Industry-Level Analysis of Information Technology Return and Risk: What Explains the Variation?

Abstract

Motivated by the wide dispersion in the returns on the use of information technology (IT) across industries, we conduct an industry-level examination of IT return and risk, focusing on the moderating role of industry competition, regulation, and technological change. We address the following research questions: What is the impact of IT investment on the return and risk dimensions of industry financial performance? How do industry characteristics moderate the relationship between IT investment and industry performance? Our analysis of these questions finds that higher levels of industry competition are associated with higher IT productivity (contribution of IT to value-added output), lower IT profitability (contribution of IT to industry average return on assets [ROA]), and higher IT risk (contribution of IT to ex ante variability of ROA). This is consistent with the notion that competition induces riskier IT investments, despite the fact that returns tend to be competed away. Higher levels of industry regulation are associated with lower IT returns in both productivity and profitability, but also lower IT risk. Finally, a higher rate of technological change induces both higher IT returns and higher IT risk. A variety of tests indicate that our results are robust. Our results shed light on factors that drive variation in IT performance across industries, and provide useful industry-level performance benchmarks of the return and risk impacts of IT investments.

Key words and phrases:

• competition
• industry-level analysis
• information technology investment
• IT return and risk
• IT value
• productivity
• profitability
• regulation
• technological change

Appendix

Table A1. List of Variables and Constructs

Notes

1. In this widely cited study by McKinsey Global Institute [51], positive productivity growth was found in only six sectors, despite substantial and growing IT investment in most industries.

2. ROA is calculated by dividing income before extraordinary items by total assets.

3. Indeed, tests on our data indicate significant heteroskedasticity and first-order autocorrelation.

4. Due to space constraints, we present only results from PSAR1. Results from AR1 are quite consistent with PSAR1, and are available upon request.

5. Two industries in the BEA list are dropped: Management of Companies and Enterprises cannot be matched with our other data sources, and Real Estate has extraordinarily high levels of capital investment.

6. Goh and Kauffman [33] used BEA data from an earlier time period, 1992 and 1997, before the redefinition of some of the data by the BEA in 1998.

7. To ensure robustness, we also use HHI to classify industry competitiveness and rerun all the empirical analyses.

8. We present IT productivity robustness checks for the full sample and the industry competition subsamples. Due to space constraints, results from regulation and technological change subsamples are not presented.

9. Omitted variables and self-selection are other reasons for endogeneity. However, we believe they are less of an issue in our research setting because PSAR1 and fixed effects can relieve the problem of omitted variables [15] and our data set covers all industries and each industry in its entirety, relieving the problem of self-selection.

10. Note that in our profitability and risk model specifications, independent variables including it are at year t and dependent variables are at year t + 1, making them less subject to the simultaneity problem.

11. More precisely, we used an earlier working paper version of Chwelos et al. [18].

12. In our IT productivity analysis, site-level IT capital figures were aggregated to the firm level and then aggregated up to the industry level. In IT profitability and risk analyses, firm-level IT capital is divided by firm total assets, and then an average of all available constituent firms was taken for each industry.

13. However, we cannot find a single quantitative measure for industry regulation and technological change.

Additional information

Notes on contributors

Fei Ren

Fei Ren is an associate professor of information systems at Guanghua School of Management, Peking University, China. She received her Ph.D. from the Paul Merage School of Business at the University of California, Irvine. Her research interests focus on the business value of information technology, information technology and business strategy, and electronic commerce. Her research articles have appeared in Information Systems Research and other journals and in conference proceedings.

Sanjeev Dewan

Sanjeev Dewan is professor of information systems at the Paul Merage School of Business at the University of California, Irvine. His research and teaching interests focus on the economics of information technology and on electronic commerce. He has served as a senior editor at Information Systems Research and as an associate editor at Management Science. He received his Ph.D. in business administration from the Simon School at the University of Rochester.