(Gender) Tone at the top The effects of gender board diversity on gender inequality

We address the gender wage gap in Europe, focusing on the impact of female representation in executive and non-executive boards. We use a novel dataset to identify gender board diversity across European firms, which covers a comprehensive sample of private firms in addition to publicly listed ones. Our study spans three waves of the Structure of Earnings Survey, covering 26 countries and multiple industries. Despite low prevalence of female representation and the complex nature of gender wage inequality, our findings reveal a robust causal link: increased gender diversity significantly decreases the adjusted gender wage gap. We also demonstrate that to meaningfully impact gender wage gaps, the presence of a single female representative in leadership is insufficient.

1. (Gender) Tone at the top
(Gender) Tone at the top
The effects of gender board diversity on gender inequality
Bram Timmermans
European Association for Labour Economics
September 2024

2. (Gender) Tone at the top
Motivation
Our contribution
We establish a causal link:
gender board diversity→adjusted gender wage gaps

3. (Gender) Tone at the top
Motivation
The existing literature
A blossoming literature linking managers’ gender and inequality within firms in firms
Existing results are mixed
No evidence of spill-overs at the top of the firm
Abendroth et al. (2017); Bertrand et al. (2019); Maida and Weber (2022)
Positive effects from managers to employees, but negative within ranks
Hensvik (2014); Kunze and Miller (2017)
Wages of men and women react differently to diversity
Cardoso and Winter-Ebmer (2010); Flabbi et al. (2019)
Current results have limited external validity
Results are commonly country-specific
Often restricted to stock listed firms

4. (Gender) Tone at the top
Motivation
The existing literature
A blossoming literature linking managers’ gender and inequality within firms in firms
Existing results are mixed
No evidence of spill-overs at the top of the firm
Abendroth et al. (2017); Bertrand et al. (2019); Maida and Weber (2022)
Positive effects from managers to employees, but negative within ranks
Hensvik (2014); Kunze and Miller (2017)
Wages of men and women react differently to diversity
Cardoso and Winter-Ebmer (2010); Flabbi et al. (2019)
Current results have limited external validity
Results are commonly country-specific
Often restricted to stock listed firms

5. (Gender) Tone at the top
Motivation
What we bring: new data→new approach
Unique data #1: GBDD (Drazkowski et al., 2024)
28 milion firms (0.14% of them stock-listed)
all European countries and sectors
Unique data #2: measures of AGWG
comparable across countries and time
derived from EU SES (matched employee-employer data)
Identification from a fairly versatile instrument

6. (Gender) Tone at the top
Motivation
What we bring: new data→new approach
Unique data #1: GBDD (Drazkowski et al., 2024)
28 milion firms (0.14% of them stock-listed)
all European countries and sectors
Unique data #2: measures of AGWG
comparable across countries and time
derived from EU SES (matched employee-employer data)
Identification from a fairly versatile instrument

7. (Gender) Tone at the top
Motivation
What we bring: new data→new approach
Unique data #1: GBDD (Drazkowski et al., 2024)
28 milion firms (0.14% of them stock-listed)
all European countries and sectors
Unique data #2: measures of AGWG
comparable across countries and time
derived from EU SES (matched employee-employer data)
Identification from a fairly versatile instrument

8. (Gender) Tone at the top
Motivation
What we bring: new data→new approach
Unique data #1: GBDD (Drazkowski et al., 2024)
28 milion firms (0.14% of them stock-listed)
all European countries and sectors
Unique data #2: measures of AGWG
comparable across countries and time
derived from EU SES (matched employee-employer data)
Identification from a fairly versatile instrument

