Moderative influence of compensation policy on the relationship between the personal characteristics of the CEO and the company's risk

Company size was measured as natural logarithm of companies' assets (Rogers, 2002). Share of CEO interests in the company is calculated as relation between CEO company shares and total number of company shares. In previous studies it was found that this characteristic may significantly influence company risk because the more shares CEO owns the less diversified he/she becomes which ties his interests with interests of shareholders and make him take more risks (Miller et al., 2002; Benischke et al., 2019). Companies were treated as innovative if they spend any sum of money on R&D during the year (so that dummy variable was constructed) (Humphery-Jenner et al., 2016). Furthermore, we defined growing companies as those which annual sales' growth rate is higher than median growth rate for the industry during the year (dummy variable constructed).

Table 1 shows the description of variables included in the models: their name, short name for models, measurement approach and references.

Table 1 Variables description

Models

In this study it is suggested that the structure of CEO compensation has an effect on company's risk. According to the theoretical background we assume that this effect may be either direct or moderate because we believe that the influence of personal characteristics of CEO can be changed if he/she is being paid certain proportions of bonuses and options. Furthermore, we argue that stakeholders know personal peculiarities of CEO and try to manage risk through setting compensation package. In the research we constructed eight models. The first model is basic and shows direct impact of compensation and personal traits on the risk, another seven models depict moderation effect with every characteristic of CEO. All models do not include one age group and one tenure group of CEOs as far as their inclusion will lead to full multicollinearity.

To test the specifications, we proposed to apply fixed effect model grounding on the fact that our data has panel structure and we have to catch individual effects for companies and years (as far as crisis years included).

Tested models are presented below:

(1),

, (2.1)

, (2.2)

, (2.3)

(3.1)

(3.2)

(3.3)

(4)

where: Risk - company risk;

Age2 - CEO age between 45 and 54 years;

Age3 - CEO age between 55 and 64 years;

Age4 - CEO age equal or higher than 65 years;

Tenure2 - CEO tenure between 5 and 9 years;

Tenure3 - CEO tenure between 10 and 14 years;

Tenure4 - CEO tenure equal of higher than 15years;

Overconf - CEO overconfidence;

Comp.Share - share of bonuses or options in total compensation package;

Size-company size;

CEOshare - CEO interests in company (share);

ROA - return on assets;

Innovative - innovative company;

Growth - growing company;

fe - fixed effects;

- model error.

Results

Descriptive statistics

The initial step of data analysis is the detailed view of descriptive statistics. As far as there are two types of variables in the database: continuous and binary, two approaches to their presentation were applied. Descriptive statistics for continuous variables are presented in Table 2.

Company being analyzed have a mean level of risk (annual price volatility) of 0.307 whereas there was a company which maintained the highest level of risk equals to 0.924. According to descriptive statistics compensation of CEOs' is more likely to include higher proportion of options rather than bonuses: mean value of options share in compensation package equals to 0.863 whereas mean bonus share is only 0.051. These findings support the idea that nowadays CEO's are more frequently awarded with market-based types of compensation rather than ordinary compensations based on financial statements. Companies' size standard deviation is not too high due to the reason that in the research only big traded companies were analyzed. Average CEO in UK companies owns 1.1% of overall companies' shares. Mean value of return on assets in analyzed enterprises is high enough: 1.281. Nonetheless, there are companies which had losses during 2008-2015 so that their return on asset is negative. Probably, this could happen because two crisis years (2008 and 2009) are included in the dataset. Nonetheless, the effect of crisis will be captured by fixed effects.

Table 2 Descriptive statistics of quantitative variables

There are also eleven binary variables in the dataset. Their descriptive statistics are shown in Table 3. We have divided CEOs into groups based on their age and tenure so that now there are no continuous variables depicting CEO's traits. Nonetheless, it has to be mentioned that initially CEOs presented in the dataset are aged from 35 to 70 and have a tenure from 0 to 32. In the dataset the highest portion of CEOs are aged 45-54 years (59.95 %) and 55-64 years (26.61%). The least number of CEOs belong to the groups of 35-44 years old and 65+ years old. There are almost equal proportions of CEOs who are taking this job position for 0-4 years and 5-9 years. The shares are 32.66% and 34.54% respectively. There are less more tenured CEOs in the dataset: more than one-fifth of the sample consists of those who work as CEOs for 10-14 years, 11.56% work in this position for 15 or more years. According to descriptive statistics there are not too many overconfident CEOs in the dataset: only 12.23% CEOs can be treated as overconfident. Interestingly, the majority of companies presented in sample invest in R&D and have sales' growth rate higher than median rate for the industry during the year. Correlation matrix for all variable used in the research is presented in Appendix 1, histograms of continuous variables are shown in Appendix 2.

