Bounded Attention and Attenuated Elasticities
创建于 更新于
摘要
This paper investigates how bounded attention distorts the estimation of the elasticity of substitution in a CES demand system. In a sparse-max (bounded attention) framework, equilibrium prices and expenditure shares are observationally equivalent to those generated by a fully rational market, but the elasticity identified from data equals the structural elasticity scaled by the attention weight. The standard covariance-based estimator (Feenstra, refined by Broda–Weinstein and Soderbery) identifies the attenuated elasticity, not the true elasticity, because the orthogonality condition on the supply side holds only at the attenuated value. Consequently, without external information, the estimator converges to the attenuated elasticity, implying a fundamental identification failure that is not resolved by larger samples or stronger instruments. Since attention cannot exceed one, the estimated elasticity yields a bona fide lower bound on the true elasticity. The paper derives these results in closed form for a representative consumer, validates them in simulated markets, and discusses policy instruments (e.g., salience-enhancing regulations) as potential sources to separate elasticity from attention [page::5][page::10][page::11][page::15].