9. (Gender) Tone at the top
Motivation
Theory ambiguous on the link between female managers and gender inequality
1Positive spillovers
Awareness of discriminatory practices
(Hultin and Szulkin, 1999; Cohen and Huffman, 2007)
Role model
(Linehan and Scullion, 2008; Zimmermann, 2022)
Better abilit to detect performance/talent embodied in women
(Tsui and O’Reilly III, 1989; Ridgeway, 1997)
2No or negative spillovers
Another cog in the machine
(Jia and Zhang, 2013; Torchia et al., 2011)
Queen-bee syndrome
(Staines et al., 1974; Derks et al., 2016)

10. (Gender) Tone at the top
Motivation
Theory ambiguous on the link between female managers and gender inequality
1Positive spillovers
Awareness of discriminatory practices
(Hultin and Szulkin, 1999; Cohen and Huffman, 2007)
Role model
(Linehan and Scullion, 2008; Zimmermann, 2022)
Better abilit to detect performance/talent embodied in women
(Tsui and O’Reilly III, 1989; Ridgeway, 1997)
2No or negative spillovers
Another cog in the machine
(Jia and Zhang, 2013; Torchia et al., 2011)
Queen-bee syndrome
(Staines et al., 1974; Derks et al., 2016)

11. (Gender) Tone at the top
Data and methods
Databases
We link two databases at the industry×country×year level (cells)
1GBDD→Gender Board Diversity Database
2EU-SES→data on (adjusted) gender wage gaps

12. (Gender) Tone at the top
Data and methods
GBDD: Gender Board Diversity Database
Draws on Orbis data
Recovers registry information: board members
Data available since early 1990s for 44 European countries
Two challenges addressed by (Drazkowski et al., 2024)
1Identifying board members in each firm / year
2Assigning gender to board members
Two measures of gender board diversity
1Share of firms with at least one woman on board
2Average share of women among all board members

13. (Gender) Tone at the top
Data and methods
GBDD: Gender Board Diversity Database
Draws on Orbis data
Recovers registry information: board members
Data available since early 1990s for 44 European countries
Two challenges addressed by (Drazkowski et al., 2024)
1Identifying board members in each firm / year
2Assigning gender to board members
Two measures of gender board diversity
1Share of firms with at least one woman on board
2Average share of women among all board members

14. (Gender) Tone at the top
Data and methods
GBDD: Gender Board Diversity Database
Draws on Orbis data
Recovers registry information: board members
Data available since early 1990s for 44 European countries
Two challenges addressed by (Drazkowski et al., 2024)
1Identifying board members in each firm / year
2Assigning gender to board members
Two measures of gender board diversity
1Share of firms with at least one woman on board
2Average share of women among all board members

15. (Gender) Tone at the top
Data and methods
New stylized facts (and high quality of the data!)
Mean SD P10 P50 P90
Share of firms with at least one woman on boardi,c,t 0.378 0.150 0.197 0.361 0.582
Average share of women among all board membersi,c,t0.255 0.115 0.137 0.232 0.417
Sample used to obtain the measures (GBDD)
N. of firmsi,c,t 12065 24263 97 3463 28960
N. board membersi,c,t 20704 39009 179 6221 55468
N. female board membersi,c,t 4997 9187 49 1548 13868
Observations 1284
In an average cell, 60%+ of firms did not have a women on board!!
Around 25% of all aboard members are women

16. (Gender) Tone at the top
Data and methods
New stylized facts (and high quality of the data!)
Mean SD P10 P50 P90
Share of firms with at least one woman on boardi,c,t 0.378 0.150 0.197 0.361 0.582
Average share of women among all board membersi,c,t0.255 0.115 0.137 0.232 0.417
Sample used to obtain the measures (GBDD)
N. of firmsi,c,t 12065 24263 97 3463 28960
N. board membersi,c,t 20704 39009 179 6221 55468
N. female board membersi,c,t 4997 9187 49 1548 13868
Observations 1284
In an average cell, 60%+ of firms did not have a women on board!!
Around 25% of all aboard members are women

17. (Gender) Tone at the top
Data and methods
EU-SES Structure of Earnings Survey
A large and comprehensive database on individual workers with data on earnings
Available for (almost) all EU countries every 4 years→focus on years 2010, 2014 and 2018.
A survey of firms
Detailed data on wages, hours, occupation, tenure, etc
Missing information on household (children, marital status).
˜
Nopo (2008) as a non-parametric decomposition method
Recovers the adjusted gap for workers in common support
No need to specify functional form
Adjust for: age, education, position (ft/pt), sector, occupation, size of firm.

