SGER: Dynamic Economic Models: Calibration, Simulation, and Testing

Arizona State University, Scottsdale AZ

Investigators

Abstract

This Small Grant for Exploratory Research addresses a central problem in quantitative economics: the estimation of parameter values in non-linear optimization models, where there seems to be a shortage of general, reliable estimation procedures. This research project is made up of two projects. The first develops a general framework for the estimation and evaluation of dynamic economic models. The second is concerned with the simulation of a dynamic principal-agent model to address some important issues in top executive compensation. Growth theory can provide an appropriate framework to study a wide variety of problems in macroeconomics and asset pricing. But an unsatisfactory aspect of this research endeavor is that model predictions are usually confronted with data in a rather casual way. The purpose of this project is to develop a general framework that allows for a rigorous application of statistical analysis for the evaluation, calibration and estimation of dynamic optimization models. Part of this analysis builds on important contributions by Lee and Ingram (1991) and Duffie and Singleton (1993). As these authors acknowledge, however, the conditions leading to their proofs of strong consistency and asymptotic normality of their simulated moments estimators are quite restrictive. The investigator's goal is to offer a general proof of strong consistency and to explore the properties of the stochastic solution in order to establish asymptotic normality of the estimator. This work is based on some approximation results, which are also useful to evaluate numerical simulations. Further progress in quantitative policy research requires the development of a general framework, as proposed here, for reliable estimation and evaluation procedures. The second part of this research project focuses on the simulation of a dynamic principal- agent model to address several issues on executive compensation. Even though the seminal papers by Abreu, Pierce and Stacchetti (1986, 1990) and Spear and Srivastava (1987) seem to have laid the foundations for dynamic contracting, it is widely agreed that the empirical implications of this theory have not been explored in a systematic way. Because of some stringent assumptions in some of these models, the evolution of the distribution of consumption and earnings becomes degenerate, which seems counterfactual. Moreover, relatively little work has been devoted to the calibration and simulation of dynamic principal-agent models and to the evaluation of their predictions regarding the structure of compensation, the sensitivity of pay to performance, and the costs and benefits of stock option contracts. The ultimate objective of this research is then to enhance our understanding of the workings and performance of some compensation schemes for the provision of insurance and incentives