Introduction
This report draws upon matched employer-employee data collected as part of the OECD LinkEED project. Bringing together employer and employee data in a single framework allows for the analysis of the role of the firm in determining workers' wages, as well as the role of worker characteristics such as skills and gender for firm-level outcomes.
In the context of this report, the focus is on the extent of gender bias that exists in the energy sector – a sector that has been lacking in gender diversity since its creation and that is currently making considerable efforts to change this imbalance. [See for example the Reference to C3E initiative (now the Equality in Energy Transitions Initiative) and the Clean Energy Ministerial’s Equal by 30 Campaign.]
The data cover Austria, France, Germany, Portugal and Spain.1 Unless otherwise stated, the figures are aggregated data of all five countries. For Austria and Portugal, the data cover the full population of firms and workers in the corporate sector, whereas for France, Germany and Spain, the data cover a large representative sample of workers. For the purposes of this paper, the energy sector is defined at the three-digit level using ISIC and NACE classifications.2
The years covered in the data are: Austria (2000‑2018), France (2002‑2018), Germany (2002-2018), Portugal (2002‑2017), and Spain (2006-2018).
Firms in the energy sector are systematically compared with other firms, referred to as the non-energy sector. The firms in the non-energy sector represent 98% of firms in the full sample. Energy firms are approximately 50% larger on average than non-energy firms, in terms of employment, and wages are 9% higher. The wage gap between the sectors cannot be explained by the skill composition of the workforce, since controlling for skills actually increases the difference to 12%. High wages in energy firms instead reflect the importance of rents shared with workers. Wage premia in energy firms, i.e., the average firm wage after abstracting from workforce composition, are 8% greater on average than for the economy as a whole.3 Dispersion of the wage premia is also greater.
The relative importance of wage premia in energy firms may reflect a variety of factors, including the employees’ average performance related to productivity, capital-intensity or profits, the wage-setting power of firms, and the nature of wage-setting institutions. Significantly, the ability to set wages differently from their competitors for similarly qualified workers increases the potential for firms to discriminate between workers with different characteristics, including demographic attributes such as gender, underlining the importance of looking at potential bias in the energy sector.
The energy sector is male-dominated and women earn lower wages than men
In line with previous studies, we find significantly fewer women working in the energy sector compared to men. In relative terms the gap is more than twice as large as it is the case in the non-energy sector. In addition, we see that wages for female employees are almost 20% lower than for male employees, with the gap being somewhat greater than in non-energy firms. Significantly, the wage gap remains approximately the same when the effects of skill composition in terms of ability, education and potential experience are accounted for, indicating that the gap is not a function of gender differences in skill levels within firms.4
Average gender wage and employment gaps by sector
OpenExploiting the linked nature of the dataset we are able to assess the extent to which the wage gap between similarly skilled men and women reflects the tendency of men and women to work in different firms offering different wage premia (“between firms”) or differences in pay between similarly skilled men and women within individual firms (“within firms”).5 It is interesting to note that while the within-firm component is relatively more important, it is approximately the same for energy and non-energy firms. However, the between-firm component is considerably larger for energy firms than for non-energy firms. In other words, women in the energy sector are working in firms with relatively lower wage premia than men, and more so than women working in non-energy firms.
Average gender wage gaps within and between firms, non-energy vs. energy sectors
OpenIt is also possible to compare the importance of wage gaps for different ages of employees. There is a generally increasing gender wage gap with age for both energy and non-energy firms amongst workers 40 years of age and younger. However, in the energy sector the gender wage gap continues to increase with age until almost 50 years of age. The evolution of the gender wage gap with age is a function of both “within-firm” and “between-firm” channels. In both the energy and non-energy sectors, for younger employees the within-firm component has the strongest effect on the wage gap, while for older employees the between-firm component is more important
Gender wage gaps arise from wage structures within firms and from access to firms that pay high wages
The finding that the between-firm component is relatively more important for energy firms motivates an interest to identify the mechanisms through which such “sorting” of employment between firms arises, building off the work of Card et al. This sorting can arise either through differences in the percentage of male and female employees in firms with different average wages (allocation component), and/or through differences in the gender-specific wage structure of firms (dispersion component). In order to disentangle the effects of these two different channels, a counterfactual was created based on the mapping of the gender composition of the workforce in non-energy firms onto energy firms. In effect the objective is to measure the sorting component in the energy sector if the firm wage premium dispersion was the same as on average in the country, holding constant the allocation of men and women.
