More Women in the Labor Force

During the past few decades, European female labor force participation has increased substantially. From participation rates of at most 40 percent in the early 1980s in a number of advanced European countries, including Spain and Ireland (where participation rates were then below 40 percent), most advanced countries now stand at about 80 percent for women ages 25–54 years—the EU 2014 average. As a result, European female labor force participation is now almost on par with that of North America and East Asia and is well above that of South Asia and the Middle East and North Africa (Figure 4).

View Full Size

Progress in Female Labor Force Participation Rate

Citation: Departmental Papers 2016, 002; 10.5089/9781513562513.087.A001

Sources: Eurostat and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

However, progress has been uneven across countries and has stalled in recent years. While the increase in women’s participation in Spain and Luxembourg has continued at a broadly constant pace during the past three decades and Malta’s female participation rate almost doubled in the past 15 years (albeit from a very low level), progress in Italy has been much slower.⁶ In contrast, labor force participation among prime working-age women in some emerging and northern European countries has traditionally been about 80 percent or higher, leaving less space for further significant increases. In fact, most of these countries have remained at broadly unchanged levels during the past two decades, and participation in Romania has declined moderately. In turn, whereas participation rates appear to be converging to Nordic levels, significant dispersion in participation across Europe remains, ranging from about 65 percent in Italy to about 90 percent in Lithuania, Slovenia, and Sweden (Figure 4).

Gender gaps in participation also remain. While there has been a marked increase in female labor force participation rates, they remain well below male participation rates. As of 2014, the gender participation gap was above 10 percent in a majority of countries, and above 20 percent in Malta and Italy. In contrast, the gap was only about 5 percent in Sweden and Norway and virtually closed in Lithuania. The gender gap also varies across age groups and education levels. In Italy, the participation gender gap is most prevalent among people older than 30, while in Poland it narrows for people in their 40s and 50s, when women are past their prime childbearing years. Gender gaps also tend to narrow with higher education levels.

In addition, many working women are employed at less than full time. The average number of hours worked per week has remained broadly stable over the past decade for the average EU country. However, this masks substantial variation across countries. In the Netherlands, a high female labor force participation rate coincides with a considerable gap in hours worked between women and men, as more than half of women between the ages of 25 and 54 are employed part time. In Germany, women work about 30 hours per week, while men work for nearly 40 hours per week. On the contrary, women and men in Bulgaria work equally long workweeks of about 40 hours. While enhanced opportunities to work part time can lift participation rates through a reconciliation of family life and employment, part-time employment may also result from policy-induced constraints to taking up full-time work (for example, taxation or under-provision of childcare).⁷

But Still Few Women in Corporate Leadership Roles

More European women have entered the corporate boardroom. Since 2003, when Norway passed a law mandating at least 40 percent representation of each gender on the board of publicly listed companies, many European countries have followed suit (Profeta and others 2014). Most recently, Germany passed a law that requires publicly listed companies to have women occupy at least 30 percent of supervisory seats as of 2016. Overall, the introduction of quotas has supported a substantial rise in the share of women on the boards of Europe’s largest publicly listed companies (Figure 5).

View Full Size

Women in Senior Positions

Citation: Departmental Papers 2016, 002; 10.5089/9781513562513.087.A001

• Download Figure
• Download figure as PowerPoint slide

However, most countries are still a long way away from gender parity in senior corporate positions. While legal requirements have boosted the share of women in the boardroom to about 18 percent, only 12 percent of executive positions among Europe’s 620 largest listed companies were held by women in 2015.⁸ In the broader corporate sector, women have made greater strides. Analysis of the gender composition of senior positions—both in management and on corporate boards—of more than 2 million companies in 34 European countries reveals that almost a quarter of such positions are held by women.⁹ However, the cross-country variation is large. Further, in all countries, a sizable gap remains between the gender composition of the workforce and the gender composition of senior positions.

