Social Economy, Gender Equality at Work and the 2030 Agenda: Theory and Evidence from Spain

: The improvement in women’s labor conditions and the elimination of segregation and other forms of direct or indirect discrimination have become one of the major challenges of the international political agenda, and as so have been included in the Sustainable Development Goals (SDGs) launched by the UN in the 2030 Agenda for Sustainable Development. At the same time, there is an increasing interest in the e ﬀ ects that the Social Economy (SE) might have on the achievement of the SDGs, as a consequence of its distinguishing of people-oriented principles. The goal of this paper is to analyze the speciﬁc contribution of SE entities to the reduction of gender inequalities in the labor market. We conduct an impact analysis with quasi-experimental counterfactual techniques, in which we compare one experimental group (the SE) with a control group (proﬁt-seeking ﬁrms) using labor data from Spain for the period 2008–2017. The results indicate that social economy entities signiﬁcantly contribute to the achievement of SDGs 5, 8 and 10, showing higher female participation, more stable jobs, and a lower degree of the glass-ceiling phenomenon.


Introduction
Female discrimination at work continues to be pervasive around the world. For example, the labor force participation for women aged 25−54 is 63%-compared to 94% for men [1]. This is in great part the consequence of women spending around 2.5 times more time on unpaid care and domestic work than men [2]. Furthermore, for those women who have a paid job, the gender wage gap is estimated to be 23% [1]. In addition, two billion workers-representing 61.2% of the world's employed population-are in informal employment, but if we only consider developing countries and disaggregate by gender, 92% of all female workers work in the informal sector in these countries, compared to 87% of male workers [3].
Within this reality, the improvement in women's labor conditions and the elimination of discrimination at work have become major challenges of the international political agenda, and as so have been included in the Sustainable Development Goals (SDGs) launched by the United Nations (UN) in the 2030 Agenda for Sustainable Development. Specifically, women's conditions in the labor market are considered in the 5th SDG, which is focused on gender equality and women's empowerment; in SDG 8, which regards economic growth, full employment and decent work for all; and ultimately, from a more global perspective, also in SDG 10, which targets the reduction of inequality within and among countries.
At the same time, there is an increasing interest in the effects that the Social Economy (SE) might have on the achievement of the SDGs [4]. In this sense, some literature argues that the SE can, among other things, not only improve labor conditions for both women and men, but also foster gender equality at work [5][6][7]. The aim of this article is to assess the potential contribution of SE entities to SDG 5, SDG 8 and SDG 10 from a gender perspective. To this end, we first need to construct the theoretical framework that links the SE to those SDGs. For the purpose of clarification, we can think of a two-step argument (Figure 1) in which the SE contributes to higher gender equality at work and, through that channel, has a positive impact on the mentioned SDGs from a gender perspective.
Sustainability 2020, 12, x FOR PEER REVIEW 2 of 19 SDG 5, SDG 8 and SDG 10 from a gender perspective. To this end, we first need to construct the theoretical framework that links the SE to those SDGs. For the purpose of clarification, we can think of a two-step argument (Figure 1) in which the SE contributes to higher gender equality at work and, through that channel, has a positive impact on the mentioned SDGs from a gender perspective. In this context, the relationship between gender equality at work and SDGs 5, 8 and 10, addressed in link 2 of Figure 1, is partly related to the fact that the labor market is one of the dimensions through which women can improve their equality and empowerment. Thus, achieving higher gender equality at work contributes positively to SDG 5: "gender equality and women empowerment". Gender equality at work is also part of the more global objective of "full employment and decent work for all" mentioned in SDG 8, as women are part of the labor force and thus an improvement in their participation and labor conditions is directly related to the achievement of that SDG.
SDGs 8 and 10 also address broader goals related to economic growth and the reduction of inequality within and among countries. In this sense, the relationship between gender equality and long-run economic growth seems to be an endogenous process with two-way causality [8].This means that reducing gender inequalities contributes to economic growth, and the other way around, economic growth contributes to the reduction of gender inequalities. The dynamics of the relationship between gender equality and economic growth are complex and beyond the scope of this paper, focused on link 1. For a deeper discussion on the topic see for example Cuberes and Teignier [8].
Focusing on the direction of causality from gender equality to economic growth, long-run growth models have shown how gender equality enhances human capital endowment, including future generations [9,10]. Empirical evidence also shows that gender equality boosts economic growth through the impact on the size of the labor force and on productivity [10][11][12], among other things, because it makes better use of the existing talent pool, especially from new incomers [12,13]. Higher gender equality also leads to higher income equality [12,14], which in turn can improve the sustainability of economic growth [15,16]. The literature also shows that companies with genderdiverse boards yield higher corporate profits [17,18] and higher performance resilience [19]; this is because these boards may lead to more innovative thinking, and because organizations that respond to women's needs may be more likely to attract talented men and women, and better understand possible customer segments by gender [20][21][22].
Thus, the available literature supports the positive impact that achieving higher gender equality in the labor market has on sustainable development, something that has been acknowledged by most public international organizations, and that constitutes one of the key aspects included in the 2030 Agenda. On the one hand, SDGs 5, 8 and 10 directly include targets related to gender equality at work, as we have seen above. On the other hand, gender equality is considered to have a strong positive cross-cutting effect on all the other SDGs that the Agenda recognizes in paragraph 20: "Realizing gender equality and the empowerment of women and girls will make a crucial contribution to progress across all the Goals and targets. The achievement of full human potential and sustainable development is not possible if one half of humanity continues to be denied its full human rights and opportunities. Women and girls must enjoy equal access to quality education, In this context, the relationship between gender equality at work and SDGs 5, 8 and 10, addressed in link 2 of Figure 1, is partly related to the fact that the labor market is one of the dimensions through which women can improve their equality and empowerment. Thus, achieving higher gender equality at work contributes positively to SDG 5: "gender equality and women empowerment". Gender equality at work is also part of the more global objective of "full employment and decent work for all" mentioned in SDG 8, as women are part of the labor force and thus an improvement in their participation and labor conditions is directly related to the achievement of that SDG.
