The Impact of Information Technology

Evidence from the Healthcare Industry

Nirup M. Menon

• Publisher: Taylor & Francis
• Year: 2020
• Print ISBN: 9780815334477
• eBook ISBN: 9781136752292
• Edition: 1
• Topic: Business
• Subtopic: Business General
• Index: Business

First published in 2000. This book addresses the measurement of the effect of information technology (IT) investments on a firm's productivity. Determining a quantifiable impact of a firm's IT has plagued senior executives, researchers, and policy-makers for several years, as evidenced by articles in trade magazines such as Fortune and Businessweek and in academic journals such as Management Science. Simple statistical techniques for measuring IT impact in a firm are fraught with methodological problems, as these techniques do not account for either the causal direction in managerial decision making or the behavioral assumptions about firms. Therefore, such studies have led to results and inferences that are not generalizable. While studies that measure the satisfaction of people who use IT are important, management typically would like to know whether IT has reduced operation costs by streamlining processes or increased revenues by increasing the demand-meeting capability of the firm. This book attempts to determine cost-reduction or output-enhancement that may be linked to IT investments through methodological sophistication. The healthcare industry presents an important and interesting context in which to study IT impacts for several reasons. First, since the implementation of the Prospective Payment System (PPS) by Medicaid, most hospitals adopted cost containment measures, and hence capital investments in hospitals have come under greater scrutiny than ever before. Second, hospitals have been more thorough in reporting capital and labor expenses and revenues (due to state regulation) at a level of detail that makes it possible to aggregate IT and other capital investments without serious measurement error. Most non-healthcare firms do not collect or report such data in their financial statements. Finally, though hospitals were slow in IT adoption, most hospitals have been acquiring sophisticated hardware and software over the past few years. Results of the analysis bear evidence of the positive impact of IT on production of healthcare services. It also shows how methodological differences can lead to conflicting results. The effect of PPS determined in a comparative way shows that the economic behavior in the post-PPS differs from that in the pre-PPS years.

CHAPTER 1
Introduction

Information technology (IT) has become important to the very survival of organizations. Information technology spending in the U.S. has increased from a few million dollars in 1970 to over $50 billion in 1990 (Jorgenson and Stiroh, 1995). Industry analysts and economists have observed a “co-incidental” slowdown in the productivity of the U.S. economy during the same period (e.g., Bell et al., 1991). While it is possible that the slowdown may have been caused by a variety of reasons other than IT-related phenomena, the possibility that IT may be one of reasons for this slowdown has led researchers to attempt to extract a measure of the value of IT in the production of goods and services (Brynjolofsson and Hitt, 1994). It has been suggested that the slowdown in the productivity of the economy is a contrivance of the techniques used for observation, measurement, and analysis and that it is not a real phenomenon in itself (Romer, 1987). Studies also showed that investments in IT may have been misspent. These studies do so by showing that the productivity contribution of IT was empirically shown to be negative (e.g., Roach, 1987). Other explanations of the IT effect on general productivity include the finding that IT creates surplus for consumers but not for producers, thereby diluting the productivity contribution of IT (Breshnahan, 1986; Brynjolofsson and Hitt, 1996). These latter studies have added much to the understanding of the business value of IT.

In understanding the impact of IT on the economy, history lends insight when we compare the technological revolution of computers and information systems with the technological revolution that stemmed from innovations such as steam engines (Simon, 1986) or the dynamo (David, 1990). The instantaneous productivity improvements that were expected from the steam engine and the dynamo were not observed partly because of the need to re-organize the production operations. Introduction of IT has led to similar attempts of reorganization (Hammer and Champy, 1992). The two factors—(1) the need for observing the changes in the use of new technology over time and (2) the complexity of analytical treatment of IT impact on firms—lead to the conclusion that the effect of IT on the productivity of the firm, industry, and economy is best judged using longitudinal, empirical studies.

In general, productivity growth and the contribution of inputs to productivity growth has been empirically measured using index numbers (Harris and Katz, 1991) or econometric techniques (Schmidt and Lovell, 1979; Kumbhakar, 1990). Among the several major differences between the two techniques, the index numbers technique assumes away inefficiency in production which econometric techniques are capable of modeling, whereas econometric techniques make several assumptions to simplify computation and are subject to specification error. Since productivity and efficiency are inextricably linked, the productivity approach is a more appropriate technique than the index number approach for modeling inefficiency of production that results from bounded rationality, regulation and labor unionization. Managerial inefficiency can take two forms—technical inefficiency that results from the inability of the production process to produce the maximum output for a level of inputs used (due to poor engineering design, regulation, etc.) and allocative inefficiency that results from selecting a technically efficient input mix that is not optimal in cost (due to labor unionization, the inability to observe prices of inputs, etc.).

