What stands behind the (lack of) women in boards? Can it all be explained by cultural factors? Are there any differences between public and private firms? We provide an answer to these key questions using the most comprehensive data available.

1. Gender diversity on management and supervisory boards
Joanna Tyrowicz (FAME|GRAPE, UoW, IAAEU and IZA)
Siri Terjesen (American University)
Jakub Mazurek (FAME|GRAPE)
Western Economic Association Meeting, 2019, San Francisco
1

2. Motivation

3. Established in the literature
• Fun fact: More men by the name John than women in NYSE boards
• Glass ceilings:
• Stock listed companies lose market value subsequent announcing female
CEOs (Chapple & Humphrey, 2014) even if no evidence of weaker
performance (Wolfers, 2006; Terjesen, forthcoming)
• Different management strategies: e.g. risk-taking (Nakano and Nguyen,
2012; Berger et al, 2014; Facio et al, 2016) M&A (Levi et al, 2014), etc.
• Adams & Kirchmaier (2013, and later): women-friendly countries have
more women on boards
• Employment flexibility
• Access to alternatives (↑ equality → ↓ women in STEM)
• Problems in this literature
• Typically data only for stock listed companies
• Choice of board(s) members is not random
• Decision-makers own resentiment vs perception of shareholders tastes
2

4. How can we innovate relative to the literature?
• Data:
• Go beyond stocklisted firms (→ public monitoring)
• Cross-country: delineate between supervisory to management boards
• Over time: increasing presence of women
Use (8 waves of) Amadeus data (100 mio firm-years over nearly 3
decades) Provide a novel method for gender assignment
• Question: test (Adams & Kirchmaier) intuitions on “good-for-all” drivers
H: ↑ gender equality more likely to have ↑ presence of W on S + M boards &
across F
H: cultural values changing towards rationality and freedom of self-expression
more likely to ↑ presence of W on S + M boards & across F
H: ↑ competition & innovativeness → ↑ presence of W on S + M boards
3

5. Table of contents
Motivation
Data
Identifying women and board members
Women on boards: stylized facts
Results
Conclusions
4

6. Contents
Motivation
Data
Identifying women and board members
Women on boards: stylized facts
Results
Conclusions
5

7. Amadeus data
• Comes from national information providers
• Typically full registry data: ownership details, NACE and board(s)
• Often: employment
• For many firms: balance sheet and profit/loss statement
• Kalemli-Ozcan et al (2015): standard for cleaning the data
• This study: no use of financial data → all available firms
6

8. Data
Table 1: Final samples
Full data With country-level institutional measures available
People Firms People Firms
Total # 112,351,222 69,327,072 28,946,738 17,666,148
Total unique 20,873,827 11,924,905 6,917,093 3,657,692
# Men 87,064,480 - 22,674,348 -
# Women 25,286,742 - 6,272,390 -
In firms which should have a supervisory board (*)
Total # 59,907,648 37,680,656 –——– –——–
Total unique 10,825,012 6,324,058 –——– –——–
# Men 45,988,164 - –——– -
# Women 13,919,484 - –——– -
In firms with data on supervisory board members (**)
Total # 1,960,606 463,872 625,192 134,399
Total unique 317,812 67,914 194,567 32,327
# Men 1,532,492 - 521,825 -
# Women 428,114 - 103,367 -
7

9. Measurement
Data features
• Growing availability over time
• ≈ 50% in market services (8.5m f-y in manuf + 6m f-y in construction)
• Constraining factor: country-level institutional variables (18/70m f-y )
How to measure gender board diversity?
• a firm level share of women on board (unweighted average)
Matsa & Miller (2011); Ahern & Dittmar (2012); Adams & Kirchmaier (2016)
• sum of women on boards (relative to men = weighted average)
Wolfers (2006); Adams & Ferreira (2009)
• fraction of firms that do not have women on board
novel indicator
8