SDGs 8 and 10 also address broader goals related to economic growth and the reduction of inequality within and among countries. In this sense, the relationship between gender equality and long-run economic growth seems to be an endogenous process with two-way causality [8].This means that reducing gender inequalities contributes to economic growth, and the other way around, economic growth contributes to the reduction of gender inequalities. The dynamics of the relationship between gender equality and economic growth are complex and beyond the scope of this paper, focused on link 1. For a deeper discussion on the topic see for example Cuberes and Teignier [8].
Focusing on the direction of causality from gender equality to economic growth, long-run growth models have shown how gender equality enhances human capital endowment, including future generations [9,10]. Empirical evidence also shows that gender equality boosts economic growth through the impact on the size of the labor force and on productivity [10][11][12], among other things, because it makes better use of the existing talent pool, especially from new incomers [12,13]. Higher gender equality also leads to higher income equality [12,14], which in turn can improve the sustainability of economic growth [15,16]. The literature also shows that companies with gender-diverse boards yield higher corporate profits [17,18] and higher performance resilience [19]; this is because these boards may lead to more innovative thinking, and because organizations that respond to women's needs may be more likely to attract talented men and women, and better understand possible customer segments by gender [20][21][22].
Thus, the available literature supports the positive impact that achieving higher gender equality in the labor market has on sustainable development, something that has been acknowledged by most public international organizations, and that constitutes one of the key aspects included in the 2030 Agenda. On the one hand, SDGs 5, 8 and 10 directly include targets related to gender equality at work, as we have seen above. On the other hand, gender equality is considered to have a strong positive cross-cutting effect on all the other SDGs that the Agenda recognizes in paragraph 20: "Realizing gender equality and the empowerment of women and girls will make a crucial contribution to progress across all the Goals and targets. The achievement of full human potential and sustainable development is not possible if one half of humanity continues to be denied its full human rights and opportunities. Women and girls must enjoy equal access to quality education, economic resources and political participation as well as equal opportunities with men and boys for employment, leadership and decision-making at all levels" [1] (p. 10).
Once the second link has been explained, we turn to the first link in Figure 1, which is the link of interest in our empirical study: why do we expect the SE to improve gender equality at work? The answer must come from the main distinguishing features that characterize the SE sector: priority of social and/or environmental objectives and principles over profits, and democratic governance (most SE types of entities have democratic governance systems, but not all, as in the case of foundations). Firstly, as Rivas and Sajardo [23] argue in their theoretical argumentation for male and female participation in SE entities, social economy values and principles are directly related to the pursuit of equality among all people, including men and women. Thus, at least in terms of their philosophy, SE organizations reduce female discrimination at work, as it is one of their explicit goals or principles to do so. The equality principle is reinforced with the governance systems of SE entities, which allow female members to have better access to the necessary formal decision structures to make sure this principle is accomplished. In addition, through their participation in the governance of these organizations, women have brought topics such as children, family, women's health, gender violence and discrimination into the social agenda [24,25], a pattern that should reduce gender inequalities at work. However, some studies have called attention to the reproduction of some inequalities within the SE sector that are present in the broader labor market, such as the higher prevalence of partial contracts among women [7,26]. According to feminist theories, the presence of some gender inequalities can be explained by the fact that SE organizations develop all their activity within the existing cultural context and can therefore be influenced by its customs and traditions [23] (p. 97).
A second argument for link 1 builds on the SE's strategies to attract and retain talent. The theory of the non-profit sector [27][28][29] considers that in SE entities, wages are lower than in profit-seeking firms, as workers obtain a series of non-monetary benefits that go beyond their salary, such as identifying themselves with the entity's social purposes or maximizing and guaranteeing employment [30]. The dynamics by which firms and workers match in the labor market is complex and exceeds the focus of this theoretical framework, but as analyzed by Lafranchi and Narcy [31], the reasons for the higher participation of women in the non-profit sector is a combination of supply and demand factors. The authors point out the mentioned benefits and policies beyond monetary remuneration as one of the reasons, along with the industry specialization of SE entities and access to part-time jobs and shorter workweeks, for the overrepresentation of women in these types of entities. In fact, the working environment that those human resource policies might create seems to be more attractive to women than to men, who traditionally respond to economic incentives such as higher wages, while women are closer to pursuing social values [32,33].
Besides the specific conditions of the job position, labor stability is a key issue from the worker perspective and, thus, a more equal environment should include gender equality in labor stability in the SE. Borzaga et al. [5] argue that SE organization show a higher resilience in times of crisis, for both women and men, increasing the stability of jobs in crucial times for workers. The authors explain this behavior by arguing that their ultimate goal is not to make profits but to help their members and society. During an economic crisis, we can then expect cooperatives and other SE organizations to work hard to maintain jobs, even if this means reducing salaries or running a business deficit.
