Therefore, there could be bias in the estimates if we were to use the ratio of bank loans normalized by total assets. To address this problem, we also specify the growth rate of bank loans, which is defined as [(total borrowings in year t − total borrowings in year t − 1)/total assets in year t − 1] as a proxy for Bank Loans i,t. Guaranteei is a dummy variable that takes a value of one if a firm satisfies the requirements for inclusion in the ECG program. i,tis the error term for firm i in year t from 2006 to 2009. is firm is fixed effects. We regard the observations for firms where the guarantee dummy is equal to one as the treatment group.