Demand Estimation in Models of Imperfect Competition

Alexander  MacKay; Nathan H. Miller

Working Paper | HBS Working Paper Series | 2018

Demand Estimation in Models of Imperfect Competition

by Alexander MacKay and Nathan H. Miller

Abstract

We consider the identification and estimation of demand systems in models of imperfect competition. Under standard assumptions about demand and supply, the bias that arises from price endogeneity can be resolved without the use of instruments. We provide a constructive identification result where the causal price parameter can be expressed as a function of the covariance of unobserved shocks. The function is estimated efficiently by the output of ordinary least squares regression. Thus, with a covariance restriction on unobservable shocks, structural parameters can be point identified. Further, it can be possible to place bounds on the structural parameters without imposing a covariance restriction. We illustrate the methodology with applications to ready-to-eat cereal, cement, and airlines.

Keywords: demand estimation; identification; endogeneity bias; covariance restrictions; ordinary least squares;

Language: English

Format: Print

61 pages

SSRN Read Now

Citation:

MacKay, Alexander, and Nathan H. Miller. "Demand Estimation in Models of Imperfect Competition." Harvard Business School Working Paper, No. 19-051, October 2018.

The duration of a vertical relationship depends on two types of costs: (i) the transaction costs of re-selecting a supplier and (ii) the cost of being matched to an inefficient supplier when the relationship lasts too long. For commodified goods and services, this tradeoff can be the primary determinant of the duration of supply contracts. I develop a model of optimal contract duration that captures this tradeoff, and I provide conditions that identify underlying costs. Latent transaction costs are identified even when the exact supplier selection mechanism is unknown. I estimate the model using federal supply contracts and find that transaction costs are a significant portion of total buyer costs. I use the structural model to estimate the value of the right to determine duration to the buyer, compared to a standard duration. Finally, a counterfactual analysis illustrates why quantifying transaction costs is important for the accurate analysis of welfare.

Economic theory suggests that demand is more elastic in the long run relative to the short run, but evidence on the empirical relevance of this phenomenon is scarce. We study the dynamics of residential electricity demand by exploiting price variation arising from a natural experiment: the introduction of an Illinois policy that enabled communities to select electricity suppliers on behalf of their residents. Using a flexible difference-in-differences matching approach, we estimate a one-year price elasticity of -0.16 and three-year elasticity of -0.27. We also present evidence that consumers increased usage ahead of these announced price changes. Finally, we project that the price elasticity converges to a value between -0.30 and -0.35 after ten years. Our findings highlight the importance of accounting for consumption dynamics when evaluating energy policy.

This article discusses the empirical challenges that researchers face when demonstrating the existence and effects of resale price maintenance (RPM). We outline three approaches for finding price effects of RPM and the corresponding hurdles in data and methodology. We show that the quantity test that was suggested by Posner (1977; 1981) does not identify the change to welfare when demand-enhancing effects are considered generally. Finally, we present some solutions to the challenge of identifying welfare effects, and we suggest guidelines for future research.

Faculty & Research

Demand Estimation in Models of Imperfect Competition

Abstract

Citation:

About the Author

More from the Author