What prevents women from reaching the top?

• Published date: 01 September 2022
• Author: Matti Keloharju,Samuli Knüpfer,Joacim Tåg
• Date: 01 September 2022
• DOI: http://doi.org/10.1111/fima.12390

DOI: 10.1111/fima.12390

ORIGINAL ARTICLE

What prevents women from reaching the top?

Matti Keloharju1,2,3Samuli Knüpfer3,4Joacim Tåg3

1Aalto University School of Business, Espoo,

Finland

2Centre for Economic Policy Research,

London,UK

3Research Institute of Industrial Economics

(IFN), Stockholm, Sweden

4BI Norwegian Business School, Oslo, Norway

Correspondence

Matti Keloharju,Aalto University School of

Business, P.O. Box21220, FI-00076 Aalto,

Finland.

Email: matti.keloharju@aalto.fi

Abstract

We use rich data on all business, economics, and engineer-

ing graduates in Sweden to study the lack of women among

chief executive officers (CEOs). A comprehensivebattery of

graduates’ characteristicsexplains 40% of the gender gaps in

CEO appointments and 60% among graduates with children.

The explanatory power mostly comes from absences and

unemployment, which are about twice as likely for women

as men. These gender differences increase following child-

birth, and they persist in the long run. We present and dis-

cuss potential explanations to the explained and remaining

gaps. Although the large unexplained share makes it hard to

pinpoint the exactreason for the gender gap in CEO appoint-

ments, the large contribution of labor market attachment to

the explained share suggests work–family trade-offs are an

important part of the story.

KEYWORDS

CEOs, family, gender gap, labor market attachment, qualifications

JEL CLASSIFICATION

G34, J16, J24, J31

1INTRODUCTION

Womendo not fare well in the executive labor market. For example,in S&P 500 companies, they account for 45% of the

workforce but hold only 27% of the executiveand senior-level official and manager positions. The fraction of women

is even smaller at the very top of the organization: they account for 6% of the chief executiveofficer (CEO) positions

(Catalyst, 2020). This low representation of women at the top is often referred to as the glass ceiling.

Thisis an open access article under the terms of the Creative Commons Attribution- NonCommercial-NoDerivs License, which permits

use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or

adaptations are made.

© 2022 The Authors. Financial Management published by Wiley Periodicals LLCon behalf of Financial Management Association Inter-

national.

Financial Management. 2022;51:711–738. wileyonlinelibrary.com/journal/fima 711

712 KELOHARJUET AL.

What prevents women from reaching the top? We study this question by following the careers of 14 cohorts of

business, economics, and engineering graduates—the three most common fields of education for CEOs—over a quar-

ter century and ask how their educational qualifications, family background, labor market and executiveexperience,

and absences explain their career success. Our data coverthe entire adult population of Sweden and all its employers,

including private firms and the public sector,resulting in an exceptionally large sample of 143,000 graduates. We col-

lect a comprehensive battery of characteristicsof the graduates and their families, which allows us to analyze a host of

gender differences in individual characteristics that likely are consequential for making it to the top. Wecomplement

the dataset with survey responses on the monthly hours worked byg raduates.Almost all of our data come from offi-

cial government registries and thus are likelymore reliable than the biographical and self-reported data used by many

studies on top executives.

We first analyze gender differences in individual characteristicsrelevant for making it to the top. We find graduate

women are about twice as often as men absent or not employed during the sample period, which spans about one-half

of the duration of their working life, including the years when they would be most likelyto progress on their careers.

Women work on average15% shorter hours than men. Women are also less likely to have experience as an executive

than men, in particular from executive assignments in production and operations, and sales and marketing,the two

most common functional backgrounds for CEOs. At the same time, women are more likely to have a business or eco-

nomics degree than men, the fields most predictive of making it to the top. Because of these different and potentially

countervailing forces, it is not ex ante obvious how much graduate characteristics are able to explain CEO appoint-

ments, or even which gender the marketviews to be on average more qualified for a CEO job.

Armed with these findings, we study the joint role of educational qualifications, family background, labor market

and executiveexperience, and absences in explaining top executive outcomes. A Blinder–Oaxaca (Blinder,1973; Oax-

aca 1973) decomposition allows us to estimate how the labor market for CEOs values the bundle of characteristics

each graduate possesses and how the value of this bundle differs by gender.This decomposition suggests the gender

gap in CEO appointments is in part explained by careers, in particular by prior executive-level experience men are

more likely to secure. At the same time, women’s higher likelihood of having a business or economics degree means

that educational qualifications do not explain the gap, but ratherwiden it.

Characteristicsrelated to labor market attachment—parental leave, sick leave, unemployment, and being outside of

labor force—are by far the most important contributor to gender gaps in CEO appointments. The attachment explains

26% of gender differences in CEO appointments. Among graduateswith children, attachment explains as much as 41%

of the gender gap in CEO appointments, that is, more than the 40% of the gap we cannot explain. The gaps in labor

market attachment primarily arise following the birth of the first child, when women’s absences increase and working

hours decline relative to those of men. The gaps narrow as the children grow up, but importantly,they do not return to

their prechildbirth levels during the 10 yearsfollowing the birth of the first child.

Our empirical setting does not afford us the luxury to claim that labor market attachment or career decisions are

independent of the graduates’ probability of reaching the top. Forexample, women may choose to differentially invest

in their career in anticipation of labor market discrimination, which may then affect their ultimate career outcomes.

Moreover, despite of our comprehensive battery of characteristics, a large share of the gender gap remains unex-

plained.We discuss three broad classes of explanations for this unexplained gap at the end of the paper—gender differ-

ences in job preferences; gender differences in leadership skills; and discrimination—but, like the literaturein general,

we lack the tools to differentiate between them. Instead, we offer additional evidence and arguments that allow the

reader to better interpret our findings.

We give perspectiveto the discrimination hypothesis by drawing from Statistics Sweden’s Work Environment Sur-

vey(wave of 2015), which asks its respondents about their experienced gender discrimination at work. The fraction of

reported gender discrimination among female (male) respondents in our sample is relatively small, 9% (2%). Because

the (unexplained) gender gap is so large, it is unlikely that discrimination can directly account for it. If discrimination

is to explain a substantial share of the CEO gender gap, the most plausible path would appear to go through women’s

underinvestment in their careers due to anticipated discrimination ratherthan directly.