The 2007-2009 crisis stressed the importance of liquidity for banks. Using a risk factor model, we propose a measure of bank exposure to liquidity risk based on their sensitivity to aggregate liquidity conditions. Results indicate that liquidity risk is a specific risk. Moreover, this measure sheds light on the heterogeneity among banks in terms of exposure to liquidity risk. Banks benefit, lose or are insensitive to liquidity conditions, and we document large variation in exposure across the 2008 and 2011 crises. Larger size and capital levels tend to insulate banks from aggregate liquidity risk. However, deposit share, reliance on wholesale funding and funding gap impact only banks whose risk decreases with increasing aggregate liquidity risk. These ratios indicate the level of liquidity production by banks. This suggests that market discipline applies to liquidity production but only on the less risky banks in case of a liquidity crisis. Thus market discipline appears to be one-sided. To that extent it reinforces the necessity to impose liquidity requirements to all banks, as through the Basel III liquidity ratios.

Trust towards banks plays a central role in theoretical literature. Diamond and Dybvig (1983) argue that in a trustworthy environment banks can easily collect deposit foster banking activity and asset transformation. Diamond and Rajan (2001) posit that a high trust environment discourages banks from creating liquidity. To address these conflicting views, the current study measures liquidity creation using Berger and Bouwman's (2009) methodology, then assesses the level of trust in the environment with four proxies and two additional instruments deployed in previous research. The results confirm a positive effect of trust in banks on liquidity creation, especially for small or state-chartered banks and during economic downturns. The results are robust to time effects and potential endogeneity concerns.