Le rôle de la garantie des dépôts dans la prévention des paniques bancaires : fondements théoriques et études empiriques

This paper reviews and discusses theoretical litterature on banking panics and the different methods to prevent them. Deposit coverage is compared to alternative solutions, some of which have been implemented in the past: partial or total suspension of deposit convertibility (whether or not carried out by a clearinghouse association) and narrow banking. Two major theories on bank runs stand out. On the one hand, purely self-fulfilling banking panics resulting from a coordination failure between depositors. Such runs are ineffective in so far as they may drive sound banks to bankruptcy and require a high deposit coverage to be prevented. On the other hand, ?information-based runs? theory shows that depositors, being informed of upcoming difficulties regarding their bank (information on fundamentals) run preventively to withdraw their funds. Policy implications will then be different. In order to test existing theories, emprirical works have focused first on historical studies about past banking panics, notably in the United States. This has lent support to the ?information-based runs? theory, even if the results are discussed. However, experimental research on self-fulfilling banking panics shows that such runs are recurrent and persistent phenomena. Besides, only a full coverage deposit insurance is an effective method to prevent such panics but a partial coverage is sufficient to curb ?information-based runs?. Finally, empirical studies on recent banking crises find that explicit deposit insurance can have an adverse impact on bank stability, unless there exists a quality institutional environment and banking regulation.