18. (Gender) Tone at the top
Data and methods
EU-SES Structure of Earnings Survey
A large and comprehensive database on individual workers with data on earnings
Available for (almost) all EU countries every 4 years→focus on years 2010, 2014 and 2018.
A survey of firms
Detailed data on wages, hours, occupation, tenure, etc
Missing information on household (children, marital status).
˜
Nopo (2008) as a non-parametric decomposition method
Recovers the adjusted gap for workers in common support
No need to specify functional form
Adjust for: age, education, position (ft/pt), sector, occupation, size of firm.

19. (Gender) Tone at the top
Data and methods
Rather high dispersion of adjusted gender wage gaps
Mean SD P10 P50 P90
Adjusted gender wage gapi,c,t 0.142 0.086 0.048 0.128 0.253
Matched meni,c,t(share) 0.896 0.117 0.741 0.936 0.988
Matched womeni,c,t(share) 0.941 0.061 0.864 0.958 0.994
Sample used to obtain measures(SES)
Number of women in celli,c,t 13895 33864 625 3820 28542
Number of men in celli,c,t 13345 22676 1211 5492 35705
Observations 1284

20. (Gender) Tone at the top
Data and methods
Rather high dispersion of adjusted gender wage gaps
Mean SD P10 P50 P90
Adjusted gender wage gapi,c,t 0.142 0.086 0.048 0.128 0.253
Matched meni,c,t(share) 0.896 0.117 0.741 0.936 0.988
Matched womeni,c,t(share) 0.941 0.061 0.864 0.958 0.994
Sample used to obtain measures(SES)
Number of women in celli,c,t 13895 33864 625 3820 28542
Number of men in celli,c,t 13345 22676 1211 5492 35705
Observations 1284

21. (Gender) Tone at the top
Data and methods
Methods
We estimate
AGWGi,c,t=β0+β
OLS
1GBDi,c,t+γc+γt+γs+ϵi,c,t (1)
where
AGWGi,c,tis the adjusted gender wage gap withinindustry,country and period (t)
GBD is a measure of gender board diversity:
1Share of firms with at least one woman on boards
2Average proportion of women among board members
However, endogeneity likely an issue:
unobserved time varying variables (→e.g. use of flexible work arrangements)
reverse causality

22. (Gender) Tone at the top
Data and methods
Methods
We estimate
AGWGi,c,t=β0+β
OLS
1GBDi,c,t+γc+γt+γs+ϵi,c,t (1)
where
AGWGi,c,tis the adjusted gender wage gap withinindustry,country and period (t)
GBD is a measure of gender board diversity:
1Share of firms with at least one woman on boards
2Average proportion of women among board members
However, endogeneity likely an issue:
unobserved time varying variables (→e.g. use of flexible work arrangements)
reverse causality

23. (Gender) Tone at the top
Data and methods
Identification
Candidate instrument: share of household consumption in final output (in celli,c,t)
Why?Theodoropoulos et al. (2022) argue that direct contact with customers requires understanding
fieminine” needs
Exclusion restriction:uncorrelated with adjusted gender wage inequality→once we account for
differences in characteristics
The IV specification is:
AGWGi,c,t=β0+β
IV
1
[GBDi,c,t+γc+γt+γs+ϵi,c,t
GBDi,c,t=α0+α1HH. cons.i,c,t+δc+δt+δs+υi,c,t

24. (Gender) Tone at the top
Data and methods
Does the first-stage work?.32
.34
.36
.38
.4
.42
Share of firms with some women on board
i,c,t
0 .1 .2 .3 .4 .5
Share of household expenditure
in total outputi,c,t
.2
.22
.24
.26
.28
.3
Avg. share of female in boards
i,c,t
0 .1 .2 .3 .4 .5
Share of household expenditure
in total outputi,c,t
Notes:Binned scatter residualized on sector, country and year.

