The credit rating industry has played a pioneering role in the broader opening of Chinese financial markets in recent years, but its fairness and consistency have been frequently doubted. This research studies rating consistency across Chinese credit rating agencies and discusses its underlying mechanism, motivation, and impacts. The results show that bonds with identical ratings given by different agencies differ in credit risk, with significant differences in issuance spreads, default rates, yield spreads, yield volatility, and probabilities of rating adjustments after issuance. However, given observable characteristics, ratings do not vary across agencies. This indicates that the rating inconsistency across agencies is a separating equilibrium under information asymmetry: Issuers with similar observable characteristics but differing credit risk choose different rating agencies, so that their choice of rating agencies consequently signals the bonds’ unobservable risk. Moreover, it is shown that the threshold rating requirements set by regulations motivate the rating inconsistency, which enables ineligible firms to obtain inflated ratings from aggressive agencies and thus successfully access bond markets. Using propensity score matching and Heckman two-stage approach to address endogeneity, further analysis reveals that this rating-inflated financing for regulatory arbitrage does not result in subsequent outperformance by bond issuers.