Greater gender parity in senior corporate positions has been achieved in countries where women are attached to the labor force full time. Various hypotheses have been offered for the underrepresentation of women in leadership positions in the corporate world as well as in politics: from demand-side constraints, such as preexisting social norms and gender stereotypes that create a glass ceiling for women, to supply-side explanations, such as women’s shorter work hours. For instance, preexisting social norms and gender stereotypes may serve to bias bosses and voters against appointing women as managers and leaders (Huddy and Terkildsen 1993; Eagly and Karau 2002). Lack of exposure to female leaders, in turn, may perpetuate biased perceptions of women’s effectiveness in leadership roles (Beaman and others 2009). Women themselves might not believe in their ability to lead, since they rarely see other women succeed in such positions (Beaman and others 2012). They may also leave high-power career tracks to have children (Bertrand, Goldin, and Katz 2010).

While pinning down the causes for the underrepresentation of women in senior positions is beyond the scope of this paper, two stylized facts emerge from our analysis. The overall female labor force participation rate is not a good predictor of the representation of women in senior positions in the broader corporate sector. However, a very strong negative association exists between the incidence of part-time employment among working women and the share of women in senior corporate positions, lending support to the supply-side explanations for the gender gaps in senior positions (Figure 5, panels 5 and 6).

In Addition to Individual Choice, Policies Affect Employment Decisions

Both individual characteristics and policies likely affect a woman’s decision to work (Figure 6).¹⁰

• Individual characteristics. When deciding whether to join the labor force, women compare the value of home production relative to the return to working outside the house (Becker 1965). For example, the return derived from household work increases with the number of children, while higher education credentials strengthen incentives for labor force participation through higher potential earnings. Gender attitudes or beliefs about women’s role in society are also important, as they determine the disutility of market work from violating personally held beliefs or social norms (Fernandez 2013). In this respect, both demographics and women’s attitudes have become more favorable toward women working in Europe. Among demographic factors, women now tend to have fewer children and obtain more education than three decades ago, virtually closing the education gap vis-à-vis men (Bloom and others 2009; Eckstein and Lifshitz 2011). At the same time, social norms and attitudes toward women working have become more favorable, and the rising number of women in positions of power provides role models for new generations. In 2014, women accounted for 28 percent of European parliamentarians, up from 18 percent about two decades earlier.

• Policies. Policies can create substantial (dis)incentives for women to work, in particular for women with children. First, the tax system can create disincentives to work, or to work full time, for the second earner in a family (often a woman) through a relatively high marginal tax rate (Bick and Fuchs-Schündeln 2014; Dao and others 2014; Colonna and Marcassa 2015). However, there has been no clear direction of change in this area, and taxation for married couples across countries varies from completely joint to separate. In contrast, specific family-oriented policies have generally moved in the direction of supporting women’s participation in the workforce (Carta and Rizzica 2015). Public spending on early education and childcare has increased across most countries since the early 1990s, facilitating mothers’ return to work (Jaumotte 2003; Steinberg and Nakane 2012; Thévenon 2013). At the same time, family allowances in the form of cash lump-sum transfers have been generally reduced. While parental leave policies are adjusted only infrequently, a number of countries, including the United Kingdom, Ireland, and Slovak Republic, now provide more than 30 weeks of maternity leave for women, supporting the return of mothers to work (Ondrich and others 2003; Edin and Gustavsson 2008).¹¹

View Full Size

Individual Characteristics and Policies

Citation: Departmental Papers 2016, 002; 10.5089/9781513562513.087.A001

• Download Figure
• Download figure as PowerPoint slide

Our analysis confirms that both individual characteristics and policies are important for understanding the employment decision for women (Figure 7; Box 1). A substantial literature has examined the drivers of female labor force participation. However, without micro-level data it is difficult to fully account for individual attitudes and choice and establish the role of changes in policies. Using such data on individuals, Christiansen and others (2016a) examine the role of both individual characteristics and policies.¹² Specific policy recommendations would vary, however, depending on each country’s circumstances.

