We study how bank liberalization affects analyst forecast accuracy using two interest rate deregulations in China—the removal of the cap on bank lending rates in 2004 and the removal of the floor in 2013—as quasi-natural experiments. Our results show that the analyst forecast accuracy for high-risk firms decreases significantly after the removal of the lending rate cap, whereas analyst forecast accuracy for low-risk firms increases significantly after the removal of the lending rate floor. Moreover, interest rate liberalization affects forecast accuracy through operational risk and information asymmetry channels. Furthermore, the impact was concentrated on firms whose actual performance fell short of performance expectations and those that received more bank loans. Our findings imply that interest rate liberalization policies may have unintended consequences for analyst forecasts.

Utilizing the removal of the bank lending interest rate floor (IRFD) in China as an exogenous shock of banking liberalization, we find that IRFD leads to a significant rise in firms’ cost of equity capital, which is consistent with the prediction from the MM theory. The identified effects are more pronounced among firms with weaker ex-ante corporate governance and more severe ex-ante agency problems. We also find that IRFD witnesses an increase in the amount of acquired bank loans, a decrease in the average interest rate, and an increase in free cash flow. Further evidence also suggests IRFD provokes a drop in firms’ investment quality. Overall, our findings highlight an unexplored role of banking sector deregulation on firms’ cost of equity capital.

We propose a model of bank monitoring and borrower financial misreporting. Using the staggered liberalization of the banking sector in China as a natural experiment, we find that, consistent with the model’s prediction, entry by more efficient foreign banks reduces corporate misreporting fraud. Fraud reduction is greatest among borrowers of foreign banks, but fraud also drops among borrowers of domestic banks, suggesting a spillover effect. As predicted by the model, fraud reduction is greatest for borrowers with higher levels of fixed assets or lower levels of working capital. Our evidence suggests that improved bank monitoring reduces financial misreporting.