Productivity analysis of IT can be conducted at an economy, industry, firm or process level. While economy- and industry-level analyses may confound results due to aggregation of data and heterogeneity of units of analysis, process-level analyses may suffer from other equally serious problems such as insufficient sample size and the validity in the generalization of results. Since IT productivity in a firm cannot be determined in isolation but has to be determined in relation to other factors such as human resource strategy and external economic institutions (Osterman 1990), a multifactor, firm-level productivity analysis seems to be the most appropriate line of inquiry. In a firm-level analysis, a sample of homogeneous firms should be chosen to minimize confounding from other factors when eliciting productivity contribution of IT. For example, the firms should belong to one industry.

One way to empirically elicit productivity contribution of inputs is to assume a functional form for the technology that transforms inputs into outputs. The second way is to compare the amounts of input used and outputs produced for each firm relative to other firms in the sample without specifying a functional form for the production technology. The former technique is called the parametric technique and the latter is called the nonparametric technique. In this study, the productivity of various input factors and the various efficiencies are determined by both parametric and nonparametric techniques. The results from the various models from both techniques are triangulated, and the interplay between the efficiencies and the organization’s environment is also studied.

The setting for the study is the healthcare industry. This industry makes an interesting arena because of the recent increase in hospital IT spending and cost-cutting initiatives in the industry. The healthcare industry is a highly regulated industry which allows the studying of regulation effects on investment patterns with possible insight for policymakers. Because the industry is heavily regulated, hospitals collect and maintain data on spending and earnings. This data is very suitable for a firm-level study. The organizational choices of hospitals, such as teaching status, also provide a context for studying the relation between organizational choice and IT productivity.

This book is organized as follows. In Chapter 2, I review literature pertinent to IT impact issues discussed above and in the rest of the book. In Chapter 3, I review the history and background of the healthcare industry. The field data used in the analysis is explained in Chapter 4. Chapter 5 contains the production function approach to studying IT impact whereas Chapter 6 contains a cost function approach to studying the effect of regulation on capital investments including IT investments. Chapter 7 contains a nonparametric approach to IT productivity and efficiency studies in an effort to triangulate results from parametric techniques in Chapters 5 and 6. Chapter 8 contains a discussion of results and concludes the book with remarks on future research directions.

CHAPTER 2
Productivity Literature

2.1 BUSINESS VALUE OF IT

A manager of a management information systems (MIS) department in a firm is concerned with measuring the success of the department. He/she can evaluate a specific project’s success or the department’s success from various perspectives such as system quality, information quality, the rate of use of the system, user satisfaction, individual impact and organizational impact (DeLone and McLean 1992). The former five are considered internal to the firm in the sense that these are indicators of performance of intermediate business/social processes that lead to the realization of one or more final organizational goal(s). On the other hand, organizational impact is a direct indicator of the impact of IT on the realization of the final organization goal and subsumes the effects from the other five success measures. Organizational impact of IT indicates the impact of IT investments or usage in relation to the organization’s external environment such as markets, consumers, competition, etc. Among the five perspectives listed above, organizational performance has been rated higher than the others by information systems (IS) practitioners as an indicator of IS performance (Brancheau and Wetherbe 1987). Organizational impact is a broad term that captures several dimensions of organizational performance. Three primary dimensions are productivity, profitability and consumer surplus (Brynjolfsson and Hitt 1996). In this study, I focus on the productivity dimension of IT as a measure of the business value of IT.

2.2 THE MODERN PRODUCTIVITY PARADOX

Before reviewing IT productivity literature, a foray into the modern productivity paradox literature is warranted. The paradox has come about as a result of the contradiction between the commonly held perception that technological innovations lead to higher productivity and the observed fact that the productivity of the advanced economies is slowing down (e.g., Bell et al., 1990). Since the initial discovery of this contradiction in the early 1980s, researchers and policymakers have devoted a large amount of time, several journal articles and seminars attempting to determine whether the contradiction is a statistical artifact or a real phenomenon. The unravelling of the paradox requires understanding the computer revolution and the impacts of IT in organizations.