25. (Gender) Tone at the top
Data and methods
Does the first-stage work?
Average share Share of firms
of women on boards with 1+ women
Share of household consumption 0.145*** 0.154***
in final output (0.0176) (0.0212)
FE: sector Yes Yes
FE: country Yes Yes
FE: year Yes Yes
N 1284 1284

26. (Gender) Tone at the top
Results
In OLS: results are inconclusive.12
.13
.14
.15
.16
Adjusted GWG
.1 .2 .3 .4 .5
Average share of women
on boards
.12
.13
.14
.15
.16
Adjusted GWG
.2 .3 .4 .5 .6
Share of firms with
at least a woman on boards
Notes:Binned scatter residualized on sector, country and year
AGWG with age, education, position(ft, pt), ISCO 08 (1 digit), ownership, and firm size.

27. (Gender) Tone at the top
Results
IV provides conclusive estimates
Average share of women Share of firms with 1+ women
OLS IV OLS IV
Gender board diversityi,c,t -0.00683 -0.302*** 0.00887 -0.285***
(0.0374) (0.110) (0.0279) (0.105)
FE: sector Yes Yes Yes Yes
FE: country Yes Yes Yes Yes
FE: year Yes Yes Yes Yes
N 1284 1284 1284 1284
First stage F-statistic (w/o FE) 68.01 52.97
One SD increase in share of firms with women on board (0.15) reduces AGWG by 4 percentage points
(−0.285∗0.15))

28. (Gender) Tone at the top
Results
Gauging the effect: European gender quota directive
EU Directive→at least 33% of women on boards of firms with 100+employees
Simulate the average share of women if firms below threshold were to adopt these measures
Predict the AGWG with new values of GBD
Mean SD
Female share on boards
As observed 0.268 0.119
Predicted 0.461 0.078
Gender Inequality measures
As observed 0.140 0.082
Predicted 0.082 0.052
Country x sector cells 430

29. (Gender) Tone at the top
Results
Gauging the effect: European gender quota directive
EU Directive→at least 33% of women on boards of firms with 100+employees
Simulate the average share of women if firms below threshold were to adopt these measures
Predict the AGWG with new values of GBD
Mean SD
Female share on boards
As observed 0.268 0.119
Predicted 0.461 0.078
Gender Inequality measures
As observed 0.140 0.082
Predicted 0.082 0.052
Country x sector cells 430

30. (Gender) Tone at the top
Results
Robustness checks
Focusing on senior managers instead of all boards
Controlling for the women to men ratio in the industry
Controlling for differences in workforce composition across industries
Focusing on the subsample of cells with enough people in common support

31. (Gender) Tone at the top
Results
How many women to make a difference?

32. (Gender) Tone at the top
Results
How many women to make a difference?
We estimate the following regression
AGWGi,c,t=β0+βShare of firms withN= 0 womeni,c,t+γs+γc+γt+ei,c,t (2)
AGWGi,c,t=β0+βShare of firms withN= 2+ womeni,c,t+γs+γc+γt+ei,c,t (3)
AGWGi,c,t=β0+βShare of firms withN= 3+ womeni,c,t+γs+γc+γt+ei,c,t (4)

33. (Gender) Tone at the top
Results
How many women to make a difference?-1
-.5
0
.5
1
b
IV
and 90% CI
Main
specification
Share with
no women
Two or more
women
Three or more
women

