Ebook The Structure of Adjustment Costs in Mainframe Computer Investment
Adjustment costs are a central analytical concept of investment models because they bring an intertemporal logic to a firm's input decisions. Prior investment models, however, give only ad hoc treatment to the concept of adjustment costs and explain that adjustment costs are costs of firm specific capital installation activities. This paper provides an empirical investigation into the source of investment adjustment costs by paying serious micro-level attention to the capital installation processes and their dynamics. I estimate a new structural model of lumpy investment using micro level data and a dynamic programming technique.
The model analyzes investment by its frequency and size and allows me to identify lump-sum adjustment costs as a set of structural parameters. I further investigate how the adjustment costs differ across organizations. The results show that adjustment costs arise from the specificity of the complementary assets.
The subject of this empirical study is mainframe computer investment. With observations drawn from a panel data set of business data-processing sites in the United States in particular, this paper analyzes computer capital installations in relation to their complements such as software, human capital, and other forms of organizational knowledge that support the information systems. I identify two classes of activities that support computer implementations. One is engineering activity that produces in house application software and customized databases.
The other is organizational accommodation needed for implementing new computer systems and applications, such as changing users' work routines and the information or incentive structure of the firm. I infer the extent to which these two sets of activities complement adjustment processes from different combinations of software in use at these sites. I empirically investigate whether these organizational differences explain the variation in size and frequency of lumpy investments that are observed in micro-level panel data.
For each establishment, adjustment size and frequency of lumpy investment jointly reveal information about investment hysteresis and lump-sum adjustment costs. So, I construct a model that predicts investment on both dimensions. This model and its estimates are based on the optimal investment rules that economic theory predicts under lump-sum adjustment costs and uncertainty. This model enables me to characterize the structural relationship between investment patterns and adjustment costs which neither cross-section analysis nor duration dynamics alone can capture.
The optimal decision rule predicts lumpy action after a sequence of inactive periods. This rule captures the essential relationship between the option value and investment timing in irreversible decisions whose economic relevance has been argued by Dixit and Pindyck (1994). In order to apply the model to a panel data of investment, I derive the optimal investment rule by dynamic programming and estimate parameters characterizing the heterogeneity of lump-sum adjustment costs.
Download
PDF Ebook The Structure of Adjustment Costs in Mainframe Computer Investment
Posted in :
