How do Managers use Management Control Systems in Response to Shareholder Activism?

Abstract

We study how managers interactively use management control systems (MCS) in response to shareholder activism in a context of performance declines. We predict that broad-scope MCS design is a key factor that helps explain variation in the level of interactive use. In addition, we argue that the extent to which diagnostic use of MCS alleviates shareholder pressures depends on analysts’ recommendations. We empirically test our hypotheses with survey and archival data. Data on shareholder activism is hand-collected. The results provide support for our predictions, suggesting that managers use MCS interactively to assuage shareholder activism in the context of performance declines. This effect is less pronounced for managers equipped with broad-scope information from their MCS. Our findings indicate that in a setting of favorable analysts’ recommendations, managers use MCS diagnostically instead of interactively to cope with shareholder activism. Overall, the findings may help inform our understanding of how firms manage shareholder activism from a management control perspective.

Keywords:

• Interactive use
• Diagnostic use
• Management control systems
• Shareholder activism
• Analyst recommendation

Acknowledgements

We appreciate the comments and suggestions of Matthias Mahlendorf (editor) and the two anonymous reviewers. We also benefited from comments, assistance, and suggestions received from Beatriz Garcia Osma, Ricardo Malagueño, Facundo Mercado, Januario José Monteiro, Jose Luis Ucieda, and seminar participants at the 2021 AECA Annual Meeting. We acknowledge the financial contributions from the Spanish Ministry of Science and Innovation (PID2019-104163RA-I00), Santander Financial Institute, Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) (grants 421218/2018-5 and 305812/2018-1), Catedra UAM-Auditores Madrid, and the Madrid Government (within the framework of the multi-year agreement with the Universidad Autonoma de Madrid on Line 3: Excellence for University Staff - PRICIT). Additional materials are available in an online Supplement at the journal’s Taylor and Francis website.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental Data and Research Materials

Supplemental data for this article can be accessed on the Taylor & Francis website, https://doi.org/10.1080/09638180.2022.2063152

Appendix A. Respondent demographics.

Appendix B. Main variables scales.

Appendix C. Descriptive statistics on the main constructs at the item level.

Appendix D. Discriminant validity on the main survey variables at the item level.

Appendix E. Multitrait matrix.

Appendix F. Robustness tests and additional results.

Notes

1 The LoC framework comprises four levers: interactive, diagnostic, beliefs and boundary systems. Due to our research question and hypothesis development, we focus mainly on the interactive and diagnostic uses. Shareholder activism and earnings pressures occur over a short period (a quarter or a year); thus, it is unlikely that there will be enough time for changes in these two value systems (i.e., beliefs and boundaries) to affect firm performance. For robustness, we include both beliefs and boundaries as control variables in our analyses.

2 For completeness, we also test the effects of the three additional features in our model. We also include all these information dimensions as control variables in our analyses.

3 In his work, Bedford (2020) explains that management control is more effective when MCS are designed and adopted prior to initiating an interactive or diagnostic use.

4 This timing is the main reason we do not expect that broad-scope design is potentially a choice variable (i.e., shareholder activism as an antecedent of MCS design). Non-tabulated results confirm this view. We are grateful to an anonymous reviewer for this suggestion.

5 Mundy (2010, p. 507) provides qualitative evidence related to this phenomenon by indicating that diagnostic control is used to monitor customer satisfaction but, if customers express dissatisfaction and the project is at risk, managers are encouraged to use interactive controls. Similarly, Garcia Osma et al. (2018) indicate that the diagnostic use is unlikely to help when the pressure comes from an external stakeholder (i.e., lender). As the diagnostic use focuses on deviations, without proper focus, manager involvement and open debate, ‘red lights going off across the board would only distract’ managers (p. 42).

6 Prior work has largely analyzed how managers react to pressures from analyst recommendations, mainly (i) by ignoring them; (ii) managing capital markets’ expectations using, for example, earnings guidance; or (iii) using both earnings management strategies, real and accrual (Burgstahler & Eames, 2006; Graham et al., 2005). However, this literature remains silent about how firms internally manage these pressures, which would be a step previous to the actions described above.

