The growth of healthcare expenditure provokes constant comments and discussions, as countries battle the issues on cost containment and cost effectiveness. Prior to 1978, medical institutions in China were either state-owned or were collective public hospitals. Since 1978, China has been trying to rebuild its healthcare system, which was destroyed during the ‘cultural revolution’, allowing private medical institutions to deliver healthcare services. As a result, private medical institutions have grown from 0% to 28.57% between 1978 and 2010. In this context, we compare outpatient healthcare expenditures between public and private medical institutions. The central problem of this comparison is that the choice of medical institution is endogenous. So we apply an instrumental variable (IV) framework utilizing geographic information (whether the closest medical institution is private) as the instrument while controlling for severity of health and other relevant confounding factors. Using China's Urban Resident Basic Medical Insurance Survey 2008–2010, we found that there is no difference in expenditure between public and private medical institutions when IV framework is used. Our econometric tests suggest that our IV model is specified appropriately. However, the ordinary least square model, which is inconsistent in the presence of endogenous regressor(s), reveals that public medical institutions are more expensive. Copyright © 2013 John Wiley & Sons, Ltd.