Table 3 Descriptive statistics for dummy variables

3. Results of models' estimation

We have constructed four basic models where the first one represents the direct influence of CEO traits and compensation types on the risk, next six models capture the moderation effect for every group age and tenure and the last model is constructed to test moderation effect of compensation on the link between overconfidence and risk. Furthermore, all models are tested for both options and bonuses.

Results of models' estimations for each type of compensation presented in Tables 4 and 5. Table 4 includes results for models with options whereas Table 5 depicts results of models' estimation for bonuses.

Table 4 Results of models' estimation. Options share

Table 5 Results of models' estimation. Bonuses share

According to the results presented in Table 4 and Table 5, on average, other thing being equal, all age groups included in the models have direct negative impact on the risk compared to the youngest CEOs included in the first age group. The largest negative direct influence is observed for CEOs aged 65 and more and those which are 55-64 years old. These results are consistent with the theory that older CEOs tend to be more cautious and conservative. As for tenure, the results are not that unequivocal: all models prove that based on the direct influence CEOs with tenure from 5 to 9 years are less risk compared to other groups. Furthermore, the results support the idea about nonlinear connection between CEO's tenure and risk as far as the coefficients before CEOs' group tenured 5-9 years is the lowest, after that the coefficient before more tenured CEOs (10-14 years) is getting higher in all models and decreases for the last group of CEOs' tenure. Only one model out of sixteen prove positive significant influence of CEO overconfidence on corporate risk. This result in consistent with findings revealed by Chava and Purnanandam (2010), Kim and Lu (2011), Liu et al. (2013).

The results depict that, on average, other things being equal, options share in total compensation structure of CEOs have direct positive influence on the risk (two out of eight models support this notion). This finding is in line with previous research (Jin, 2002; Armstrong et al., 2012).

Furthermore, all models, except 2.1, show significant direct negative influence of bonuses share in total compensation of CEOs on corporate risk, which is consistent with the results of previous studies (Vallascas & Hagendorff, 2013).

Moreover, the results of models' estimations in Table 4 depict positive moderation effect of options share on the link between corporate risk and CEOs' age from 45 to 54 and negative moderation effect on the link between companies' share price volatility and CEOs' age between 55 and 64. These findings can be explained by the fact that younger CEOs (from 45 to 54 years old) are less risk averse compared to older ones because they are in the beginning or in the middle of their career as CEO and still strive to build strong long-run reputation and reach corporate goals. For them the increase in options share forms induces for riskier decisions which can lead to welfare increase as company's value grows. These results support classic agent theory (Benischke et al., 2019). Remarkably, for older CEOs (55-64 years old) prospect theory can be applied. Probably, these CEOs are more risk averse which forms their cautious attitude to options being awarded because they care more about their welfare rather than about long-run career perspectives compared to younger CEOs. Furthermore, such CEOs already think about their retirement and prefer stability. They understand that the increase of risk may lead to either welfare growth or decline and overestimate bad outcomes and possible losses rather than believe that they can benefit from the situation (Martin et al., 2009).

In addition, according to the obtained results, on average, other thing being equal, the increase in options share decreases the influence of CEOs' overconfidence on corporate risk. These findings show that with the increase of probability to lose welfare even overconfident CEOs seek to reduce the overall risk level of the company in order to save their money.

In the research options share does not have any influence on the link between CEOs' tenure and corporate risk.

Table 5 includes results for interaction effect of bonuses on the link between CEOs' traits and corporate risk. Bonuses share in total CEOs' compensation have negative moderation effect on the link between second age group (45-54 years old) and share price volatility and positive impact on the link between CEOs aged 55-64 years.

According to theoretical background, bonuses represent incentives for short-run decisions that is the reason why older CEOs increase the risk of the company when they are being awarded more bonuses. This group of CEOs striving to get as much as they can in the short run perspective (they want to have stable income in the short run because of retirement closeness) and having clear understanding of what they should do for this (compared to the situation when CEOs are awarded with options and need to make more complicated decisions to rise the welfare), make risky, sometimes ill-conceived, short-run corporate decisions which increase overall corporate level of risk. As for younger CEOs, for them the increase of bonuses share in compensation package decreases the interest to rise overall corporate risk because they suggest getting more from market-based types of compensation due to the fact that they have desire, time and opportunity to make complicated decision for the long-run perspective.

