Using the 2006 wave of the Chinese General Social Survey, the present paper examines the effect of risk attitudes on the likelihood of entrepreneurship in China. Our results show that risk attitudes have a nonlinear effect on the likelihood of being entrepreneurs. Risk neutral people are most likely to be entrepreneurs, while both risk averse and risk seeking people prefer to work for wages. When we further divide entrepreneurs into necessity and opportunity entrepreneurs, we find only a marginal difference in risk attitudes between wage workers and necessity entrepreneurs, while less risk averse individuals tend to be opportunity entrepreneurs. Our results have important policy implications for the government's efforts to promote entrepreneurial activities. Caliendo et al., 2009; Dohmen et al., 2011). 2 To the best of our knowledge, there exist few studies (e.g. Djankov et al., 2006;Yueh, 2009;de Mel et al., 2010) regarding the impact of risk attitudes on the choice of self-employment in developing countries. One reason may be the lack of good-quality data on individual risk attitudes in developing countries. Nevertheless, risk attitudes may play a much larger role in entrepreneurship in developing countries as a result of the instability of political and economic environments and imperfect markets (de Mel et al., 2010).Although there is general agreement in the existing literature that there is a negative relationship between risk aversion and self-employment, some studies (e.g. Tucker, 1988;Parker, 2008;de Mel et al., 2010) also report no significant effect of risk attitudes on the possibility of self-employment. One explanation for the inconsistent findings is that previous studies examine different subgroups of entrepreneurs (Aguilar et al., 2013). If risk preferences are heterogeneous across subgroups, we may not be able to justify any general conclusion regarding the relationship between risk attitudes and entrepreneurship as a whole. 3 Another possibility is that the existing literature only pays attention to the linear risk attitudes effect. This practice might be problematic because personality traits including risk attitudes can have a nonlinear effect on labor market outcomes (Heckman and Kautz, 2012). 4 As the largest developing country in the world, China is an ideal case to study the