Limits Points of Endogenous Misspecified Learning

Fudenberg, Drew; Lanzani, Giacomo; Strack, Philipp

doi:10.2139/ssrn.3553363

Cited by 15 publications

(9 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The “if” part is enough for all our results except the nonmyopic version of Theorem 1. In the Supplemental Material (Fudenberg, Lanzani, and Strack (2021)), Section B.2, we show how this result can be extended to the case where the “only if” part is not satisfied.…”

mentioning

confidence: 86%

Limit Points of Endogenous Misspecified Learning

Fudenberg¹,

Lanzani²,

Strack

2021

ECTA

Self Cite

View full text Add to dashboard Cite

We study how an agent learns from endogenous data when their prior belief is misspecified. We show that only uniform Berk–Nash equilibria can be long‐run outcomes, and that all uniformly strict Berk–Nash equilibria have an arbitrarily high probability of being the long‐run outcome for some initial beliefs. When the agent believes the outcome distribution is exogenous, every uniformly strict Berk–Nash equilibrium has positive probability of being the long‐run outcome for any initial belief. We generalize these results to settings where the agent observes a signal before acting.

show abstract

mentioning

confidence: 86%

Limit Points of Endogenous Misspecified Learning

Fudenberg¹,

Lanzani²,

Strack

2021

ECTA

Self Cite

View full text Add to dashboard Cite

show abstract

“…Thus, the question of convergence remains open at this level of generality. Recently there has been progress tackling convergence, but all in the context of static environments where the only relevant state variable is the agent's belief (Fudenberg et al 2017, Heidhues et al 2018, 2021, Esponda et al 2019, Frick et al 2020a, and Fudenberg et al 2020).…”

Section: Equilibrium Foundationmentioning

confidence: 99%

Equilibrium in misspecified Markov decision processes

Esponda¹,

Pouzo²

2021

View full text Add to dashboard Cite

We provide an equilibrium framework for modeling the behavior of an agent who holds a simplified view of a dynamic optimization problem. The agent faces a Markov decision process, where a transition probability function determines the evolution of a state variable as a function of the previous state and the agent's action. The agent is uncertain about the true transition function and has a prior over a set of possible transition functions; this set reflects the agent's (possibly simplified) view of her environment and may not contain the true function. We define an equilibrium concept and provide conditions under which it characterizes steady‐state behavior when the agent updates her beliefs using Bayes' rule.

show abstract

“…Complementary to their paper, we focus on beliefs rather than action frequencies, and we derive sufficient conditions for belief convergence that can be checked from static primitives without considering dynamics (such as their differential inclusion). In more recent work, Fudenberg, Lanzani, and Strack (2020) also study a general finite-action model and focus on action convergence instead of beliefs. They propose criteria for the stability of Berk-Nash equilibrium actions, partly building on the p-dominance method we introduce in this paper.…”

Section: Related Literaturementioning

confidence: 99%

Stability and Robustness in Misspecified Learning Models

Frick¹,

Iijima²,

Ishii

2020

SSRN Journal

View full text Add to dashboard Cite

We present an approach to analyze learning outcomes in a broad class of misspecified environments, spanning both single-agent and social learning. Our main results provide general criteria to determine-without the need to explicitly analyze learning dynamics-when beliefs in a given environment converge to some long-run belief either locally or globally (i.e., from some or all initial beliefs). The key ingredient underlying these criteria is a novel "prediction accuracy" ordering over subjective models that refines existing comparisons based on Kullback-Leibler divergence. We show that these criteria can be applied, first, to unify and generalize various convergence results in previously studied settings. Second, they enable us to identify and analyze a natural class of environments, including costly information acquisition and sequential social learning,

show abstract

Limits Points of Endogenous Misspecified Learning

Cited by 15 publications

References 30 publications

Limit Points of Endogenous Misspecified Learning

Limit Points of Endogenous Misspecified Learning

Equilibrium in misspecified Markov decision processes

Stability and Robustness in Misspecified Learning Models

Contact Info

Product

Resources

About