The figure below shows that the relatively greater role of sorting in energy firms in comparison with non-energy firms is largely, but not exclusively, explained by the fact that in the energy sector wage premia differences between firms are more pronounced (dispersion component), rather than by the tendency of women to work in low-wage firms (allocation component). This result is confirmed with a further comparison in which the counterfactual is based on the mapping of the wage structure from non-energy firms onto energy firms.
Gender wage gap decomposition by sorting, dispersion and allocation for the energy and non-energy sectors
OpenReturning to the role of skills using the International Standard Classification of Occupations (ISCO) we see that the gender wage gap is higher for high-skilled workers. This may reflect both general access barriers to high-wage occupations within firms (the glass ceiling), but also differences in wages for work of equal value due to the greater importance of individual bargaining, discrimination and networking, with women at a disadvantage. In addition, the between-firm component of the wage gap is much more important for low-skill workers than it is for high-skill workers in the energy sector. This suggests that access to desirable energy firms is particularly an issue for low-skilled women.
Women are more likely to leave energy sector firms for jobs outside the sector
In light of these results, we now decompose the gender wage gap within firms, following Card et al., into differences in tasks and responsibilities and differences in bargaining and discrimination.6 This is a critical distinction because the public and corporate policies to address the two cases may be very different. On the one hand, the “division of responsibilities” has important implications for both recruitment and advancement practices. On the other hand, issues related to negotiations and wage-setting stem most notably from gender norms and biases of both employers and employees. It is interesting to see that the bargaining and discrimination component is relatively more important in the energy sector than in the non-energy sector.
Gender wage gap decomposition by tasks and responsibilities, bargaining and discrimination, and sorting for the energy and non-energy sectors
OpenAs a final contribution, we focus on career trajectories. The data show that women in the energy sector are less likely to remain at a given firm than men at all ages, a trend which is not present in the non-energy sector. For employees that leave their firms there are three possible destinations: i) leaving the labour market; ii) going to another firm in the energy sector; and, iii) going to a firm in the non-energy sector. In the first instance we find that women are more likely than men to leave the labour market at all ages in both the energy sector and non-energy sector, though the gender gap is more visible and consistent for energy.
In addition, women employed in the energy sector are much less likely than men to leave for another firm within the sector at all ages, something which is less evident in the non-energy sector. Men advance in their careers by moving to more productive and high-wage firms within the sector, while women tend to stay behind, increasing the gender wage gap between firms with age. We also see that there is a consistent (albeit minimal) gap showing that a higher share of women are leaving the energy sector for other sectors compared to men. It is interesting to note that the opposite is true for the non-energy sector.
Conclusions
The evidence presented indicates that the energy sector has a relatively low percentage of female workers relative to other parts of the corporate sector, signalling issues of attracting and/or retaining a diverse workforce. It also reveals that rents are relatively high, underscoring the importance of assessing how these rents are shared with men and women in the same firm, and how sorting of employees across firms with different rents and wage-setting practices contributes to the gender wage gap. The key findings presented can be summarised as:
- Gender wage gaps are greater in the energy sector than in the non-energy sector. Much of the gender wage gap is explained by differences in pay between men and women with similar skills within firms, but it is interesting that the between-firm component is markedly greater than it is in the non-energy sector due to the concentration of women in low-wage energy firms.
- Moreover, the between-firm element of the gender wage gap is particularly important for low-skilled workers in the energy sector whereas it is of relatively little importance for workers in the non-energy sector, regardless of skill level.
- A counterfactual experiment reveals that wage dispersion between firms plays a greater role than differences in gender shares in explaining the gender wage gap between firms.
- While it is found that differences in tasks and responsibilities have a significant impact on the gender wage gap in all firms, in the energy sector the role of bargaining and discrimination (and thus firm-level wage-setting practices) is markedly greater than in the non-energy sector.
- Women are more likely than men to leave the labour market at all ages in the energy sector. In addition, women employed in the energy sector are much less likely than men to leave for another firm within the sector, and they are more likely to leave the energy sector for another sector.
What do these results mean for policymakers?
