Most dynamic general equilibrium models assume that households can perfectly observe the state variables. Complete markets rationalize this assumption: in a decentralized equilibrium households learn about aggregates through participating in markets, so if markets are complete so too will be information. However if markets are incomplete, households will in general be imperfectly informed about the aggregate economy, and hence about other agents. This means that rational households have to form expectations of aggregate states, and of other households’ behaviour, leading to an in?nite regress of expectations (Townsend, 1983, Woodford, 2003, Nimark 2007a,b).
To investigate these issues, we describe a simple dynamic general equilibrium model in which households are heterogenous because they face an idiosyncratic productivity shock in addition to an aggregate productivity shock. If capital is the only tradeable asset, households’ information is limited to a knowledge of their own allocations, along with the prices they observe from participating in labour and capital markets. We describe such households as solving a signal extraction problem using a version of the Kalman ?lter that allows for endogeneity of the states (Baxter, Graham and Wright, 2007) and explicitly model higher-order expectations using techniques developed by Nimark (2OU7a).
We make two main contributions to the dynamic general equilibrium literature:
- 1. We show how to model information consistently in a standard dynamic general equilibrium framework and that noise is not necessary to motivate imperfect information. In our model, the informational problem arises from the structure of the real economy.
2. We provide analytical and numerical results to show that incomplete markets can dramatically change the dynamics of DGE models when their informational implications are considered.
Download
Information, heterogeneity and market incompleteness
