The stylized fact for economies experiencing financial crises, that slow economic reform is followed by persistent stagnation, is usually explained as follows: Forbearance policy (i.e., an implicit subsidy to inefficient sectors) distorts resource allocation and causes a supply shortage of resources to the productive sectors. Since the Japanese economy has stagnated for over a decade even though Japan has had sufficient resources to allocate to productive sectors, I propose another explanation: Forbearance impedes the recovery of confidence that is lost during a financial crisis.
If confidence is restored through Bayesian learning by economic agents based on observations of government actions, then the inaction of the government (forbearance) impedes Bayesian learning.
Since it seems plausible to assume that economic agents face Knightian uncertainty (i.e., a non-unique prior) after a financial crisis, I use the framework of uncertainty aversion developed by Gilboa and Schmeidler (1989). The Bayesian update rule for a non-unique prior is proposed and used for the analysis of confidence building.
The model shows that the forbearance policy hinders the convergence of a non-unique prior and delays the economic recovery.