While from a theoretical perspective, the philosophy of SE entities should ensure more equal labor environments, empirical evidence shows a disparity of results. Regarding participation, applied research suggests that women tend to have an equal participation or even become a majority in SE organizations. For instance, the share of female workers in social enterprises was 70% in Belgium and 67% in France [34]; 67% in the UK third sector versus 40% in the private sector [35]; 77% in the Portuguese SE sector [36] and 48% in Spanish SE companies [37]. However, in order to properly interpret these results, it is important to take into account that SE entities have a strong presence in some activities that are related to traditionally female roles, such as care or education. Thus, the comparison with the rest of the economy requires controlling for industry; otherwise, the prevalence of women might reflect this specialization. We also know that women have a high presence in SE entities in decision-making positions, although the results depend on the definition of the group analyzed.
In the UK, 50% of higher managers in the third sector are women, compared to 24% in the private sector [38]. In the case of Spain, the presence of women in very highly-qualified positions is 58.1% in SE organizations, versus 51.2% in other private organizations [37]. However, in Portugal only 25% of non-governmental organizations are chaired by women [38].
Evidence also support a higher resiliency in times of economic crisis observed in cooperatives around the world, providing more stable labor environments for workers [39,40]. In the case of Spain, Díaz and Marcuello [41] found the level of employment in cooperatives to be less sensitive to variations in the GDP, and therefore concluded that cooperatives offer job protection during recessions for both women and men. However, we do not have any empirical evidence about the effect that the SE can have on the equality of job stability between women and men. Despite this, in Spain, Calderon and Calderon [42] showed a higher quality of employment in cooperatives and labor societies based on some of the employment's critical dimensions, such as gender equality, flexibility, safety, inclusion, diversity and non-discrimination.
Finally, even though wages might not be considered a priority for people who work in SE, evidence supports the prevalence of less wage inequalities in the SE than in profit-seeking firms. Ben-Ner et al. [43], for example, found similar wages and less wage inequality in non-profit organizations in Minnesota; Leete [44] in the US, and Narcy [45] in France found the wage gap to be smaller in social enterprises. Specifically focusing on gender differences, Etienne and Narcy [46] showed that the unexplained gender wage gap was smaller in the non-profit sector in France, and that this was primarily due to lower levels of occupational segregation in the French non-profits. Teasdale et al. [35] found that women were paid on average 16% less than men in the third sector in the UK, compared with a gap of 33% in the private sector. In Spain, Santero and Castro [7] observed that wages were higher on average in the private sector, but so too was the gender wage gap: women in the SE sector earned 26.4% lower wages, versus 34.4% in the private sector. These authors also estimated that from this total wage gap, women earned 15.3% less than men in the SE sector due to direct gender discrimination-the wage differential that cannot be explained by observable factors. This discrimination went up to 29.1% for the private sector.
Overall, we can conclude that we can learn much about the impact of the SE on sustainable development by studying the effects of the SE on gender equality at work. From the available evidence, it seems that the SE tends to produce less wage inequalities and possibly higher employment stability for both women and men. However, we need more evidence to see if SE entities really foster equal labor participation, equal leadership opportunities, and equal job stability between women and men. These are the goals of this paper. In other words, our research intends to answer three important questions, always controlling for type of industry and size of entity, and always in comparison with profit-seeking firms: (1) do Spanish SE entities hire proportionally more women?; (2) do they offer more equal opportunities to reach high management positions?; and (3) do they offer women more job stability?
From this perspective, the first two research questions are addressed in our first working hypothesis that states that the SE will show higher female participation levels in the labor market and lower vertical segregation levels of women compared to the behavior of profit-seeking firms. These characteristics will contribute to SDG 5.5: "ensure women's full and effective participation and equal opportunities for leadership at all levels of decision making in political, economic and public life". This analysis will be completed by detailing personal characteristics to identify the differences in some vulnerable groups, i.e., people at risk or in a situation of social and labor exclusion: people with disabilities, people over 55 and people with low qualifications [47].
Our second working hypothesis, dealing with the third research question, states that the SE will show smaller gender differences in the stability of job positions than that observed in profit-seeking firms. This effect is closely linked to SDG 8.5 regarding " . . . full and productive employment and decent work for all women and men . . . ". Stability will be measured in terms of the existence of part-time jobs, unstable jobs and a higher number of and/or longer unemployment experiences.
Together, the realization of both working hypotheses also entails an improvement of SDG 10.2 (" . . . empower and promote the social, economic and political inclusion of all irrespective of age, sex . . . "), due to the reduction of the participation, promotion and stability gaps between women and men.
We focus on the Spanish case, as it is a country committed to the reduction of gender differences in labor conditions with the development of a legal and institutional framework, along with specific plans and policies, including the Spanish Organic Law 3/2007 for Effective Equality between Women and Men. However, regardless of these efforts, there is still a long way to go to reach gender equality in the labor market. Furthermore, Spain has been a pioneer in making visible the role of the social economy in society by the development of a specific law, Law 5/2011 for Social Economy and, more recently, the inclusion of the term Social Economy in the name of the former Employment Ministry (now the Ministry for Employment and Social Economy). This recognition has come along with the development of policy strategies, both at the national and regional levels, such as the current Spanish Strategy for the Social Economy 2017-2020. Moreover, the size of the Spanish economy and the availability of a database that permits the identification of Social Economy entities beyond cooperatives allow obtaining statistically significant results with interesting policy implications.
