Safety has been a consistent issue with the rise of transportation network companies (TNCs), especially in China with annual revenue exceeding 1.5 trillion CNY. In this paper, we explore intrinsic factors that can influence the risky behavior of TNC drivers by investigating the link between their economic preference parameters and traffic violations. We measure the economic preferences of 160 full-time drivers on a large Chinese TNC platform and examine their violations over the previous 13 months. We have four major findings. First, more risk-averse drivers have less violations. Second, present bias and patience do not affect drivers' risky behavior except that more patient drivers commit more direction|sign violations, which may have been caused by the higher expectation of the gain from such a violation of patient drivers. Third, reciprocity reduces violations of all types. Drivers' belief of other people's pro-social inclination only affects dangerous violations with long duration, i.e., speeding, in which case they take advantage of other people's attentiveness. And finally, we find no evidence of driver fatigue. These results highlight how the advantage of TNC platforms in managing drivers' incentive can affect the negative externality drivers impose through risky driving behavior, and hence provide policy implications.