Today, International Women’s Day, is an appropriate time to reflect on the role of women in corporate America. The notion of career women needing to ‘lean in’ to get ahead, while at the same time dealing with potential ‘#MeToo’ scenarios at work, has placed the issue of gender equality in corporate America squarely in the public view. It is no surprise that studies of the American executive suite have found a lack of gender diversity (Bell 2005, Matsa and Miller 2011). Similarly, scholarly articles have documented substantial compensation gaps between female executives and their male colleagues, even for similar jobs (Bertrand and Hallock 2001, Munoz-Bullon 2010, Gayle et al. 2012, Albanesi et al. 2015, Newton and Simutin 2015, Carter et al. 2017, Quintana-Garcia and Elvira 2017).
At the same time, recent literature has emphasised how a CEO’s political views affect the running of a company. This research spans a broad range of corporate decisions regarding mergers and acquisitions (Elnahas and Kim 2017), the riskiness of investments and corporate debt levels (Hutton et al. 2014), tax sheltering (Francis et al. 2016), lobbying (Unsal et al. 2016), types of litigation (Hutton et al. 2015), corporate social responsibility (Di Giuli and Kostovetsky 2014), transparency of political spending (Cohen et al. 2019), pay dispersion and diversity in the executive suite (Chin and Semadeni 2017), and dividend policy (Bayat and Goergen 2020). None of this literature examines how the political preferences of a CEO might be associated with gender equality in the executive suite.
In a previous paper with Roberto Tallarita (Cohen et al. 2019), discussed here on Vox, we document how we build a dataset inferring whether a CEO has Democratic or Republican preferences. Roughly speaking, we do so by matching CEOs with their federal election contributions, as reported to the Federal Election Commission. We also discuss why we believe that it is valid to infer a CEO’s preferences from campaign contributions, rather than interpret these contributions as being for a purely strategic motive, such as to gain access to politicians.
In a new paper (Cohen et al. 2021), we study the relationship between the political preferences of a company’s CEO and gender among top executives (excluding the CEO). We discuss four different measures of CEO political preferences. They all examine the fraction of a CEO’s political contributions that went to Republican, rather than Democratic, candidates. For this column, we will sometimes refer to a CEO as being a Democrat (likewise, Republican) if they gave only to Democratic (Republican) candidates. In the actual paper, we examine political preferences in a more nuanced manner (as in Figure 1, discussed below).
Turning towards data on corporate executives, we use the Standard & Poor’s ExecuComp database, which covers companies in the S&P 1500 index. For the top five highest-paid executives (including CEOs), ExecuComp provides total compensation, stock compensation, age, title, and gender. Approximately 9% of non-CEO executives in this dataset are women.
We complement this dataset with Form 4 filing data from the Securities and Exchange Commission, accessed via EDGAR. These are reports made in compliance with Section 12 of the Securities Exchange Act of 1934, which requires every director, officer, or owner of more than 10% of a company’s equity to report to the Securities and Exchange Commission his or her relationship to the company and information on any acquisitions or dispositions of company securities.
Under the assumption that all officers transact in the company stock, these data allow us to paint a complete picture of the officers in a firm. The Form 4 data cover approximate 10 top executives in each firm. Approximately 12% of non-CEO executives in this combined dataset are women.
Figure 1 shows raw data connecting the fraction of a CEO’s political contributions that go to Republican candidates and the fraction of their executive suite that is female, using either the ExecuComp dataset or the union of that dataset with Form 4 data. Companies run by CEOs who lean more Republican employ fewer women.
Figure 1 Fraction of CEO’s political contributions to Republican candidates and fraction of executive suite that is female
Notes: The horizontal axis shows the fraction of a CEO’s political contributions that go to Republican candidates. The vertical axis shows the fraction of the executive suite who are women. ‘ExC’ uses the ExecuComp data set to define the executive suite, while ‘F4’ complements this with data from Form 4.