The present work contributes to the literature in three dimensions. First, it expands the literature on the impact that SE organizations can have on the realization of the SDGs related to gender equality. It does so by investigating the differences between standard profit-seeking firms and SE organizations in terms of women's access to employment and to high rank positions, further analyzing the added difficulties when they belong to other vulnerable groups. As we saw above, there is little empirical evidence for these important aspects. This article also examines the difference in the gender gap on job stability by comparing the career paths of women and men during 2008-2017. This analysis is important because most studies have focused on the temporary and part-time nature of women's contracts, leaving aside other aspects related to rotation, the number of contracts and their duration.
Our results indicate that SE entities show higher gender equality in terms of access to paid jobs, to high management positions and to more stable jobs.
Second, the scientific literature mostly focuses on particular entities like cooperatives, while this article studies a broader group of organizations that better resemble the international conceptual definition of the SE. Specifically, this study includes all those entities that are defined as belonging to the SE under Spanish Law 5/2011 (detailed description in next chapter). This allows us to have a clearer understanding of the SE's total contribution to sustainable development.
Finally, probably the most important contribution of this article lies in its methodological approach. We conducted an impact analysis with quasi-experimental counterfactual techniques, which permits the comparison of different gender equality indicators between SE and a control group formed by profit-seeking entities, controlling for other factors that might blur the specific contribution of SE to the SDGs analyzed. After reviewing the most relevant literature, posing the goals and specifying the working hypothesis, the next section describes the data and methodology used. The article then presents the main results obtained, and finally draws conclusions and makes recommendations.

Methodology and Database
The methodology used is based on counterfactual techniques for simulation exercises, which permit the comparison of a group exposed to a treatment with a group in which there is an absence of that treatment. In our case the objective is to compare the situation of people exposed to the treatment of "working in the SE", and thus affected by the distinguishing principles that characterize those entities, to people in the control group entities, who are not exposed to the treatment (limited liability companies and corporations). The social economy literature uses different but mostly equivalent terms to refer to traditional firms whose ultimate goal is to maximize profits, and refers to them as profit-seeking firms, capitalist firms or traditional firms, among other terms. We will use profit-seeking firms in this paper. In particular, when we compare both groups we focus on the effect of the "treatment" on gender differences in labor conditions, testing whether the treatment has any effect on these differences. Given that people are not randomly assigned to the SE or to profit-seeking firms, we used a quasi-experimental approach to the impact analysis [48].
The control group was obtained by stratified random sampling, in order to control for other factors that can also affect gender equality at work. In this sense, the comparative analysis was made with the hypothesis that the size of an entity and the industry in which it conducts its activity are the key global factors in terms of the differential behavior of an enterprise regarding labor conditions. Thus, a profile of the social economy organizations sample was obtained with regard to their sectors of activity and size, which was then reproduced when defining the stratification of the control group (CG), with the same cross distribution per business size and sector of activity as the SE entities sample. Thus, the samples have the following format-Equation (1): where j is the economic sector, i is the business size, SE accounts for the SE sample, and NSE accounts for the Non-Social-Economy or profit-seeking entities randomly included in the sample of the control group. A sub-sample for both the treatment and the control group was designed for the analysis of working lives during the period 2008-2017, as we excluded people from the SE sample that had any employment experiences outside the SE during the 10 years analyzed. The goal was to capture the differences between people that always worked in the SE in comparison to those who always worked in profit-seeking entities.
The differences between our treatment and control groups in terms of vertical segregation, leadership opportunities and job stability are analyzed with non-parametric (Chi-square test) statistical tests for independence between groups' behavior. We also compare the association between sex and categorical variables (distribution of workers in the analyzed variable) using a test of association, which allows us to compare the degree of association, and thus the influence of sex in each group considered (SE and CG). For continuous variables, in particular regarding stability at work, we compare the average situation between the SE and the control group (CG) for men and women using a multivariable analysis-two way ANOVA techniques-that allows us to jointly analyze the effect of the two variables considered on stability: sex and type of entity (SE or CG). Treatment of the database and statistical procedures were conducted using the SPSS Statistics Software (version 26, IBM Corporation, Armonk, NY, USA) and the STATA Software (version 11, StataCorp LLC, College Station, TX, USA).
The data used in the analysis come from the Continuous Sample of Working Histories (CSWH), a dataset provided by the Spanish Ministry of Labor and Social Security on a yearly basis. The CSWH is a micro-level dataset built upon Spanish administrative records. By means of a simple random sampling system, it consists of a representative sample (4%-1.2 million individuals) of the population registered with the social security administration during the sampling year, 2017 in our case, including active workers and recipients of unemployment benefits, among others. Moreover, the dataset provides longitudinal information for any individual included in the sample regarding all relationships with the Social Security System, which constitutes key information for the analysis of employment stability. This database can be obtained at zero cost through a request on the Ministry´s website.
The delimitation of the SE for the Spanish case was carried out considering the definition contained in Law 5/2011, of Social Economy, and those used by the scientific association CIRIEC-SPAIN (Centre International de Recherches et d'Information sur l'Economie Publique, Sociale et Coopérative) and the Spanish Business Confederation of Social Economy (CEPES). The following types of entities were identified: Cooperatives, mutual benefit societies, employee-owned companies, sheltered work centers, agricultural processing companies, fishermen's associations, associations and foundations.
The original CSWH dataset was refined by selecting those people that were formally working for some organization on 1 January 2018, so that we could identify all the enterprises that were active at the end of 2017. This means that we discarded pensioners, recipients of unemployment benefits and the self-employed. This last group poses a challenge in our analysis, as the legislation in Spain allows participants in a cooperative to be either all self-employed workers or employees. The implementation of the counterfactual analysis and the available data do not allow the inclusion of cooperatives with self-employed workers. The construction of the control group is not possible for this type of cooperative because the control group would be formed by individual entrepreneurships while in cooperatives, self-employed workers act jointly as a unique firm. Thus, we understand that it would be necessary to adopt a different approach to the one used in this paper to analyze labor conditions in that type of firm using counterfactual analysis.
