In recent years, many lower to middle income countries have looked to insurance as a means to protect their populations from medical impoverishment. In 2003, the Chinese government initiated the New Cooperative Medical System (NCMS), a government-run voluntary insurance program for its rural population. The prevailing model of NCMS combines medical savings accounts with high-deductible catastrophic hospital insurance (MSA/Catastrophic). To assess the effectiveness of this approach in reducing medical impoverishment, we used household survey data from 2006 linked to claims records of health expenditures to simulate the effect of MSA/Catastrophic on reducing the share of individuals falling below the poverty line (headcount), and the amount by which household resources fall short of the poverty line (poverty gap) due to medical expenses. We compared the effects of MSA/Catastrophic to Rural Mutual Health Care (RMHC), an experimental model that provides first dollar coverage for primary care, hospital services and drugs with a similar premium but a lower ceiling. Our results show that RMHC is more effective at reducing medical impoverishment than NCMS. Under the internationally accepted poverty line of US$1.08 per person per day, the MSA/Catastrophic models would reduce the poverty headcount by 3.5-3.9% and the average poverty gap by 11.8-16.4%, compared with reductions of 6.1-6.8% and 15-18.5% under the RMHC model. The primary reason for this is that NCMS does not address a major cause of medical impoverishment: expensive outpatient services for chronic conditions. As such, health policymakers need first to examine the disease profile and health expenditure pattern of a population before they can direct resources to where they will be most effective. As chronic diseases impose a growing share of the burden on the population in developing countries, it is not necessarily true that insurance coverage focusing on expensive hospital care alone is the most effective at providing financial risk protection.