• Individual characteristics. The analysis highlights that while more education is associated with a higher probability of a prime-aged woman working, education does not help explain the extent of full-time versus part-time work. In contrast, while marriage in itself does not significantly alter the employment decision for women, among working women the data reveal that married women do tend to work shorter weeks than unmarried ones. Also, each additional child is associated with a lower probability of a woman working. Importantly, a woman’s self-reported attitude toward working, which helps capture her personal employment choice, is a strong predictor of whether she is working. Likewise, intergenerational patterns should not be dismissed. Women who grew up with working mothers are more likely to work themselves, suggesting that the gender gap can be gradually closed over time to the extent that policies do not discriminate against women working today.

• Policies. The significance of the coefficient on taxation underscores that as working women often earn the secondary income in a family unit, higher relative tax rates on the secondary earner discourage women from participating in the labor force (in particular in advanced Europe) and from working full time. However, the positive association between the probability of employment and public spending on childcare and early childhood education (in particular in emerging Europe) supports the hypothesis that public spending can facilitate the return to work after childbirth. In contrast, lump-sum cash transfers may lessen the necessity for a woman to work, given the associated increase in non-wage household income. While excessive parental leave may deter a woman from returning to work full time, more parental leave is associated with a higher likelihood of employment. Finally, the finding that changes in these policies matter more for women than for men underscores that removing disincentives created by policies can help narrow the gender participation gap.

View Full Size

Marginal Effects of Individual Characteristics and Policies on Female Employment

(Percentage points)

Citation: Departmental Papers 2016, 002; 10.5089/9781513562513.087.A001

Source: Christiansen and others (2016a).Note: Impact per one-standard-deviation increase (during 2002 - 12 across countries) in the given variable. Coefficient on “Married” is insignificant.

• Download Figure
• Download figure as PowerPoint slide

Recent changes in policies have supported female employment in a number of countries. Using the results of the empirical analysis described in Box 1, a decomposition of the actual change in employment rates across countries between 2002 and 2012 suggests that the positive evolution of attitudes toward women working have helped lift women’s employment rates (Figure 8). However, even after accounting for demographics and personal choice, policies have had significant influence. In particular, lower taxes on the second earner in a number of countries, including Norway and the United Kingdom, have helped support female employment. Across a number of countries, including the Czech Republic, Poland, and Norway, increased spending on childcare and reduced family allowance have also positively contributed.

View Full Size

Decomposing the Change in the Female Employment Rate, 2002–12

Lower relative tax has broadly supported employment, as have childcare spending and reduced family allowance.

Citation: Departmental Papers 2016, 002; 10.5089/9781513562513.087.A001

Source: IMF staff calculations.1/ Captures time dummy and other macro controls.

• Download Figure
• Download figure as PowerPoint slide

^{Box 1.}

What Shapes the Employment Decision for Women?

Over the past decades, there has been a substantial increase in the share of women working in EU countries. A large literature has examined the drivers of female employment in advanced economies, largely focusing on supply-side explanations including the role of policies. However, two questions remain: (1) to what extent do women’s decisions to work simply reflect individual preferences? and (2) can removing policy distortions and improving public provision of services that help women reconcile work and family life further boost female labor force participation in Europe?

In an accompanying background paper (Christiansen and others 2016a), we address these two questions using individual-level data from the 2000 and 2012 rounds of the “Family and Changing Gender Roles” module of the International Social Survey Programme (ISSP) across 24 European countries. The ISSP data provide detailed measurement of individuals’ beliefs concerning the relative roles of men and women in society. These beliefs include, among others, whether working mothers can ensure a warm and secure relationship with their children, whether family life and children suffer when mothers are employed outside the home, and whether women find work in the domestic sphere more fulfilling than work outside the home. With an exclusively European focus, the analysis takes advantage of these micro-level data to disentangle the effects of individual choice and macro-level policies on women’s employment decisions, exploring both extensive (whether a woman is employed) and intensive (number of hours worked) margins of female employment.

The estimated model of female labor supply assesses the relative importance of various demographic characteristics, women’s self-reported attitudes toward women working (from the principal component of responses to a total of nine questions in each of the two rounds of the ISSP surveys), and policy variables that are likely to affect women’s employment outcomes.

