Software-Vendor Diversification: A Source of Organizational Rigidity in Adversity?

ABSTRACT

Firms often assemble digital infrastructures using continuously evolving software applications sourced from a multitude of vendors. Using the theoretical lens of the threat-rigidity thesis, we raise the possibility that during adverse environmental conditions, software-vendor diversification can be a source of organizational rigidity that may dampen firm performance. Empirical analysis using data on 918 large public U.S. firms operating during two severe environmental shocks, the global financial crisis and the burst of the dot-com bubble, lends strong support to our thesis. Results indicate that a variety of firm performance indicators (e.g., stock return and operating income measures) are negatively associated with software-vendor diversification during crisis periods. Mediation analysis highlights the role of IT-related material weakness in firms’ internal controls in transmitting threat-rigidity effects that decrease performance. These results underscore the importance of software portfolio optimization for countering the dysfunctional effects of software-vendor diversification during adverse environmental shocks.

KEYWORDS:

• Software-vendor diversification
• IT-related control weakness
• internal controls
• digital infrastructures
• software portfolio
• economic shocks
• firm performance
• organizational rigidity

Acknowledgement

We thank Anand Gopal, Steve Maex, Sunil Mithas, and Siva Viswanathan for their helpful comments on an earlier version of the paper. We thank the Editor-in-Chief and two anonymous reviewers for their constructive suggestions throughout the review process.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/07421222.2023.2196774

Notes

¹ According to the Public Company Accounting Oversight Board (PCAOB), an effective internal control “provides reasonable assurance regarding the reliability of financial reporting and the preparation of financial statements for external purposes” [74]. An material weakness in internal control is defined as “a deficiency, or a combination of deficiencies, in internal control over financial reporting, such that there is a reasonable possibility that a material misstatement of the company’s annual or interim financial statements will not be prevented or detected on a timely basis” [74].

² The most recent trough in the U.S. economy’s business cycle occurred in April 2020, but the corresponding next peak is yet to be determined due to the ongoing impacts of the Covid-19 pandemic.

³ Section E of the Online Supplemental Appendix provides more details, examples, and additional results related to the three types of ITMW.

⁴ We include hardware-vendor diversification as a control because it is correlated with software-vendor diversification (correlation = 0.317, p-value < 0.001). However, hardware-vendor diversification could have no effect on performance due to the possible commodification of hardware, with insignificant differences in features between hardware assets sourced from different vendors [12, 62]. Our results are consistent with this expectation.

⁵ We tested the mediation effect of ITMW for our main cross-sectional regression only because ITMW data are reported annually. Thus, we could not construct a meaningful panel to combine the mediation analysis with the DID analysis as the dependent variables of the DID analysis (Eq. 3)(3) $\begin{array}{rl}& Performanc{e}_{it}={\eta }_i+{\mu }_t+{\beta }_{1}Software-VendorDiversificatio{n}_{it}\\ & +{\beta }_{2}Software-VendorDiversificatio{n}_{it}\times Crisi{s}_t+\beta Control{s}_{it}+{\in }_{it}\end{array}$(3) are measured in monthly or quarterly intervals. We did not perform the mediation analysis for the dot-com bubble burst either, because ITMW data are available only after the Sarbanes-Oxley Act of 2002.

⁶ In a robustness check, we added these firms back into our sample and found consistent results.

⁷ For brevity, Table A2 in the Online Supplemental Appendix classifies firms into industries following the Fama-French 12-industry classification. In our empirical analysis, however, we used a more granular definition of industry based on the Fama-French 48-industry classification. The results are robust if we used the Fama-French 12-industry classification instead.

⁸ Average value-weighted annual stock return is calculated using data from Kenneth French’s data library (http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html#Research).

⁹ We still included size as a control variable in this regression to control for within size decile differences.

¹⁰ We further included Hardware-Vendor Diversification×Crisis as an additional control to ensure that our results are not driven by hardware-vendor diversification during crisis. We found consistent results if this interaction is not added.

¹¹ We matched without replacement and set a caliper of 5 percent, that is, the difference in the propensity scores of the treatment and control firms cannot exceed 5 percent. Setting a caliper helps us remove the treatment firms that are substantially different from the control firms, and vice versa, which improves compatibility. Our results are robust to alternative calipers such as 10 percent.

¹² For stock return, the significance level of the coefficient declines to 10 percent likely because the size of the matched sample is much smaller compared to the full sample (480 vs. 918), and hence, the power of test has likely diminished. The magnitude of the coefficient is, however, similar to that from the full sample analysis (-0.258 vs. -0.255).

¹³ For R²(ITMW) × R²(performance) = 0.09, the confounder needs to explain, for example, 30 percent of the remaining variances in both ITMW and performance to make ACME zero.

Additional information

Notes on contributors

Jing Gong

Jing Gong ([email protected]) is an Associate Professor of Commerce at the McIntire School of Commerce, the University of Virginia. She holds a PhD degree in Information Systems and Management from Carnegie Mellon University. Dr. Gong’s research interests include digital marketing; digital platforms; and economics of information systems. Her work has been accepted for publication by journals including Journal of Management Information Systems, Information Systems Research, MIS Quarterly, The Accounting Review, Manufacturing & Service Operations Management, and Journal of Retailing.

Yi Liang

Yi Liang ([email protected]) is an Assistant Professor of Commerce at the McIntire School of Commerce, the University of Virginia. He received his PhD degree from Carnegie Mellon University. Dr. Liang’s research interests include the interface of IT and accounting; accounting and corporate laws; and financial institutions. His papers have been accepted for publication by journals including The Accounting Review, Journal of Accounting Research, Management Science, and Journal of Management Accounting Research.

Narayan Ramasubbu

Narayan Ramasubbu ([email protected]; corresponding author) is a Professor of Business Administration at the University of Pittsburgh. He received his PhD degree from the University of Michigan, Ann Arbor. His research interests include software product development and services delivery; information technology-driven innovation; user behavior within digital platforms; and the design, implementation, and governance of enterprise information systems. Dr. Ramasubbu has served on the editorial boards of Journal of AIS, Information and Technology Management, Management Science, and MIS Quarterly.