How to Solve Dynamic Stochastic Models Computing Expectations Just Once

By Kenneth Judd, Lilia Maliar and Serguei Maliar

http://d.repec.org/n?u=RePEc:nbr:nberwo:17418&r=dge

We introduce a technique called “precomputation of integrals” that makes it possible to compute conditional expectations in dynamic stochastic models in the initial stage of the solution procedure. This technique can be applied to any set of equations that contains conditional expectations, in particular, to the Bellman and Euler equations. After the integrals are precomputed, we can solve stochastic models as if they were deterministic. We illustrate the benefits of precomputation of integrals using one- and multi-agent numerical examples.

My selection this week is more about technique than answering a question. Yet, this may be an interesting development in simplifying the computations of equilibria of models with large state spaces (many agents), especially for those with idiosyncratic and aggregate shocks.