Search

Financial crises are often associated with an endogenous credit reversal followed by a fall in asset prices and serious disruptions in the financial sector. To account for this sequence of events, this paper constructs a model where the excessive risk-taking of portfolio investors leads to a bubble in asset prices (in the spirit of Allen and Gale, "Bubbles and Crises", Economic Journal, 2000), and where the supply of credit to these investors is endogenous. We show that the interplay between the risk shifting problem and the endogeneity of credit may give rise multiple equilibria associated with different levels of lending, asset prices, and output. Stochastic equilibria lead, with positive probability, to an inefficient liquidity dry-up at the intermediate date, a market crash, and widespread failures of borrowers. The possibility of multiple equilibria and self-fulfilling crises is showed to be related to the severity of the risk shifting problem in the economy.

Four main lessons stem from the works presented in this volume. First, the rational expectation hypothesis has to go thoroughly into more general concepts. If the REH in a muthian sense seems now invalidated, this result does not mean that there is not rationality in price expectations : in the one hand, expectations may be economically rational in the sense of the advantage-cost analysis, and, in the other hand, the exchange of informations between agents through the market may involve some other types of rationalities. Secondly, it appears important to respect the individual nature of expectations both at the theoretical and empirical levels : generally, the heterogeneity is not neutral either for reaching an economic equilibrium or for econometrical estimations of expectational processes. Thirdly, expectational behaviors change over time ; both the processes and the parameters which intervene in these processes are changing. Fourthly, while expectational processes are rather extrapolative (destabilizing) when the horizon is short (less or equal to one month), they are rather regressive (stabilizing) when the horizon is long (greater than one month).