Generalized Models of Rational Inattention and Their Applications
Paper Session
Tuesday, Jan. 5, 2021 12:15 PM - 2:15 PM (EST)
- Chair: Jennifer La'O, Columbia University
Information Acquisition, Efficiency, and Non-Fundamental Volatility
Abstract
This paper analyzes non-fundamental volatility and efficiency in a class of large games (including e.g. linear-quadratic beauty contests) that feature strategic interaction and endogenous information acquisition. We adopt the rational inattention approach to information acquisition but generalize to a large class of information costs. Agents may learn not only about exogenous states, but also about endogenous outcomes. We study how the properties of the agents’ information cost relate to the properties of equilibria in these games. We provide the necessary and sufficient conditions information costs must satisfy to guarantee zero non-fundamental volatility in equilibrium, and provide another set of necessary and sufficient conditions to guarantee equilibria are efficient. We show in particular that mutual information, the cost function typically used in the rational inattention literature, both precludes non-fundamental volatility and imposes efficiency, whereas the Fisher information cost introduced by Hebert and Woodford [2020] generates both non-fundamental volatility and inefficiency.Optimally Imprecise Memory and Biased Forecasts
Abstract
We propose a model of optimal decision making subject to a memory constraint. Theconstraint is a limit on the complexity of memory measured using Shannons mutual
information, as in models of rational inattention; but our theory differs from that of
Sims (2003) in not assuming costless memory of past cognitive states. We show that
the model implies that both forecasts and actions will exhibit idiosyncratic random
variation; that average beliefs will also differ from rational-expectations beliefs, with
a bias that fluctuates forever with a variance that does not fall to zero even in the
long run; and that more recent news will be given disproportionate weight in forecasts.
We solve the model under a variety of assumptions about the degree of persistence of
the variable to be forecasted and the horizon over which it must be forecasted, and
examine how the nature of forecast biases depends on these parameters. The model
provides a simple explanation for a number of features of expectations in laboratory and
field settings, including both the over-reaction to news and inertia in elicited forecasts
documented by Afrouzi et al (2020).
The Cost of Information
Abstract
We develop an axiomatic theory of information acquisition that captures the ideaof constant marginal costs in information production: the cost of generating two
independent signals is the sum of their costs, and generating a signal with probability
half costs half its original cost. Together with Blackwell monotonicity and a continuity
condition, these axioms determine the cost of a signal up to a vector of parameters.
These parameters have a clear economic interpretation and determine the difficulty
of distinguishing states. We argue that this cost function is a versatile modeling tool
that leads to more realistic predictions than mutual information.
Discussant(s)
John Leahy
,
University of Michigan
Alessandro Pavan
,
Northwestern University
Tommaso Denti
,
Cornell University
Alexander Frankel
,
University of Chicago
JEL Classifications
- E7 - Macro-Based Behavioral Economics
- D8 - Information, Knowledge, and Uncertainty