We show that these results are robust to more nuanced statistical analyses. In particular, we find that companies run by a CEO who only donates to Democrats employ a 15–25% higher fraction of women in the executive suite than those run by CEOs who only donate to Republicans and that this difference is statistically significant. These findings hold true when comparing companies that are similar in many dimensions, such as looking at the same company over time under different CEOs.
To get a better idea of the direction of causality that this relationship represents, we also perform an event-study analysis. We show that companies that replace a Republican CEO with a Democratic CEO, as compared with similar companies that replaced their Republican CEO with another Republican, did not have any noticeable trend in gender diversity in the executive suite before the change in CEO.
However, within three years of appointing the Democrat, women’s representation in the executive suite increases by as much as 60%, with the increase statistically significant. The opposite exercise, replacing a Democratic CEO with a Republican CEO does not affect female representation in the executive suite. Breaking down these results, we show that the increase in the fraction of executives who are female, when replacing a Republican with a Democrat, is due to an increase in the number of women in the executive suite, rather than a change in the number of executives.
Next, we examine gender wage gaps in the executive suite. First, we find that female non-CEO executives are paid 7–10% less than men for similar jobs and in similar firms (or even in the same firm). This result is similar to existing estimates in the literature. We then show that much of this gender wage gap disappears when the CEO of a company is a Democrat. Depending on our exact measure of CEO political preferences, we cannot reject the hypothesis that female executives are paid the same as their male colleagues when the CEO of the company is a Democrat.
We then turn to the composition of compensation and, in particular, focus on incentive pay. We look at three different measures. The first is the ratio of salary and bonuses to total compensation (the ‘cash ratio’). When this measure is higher, employees are mostly compensated in cash, rather than stock, and thus see less incentive pay. We find that female executives indeed have a higher cash ratio than similar male executives and that this gap can be entirely accounted for by the political preferences of the company’s CEO.
The second measure is the sensitivity of an executive’s stock option package to stock prices (‘delta’). Delta is indeed lower for female executives than their male counterparts, though this gap is quantitatively and statistically eliminated when the CEO of the company is a Democrat.
The third measure is the sensitivity of an executive’s stock option package to the volatility of stock prices (‘vega’). Again, we find that female executives receive lower vega, but cannot reject the hypothesis that this gap can be accounted for by the political preferences of the company’s CEO.
Our findings suggest that understanding the personal preferences of CEOs, such as their political preferences, can go a long way towards understanding the (lack of) gender diversity in corporate America, and the persistence of compensation differences between male and female executives. More research is needed to disentangle the underlying mechanisms driving our results.
Albanesi, S, C Olivetti and M J Prados (2015), “Gender and dynamic agency: Theory and evidence on the compensation of top executives”, in Solomon W Polachek, Konstantinos Tatsiramos, and Klaus F Zimmermann (eds.), Gender in the Labor Market, volume 42 of Research in Labor Economics, Emerald Publishing Ltd, 1–59.
Bayat, A, and M Goergen (2020), “CEO political ideology, shareholder primacy and dividend policy”, European Corporate Governance Institute Working Paper Series in Finance No. 650/2020.
Bell, L A (2005), “Women-led firms and the gender gap in top executive jobs”, IZA DP No. 1689.
Bertrand, M, and K F Hallock (2001), “The gender gap in top corporate jobs”, Industrial and Labor Relations Review 55: 3–21.
Carter, M E, F Franco and M Gine (2017), “Executive gender pay gaps: The roles of female risk aversion and board representation”, Contemporary Accounting Research 34: 1232–64.
Chin, M K, and M Semadeni (2017), “CEO political ideologies and pay egalitarianism within top management teams”, Strategic Management Journal 38: 1608–25.
Cohen, A, M Hazan, R Tallarita and D Weiss (2019), “The politics of CEOs”, Journal of Legal Analysis 11: 1–45.