There were 569,091 people in the database formally working on 1 January 2018 in any of 235,845 organizations. Within this sample, we identified 9155 SE organizations with 28,042 workers. Using the rest of the sample, we created the NSE control group, characterized by being randomly generated, but with the same sectoral and size structure as the SE sample. We ended up with 8926 profit-seeking firms and 27,472 workers in the control group.
To begin with the empirical analysis, we show a few basic characteristics of the Spanish SE sector, such as its main type of organizations and their size and sectoral structure. According to the sample obtained for SE entities, associations and foundations are the most numerous, representing 51.4% of the total, followed by cooperatives (34.5%) (Figure 2). cooperatives with self-employed workers. The construction of the control group is not possible for this type of cooperative because the control group would be formed by individual entrepreneurships while in cooperatives, self-employed workers act jointly as a unique firm. Thus, we understand that it would be necessary to adopt a different approach to the one used in this paper to analyze labor conditions in that type of firm using counterfactual analysis. There were 569,091 people in the database formally working on 1 January 2018 in any of 235,845 organizations. Within this sample, we identified 9155 SE organizations with 28,042 workers. Using the rest of the sample, we created the NSE control group, characterized by being randomly generated, but with the same sectoral and size structure as the SE sample. We ended up with 8926 profit-seeking firms and 27,472 workers in the control group.
To begin with the empirical analysis, we show a few basic characteristics of the Spanish SE sector, such as its main type of organizations and their size and sectoral structure. According to the sample obtained for SE entities, associations and foundations are the most numerous, representing 51.4% of the total, followed by cooperatives (34.5%) (Figure 2). A second and important characteristic of Spanish SE organizations is their large average size. While 48% of profit-seeking firms have only 1 to 10 workers, this proportion drops to 41% for SE organizations. On the other hand, the proportion of medium and large firms is higher in the SE (see Table 1).

Size (nº Workers) Percentage
Micro (1-10) 41.4% Small  36.1% Medium  18.3% Large (+250) 4.1% Total 100% One of the key issues in the analysis of the SE is its specialization in terms of industry. In this sense, even though the construction of the counterfactual analysis ensures having comparable groups, controlling for size of firm and industry, it is important to have in mind the distribution of A second and important characteristic of Spanish SE organizations is their large average size. While 48% of profit-seeking firms have only 1 to 10 workers, this proportion drops to 41% for SE organizations. On the other hand, the proportion of medium and large firms is higher in the SE (see Table 1).

Size (No. of Workers) Percentage
Micro (1-10) 41.4% Small  36.1% Medium  18.3% Large (+250) 4.1% Total 100% One of the key issues in the analysis of the SE is its specialization in terms of industry. In this sense, even though the construction of the counterfactual analysis ensures having comparable groups, controlling for size of firm and industry, it is important to have in mind the distribution of employment in the SE. Table 2 shows the distribution of the employees working in SE entities, showing that in terms of employment, the key sectors are related to social activities, including social work, education and health care services. However, employment is to some extent distributed among other sectors, like business-related services, trade, manufacturing, and agriculture.

Results
The results are presented in two large blocks, each one related to one of the research hypotheses described in the introductory section. Our first working hypothesis states that the SE will show a greater female participation in the labor market and a lower vertical segregation of women compared to the observed behavior in profit-seeking firms. We focus first on women's participation in the labor market, including the analysis of female participation in vulnerable groups. Secondly, for the analysis of vertical segregation, we will compare women's participation in positions of greater responsibility and the evolution of labor trajectories in terms of qualification levels. We will therefore try to detect the possible presence of a "glass ceiling" phenomenon posing an invisible barrier that prevents women from climbing to upper corporate rungs.
Our second working hypothesis is that the SE will show smaller gender differences in the stability of job positions than that observed in profit-seeking firms. The second part then analyzes workers' trajectories, allowing us to address employment stability from various perspectives, including the number of working experiences, their duration, the working day and their experiences of unemployment and inactivity.

Female Participation in the Labor Market
We begin analyzing female participation (degree of feminization) in the labor market by measuring the percentage of working women in relation to the total (Hypothesis 1a). Then, we introduce a specific working hypothesis (Hypothesis 1(a1)) to study the situation of women within groups at risk of exclusion from the labor market: disabled people, people over 55 years of age and people with low qualifications.
In statistical terms, we will analyze the degree of independence of the samples, our objective being to prove that they are not independent, and that the degree of female participation depends on the analyzed group (SE or CG). To do this, we use the Chi-square statistic, and if the level of significance is less than 0.05, the independence of the samples (SE and CG) is not accepted, and therefore, the differences observed can be said to be statistically significant.
Hypothesis 1a. The degree of feminization of SE companies is higher than that of profit-seeking companies.