Firstly, the relatively low employment rates of women in energy firms and the concentration of women in low-wage firms (sorting) and occupations (segregation) underscore the importance of ensuring that recruitment and promotion practices in the sector are unbiased. These trends are interlinked, since higher representation of women in low-wage firms and occupations is likely to be contributing to the low employment levels of women in the energy sector more generally. In particular, the findings suggest a lack of career mobility and advancement for women in energy jobs compared to those in other sectors, which will affect both the attraction and retention of a diverse workforce. If women are not advancing in the energy sector, there will be fewer female role models and mentors to attract more women. Furthermore, if women working in the energy sector are unable to advance in their careers, they will be motivated to change sectors.
The low employment rate of women in the energy sector is also certainly in part due to a pipeline problem, since there is still a disproportionately low number of women with STEM degrees, reducing the pool of potential female applicants for some firms and occupations. Specialisation differences between men and women can be explained most notably by gender norms and expectations, which are further enforced by the lack of female role models. Thus, measures taken to ensure transparency and fairness in recruitment and performance assessment will contribute to both attracting and retaining more women in the energy sector. Corporate training on unconscious gender bias, particularly amongst managers, can also play an important role in establishing equitable recruitment practices and performance assessments.
Secondly, elements related to the quality of the work environment can also have significant negative impacts on the career trajectories of female employees in the sector. Measures taken – above and beyond wage incentives – to help women meet and balance their career and personal objectives are essential. Such measures include focussing on improving work-life balance through the provision of adequate parental leave, affordable and quality child care, flexible working arrangements, but also creating a more supportive environment through the use of mentoring programmes, career guidance and training. Awareness building on unconscious gender bias and appropriate workplace behaviour can also be key to ensuring a safe and attractive workplace for women and men. In addition, effective sexual harassment policies are essential. While many such measures may be nominally gender-neutral they can have disproportionately more important benefits for female employees.
And finally, the relatively high pay gap in the energy sector between men and women with similar skills due to discrimination and bargaining underscores the importance of ensuring that workers (and others) are guarded against discriminatory practices. As a high-rent sector, there are greater returns on bargaining efforts and scope for discrimination. While we are not in a position to assess the importance of the factors that drive gender differences in bargaining and discrimination, gender bias in corporate practices related to wage setting may play a role. A number of countries have introduced “pay transparency tools” to help close the gender wage gap. In some cases, these focus on information provision, but in other cases sanctions can be imposed for firms in which the gap is particularly significant. In addition, collective bargaining can play an important role in reducing the scope for discriminatory policies, including those related to gender.
There is also evidence that women are less likely to demand wages that reflect their true contributions to firm rents. Looking at the employer side, a recent study shows that while women and men are equally likely to request a raise, men are more likely to succeed with their request even when controlling for background factors like education, tenure and contract type. These explanations apply to the entire workforce, thus it is worth investigating why such issues might be more prominent in the energy sector compared to the non-energy sector.
In conclusion, this paper has reported on the analysis of quantitative data related to employee and firm characteristics and market structure on gender gaps in employment, career trajectories and wages. However, assessing the implications of discriminatory practices as well as the benefits of policies to address them requires complementary work, eliciting qualitative information directly from those affected. Such research was recently conducted by the Nuclear Energy Agency, which surveyed over 8 000 women working in the nuclear energy sector across the world in 2021. The reported key barriers facing women in the sector are a lack of pay transparency and fairness, workplace sexual harassment, a lack of female role models, and absence of measures to help with work-life balance, particularly concerning pregnancy and family related-duties. All of the barriers were identified to have a negative impact on gender wage gaps and career trajectories. These findings are consistent with those of other sectors that have historically employed mostly men and that are also facing a challenge of retaining women.
The authors responsible for this research are Antton Haramboure and Alex Hijzen from the Employment, Labour and Social Affairs Directorate of the OECD, as well as Ashley Acker and Nick Johnstone from the Energy Data Centre of the IEA.
References
France
This work is supported by a public grant overseen by the French National Research Agency (ANR) as part of the “Investissements d’Avenir” programme (reference: ANR-10-EQPX-17 - Centre d’acces securise aux donnees – CASD).
Germany
The data access to the SIEED was provided via on-site use at Centre Secure Data Access Center (CASD) of the National Institute of Statistics and Economic Studies (INSEE) and subsequently remote data access via the Josua platform from the Research Data Centre (FDZ) of the German Federal Employment Agency (BA).
