Examining Complementary Effects of IT Investment on Firm Profitability: Are Complementarities the Missing Link?

Abstract

Prior research has documented that IT investment increases market returns. Economic theories predict such returns to be recognized in accounting profitability; this relationship remains ambiguous in prior literature. We reexamine the relationship between IT investment and firm profitability. Our approach is unique in that we examine complementarities between distinct IT components. We document that a firm’s investments in IT components exhibit different impacts on its profitability conditional on the level of investments in complementary components.

Keywords:

• IT investment
• firm profitability
• IT business value

Notes

1. It is important to note that prior studies do document a positive association between IT investment, financial market returns, and productivity (Chatterjee et. al. 2001; Subramani & Walden, 2001; Hitt & Brynjolfsson, 1996).

2. The increase on stock returns could also be partially explained by the decrease on firm-specific risks without changing the expected cash flow due to the improved competitive position.

3. Prior research used International Digital Group (IDG) and CII extensively. The data is collected through questioners.

4. Two firms were newly established in year 2000; the other four firms have an unusually low number of employees. The numbers of employees were 3, 6, 10, and 15 for these four public firms.

5. A company’s IT stock is the accumulated value of IT assets.

6. Hardware includes new purchase and depreciation on computer hardware; if hardware is rented, the rental costs for the current year is used; For manufacturing industry, general administration for factory offices IT expenses are included; For financial services firms, expenses for all branch, distribution office, and operations are included.

7. This problem is partially addressed in the Computer Intelligence InfoCorp database where mainframe prices are adjusted according to age. However, the problem persists with PCs and servers.

8. Three years is the standard accounting convention for the useful economic life of computer hardware assets. We sampled 64 publicly-traded firms on their depreciation policy for personal computers. All 64 firms reported that a personal computer is fully depreciated and retired within three years after the purchase. Further, guidelines from the Taiwanese SEC suggest that the useful life of a personal computer ranges from 1.5 to 5 years (IBM).

9. Software expenses includes new purchase, new development, new contract development, package software purchase expenses; If system software has been included in the hardware price, the cost is not counted as software. If application software is developed in-house and the cost has been included in the human resources or training expenses, it is counted as software. For manufacturing industry, general administration for factory offices IT expenses are included; For financial services firms, expenses for all branch, distribution office, and operations are included.

10. Prior research (Brynjolfsson & Hitt, 1996, Hitt & Brynjolfsson, 1996) transforms IT-related labor expenses to IT labor assets by multiplying the expenses by a factor of three. We perform a sensitivity analysis by taking the same approach with a factor from two to five on these expense items (software, training, and IT labor) and find similar results reported in Table 4.

11. The R&D stock is calculated at a discount rate of 15%. According to Hall (1990), the stock number is not very sensitive to different discount rates.

12. One significant difference is noted in models including outliers. The CH Rate in the third ROA model becomes statistically significant with outliers included.

13. AICs for ROE two-way, three-way, four-way models are 3541.41, 3629.86, and 3629.86 respectively.AIC for ROA two-way, three-way, four-way models are 6046.91, 6047.80, and 6047.78 respectively.

14. Four-way and three-way interactions between IT components do not provide statistically significant improvement on the model quality. Models with two-way interactions are selected due to efficiency.

15. Every additional unit invested in training will reduce ROA by .0956 and ROE by 1.31. When allocating this additional unit of investment in hardware, firms will increase ROE by .1331. We also examined the economic significance and found that increasing the hardware investment by one standard deviation will improve ROA by 34%. However, increasing the investment in training by one standard deviation will reduce ROA and ROE by 125% and 130%, respectively.

16. Note we only include statistically significant coefficients when taking derivatives.

17. For example, the marginal impact of hardware on ROA when its complementary investment (training) is at the mean level is calculated by substituting the mean of training investment (0.0335) into the marginal impact formula.

18. F-tests confirm that CT at minimum, CT and CS at mean, and CS, CT, and CL at maximum are statistically significantly different from zero for correctness.

¹⁹ We checked the assumption of investment stability by comparing hardware investments by year for the 213 firms with data in both years of our sample. For these firms, we found no statistically significant difference in hardware investment levels within the time window examined.

²⁰ As defined by MIC, servers include mainframe-range servers (e.g., IBM^® AS/400 and R6000 computers).

Additional information

Notes on contributors

Johnny Jiung-Yee Lee

Johnny Jiung-Yee Lee is an associate clinical professor of Accounting at LeBow College of Business, Drexel University. He earned his PhD in Accounting and Information Systems from the University of Utah. He has published in journals, such as Communications of the Association for Information Systems, Communications of ACM, Journal of Business Research, Journal of Real Estate Portfolio Management, Accounting Perspectives, Management Research Review, and Information Systems Management. His current research interests are in accounting information systems, management accounting, e-commerce, supply-chain, consumer behavior, and business value of information systems.

Taylor Randall

Taylor Randall received his undergraduate degree with honors in accounting at the University of Utah and his MBA and PhD in operations and information management from the Wharton School of Business at the University of Pennsylvania. Dr. Randall’s area of research interest includes issues at the interface of accounting and operations with publications in Management Science, Operations Research, Accounting Organizations and Society, and Marketing Science, as well as other journals. Dr. Randall has served as a director for the University Venture Fund since 2003, during which time the UVF has become the largest independent student-run venture in the country.

Paul Jen-Hwa Hu

Paul Jen-Hwa Hu is David Eccles Professor of Information Systems at the University of Utah. He receives his PhD degree from the University of Arizona. His research interests include information technology in healthcare, e-commerce, digital government, and knowledge management. Hu has published papers in Information Systems Research, Journal of MIS, Journal of AIS, Decision Sciences, Journal of the American Society for Information Science and Technology, Communications of the ACM, and various IEEE Transactions and journals.

Anne Wu

Anne Wu is a chair professor at National Chengchi University in Taiwan and is the director of Taiwan Intellectual Capital Research Center. She received her PhD from George Washington University. Professor Wu’s current research interests include activity-based costing/activity-based management, balanced scorecard, compensation management, and intellectual capital. She has published articles in many journals including Accounting, Organizations, and Society, Contemporary Accounting Research, Journal of International Business Studies, Strategic Management Journal, and Leadership Quarterly. She has won several best paper awards from international conferences and has received outstanding research awards thrice from the National Science Council (NSC) in Taiwan.