The purpose of this paper is to provide a specific test of Boucher, Carter et al. (2008) framework on risk rationing. The data were collected through a survey of 730 farm households in Shaanxi province conducted in November 2010. We compare factor associated with risk rationed, quantity rationed and price rationed farmers. Seemingly unrelated regressions are performed using risk rationing, quantity rationing and price rationing measure as the dependent variable and measures of demography, wealth, income, year of farming and risk aversion as independent variables. We apply seemingly unrelated regression, cluster analysis and cross tabulation in the study. According to a seemingly unrelated regression, we find existing risk rationing is due to risk-based behavior by borrowers. A cross tabulation results support the proposition by Boucher, Carter et al showing the financial wealthy is risk rationed and relatively land-poor is risk rationed. This paper is believed to be among the first empirical validation of the risk rationing theory.