10. Contents
Motivation
Data
Identifying women and board members
Women on boards: stylized facts
Results
Conclusions
9

11. Gender: exact names of board(s) members
In Amadeus as of 2014
• Heuristics to identify gender
1. H1: in some languages gender directly identifiable
e.g. vowel ending names in some Slavic languages, -ova in Czech, etc.
2. H2: the books of names
e.g. dedicated lists for each of the Scandinavian languages
• Resolving conflicts & dropping “impossible” countries
e.g. the Netherlands
• Manipulation check: 2010 & 2014 waves of Amadeus have salutations →
compare our gender assignment to salutations
total name-type-observations assigned: 16,254,928;
total with Amadeus confirmed gender: 15,371,479;
total men attributed as men: 10,074,034;
total women assigned as women: 4,048,932;
total men assigned as women: 10,963;
total women assigned as men: 10,626 so I think we are ok
10

12. Gender: exact names of board(s) members
Year % men in Amadeus attributed as % women in Amadeus attributed as
men women men women unassigned
2000 0.826 0.002 0.004 0.815 0.18
2001 0.824 0.002 0.005 0.808 0.187
2002 0.824 0.002 0.004 0.812 0.184
2003 0.823 0.002 0.004 0.809 0.187
2004 0.825 0.003 0.005 0.809 0.186
2005 0.825 0.002 0.005 0.810 0.185
2006 0.824 0.003 0.005 0.806 0.188
2007 0.835 0.003 0.005 0.815 0.179
2008 0.898 0.001 0.002 0.890 0.107
2009 0.990 0 0 0.985 0.015
2010 0.990 0 0 0.980 0.02
2011 0.989 0 0 0.981 0.019
2012 0.980 0 0 0.979 0.021
11

13. Board members: exact names of positions
In Amadeus as of 2014
• Countries differ in legal forms & positions
1. Catalogue positions in Amadeus and match them to law
2. Country by country
3. Assign no position in case of doubt
• Iterate this process
Manipulation check: 2014 wave of Amadeus → compare attribution
Amadeus Our heuristics
attribution No function S M # Total
No function 2,488,540 17,535 998,736 3,504,811
S 229,861 106,848 14 336,723
M 1,359,167 3,796 4,861,980 6,224,943
# Total 4,077,568 128,179 5,860,730 10,066,477
12

14. Contents
Motivation
Data
Identifying women and board members
Women on boards: stylized facts
Results
Conclusions
13

15. Prevalence
Unweighted average
Management boards (*) 22.55% 20.62%
Management boards (**) 16.80% 13.45%
Supervisory boards 23.30% 16.72%
Weighted average
Management boards (*) 14.13% 14.49%
Management boards (**) 16.99% 14.45%
Supervisory boards 32.48% 21.54%
% of firms with no women on board
Management boards (*) 70.86% 72.28%
Management boards (**) 77.80% 78.96%
Supervisory boards 59.11% 68.35%
14

16. Prevalence: management boards
15

17. Prevalence: supevisory boards
16

18. Time trends over time: management boards
17

19. Contents
Motivation
Data
Identifying women and board members
Women on boards: stylized facts
Results
Conclusions
18

20. Hypothesis 1 & 2
19

21. Hypothesis 3
20

22. Hypothesis 3
21

23. Summary of insights
• Adams & Kirchmaier (2016): general openness to women should make it
easier for them to be on supervisory boards (⇒ management boards)
• Nothing robust about womens participation in the labor markets
• Only gender wage inequality works consistently, but weak
• Not much evidence for Ingelhart either
• No consistent patterns for competition or knowledge intensity
• Too much of a leap
22

24. Contents
Motivation
Data
Identifying women and board members
Women on boards: stylized facts
Results
Conclusions
23

25. Instead of conclusions
• The data beyond stoclisted show more women in M than in S
• The institutional “equality all around” story is not robust
• Documented patterns: key role of firms with no women on boards
• Women are becoming more numerous (and less “infrequent”) → changes
in selectivity patterns or changes in economy structure?
• Changes in corporate Europe = changes in institutional Europe
24

26. I am happy to take questions
Thank you for your comments!
w: grape.org.pl
t: grape org
f: grape.org
e: j.tyrowicz@grape.org.pl
25