Masterclass by Robert A. Miller, The econometrics of optimal contracting , 12-13 March 2020

Econometricians are long familiar with the instrumental variables approach to estimating demand and supply curves, and through revealed preference theory recovering the utility functions of consumers and the production functions of firms, in partial and (to a lesser extent) general equilibrium models. However as workplace arrangements become more flexible, and as the institutional details of financial markets and auctions become more transparent, representing markets by simple demand and supply curves becomes a less appealing paradigm. Given the availability of larger and more detailed data sets, empirical methods for quantifying the legal and informal business infrastructure that supports transactions, is an increasingly important research area. This course analyzes identification and estimation in models of market microstructure and optimal contracting, with one eye on future reseach. It covers applications in labor supply, managerial compensation, procurement contracts and nonlinear pricing. To convey a sense of the problems planners face course participants will also play each other in several real time experiments.

Our point of departure is to exploit a well-known equivalence between a social planner’s problem and a competitive equilibrium with complete markets as the basis for estimation. We provide several examples of where this approach has been applied, and then turn to a pressing question: how should estimation proceed when there are market imperfections? We show some progress can be made by incorporating commonly acknowledged externalities within the estimated structural models used for policy prescription. However there is a practical side to the theorem that competitive equilibrium is not implementable (or in layman’s language there is no real world counterpart to the mythical Walrasian auctioneer): the equilibrium of estimated structural models characterizing financial limit order markets (where billions of dollars are traded each day) typically exhibit sizable deviations from a model where all gains from trade are realized. How serious are these losses and what can be done?

In situations where market failure occurs because payoff relevant contingencies are not contractible, typically the case when there is asymmetric information, a natural extension to the neoclassical approach is to assume that principals (such as shareholders and procurement agencies) design institutional arrangements that extract the maximal gains from trade with their counterparties given the evolving information and technology. This is called optimal contracting. The restrictions arising from the resulting equilibrium allocations and payments provide a basis for identifying and estimating the preferences, information and technology available to the agencies involved. We investigate situations where there is hidden information about demand (and the potential for price discrimination), moral hazard (that arises in the corporate world because shareholders and the managers or stewards of their wealth have different objectives), and adverse selection (perhaps because sellers of different cost and quality are not observed by a buyer), all potentially intertwined with considerations of specific human capital (in which some workers learn faster than others and firms cull from a selection of new recruits).

The econometrics of optimal contracting