The model has the following general form:

$E_{jit}=\alpha +Z_{jit}\beta +X_{it}\delta +{\theta }_i+{\rho }_t+{\varepsilon }_{ijt},$

where E_jit takes the value of 1 if individual j living in country i is employed (Table 1, models 1 and 2) or working full time (Table 1, model 3) at time t; Z_jit includes individual characteristics, such as the number of children living in the family, marital status, education, age, whether the mother worked, and attitudes toward gender roles.¹ X_it reflects country policies, including the relative tax rate on the second earner, public spending on early child education and childcare (in percent of GDP), public spending on family allowance (in percent of GDP), and number of weeks of parental leave for women with job protection. θ_i and ρ_t are country and time fixed effects, respectively. All regressions also include the male employment rate and the output gap to control for gender-neutral labor market policies that affect the overall level of employment and cyclical demand-side factors. The regressions are estimated using a linear probability model where robust standard errors are clustered at the country-year level.

^{Table 1}.

Individual Characteristics, Policies, and Employment

Source: IMF staff estimates. Note: Robust, clustered (at the country-year level) standard errors in parentheses. Significance is denoted as follows: *** p<0.01, ** p<0.05, * p<0.1. All regressions include time and year fixed effects and controls for the output gap and the male employment rate.

^{Table 1}.

Individual Characteristics, Policies, and Employment

		Whether or not employed All women		Full-time vs. part-time Employed women
		(1)	(2)	(3)
Structural	Number of children	−0.0562***	−0.0565***	−0.0659***
		(0.0072)	(0.0072)	(0.0099)
	Education (years)	0.0378***	0.0391***	0.0047
		(0.0069)	(0.0071)	(0.0081)
	Education (years), squared	−0.0010***	−0.0010***	−0.0001
		(0.0002)	(0.0002)	(0.0003)
	Age (years)	0.0624***	0.0618***	0.0209***
		(0.0105)	(0.0104)	(0.0060)
	Age (years), squared	−0.0008***	−0.0008***	−0.0003***
		(0.0001)	(0.0001)	(0.0001)
	Mother working	0.0220**	0.0227**	0.0175
		(0.0108)	(0.0108)	(0.0115)
	Married	−0.0001	0.0011	−0.0626***
		(0.0126)	(0.0127)	(0.0125)
	Attitude	0.0439***	0.0437***	0.0281***
		(0.0045)	(0.0046)	(0.0038)
Policies	Tax on the second earner	−0.7531***	−0.6557***	−0.1016
		(0.0781)	(0.0520)	(0.0816)
	Childcare spending	0.1168**	0.1025***	0.0336
		(0.0441)	(0.0363)	(0.0299)
	Family allowance		−0.1048***	0.0328
			(0.0280)	(0.0326)
	Parental leave	0.1548***	0.2079***	−0.1338***
		(0.0174)	(0.0200)	(0.0211)
	Parental leave, squared	−0.0004***	−0.0007***	0.0005***
	(0.0001)	(0.0001)	(0.0001)
	Observations	10,495	10,495	8,174
	Adjusted R-squared	0.1524	0.1535	0.1575

Source: IMF staff estimates. Note: Robust, clustered (at the country-year level) standard errors in parentheses. Significance is denoted as follows: *** p<0.01, ** p<0.05, * p<0.1. All regressions include time and year fixed effects and controls for the output gap and the male employment rate.

View Table

^{Table 1}.