The degree of feminization in the SE is 56.2% compared to 54.6% in the control group ( Table 3). The Chi-square test rejects the null hypothesis of equal distributions by sex (Table 3), so the differences in the degree of women's participation between both groups are statistically significant. In particular, a slightly better situation is observed in the SE in terms of contribution to the integration of women in the labor market with a higher participation. Therefore, from a global perspective, we can accept our first sub-hypothesis. In this regard, it is important to emphasize that since this is a counterfactual methodology, the main conclusion is that if the social economy lost its principles, women would have fewer opportunities to participate in the labor market, regardless of the productive specialization of the economy, since the control group has the same specialization and shows a lower participation of women. Women belonging to vulnerable groups may have twice the difficulty to access a job [49]. We will analyze the difference in participation of women over 55 years old with disabilities, and those with low qualifications. These groups can be analyzed by stating the following hypothesis: Hypothesis 1(a1). We expect to find a greater degree of feminization of vulnerable groups within SE companies than in the CG.
Considering workers with disabilities, Table 4 includes their distribution both by sex and by organization type (SE or CG), as a percentage of the total number of people with disabilities working in any of the two samples. The results show that the participation of this group is much higher in the SE than in the CG (3680 people compared to 452, respectively). In this sense, it is worth mentioning that the SE includes "Sheltered Employment Centers", whose mission is to encourage the integration of people with disabilities into the labor market. The distribution of employment in this group shows that 32% of disabled employees are women in the SE, compared to 4.8% of women in the control group. The contrast of equal distributions by sex between the CG and the SE shows that equality cannot be accepted (Table 4). Therefore, Hypothesis 1(a1) is accepted for workers with disabilities. In this sense, the SE contributes significantly to equalizing opportunities for this group through its principles, as opposed to the control group, for which an overall percentage is observed that is slightly higher than that established in the legislation for companies with more than 50 workers. In this respect, it is important not to be confused by a high degree of compliance on the part of the companies, since in our case it is the percentage of disabled workers considering the total number of workers, not by company. Regarding the relationship between age and gender equality in labor participation, differences between the SE and the CG distributions are statistically significant, positive and strong when workers are over 55 years old, which is the segment with the greatest risk of labor exclusion. Employment of over-55s is higher in the SE than in the CG. Of the total employment of the 55+ age group (Table 4), women in the SE represent 31.2% (compared to 21.4% in the CG). The gender distribution of this type of employee is statistically different between the SE and the CG according to the Chi-square test (Table 4). Therefore, Hypothesis 1(a1) is accepted for the group of people over 55 years old, implying that the degree of feminization of people over 55 is higher in the SE.
The last vulnerable group we will consider is the low-skilled group. People with low qualifications have greater difficulty to access the labor market, with higher rates of unemployment and greater job instability [50]. Qualifications will be approximated by the tax contribution group associated with each contract. The results show that women with low qualifications participate slightly more than men, both in the SE and in the CG. However, the participation of low-qualified women in the SE is 28.4% compared to 27% in the CG. These differences are statistically significant (Table 4), so we can accept Hypothesis 1(a1), as there is no independence in the distribution of the participation of workers according to sex and the analyzed group, female participation being higher in SE.

Vertical Segregation
The second aspect considered in the first working hypothesis focuses on the participation of women in positions of responsibility (Hypothesis 1(b1)) and on vertical segregation in terms of promotion possibilities (Hypothesis 1(b2)). With regard to female participation in the highest qualified group (university graduates and top management), the sub-hypothesis to be tested is :   Hypothesis 1(b1). Women have greater opportunities to participate in positions with a higher level of responsibility in the SE than in the CG.
In Table 5 we can see that there is a 10 percentage-point difference between the degree of feminization in the positions of greater responsibility in the SE (32.9%) as opposed to the CG (22.2%). Statistically, the test of equal distribution by sex for the samples of the CG and the SE indicates that there is no equality (Table 5), so we accept Hypothesis 1(b1) of our research: there are more opportunities for women to reach positions of responsibility in the SE than in the CG. Regarding the possible presence of a "glass ceiling" posing an invisible barrier that prevents women from climbing to upper corporate rungs, we analyze the trajectories of qualification levels in the last 10 years (2008-2017), for which our sub-hypothesis is :   Hypothesis 1(b2). Women have greater opportunities for promotion, understood as a change to a post associated with higher qualifications, in the SE than in the CG. Table 6 shows the distribution of the resulting subsamples according to whether people have gone down, maintained, or gone up in tax contribution groups, comparing the beginning of 2008 with the beginning of 2018. It is worth mentioning the share of workers that have had zero transitions as they had only one job during the analyzed period: in the SE it is 53.2% for women and 65.3% for men, and in the control group the percentages are 22.7% and 28.4%, respectively. Despite the conclusions in terms of stability that will be analyzed below, the following analysis only applies to the remaining set of workers, with one or more transitions. The first result we can derive from the above table is that the SE offers a more stable professional qualification environment, considering that the period includes the economic crisis. The percentage of people that have changed the qualification group of their job is higher in the control group, while in SE entities most workers remain in the qualification group. This is related to the number of jobs that people have in each group, which is smaller in the SE; this will be analyzed in the next section. If we compare the number of people that have gone up and down, in the SE, for each woman that has gone down, 3.3 have improved the category of their job, a ratio that is higher than that experienced by men (2.8 for men). Meanwhile, in the CG, for each woman that goes down, 1.7 go up, a ratio slightly higher than that for men (1.6 for men). Thus, we can accept our Hypothesis 1(b2): even though there are less workers moving, the chances of promotion (in terms of an improvement of the category) when changing job position are higher in the SE and, moreover, women are closing the gap in terms of the glass-ceiling at a faster pace than in the control group. However, according to the Chi-square test of independence and the association test, the relationship between sex and qualification changes is small, reflecting that most of the people remain in the same group, especially in the SE, where there is no significant relationship.

