer, 1996) or modest (Engen and Gruber, 2001; Lusardi, 1998) to quite large (Banks et al., 2001; Carro
ll and Samwick, 1998; Fuchs-SchÜndeln and SchÜndeln, 2005; Kazarosian, 1997). Studies in the developing countries are still in their early stages (Lee and Sawada, 2010; Meng, 2003; Zhang and Wan, 2004). Most of the studies find a substantial amount of precautionary savings. Recent studies exploit the exogenous variations of the insurance coverage caused by policy changes, including Gruber and Yelowitz (1999), Engen and Gruber (2001), and Kantor and Fishback (1996) for the US, Atella, Rosati, and Rossi (2005) for Italy, Wagstaff and Pradhan (2005) for Vietnan, and Chou, Liu and Hammitt (2003) for Taiwan. Most of these studies confirm the importance of the precautionary savings, but it is not clear whether the estimates can be applied to China, not to mention that Chinese culture remains a popular explanation for the Chinese Saving Puzzle.
The launch of public health insurance programs in China provides natural experiments to investigate  the size of the precautionary savings in China. This paper exploits one of most important policy changes in the rural areas: the introduction of the New Cooperative Medical Scheme (NCMS) since July 2003. This public health insurance program is heavily subsidized by the government, and has been introduced sequentially in different c
ounties. Households’ participation is voluntary. We focus on the double-difference comparison between the insured and the non-participants in the villages that have launched the program. The reason is that households in the same village are more comparable to each other than to households in a different village and it can reduce the bias resulted from the contemporary policy changes that were introduced simultaneously with the NCMS.
The difference-in-difference framework helps remove all the time-invariant selection bias. Selection bias on the observables is further reduced by allowing the temporal change in consumption to vary with income and health status or by applying matching difference-in-difference. The data we use combine the longitudinal Rural Fixed-Point Survey between 2003 and 2006 and a household survey on the NCMS for a subsample of the 2006 round of the Rural Fixed-Point Survey.
The results indicate that household consumption other than health expenditure has increased by about 5.6 percent or 147.7RMB owning to the health insurance coverage. T
he magnitude is much larger than the average cost of the insurance that was mostly 30 RMB in 2003 and 50RMB in 2006. This is also consistent with the time trend of the saving rate at the aggregate level: the saving rate in the rural area has declined sharply since 2005 (Prasad, 2009). The result is quite robust to different specifications that try to control the difference in the counterfactual trend of consumption between the insured and the non-participants.

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。