Working Papers

Works in Progress

  • Elasticity Estimation in Discrete Choice Models with Potential Demand Misspecification (with E. Segarra)
    Abstract: Estimates of elasticities are a common output of interest in discrete choice models. These estimates can be significantly biased when the potential demand is misspecified. This paper decomposes the bias in elasticity estimates of the logit model into a direct effect, from the misspecified potential demand, and an indirect effect, from bias in the structural parameter estimates induced by the misspecification. These effects can go in opposite directions so removing the bias from the indirect channel will have an indeterminate effect on the total bias in the elasticity. We provide a complete characterization of when correcting for the indirect effect will mitigate versus exacerbate the elasticity bias. Using this characterization, we provide two prescriptions for empiricists that dictate when removing bias from the indirect effect will reduce total bias in own-price elasticity estimates.

  • How Costly is Switching Auditors? Evidence from a Structural Estimation (with F. Celentano)
    Abstract: This paper studies the economic costs that public firms incur from switching auditors. We find that firms switching auditors typically gain a reduction in their fees. However, public firms often do not change auditors because of switching costs. In addition to typical switching costs associated with searching for and onboarding an auditor, firms also incur a reputation cost when replacing their auditors. The reason for the reputation cost is that investors view auditors as performing experiments to determine if a firm’s books are accurate and the audits are less likely to miss poor accounting practices when the auditor is more familiar with the books. In contrast to the literature which looks at stock prices and borrowing costs to understand switching costs, we plan to measure the costs by estimating a dynamic structural model using data on firms’ decisions of whether to change their auditors. This structural approach will contribute to the literature by measuring costs using auditor choice data, an arguably better approach, and by enabling us to measure what portion of the switching cost is attributable to the reputation effects. The model will also allow us to evaluate how proposed policy changes in the U.S., such as requiring firms to switch their auditors periodically, will affect firms’ costs.