Decent Work and Stable Jobs: Female Workers in the SE vs. the CG
The scientific literature tends to associate the SE with more stable jobs as a logical consequence of their people-oriented principles. Our second working hypothesis states that the SE will show smaller gender differences in the stability and quality of job positions than that observed in profit-seeking firms.
Stability was analyzed by obtaining the number of episodes of employment, unemployment and those not covered by social security (due to periods of unemployment after consuming the period of unemployment benefit or exits from the labor force) for both groups, and the duration of these periods. In this respect, we have controlled for periods in which a person might have two jobs or even be working part-time while experiencing a period of unemployment benefit associated with another part-time job, or other overlapping situations. Thus, durations are net in this sense, permitting the accurate calculation of the length of time that a person worked or was unemployed during the 10 years analyzed. Given the high proportion of workers that experienced only one employment episode and no unemployment periods, together with an important group of people with a high volume of short-term employment episodes, the standard deviations tend to be high, and thus it is convenient to combine the use of mean and median values. Table 7 shows the percentage of people that satisfy the accompanying condition for each gender group in the SE and the CG. As discussed below, the SE shows a higher percentage of people with only one job during the considered period and those who have not experienced unemployment periods (including episodes not covered by social security), even if they have more than one job. The magnitude of the differences clearly shows that the SE provides more stable jobs than the CG, for both men and women. From a gender perspective, women are always worse off than men, although their situation is significantly better in SE entities. Despite this, the SE shows greater gender differences than CG, although a detailed study of the distribution will show a more complete and robust picture of the gender gap. Table 7. Men (women) that satisfy each condition as a percentage of men (women).  Table 8 shows the average and median number of episodes of employment, unemployment and periods not covered by the social security system, their average length and two specific indicators of job stability: the percentage of time that a person has been working during the analyzed period, and the percentage of time working in full-time jobs. The results obtained for our samples confirmed the higher resilience that SE employment has in periods of economic crisis and recovery. During 2008-2017, SE workers had a lower average (both in mean and median terms) of employment episodes, and a longer duration, confirming the higher stability of employment in the SE during recessions and recovery periods. Focusing in our working hypothesis, women and men showed the same median for the number of employment episodes, although it was higher in the CG, while the gap in averages was slightly smaller in the control group, although again, with a higher number of episodes. These results can only be interpreted in terms of stability once we include the average durations. In this sense, the gap for the median is lower in the SE (a women-men ratio of 0.5 compared to 0.71 for the CG). This conclusion is reinforced by the observed percentage of time that a person had been working, regardless of the number of episodes. While in the SE there was no gender gap, in the control group both mean and median values were lower for women than for men. Moreover, in terms of the percentage of time working full-time versus part-time, the gender differences were smaller in the SE, both in average and median terms, and again, the median in the SE was 100% of the time working full-time for both men and women, while in GC men showed 100% but women obtained a median of 89%.

Relations with Employment
The different impact of belonging to the SE or to the CG for men and women in terms of the duration of the employment episodes was tested using Two-Way ANOVA (Analysis of Variance) techniques. The results show that both factors (sex and pertaining to the SE or to the CG) had a significant impact on the duration of employment periods (see Table 9). Moreover, the cross-effect sex*group (SE or CG) also had a statistically significant influence on explaining the differences between observations. Thus, the differences in stability observed were statistically significant.
In order to have a deeper vision of the differences in job stability, Figure 3 gives extra information about the evolution of the gender gap in terms of the duration of employment periods, showing on the horizontal axes the length of the episode and on the vertical axes the kernel density function of the distribution of observations (workers). The left side of both panels shows the precarity of workers in the CG, with a large proportion working in temporary jobs (duration less than a year), a situation that particularly affects women. Besides the more even distribution of durations in the SE sample, an interesting fact is the concentration of workers on the right side of the distribution, implying highly stable jobs. Unfortunately, there is still a gender gap in that group of workers, both in the SE and the CG. In order to have a deeper vision of the differences in job stability, Figure 3 gives extra information about the evolution of the gender gap in terms of the duration of employment periods, showing on the horizontal axes the length of the episode and on the vertical axes the kernel density function of the distribution of observations (workers). The left side of both panels shows the precarity of workers in the CG, with a large proportion working in temporary jobs (duration less than a year), a situation that particularly affects women. Besides the more even distribution of durations in the SE sample, an interesting fact is the concentration of workers on the right side of the distribution, implying highly stable jobs. Unfortunately, there is still a gender gap in that group of workers, both in the SE and the CG.

Discussion
Gender equality at work is not only a key dimension of inclusive and sustainable growth, or a "must be done" approach in terms of social justice [51]. A lot of the scientific literature shows that it is also an important factor in terms of enhancing potential economic growth, by increasing domestic endowments in terms of labor force and human capital, which might lead to improvements in innovation and productivity.
Meanwhile, the UN Inter-Agency Task Force on Social and Solidarity Economy of 2015 argued that the SE can become a suitable vehicle to facilitate the SDGs. Social economy principles and motivations suggest that social economy entities might show a smaller gender gap in labor conditions, and thus contribute to the women-focused SDGs.
In the academic field, however, there are still few studies exploring the relationship between the SE and the SDGs [52]. For the Spanish case, the contribution of SE to SDGs was studied by Calderón et al. [30] at a national level, by Mozas [53] for Andalusia and by Bastida et al. [54] for Galicia. Additional efforts are therefore needed to link the SE to progress in achieving the SDGs, and in this sense, the work presented in this paper contributes to filling this research gap.