7 The standard recommendation scale is strong buy (1), buy (2), hold (3), sell (4), and strong sell (5).

8 Because we utilize a universal theoretical model, our results are potentially not limited to Brazil. Untabulated results indicate that Brazilian listed firms show nonsignificant differences, in terms of shareholder equity or earnings per share, with listed firms from Australia, Canada, Germany, Italy or the UK, settings widely used in prior work (Henri & Wouters, 2020; Firk et al., 2019; Bedford, 2015; Janke et al., 2014), and they display significant higher values in terms of employees compared with Australia and Canada.

9 Our results are unchanged if we run our models using the full sample.

10 Appendix B shows the survey items for the main variables in the study.

11 For parsimony, descriptive statistics at the item level for the main variables in the study are displayed in Table C1 (see Appendix C).

12 As a robustness test, we run the model using four additional modified measures: (i) ROE in the prior year, (ii) change in ROE in the prior year, (iii) negative ROE in both t-1 and t-2, and (iv) ROE in prior year adjusted by industry. The results of these alterative measures triggering the interactive use of MCS are qualitatively similar than those reported in our main analysis, albeit slightly weaker.

13 We thank the editor for making this suggestion.

14 As a sensitivity test and to assuage concerns that our interactive measure captures also diagnostic use, we run our analyses using only the first three items: (i) enabling discussions in meetings, (ii) enabling continual challenge and debate, and (iii) holding a common view of the organization. According to Simons (1995), these items are very close to his understanding and argumentation of interactive use. The results (untabulated) remain unchanged using this restricted measure. We thank an anonymous reviewer for making this suggestion.

15 These items—mainly those related to focusing attention, improving understanding (interactive use), and score keeping and monitoring results (diagnostic use)—can also be traced back to Vandenbosch (1999).

16 We orthogonalize these variables before including them in the model to avoid multicollinearity problems.

17 A CMIN/DF ratio of less than 5, a RMSEA of less than 0.08, and CFI above 0.90.

18 If EIKON reports a missing value for a specific shareholder policy, we set that value to zero. Our results remain unchanged by this choice.

19 Two main points are as follows: (i) in scope, the company data facilitate shareholders’ rights to ask questions to the board or management and (ii) the policy allows shareholders to table resolutions or shareholder proposals at shareholder meetings.

20 Clan culture is comprised of five items (seven-point Likert scale) (Quinn & Spreitzer, 1991): (i) personal and familiar place; (ii) leadership by mentoring or facilitating; (iii) teamwork and consensus; (iv) loyalty and mutual trust; and (v) human development.

21 Commitment comprises the exchanges that exist between business partners who believe that an ongoing relationship with each other is important such that they maximize their efforts to justify the relationship. We measure commitment using eleven items (seven-point Likert scale) (Cropanzano & Mitchell, 2005): (i) renegotiations benefit all managers involved; (ii) similar values are shared; (iii) managers are committed to partnering; (iv) there is loyalty among managers; (v) there are long and lasting relationships; (vi) managers strive for integration; (vii) managers are willing to invest resources; (viii) managers are proud to be part of the firm; (ix) additional efforts are made only if recognized; (x) maintaining relationships is important; and (xi) managers are willing to make sacrifices. One item loading was less than 0.7 but much higher than 0.5. Thus, we opt to retain it. Dropping the item has no substantive effect on the findings.

22 The interaction terms in these untabulated results are not significant (Shareholder activism in performance declines x MCS design_Aggregation: β = 0.031, p > 0.10; Shareholder activism in performance declines x MCS design_Timeliness: β = 0.030, p > 0.10; and Shareholder activism in performance declines x MCS design_Integration: β = 0.027, p > 0.10). The results are unchanged if we run a regression with all four interactions included simultaneously.

23 Subsamples were created based on the median.

Additional information

Funding

This work was supported by PRICIT CAM-UAM; Catedra UAM-Auditores Madrid; Santander Financial Institute; Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq): [Grant Number 305812/2018-1,421218/2018-5]; Spanish Ministry of Science and Innovation: [Grant Number PID2019-104163RA-I00].