Individual Characteristics, Policies, and Employment

		Whether or not employed All women		Full-time vs. part-time Employed women
		(1)	(2)	(3)
Structural	Number of children	−0.0562***	−0.0565***	−0.0659***
		(0.0072)	(0.0072)	(0.0099)
	Education (years)	0.0378***	0.0391***	0.0047
		(0.0069)	(0.0071)	(0.0081)
	Education (years), squared	−0.0010***	−0.0010***	−0.0001
		(0.0002)	(0.0002)	(0.0003)
	Age (years)	0.0624***	0.0618***	0.0209***
		(0.0105)	(0.0104)	(0.0060)
	Age (years), squared	−0.0008***	−0.0008***	−0.0003***
		(0.0001)	(0.0001)	(0.0001)
	Mother working	0.0220**	0.0227**	0.0175
		(0.0108)	(0.0108)	(0.0115)
	Married	−0.0001	0.0011	−0.0626***
		(0.0126)	(0.0127)	(0.0125)
	Attitude	0.0439***	0.0437***	0.0281***
		(0.0045)	(0.0046)	(0.0038)
Policies	Tax on the second earner	−0.7531***	−0.6557***	−0.1016
		(0.0781)	(0.0520)	(0.0816)
	Childcare spending	0.1168**	0.1025***	0.0336
		(0.0441)	(0.0363)	(0.0299)
	Family allowance		−0.1048***	0.0328
			(0.0280)	(0.0326)
	Parental leave	0.1548***	0.2079***	−0.1338***
		(0.0174)	(0.0200)	(0.0211)
	Parental leave, squared	−0.0004***	−0.0007***	0.0005***
	(0.0001)	(0.0001)	(0.0001)
	Observations	10,495	10,495	8,174
	Adjusted R-squared	0.1524	0.1535	0.1575

Source: IMF staff estimates. Note: Robust, clustered (at the country-year level) standard errors in parentheses. Significance is denoted as follows: *** p<0.01, ** p<0.05, * p<0.1. All regressions include time and year fixed effects and controls for the output gap and the male employment rate.

In line with theoretical predictions and the findings in the literature, the analysis confirms the importance of demographic characteristics of women on their decision to work. Adding one more child to the household is associated with a 6 percentage point reduction in the likelihood of a woman working. More education increases women’s likelihood of employment. However, it does not affect the number of hours worked for those already in the labor force. Contrary to prior beliefs, married women in Europe are not less likely to work but working mothers who are married are more likely to work at less than full time. Similar to the findings of McGinn, Lingo, and Castro (2015), we document the presence of an intergenerational transmission of female employment outcomes in Europe. Women who grew up with working mothers are more likely to work outside the household as adults. Further, women with more egalitarian gender attitudes are more likely to be active in the labor market.

But policies also matter. The coefficient on the relative tax rate of the family’s second earner (usually a woman) is both statistically and economically significant, similar to previous findings in the literature (Jaumotte 2003; Thévenon 2013).² The design of tax policy should thus be mindful to minimize disincentives for women to work. Furthermore, the composition of family-friendly policies matters for female employment: higher spending on childcare services encourages women to participate in the workforce, but lump-sum allowances act in the opposite direction.³ Finally, there is an inverted U-shaped relationship between the number of weeks of parental leave and the likelihood of female employment. Here again, the design of parental leave—which generally increases female employment—should be mindful to avoid being excessive, with the optimal threshold calculated on average at 140 weeks. In sum, analysis of the ISSP microdata reveals that individual demographics, attitudes toward gender roles, and policies are all important drivers of the decision by women to supply market labor.

¹ Our findings are robust to the inclusion of a proxy for individual predicted wages, following Klasen and Pieters (2015). ² As noted in Christiansen and others (2016a), the estimates presented here suggest greater sensitivity of female labor force participation to relative taxation of the secondary earner compared to Jaumotte (2003) and Thévenon (2013). This could be due to a number of factors, including differences in methodologies, the time period covered, and the sample of countries included. ³ The availability of childcare is imperfectly captured by the variable used in the regressions (childcare spending in percent of GDP), as it does not account for differences in countries’ demographics, quality of public childcare provision, and availability of full-time childcare services. The latter is a particularly important caveat in the regression for full- versus part-time employment.

Corporate Performance May Improve

Policies that strengthen women’s attachment to the labor force could help build the pipeline of women for senior corporate positions. One of the potential causes for the persistent gender gaps in senior positions may be the limited supply of women willing or able to take on these positions. Indeed, across European countries, a strong negative correlation exists between the share of women employed on a part-time basis and the presence of women in senior corporate positions. While part-time employment is a useful entry point to the labor market for women whose labor supply is constrained by family responsibilities, policies that boost the overall labor supply of women and facilitate their eventual transition from part-time to full-time employment could help narrow gender gaps at the higher rungs of the career ladder.