David (1990, 1991) likens the computer revolution to the technological revolution fostered by the invention of the dynamo. He postulates that the postindustrial era can be divided into technoeconomic regimes punctuated by innovations in technology (David, 1991). He observes that analysts predicted dramatic changes in the technoeconomic system due to dynamos but failed to see the fruition of these changes due to the effort expended on the reorganization of factory plants and the work environment. He suggests that similarly the computer revolution will not directly and immediately translate into dramatic productivity improvements. The “slippage between advancing frontier of technology and actual practice” can only be overcome by significant reorganization and reconfiguration of productive activity over a period of time. The literature on business process re-engineering over the past decade is based on a similar vein of thought (e.g., Hammer and Champy, 1993).

Researchers have observed another surprising aspect during the productivity slow-down in economy. Since the late 1980s, the productivity of the manufacturing sector has been rising while the productivity growth of the service sector has become negative (Baily and Gordon, 1988). Adding to the conundrum is the fact that 80 or more percent of the capital in the service sector is IT capital. These observations and results intensify the debate on the impact of IT on firms. The following sections review the results of previous studies on IT productivity issues.

2.3 IT AND THE PRODUCTIVITY PARADOX

Since the inception of the use of IT in business, one primary area of IT application has been in transaction processing to speed up clerical work and to cut labor costs. That is, increasing productivity and reducing costs were primary reasons for using IT. However, until recently research has not shown that the use of IT leads to an increase in productivity. I believe that one of the problems is methodogical in that most previous studies use a growth-accounting or an index numbers approach to determining productivity wherein inefficiency in production is not modeled (which is indeed difficult to do at an aggregate economy or industry level).

In a study involving the impacts of any input (such as IT), measuring the productivity (by how much the output of a firm, industry, or economy has changed by the use of the input) is a central issue. IT productivity analysis can be approached from various levels pertaining to the unit of analysis. There have been economy-, industry-, firm-, and process-level analyses that have studied the productivity impact of IT. Economy-level studies have primarily been reported in economics journals. The primary concern of many of these studies is to investigate the decline in the productivity of the service sector in the 1970s and 1980s and to determine the reasons for the decline despite increasing investments in IT capital.

2.3.1 IT and Economy-Level Productivity

As described above, researchers were intrigued by the stagnation in the productivity of the economy since the 1970s. In particular, the productivity growth of the service sector has been negative and this has been so despite the large investments by the service sector in IT capital. One of the key measures used to study productivity has been average labor productivity at the economy level. Change in the average labor productivity (ALP) is defined as the change in the output for a unit change in labor-hours and change in multifactor productivity (MFP) is defined as the change in the output for a unit change in labor-hours and capital weighted by their shares in total cost (Baily and Gordon, 1988, p. 356). The problem with the ALP measure is that the output–labor ratio does not capture important firm-level effects such as substitution/complementarities between different types of capital and between capital and labor. Similarly, MFP suffers from problems such as the assumption of constant returns to scale of production and equality of nominal income share for each input and its output elasticity. In addition, aggregation at the economy level requires use of “aggregate” price deflators which are of highly suspect derivation and meaning.

The growth accounting framework uses a similar formulation but differs in that it is defined in terms of rates of change for output and inputs (e.g., see Oliner and Sichel, 1994). The assumptions are similar to those described in the previous paragraph as are the pitfalls. Studies in economy-level productivity suggest that we use caution in the areas regarding (1) the specification of input and output, (2) the data collection including price and quality information of complex capital such as computer and communication equipment, (3) the measurement of aggregate capital data, (4) the methodologies used such as growth accounting and index numbers, etc. A comprehensive review of the various levels of studies in IT productivity and the approaches used therein is contained in Brynjolofsson and Yang (1996).

2.3.2 IT and Industry-Level Productivity

An industry-level study of productivity is useful, especially from the viewpoint of separating the manufacturing and service sectors, because the intensity and uses of IT investments and the productivity growth in the two sectors have not been very similar. Therefore, any errors introduced in economy-level studies by the aggregation of data from dissimilar sectors can be controlled for by industry-level analyses. Most studies in either sector, such as Berndt and Morrison ...

Table of contents

• Cover
• Title Page
• Copyright Page
• Acknowledgments
• Contents
• List of Figures
• List of Tables
• Preface
• 1 Introduction
• 2 Productivity Literature
• 3 The Healthcare Industry
• 4 Empirical Data
• 5 Production Theory and Productivity Analysis
• 6 Policy Changes and IT Investments
• 7 Non-parametric Productivity Analysis
• 8 Conclusions
• Appendix A
• References
• Index