There is a positive impact of bonuses share on the link between overconfidence and companies' risk. This may be due to the fact that overconfident CEOs can take extreme risky means to increase their well-being (thereby increasing risk). Such managers are convinced that even in a short period of time they will be able to realize what they want; they take radical decisions in order to achieve results.

3.1 The issue of reverse causality between CEOs' compensation and corporate risk

Some researches argue that the relationship between CEO compensation and company risk is bidirectional: these two variables are not exogenous as far as whеn a company have a certain level of risk, shareholders may influence the compensation package of CEO to change the current level of risk in the future. So that in order to consider this peculiarity we also formed models with lag value of compensation share.

(5)

, (6.1)

, (6.2)

, (6.3)

(7.1)

(7.2)

(7.3)

(8)

These models are presented below. The logic of models' estimation as well as variables meanings are similar to those presented in the main part of the research except for lag value of compensation.

The results of models' estimation are presented in Tables 6 and 7. According to the obtained results, lag of options and bonuses shares have significant moderation effect only on the link between overconfidence and corporate risk. The results are consistent with those presented in the main part: lag of options shares have significant negative influence on the link between overconfidence and companies' share price volatility whereas lag of bonuses share positively influences this link.

In order to deepen in this particular issue further research should be made.

Table 6 Results of models' estimation. Lag of options share

Table 7 Results of models' estimation. Lag of bonuses share

Conclusion

The followers of behavioral economics assume that CEOs' strategic decisions are grounded on their values, ideology and personal experience. According to upper-echelon theory, CEOs represent one of main figures in the company and their perception of business is undoubtedly reflected in company strategy and financials. Risk is one of the most important indicators of companies' performance. According to the literature, CEO's personal traits have an impact on enterprise's level of risk (Sanders &. Hambrick, 2007). Nonetheless, most practitioners argue that correct and sound compensation policy is one of the main factors of companies' success because it influences CEOs behavior toward the level of risk which they are ready to take for the enterprise (Harvey & Shrieves, 2001). According to agency theory, managers and shareholders tend to have distinguishing interests concerning risk. Compensation committees try to set CEO compensation in such a way which allows to align interests of both sides (Gopalan et al., 2014). Nonetheless, it is difficult to say for sure how these compensations influence the natural, psychological biases of CEO's behavior which arise due to managers' personal traits.

While previous research on compensation policy do not differentiate CEO's personal traits, we explore, whether the remuneration packages may moderate the impact of CEO personal characteristics on the risk taking of the firms they run.

Our findings reveal that the increase of option shares in total compensation structure of CEO positively influences companies' risk as well as has an impact on the link between CEO's age and corporate risk. For younger CEOs this influence is positive which is in line with classic agency theory whereas for older CEOs perspective theory is more applicable: more options in compensation package make them worry about possible bad outcomes of risk increase and induce them to decrease the risk. Even overconfident CEOs tend to decrease the level of risk due to the fear of wealth loss when the proportion of options in their compensation rises.

Furthermore, in the research we found that, on average, bonuses share negatively influence companies' risk. Nonetheless, in this study we reveled that overconfident CEOs when being awarded higher bonuses share are going to increase the level of risk. The likeliest description of this result is connected with the fact that bonuses are tied with short run KPIs so that overconfident CEOs take more risk to reach them. Higher bonuses share negatively influence on the link between 45-54 years old CEOs and corporate risk as they are more motivated by market-based compensations. At the same time, older CEOs (55-64 years old) increase company risk level trying to maintain clear corporate goals to increase their welfare.

In general, the results obtained are consistent with the existing economic theory. However, since the issue of the moderating effect of compensations is poorly studied. In order to make sound conclusions and offer practical solutions for remuneration committees, it is necessary to conduct additional study. Inclusion of a larger number of companies in the sample and considering more diverse set of CEO traits could make the study more substantial. Also, the possible interrelation between risk and compensation should be considered in more detail than it was made in present research. Another limitation is connected with the usage of Black-Scholes model for options value calculation. It was assumed that companies do not have dividends, which might not be true in reality and could bias the results.

Further this study can be broadened by considering more complex relationships between the company's compensation structure and risk, dividing companies into several groups based on their risk strategies. It is possible to include other risk indicators (organizational, investment) and other characteristics of general directors. It is also worth noting that the results are based on panel data from 2008 to 2015. Accordingly, in order to be able to disseminate the results to companies in the modern economy, further research require using more recent data.