In turn, greater gender equality in senior positions could generate significant benefits at the firm level. Diversity might improve corporate productivity to the extent that it fosters complementarities in skills, generates knowledge spillovers, stimulates critical and creative thinking, makes the workplace more enjoyable, or stimulates demand.¹³ Given the existing differences in preferences and behavior along gender lines, important complementarities arise between the managerial style of men and women.¹⁴

Moreover, the economic returns to gender diversity in senior positions may have risen.

• More women in the labor force. Over the past three decades, millions of women have joined the labor force in Europe, while senior corporate positions continue to be held mostly by men. Bridging the widening gender gaps between those who hold senior positions in the corporate world and the workforce could improve firm performance.¹⁵ Women in leadership positions may be more likely to support family-friendly changes in corporate policies or serve as role models for other women, thereby raising the productivity of female workers. Women’s leadership style may also be more effective in female-dominated or female-oriented settings (Eagly, Karau, and Makhijani 1995).¹⁶

• High-tech and knowledge-intensive sectors. Relative to traditional industries, sectors characterized by complex tasks and innovative output stand to benefit more from greater diversity—including along gender lines—to the extent that it increases the set of ideas and potential solutions.¹⁷ At 40 percent of GDP, high-tech and knowledge-intensive sectors now account for a sizable fraction of economic activity in Europe.

Nevertheless, existing evidence on the impact of gender diversity on firm performance is inconclusive, often relying on small sample sizes.¹⁸,¹⁹ Influential work by McKinsey (2007) and Catalyst (2007) documented a strong positive association between the representation of women on the boards of Fortune 500 companies and corporate performance. However, later studies, which plausibly identify the causal impact on firm performance of raising the share of women in corporate boards, have challenged the early evidence (see, for example, Ahern and Dittmar 2012). Common to all studies is an important limitation: data availability typically constrains the analysis to publicly listed companies in individual countries.²⁰ The resulting small sample sizes make it hard to detect a statistically significant effect of gender diversity, particularly if its magnitude is small.

The new empirical evidence we present here suggests a strong positive association between firms’ financial performance and gender diversity in senior positions. Using a large sample of both listed and unlisted firms in Europe, we compare financial outcomes of firms within narrowly defined sectors based on the gender diversity of the senior management team and the corporate board (Christiansen and others 2016b). The analysis reveals that firms with a larger share of women in senior positions have higher return on assets (Figure 9). Adding one more woman in senior management or on the corporate board, while keeping the size of the board unchanged, is associated with an 8–13 basis points higher return on assets, or about 3–8 percent.²¹

View Full Size

Female Representation in Senior Positions and Firm Financial Performance

(Estimated ROA change from additional woman in a senior position, basis points)

Citation: Departmental Papers 2016, 002; 10.5089/9781513562513.087.A001

Sources: Orbis and IMF staff calculations.Note: Point estimate and 95% confidence interval. Return on assets computed using net income, profit before tax and earnings before interest and tax (EBIT), respectively.

• Download Figure
• Download figure as PowerPoint slide

Greater female representation could shape firm performance through two channels. Since firm performance and gender composition of its board and senior management are jointly determined, it is difficult to give a causal interpretation to the positive association. To shed light on the underlying mechanisms, we examine how sectoral characteristics shape the consequences of gender diversity (Box 2). As discussed earlier, the effect of greater female representation in senior positions is expected to be more pronounced in sectors with a larger share of women in the workforce and in sectors that demand greater creativity and innovative capacity, such as high-tech and knowledge-intensive industries. We find evidence for both of these channels at work.

• Women in the labor force. The positive correlation between gender diversity and firm financial performance is more pronounced in sectors where women form a larger share of the labor force (Figure 10). In the services sectors, where more than 50 percent of employees are women and a large gap exists between the gender composition of senior positions and the labor force, changing the composition of the board or management to include one more woman is associated with a 20 basis points higher return on assets. At the other end of the spectrum, in the construction sector, where relatively few women are employed in either the labor force or senior positions, changing the composition of the board or management to include one more woman is associated with about a 6 basis points higher return on assets—an estimate that is not statistically different from zero.

