Empirical analysis of dynamic models with multiple agents.

详细信息   

• 作者：Schmidt-Dengler ; Philipp.
• 学历：Doctor
• 年：2005
• 导师：Berry, Steven
• 毕业院校：Yale University
• 专业：Economics, General.
• ISBN：9780542394812
• CBH：3194709
• Country：USA
• 语种：English
• FileSize：5305340
• Pages：136

文摘

This dissertation consists of three essays on the empirical analysis of dynamic models with multiple agents.;The first essay studies the adoption of nuclear magnetic resonance imaging (MRI) by US hospitals. I consider a timing game of new technology adoption. The dynamic game allows me to take strategic interaction in the timing decisions into account. Using a panel data set of US hospitals, cross sectional variation in adoption times, market structure and demand is exploited to recover the profit and cost parameters of the timing game. In counterfactual experiments I decompose the cost of competition into a business stealing and a preemption effect. I find substantial changes in adoption times and industry payoffs due to competition. These changes are mostly due to a business stealing effect. Preemption accounts for a significant but small share of this change.;In the second essay (joint with Martin Pesendorfer) we consider the estimation problem in dynamic games with finite actions. We derive the equation system that characterizes the Markovian equilibria. We consider a class of asymptotic least squares estimators defined by the equilibrium conditions. This class provides a unified framework for a number of well known estimators including Hotz and Miller (1993) and Aguirregabiria and Mira (2002). We show that these estimators differ in the weight they assign to individual equilibrium conditions. We derive the efficient weight matrix. A Monte Carlo study compares the finite sample performance of alternative estimators.;In the third essay I extend earlier work by Collard et al. (2002) who develop and structurally estimate a matching model that accounts for the asymmetries in the dynamics of job flows. The estimation scheme employed by Collard et al. produces unreasonably large implied unemployment rates, unless further parameter restrictions are imposed. I first address this issue by including employment rates as additional moments to be fit without imposing parameter restrictions. Since the new results cast doubt on the validity of the model, I improve the explanatory power of the model by extending it to allow for more heterogeneity. This achieved by extending the model to allow for an arbitrarily large number of employment pools.