SIEED Citation: Berge, Philipp vom; Schmidtlein, Lisa; Seth, Stefan; Graf, Tobias; Grießemer, Stephan; Kaimer, Steffen; Köhler, Markus; Lehnert, Claudia; Oertel, Martina; Seysen, Christian (2020): "The Sample of Integrated Employer-Employee Data (SIEED): SIEED 7518, Version 1". Research Data Centre of the Federal Employment Agency (BA) at the Institute for Employment Research (IAB). DOI: 10.5164/IAB.SIEED7518.de.en.v1
Energy sector ISIC/NACE selection: 051 Mining of hard coal, 052 Mining of lignite, 061 Extraction of crude petroleum, 062 Extraction of natural gas, 072 Mining of non-ferrous metal ores, 091 Support activities for petroleum and natural gas extraction, 191 Manufacture of coke oven products, 192 Manufacture of refined petroleum products, 351 Electric power generation, transmission and distribution, 352 Manufacture of gas; distribution of gaseous fuels through mains, 353 Steam and air conditioning supply, 473 Retail sale of automotive fuel in specialized stores, 493 Transport via pipeline.
Workforce composition is controlled for through the inclusion of worker fixed effects and “flexible” experience earnings profiles.
Worker fixed effects and flexible experience earnings profiles by gender.
The sample used to generate gender wage gap breakdowns (bargaining, sorting, etc.) is more restrictive than that used to generate the overall gender wage gaps by sector. Specifically, the decomposition can only be carried out on firms with gender diverse data (i.e. at least one male and one female employee) which are connected to other firms by at least one man and one woman who changed firm during the period of time studied.
Methodologically this is done by capturing the effects of bargaining and discrimination through firm fixed effects by gender, and positing that conditional on worker fixed effects, the difference between the estimated gaps in firm fixed effects relative to a model without firm fixed effects can be attributed to differences in tasks and responsibilities.
Reference 1
France
This work is supported by a public grant overseen by the French National Research Agency (ANR) as part of the “Investissements d’Avenir” programme (reference: ANR-10-EQPX-17 - Centre d’acces securise aux donnees – CASD).
Germany
The data access to the SIEED was provided via on-site use at Centre Secure Data Access Center (CASD) of the National Institute of Statistics and Economic Studies (INSEE) and subsequently remote data access via the Josua platform from the Research Data Centre (FDZ) of the German Federal Employment Agency (BA).
SIEED Citation: Berge, Philipp vom; Schmidtlein, Lisa; Seth, Stefan; Graf, Tobias; Grießemer, Stephan; Kaimer, Steffen; Köhler, Markus; Lehnert, Claudia; Oertel, Martina; Seysen, Christian (2020): "The Sample of Integrated Employer-Employee Data (SIEED): SIEED 7518, Version 1". Research Data Centre of the Federal Employment Agency (BA) at the Institute for Employment Research (IAB). DOI: 10.5164/IAB.SIEED7518.de.en.v1
Reference 2
Energy sector ISIC/NACE selection: 051 Mining of hard coal, 052 Mining of lignite, 061 Extraction of crude petroleum, 062 Extraction of natural gas, 072 Mining of non-ferrous metal ores, 091 Support activities for petroleum and natural gas extraction, 191 Manufacture of coke oven products, 192 Manufacture of refined petroleum products, 351 Electric power generation, transmission and distribution, 352 Manufacture of gas; distribution of gaseous fuels through mains, 353 Steam and air conditioning supply, 473 Retail sale of automotive fuel in specialized stores, 493 Transport via pipeline.
Reference 3
Workforce composition is controlled for through the inclusion of worker fixed effects and “flexible” experience earnings profiles.
Reference 4
Worker fixed effects and flexible experience earnings profiles by gender.
Reference 5
The sample used to generate gender wage gap breakdowns (bargaining, sorting, etc.) is more restrictive than that used to generate the overall gender wage gaps by sector. Specifically, the decomposition can only be carried out on firms with gender diverse data (i.e. at least one male and one female employee) which are connected to other firms by at least one man and one woman who changed firm during the period of time studied.
Reference 6
Methodologically this is done by capturing the effects of bargaining and discrimination through firm fixed effects by gender, and positing that conditional on worker fixed effects, the difference between the estimated gaps in firm fixed effects relative to a model without firm fixed effects can be attributed to differences in tasks and responsibilities.