• High-tech and knowledge-intensive sectors. The positive association between gender diversity and firm performance is significantly higher in high-tech and knowledge-intensive sectors. For firms operating in these sectors, improving gender balance in senior positions is associated with a much larger increase in profitability (Figure 11).

View Full Size

High versus Low Female Intensity Sectors

Citation: Departmental Papers 2016, 002; 10.5089/9781513562513.087.A001

Source: Orbis and IMF staff calculations.Note: Gap represents the share of women in sectoral workforce less the share of women in senior positions. The diamond denotes the estimated increase in ROA from additional woman in a senior position. ROA computed using net income.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

High-Tech and Knowledge-Intensive Sectors versus Other Sectors

(Estimated ROA change from additional woman in a senior position, basis points)

Citation: Departmental Papers 2016, 002; 10.5089/9781513562513.087.A001

Source: Orbis, Eurostat, and IMF staff calculations.Note: Point estimate and 95% confidence interval. Return on assets computed using net income, profit before tax, and EBIT, respectively. Following Eurostat, industries are classified as high-tech (based on R&D expenditures) and knowledge-intensive (based on the share of workers with tertiary education).

• Download Figure
• Download figure as PowerPoint slide

^{Box 2.}

The Importance of Women in Senior Positions

There is a large literature on the effects of female representation in corporate boards on the financial performance of large publicly listed firms. However, little is known about how women in senior positions in the broader corporate sector shape the financial performance of firms. In an accompanying background paper (Christiansen and others 2016b), we provide new evidence on the link between gender diversity and firm performance in Europe.

Using the Orbis database, compiled by Bureau van Dijk, we build a sample of the unconsolidated financial statements of more than 2 million listed and unlisted companies across 34 European countries in 2013. We select firms with at least two people in the senior management team or on the corporate board, and for which information on the gender of members in senior management and corporate boards is available. Following the corporate finance literature, we capture financial performance as return on assets (ROA), which is measured in three different ways: net income over total assets, profits before taxes over total assets, and earnings before interest and taxes (EBIT) over total assets (two of which are addressed in Table 1). Gender diversity is reflected in the share of women in senior positions.

^{Table 1}.

Firm Financial Performance and Women in Senior Positions

Note: All regressions include country-industry fixed effects, indicators for the size of the firm, and control for the age of the firm, log assets, and size of the board. Robust standard errors are clustered at the industry level.

^{Table 1}.

Firm Financial Performance and Women in Senior Positions

		Correlation		Role of Female Intensity		Role of Knowledge-Intensity
ROA based on		Net income	Profit before tax	Net income	Profit before tax	Net income	Profit before tax
		(1)	(2)	(3)	(4)	(5)	(6)
Share of Women in Senior Positions		0.41 ***	0.44 ***	−0.26	−0.28	0.13	0.11
		(0.09)	(0.11)	(0.24)	(0.26)	(0.09)	(0.11)
Share of Women in Senior Positions				1.63 ***	1.74 ***
	*Female Intensity of Sector			(0.57)	(0.65)
Share of Women in Senior Positions						1.02 ***	1.19 ***
	*High-Tech/Knowledge Intensive Sector					(0.19)	(0.23)
N Obs		2,003,279	2,000,422	2,003,279	2,000,422	2,003,279	2,000,422

Note: All regressions include country-industry fixed effects, indicators for the size of the firm, and control for the age of the firm, log assets, and size of the board. Robust standard errors are clustered at the industry level.

View Table

^{Table 1}.

