By Paul de Grauwe and Yuemei Ji
http://d.repec.org/n?u=RePEc:ehl:lserod:87286&r=dge
Dynamic stochastic general equilibrium models are still dominant in mainstream macroeconomics, but they are only able to explain business cycle fluctuations as the result of exogenous shocks. This paper uses concepts from behavioural economics and discusses a New Keynesian macroeconomic model that generates endogenous business cycle fluctuations driven by animal spirits. Our discussion includes two applications. One is on the optimal level of inflation targeting under a zero lower bound constraint. The other is on the role of animal spirits in explaining the synchronization of business cycles across countries.
I guess this is more an issue of semantics, but animal spirits still need to be triggered by something, however small that may be, and this something is still exogenous to the model, right? The dream of having a fully endogenous model is a pipe dream.
NEP-DGE Blog 
Yes, of course, but the trigger can be a trivial event. As a result, one obtains a very different dynamics than in standard DSGE-models. For example, our behavioural model is sensitive to initial conditions. This implies that the transmission of shocks (e.g. an interest rate shock produced by a change in policy) depends on the exact timing of the shock.
In addition, our model generates movements in output and inflation characterised by fat tails in their distribution (without having to “import” this feature from the stochastic shocks). This feature is also observed in the data. Standard DSGE-model can only produce such a dynamics by assuming it comes from outside the model.