34. (Gender) Tone at the top
Summary
Summary
We show that improving gender board diversity decreases gender inequality
Results from most EU countries and across periods
Reduction in gender inequality is meaningful
Leverage novel database (GBDD) & new candidate instrument
Effects increase in industries with the number of female board members

35. (Gender) Tone at the top
Summary
Questions or suggestions?
Thank you!
w: grape.org.pl
t/x: grapeorg
f: grape.org
e: j.tyrowicz[at]grape.org.pl

36. (Gender) Tone at the top
Summary
Bibliography
Abendroth, A.-K., Melzer, S., Kalev, A., and Tomaskovic-Devey, D. (2017). Women at work: Women’s access to power and the
gender earnings gap.ILR Review, 70(1):190–222.
Bertrand, M., Black, S. E., Jensen, S., and Lleras-Muney, A. (2019). Breaking the glass ceiling? the effect of board quotas on
female labour market outcomes in norway.Review of Economic Studies, 86(1):191–239.
Cardoso, A. R. and Winter-Ebmer, R. (2010). Female-led firms and gender wage policies.ILR Review, 64(1):143–163.
Cohen, P. N. and Huffman, M. L. (2007). Working for the woman? female managers and the gender wage gap.American
Sociological Review, 72(5):681–704.
Derks, B., Van Laar, C., and Ellemers, N. (2016). The queen bee phenomenon: Why women leaders distance themselves from
junior women.Leadership Quarterly, 27(3):456–469.
Drazkowski, H., Tyrowicz, J., and Zalas, S. (2024). Gender board diversity across Europe throughout four decades.Nature
(Scientific Data), (87).
Flabbi, L., Macis, M., Moro, A., and Schivardi, F. (2019). Do female executives make a difference? the impact of female
leadership on gender gaps and firm performance.The Economic Journal, 129(622):2390–2423.
Hensvik, L. E. (2014). Manager impartiality: Worker-firm matching and the gender wage gap.ILR Review, 67(2):395–421.
Hultin, M. and Szulkin, R. (1999). Wages and unequal access to organizational power: An empirical test of gender
discrimination.Administrative Science Quarterly, 44(3):453–472.
Jia, M. and Zhang, Z. (2013). Critical mass of women on bods, multiple identities, and corporate philanthropic disaster
response: Evidence from privately owned chinese firms.Journal of Business Ethics, 118:303–317.
Kunze, A. and Miller, A. R. (2017). Women helping women? evidence from private sector data on workplace hierarchies.The
Review of Economics and Statistics, 99(5):769–775.

37. (Gender) Tone at the top
Summary
Bibliography
Linehan, M. and Scullion, H. (2008). The development of female global managers: The role of mentoring and networking.
Journal of Business Ethics, 83:29–40.
Maida, A. and Weber, A. (2022). Female leadership and gender gap within firms: Evidence from an italian board reform.ILR
Review, 75(2):488–515.
˜Nopo, H. (2008). Matching as a tool to decompose wage gaps.Review of Economics and Statistics, 90(2):290–299.
Ridgeway, C. L. (1997). Interaction and the conservation of gender inequality: Considering employment.American Sociological
Review, pages 218–235.
Staines, G., Tavris, C., and Jayaratne, T. E. (1974). The queen bee syndrome.
Theodoropoulos, N., Forth, J., and Bryson, A. (2022). Are women doing it for themselves? female managers and the gender
wage gap.Oxford Bulletin of Economics and Statistics, 84(6):1329–1355.
Torchia, M., Calabr`o, A., and Huse, M. (2011). Women directors on corporate boards: From tokenism to critical mass.Journal
of Business Ethics, 102:299–317.
Tsui, A. S. and O’Reilly III, C. A. (1989). Beyond simple demographic effects: The importance of relational demography in
superior-subordinate dyads.Academy of Management Journal, 32(2):402–423.
Zimmermann, F. (2022). Managing the gender wage gap - how female managers influence the gender wage gap among workers.
European Sociological Review, 38(3):355–370.