Firm Financial Performance and Women in Senior Positions

		Correlation		Role of Female Intensity		Role of Knowledge-Intensity
ROA based on		Net income	Profit before tax	Net income	Profit before tax	Net income	Profit before tax
		(1)	(2)	(3)	(4)	(5)	(6)
Share of Women in Senior Positions		0.41 ***	0.44 ***	−0.26	−0.28	0.13	0.11
		(0.09)	(0.11)	(0.24)	(0.26)	(0.09)	(0.11)
Share of Women in Senior Positions				1.63 ***	1.74 ***
	*Female Intensity of Sector			(0.57)	(0.65)
Share of Women in Senior Positions						1.02 ***	1.19 ***
	*High-Tech/Knowledge Intensive Sector					(0.19)	(0.23)
N Obs		2,003,279	2,000,422	2,003,279	2,000,422	2,003,279	2,000,422

Note: All regressions include country-industry fixed effects, indicators for the size of the firm, and control for the age of the firm, log assets, and size of the board. Robust standard errors are clustered at the industry level.

We begin by estimating the following regression model to capture the conditional correlation between a firm’s gender composition and its performance:

$\begin{array}{cc}{y}_{inc}=\beta *s{h}_{-}wm{n}_{inc}+\gamma *x_{inc}+{\alpha }_{nc}+{\varepsilon }_{inc}& \left(1\right)\end{array}$

where y_inc is the ROA of firm i, in industry n, operating in country c; sh_wmn_inc is the share of women in senior positions; x_inc are firm-specific controls (indicators for the size of the firm, firm age, the number of directors/senior managers, and tangible assets); and α_nc denotes the full set of roughly 16,000 country-industry fixed effects. Controlling for various firm characteristics, firms with greater gender diversity in senior management have consistently higher ROA (Table 1, columns 1 and 2).

Although the very granular fixed effects absorb a significant amount of heterogeneity, causal interpretation of the coefficient of interest, β, is problematic. The prevalence of women may be correlated with unobserved characteristics of the firm, which affect its financial performance. In equilibrium, it is also difficult to distinguish if the greater presence of female managers improves firm performance, or better-performing firms are simply able to attract more women.

To shed light on the causal effect of greater female participation in senior positions, we use a simple difference-in-difference strategy. In particular, our identifying assumption is that if gender diversity in management can help improve the performance of firms, its effect must be stronger in two types of industries: (1) industries that employ significantly more women in the labor force, and (2) industries with greater demand for the creativity and critical thinking that diversity in general may bring. To test these hypotheses, we estimate the following specifications:

$\begin{array}{cc}{\text{y}}_{\text{inc}}=\delta *{\text{SEC}}_{\text{n}}*{\text{sh}}_{-}{\text{wmn}}_{\text{inc}}+\beta *{\text{sh}}_{-}{\text{wmn}}_{\text{inc}}+\gamma *{\text{x}}_{\text{inc}}+{\alpha }_{\text{nc}}+{\varepsilon }_{\text{inc}}& \left(2\right)\end{array}$

Where SEC_n is alternatively (1) the female intensity of the sector to which the firm belongs and (2) an indicator for whether the sector is a high-tech or knowledge-intensive sector.

We find strong evidence for both of these channels at work. The positive association between gender equality in senior positions and firm performance is significantly stronger in sectors that employ more women in the labor force (Table 1, columns 3 and 4). For a firm that is in an industry in the top quartile in terms of female intensity, replacing one man with a woman on the board or in senior management would be associated with about a 20 basis points higher ROA. For a firm in a sector with no women in its labor force, the associated change would not be statistically different from zero or even slightly negative. Similarly, high-tech and knowledge-intensive sectors benefit significantly more from a higher share of women in senior management (Table 1, columns 5 and 6). In these sectors, the effect of adding one more woman on the board or in senior management, while keeping the size of the board unchanged, could lead to about a 30 basis points higher ROA.¹

These findings are robust to various empirical modifications. Results are robust to the treatment of outliers, they are not driven by firms in a particular country, and they are robust to using alternative measures of firms’ performance, such as labor productivity. Unfortunately, due to the cross-sectional nature of the data, we are unable to examine the association between gender diversity and volatility of returns.

¹ It is important to emphasize that, in equilibrium, it is difficult to distinguish empirically between the various theories of why greater representation of women in senior positions improves firm performance in certain sectors. Sectors that would benefit from gender diversity because of their technological characteristics already have a higher prevalence of women in the labor force, making it hard to confidently conclude whether it is the gender composition of the workforce or the nature of technology that matters for women’s value added in senior positions.