Estimating interrelated nonconvex adjustment costs

Abstract

Adjustment costs associated with firms’ acquirement or disposal of factors of production can make the individual firm unresponsive to changes in their environment. This is the reason why costs of adjustments is assumed to be one of the reasons why we observe firm and plant level adjustment patterns as unevenly distributed over time. Understanding what characterizes such costs is important for understanding firm level behaviour, but can also be important to understand dynamics of both capital and labor demand at macro level. Because several studies on adjustment costs and factor demand indicate correlation in the demand for production factors, this thesis aims to present a way to simultaneously estimate adjustment cost functions of two factors, including both convex and nonconvex components. Joint estimation also allows for possible existence of interrelations in the adjustment cost function. A likelihood function for estimation by Maximum Likelihood is derived, and results after estimation on simulated panel data is presented. The text shows how a likelihood function can be written to estimate adjustment cost parameters that can be traced directly back to a theoretical framework for adjustment costs and factor demand. It is also shown that under certain conditions, the procedure inhabits weaknesses regarding identification of model parameters, which should be improved for increased robustness.