The recent global financial crisis has intensified calls for increased bank transparency. The discussions on whether the results of banks' stress-tests should be the publicized suggest, however, that the case for increased transparency is not clear-cut. The purpose of this paper is to provide a simple setting to study the effect of increased transparency on the credit available to refinance banks' risky assets, on banks' risk taking, and on welfare. Our results suggest a reason for the elusiveness of transparency: its effects vary over the business cycle, even in the absence of any price or collateral effects.
We first show that given a bank's investment profile greater disclosure of the quality of the bank's asset portfolio, e.g., publication of stress-test results, has a procyclical effect on creditors' confidence: Disclosure encourages creditors to rollover their credit when the expected returns of the investments are high, but discourages credit rollover when the expected returns are low. This is intuitive: when the returns of banks investments are likely to be high, transparency helps spreading the good news to creditors. But if the returns are likely to be low, opacity helps limiting the negative impact of bad news on credit rollover.
We also study the effect of transparency on banks' risk taking incentives. If the expected returns of investments are increasing in the level of risk, transparency has two effects on the incentives to take risk. The first effect is one of the key justifications for increased bank transparency: increased transparency discourages risk taking as it enhances market discipline, making the banks' creditors more sensitive to the banks' risk profile. The second effect is more nuanced: the bank's risk choices may reduce or amplify the effect of transparency on creditors' confidence. In a recession, increased transparency tends to discourage creditors to rollover their credits, which may lead banks to take more risk in order to compensate this effect. In a boom, where increased transparency encourages creditors to rollover their credits, there is less need to take more risk in order to compensate for the scarcity of credit. Hence this second effect reinforces the procyclical effects of increased transparency. If the risk level does not affect the expected returns of banks' investments then the first effect is eliminated, and only procyclical elements remain. In other words, increased transparency may make banks more prudent in downturns, but only if the mean asset returns are moderately sensitive to risk level.
Irrespective of whether increased transparency leads to more or less risk taking, the socially desirable level of transparency is the minimal one in recessions and the maximal one in booms. The reason is that competitive banks' asset choices maximize their creditors' utility, and therefore banks' asset choices are irrelevant for social welfare. Hence transparency has a procyclical effect on welfare solely because it affects creditors' confidence procyclically. We show that these welfare conclusions of bank transparency are fairly robust, even if we add social costs to bank failures or to premature liquidations of banks' assets.