References

1. Antia, M., Pantzalis, C. and Park, J.C. (2010). CEO decision horizon and firm performance: An empirical investigation., Journal of Corporate Finance, 16: 288-301.

2. Armstrong, C. S., Ittner, C. D., & Larcker, D. F. (2012). Corporate Governance, Compensation Consultants, and CEO Pay Levels. Review of Accounting Studies, 17 (2), 322-351.

3. Awan H.M., Bhatti M.I., Qureshi A. (2010). How growth opportunities are related to corporate leverage decisions? Investment Management and Financial Innovations. Vol. 7, Issue 1, 2010, 90-97.

4. Bargeron L.L., Lehn K.M., Zutter C. J. (2010). Sarbanes-Oxley and corporate risk-taking. Journal of Accounting and Economics. Vol. 49, Issues 1-2, p. 34-52.

5. Barker V. & Mueller G (2002). CEO characteristics and firm R&D spending. --Management Science. 48(6). 782-801.

6. Bebchuk, L. A., & Fried, J. M. 2003. Excecutive Compensation as an Agency Problem. Journal of Economic Perspectives. Vol.17, No. 3, 71-92.

7. Bйnabou, R., 2012. Groupthink: Collective Delusions in Organizations and Markets. The Review of Economic Studies.

8. Benischke M. & Martin G. & Glaser L. (2019). CEO equity risk bearing and strategic risk taking: The moderating effect of CEO personality. Strategic Management Journal. Vol.40. p. 153-177.

9. Bernasek & Shwiff (2001). Gender, Risk, And Retirement. Journal of Economic Issues 35(2).

10. Bertrand, M. and Schoar, A. (2003). Managing with Style: The Effect of Managers on Firm Policies, The Quarterly Journal of Economics, 118(4): 1169-1208.

11. Brachinger H. W. & Brown M. & Schubert R. (1999). Financial Decision-Making: Are Women Really More Risk-Averse? American Economic Review. 89(2). 381-385.

12. Brown, R. & Sarma, N., 2007. CEO overconfidence, CEO dominance and corporate acquisitions. Journal of Economics and Business, 59(5), 358-379.

13. Bryan St. & Hwang L.& Lilien S. (2000). CEO Stock-Based Compensation: An Empirical Analysis of Incentive-Intensity, Relative Mix, and Economic Determinants. he Journal of Business, Vol. 73, No. 4. pp. 661-693.

14. Cadenillas A., Cvitanic J.,Zapatero F. (2004). Leverage decision and manager compensation with choice of effort and volatility. Journal of Financial Economics, 2004, vol. 73, issue 1, 71-92.

15. Cain, M.D. and McKeon, S.B. (2016). CEO Personal Risk-Taking and Corporate Policies, Journal of Financial and Quantitative Analysis, 51(1): 139-164.

16. Chava, S. & Purnanandam, A., 2010. CEOs versus CFOs: Incentives and corporate policies. Journal of Financial Economics, 97263-278.

17. Chen, D. and Zheng, Y. (2014). CEO Tenure and Risk-Taking, Global Business and Finance Review, 19(1): 1-27.

18. Chen, H.-L. (2013). CEO Tenure and R&D Investment:The Moderating Effect of Board Capital, The Journal of Applied Behavioral Science, 49(4): 437-459.

19. Chowdhury, J. and Fink, J. (2017). How Does CEO Age Affect Firm Risk?, Asia-Pacific Journal of Financial Studies, 46(3): 381-412.

20. Coles J.L. & Naveen L. & Daniel N.D. (2006). Managerial Incentives and Risk Taking. Journal of Financial Economics 79(2):431-468.

21. Cook M.L., Burress M.J. (2013). The Impact of CEO Tenure on Cooperative Governance. Managerial and Decision Economics 34(3):218-229.

22. Corsi C., Prencipe A. (2019). International Journal of Innovation Management. Vol. 23, Issue 05, 1-19.

23. Dittman I., Yu K., Zhang D. (2017). How Important Are Risk-Taking Incentives in Executive Compensation? Review of Finance, 1805-1846.

24. Faccio M. & Marchica M.T. & Mura M. (2016). CEO gender, corporate risk-taking, and the efficiency of capital allocation. Journal of Corporate Finance. 39.193-209.

25. Farag, H. and Mallin, C. (2018). The influence of C...