Managerial ability and tax avoidance adjustment toward the optimal level

ABSTRACT

Highlighting managerial ability as a moderator of tax adjustment behaviour, this study finds that firms led by more capable managers exhibit slower adjustment speed of tax avoidance towards the optimal level. Furthermore, our findings remain robust after employing the propensity score matching and instrumental variable approaches. Cross-sectional analysis reveals that the negative association between managerial ability and tax avoidance adjustment speed is stronger for firms facing larger transaction costs and more complex operations. This implies that competent managers are more inclined to avoid incurring costs related to unwinding previous tax strategies, even when their firms deviate from the optimal level of tax avoidance.

KEYWORDS:

• Managerial ability
• corporate tax
• tax avoidance
• optimal tax strategy
• partial adjustment model

JEL CLASSIFICATION:

• G30
• G34
• H26

Acknowledgments

Authors appreciate helpful comments from So Young Lee, Woo-Jong Lee, Masumi Nakashima (discussant), and seminar participants in 2023 Taiwan Accounting Association Annual Conference. Funding information: Sumi Jung acknowledges that this work was supported by Yonsei Business Research Institute. The authors report there are no competing interests to declare.

Disclosure statement

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

Notes

1. Although they are not at the top decision makers of the firm, managers in subsidiaries can also interact with how their compensation is linked to organisational performance. See Ortmann and Schindler (2022).

2. The authors thank Peter Demerjian for sharing the managerial ability measure at his website: https://peterdemerjian.weebly.com/managerialability.html.

3. From now on, the firm subscript i is omitted from the Tables.

4. The inference remains unchanged when we exclude MA_{t-2, t} from the first-stage regression.

5. We show the first-stage estimation results and the covariate balance tests in Appendix B.

6. We download the state-level data from the US Census Bureau (https://data.census.gov/). As we obtain the data from 2010 and the sample period starts from 2000, it was assumed that the state-level college degree percentage before 2010 is set to be same as that of 2010.

7. The metropolitan statistical area data were obtained from 10-X header database, provided by University of Notre Dame software repository for accounting finance (https://sraf.nd.edu/data/augmented-10-x-header-data/). This is merged with Compustat through matching Central Index Key identifier and filing date. The observations were deleted if the number of observations is less than five for each year-MSA.

8. We calculate firm age as the difference between the first year the firm appears in Compustat and the fiscal year of observations.

9. Loughran and McDonald (2023) suggest that common measures of firm complexity such as segments, 10-K readability, and XBRL tag diversity have limited scope. They develop a new measure by detecting words implying firm complexity through machine learning. They show that this new measure dominates previously widely-used of firm complexity.

10. We calculate the within-firm volatility of sales as the standard deviation of sales scaled by total assets during previous five years. We calculate the within-firm volatility of return-on-assets as the standard deviation of net income before extraordinary items scaled by lagged total assets during previous five years.

11. We find that the coefficient on CASHETR×MA are not significantly different between subsamples when we partition the sample based on G-index, analyst coverage, CEO duality and board independence.

12. We find that our inference does not change when we exclude specific time periods such as financial crisis around 2008 and COVID-19 period around 2020.

13. In untabulated analysis, we partition the sample based on the size of uncertain tax position (UPT). Since UPT is available from the year 2007, the number of observations reduce to 15,738. We find that the negative association between managerial ability and tax avoidance adjustment speed is stronger for firms with larger UPT. This result indicates that when competent managers initially choose aggressive tax strategies, they are more cautious in closing down the gap between the actual and optimal levels of tax avoidance.

Additional information

Funding

The work was supported by the Yonsei Business Research Institute.