The results obtained in this paper show higher percentages of women working in SE entities, and thus with more opportunities to participate in the labor market. Thus, we accept our first hypothesis regarding the higher participation level of women in SE entities compared to firms that lack their principles, in line with previous works [7,55,56]. In this sense, the work conducted in this paper contributes to the literature by giving empirical support to the key role of SE in society concerning its ability to facilitate access to employment for disadvantaged groups and socially

Discussion
Gender equality at work is not only a key dimension of inclusive and sustainable growth, or a "must be done" approach in terms of social justice [51]. A lot of the scientific literature shows that it is also an important factor in terms of enhancing potential economic growth, by increasing domestic endowments in terms of labor force and human capital, which might lead to improvements in innovation and productivity.
Meanwhile, the UN Inter-Agency Task Force on Social and Solidarity Economy of 2015 argued that the SE can become a suitable vehicle to facilitate the SDGs. Social economy principles and motivations suggest that social economy entities might show a smaller gender gap in labor conditions, and thus contribute to the women-focused SDGs.
In the academic field, however, there are still few studies exploring the relationship between the SE and the SDGs [52]. For the Spanish case, the contribution of SE to SDGs was studied by Calderón et al. [30] at a national level, by Mozas [53] for Andalusia and by Bastida et al. [54] for Galicia. Additional efforts are therefore needed to link the SE to progress in achieving the SDGs, and in this sense, the work presented in this paper contributes to filling this research gap.
The results obtained in this paper show higher percentages of women working in SE entities, and thus with more opportunities to participate in the labor market. Thus, we accept our first hypothesis regarding the higher participation level of women in SE entities compared to firms that lack their principles, in line with previous works [7,55,56]. In this sense, the work conducted in this paper contributes to the literature by giving empirical support to the key role of SE in society concerning its ability to facilitate access to employment for disadvantaged groups and socially excluded people, in line with the work of Bastida et al. [51] for cooperatives in the Spanish region of Galicia. This is the case, for instance, of women over 55, a segment with a higher risk of labor exclusion that finds more opportunities in the SE compared to the control group. SE also contributes significantly to the fight against the double discrimination faced by women with disabilities. They usually exhibit a lower participation in the labor market and poorer conditions than men [49,57]. Bearing this in mind, although active policies such as legal obligations (for instance quota requirements) or fiscal benefits can be key to the employment of workers with disabilities for both SE and profit-seeking firms [53,56], our work shows that SE principles might act amplifying the impact of these policies, making them more efficient. This effect should be taken into account by decision-makers in the design of sustainable public policies.
The employment of women by SE entities also implies a higher degree of feminization of decision-making, contributing to the reduction of the glass-ceiling phenomenon. The simulation exercise conducted estimates that women account for 33% of upper management positions, which means a difference of 10 percentage points over that obtained by firms in the control group. In fact, if we analyze working lives over the period 2008-2017, whenever a change in job position and/or firm occurred, the odds that the new position would imply an improvement in job category are higher in the SE than in the CG, contributing to the reduction of glass-ceiling type barriers for women. Consequently, one side-result that is worth mentioning is the high percentage of workers that have maintained the category of their jobs during the recovery period analyzed, which the literature links to a more resilient type of job [7,58]. It is also true that there are differences between women and men, and a more in-depth analysis of working lives is required to shed some light on these differences.
Women working in the SE experience a more stable working life than in similar firms outside the SE, in line with previous studies [7,55,59]. Both the percentage of women that maintained their employment during 2008-2017 and those that have had subsequent jobs without unemployment periods (or periods not covered by the social security system) are higher in the SE than in the CG. Moreover, the length of time that women spent working during the analyzed period is higher in the SE and the gender gap with men is smaller.
By way of conclusion, the impact analysis conducted in this paper contributes to the visibility of the SE as a key agent in the attainment of the SDGs related to decent employment and equal opportunities. It also provides rigor to the comparison of SE entities with traditional profit-seeking firms, highlighting the contribution of the distinctive principles of SE to sustainable and gender inclusive growth. This contribution should be highlighted in the policy agenda for the coming years, which will undoubtedly test countries' capacities to maintain and create decent employment in the post-COVID-19 scenario. In the incoming socio-economic context, the design of employment strategies to deal with the current crisis and economic recovery should explicitly acknowledge and support the role of SE entities, providing institutional and even fiscal support, as an example of good practices in the creation of sustainable employment from a gender perspective.
Thus, the principles that distinguish the SE from profit-seeking firms contribute directly to the achievement of three SDGs. First, SE entities show higher female participation levels and a lower vertical segregation by sex, as stated in our first working hypothesis, and in line with target SDG 5.5, which is focused on the achievement of women's full and effective participation in society.
Second, the SE generates more resilient employment and shows, in global terms, a smaller gender gap in terms of stability, as tested in our second working hypothesis. This directly contributes to the achievement of SDG 8, and in particular target 8.5: "achieve full and productive employment and decent work for all women and men, including for young people and people with disabilities". In this respect, while our first working hypothesis shows that SE provides greater chances of full employment for women, the second hypothesis tested indicates that the labor conditions are closer to the concept of decent employment in terms of stability.
Overall, the acceptance of our two working hypotheses can also be interpreted as a contribution of the SE to SDG 10, in particular target 10.2 "empower and promote the social, economic and political inclusion of all irrespective of age, sex", by improving the participation and labor conditions of women, a key factor in terms of empowerment and social promotion.
Future lines of research should try to address some methodological and analytical issues. From a methodological perspective, the use of counterfactual analysis could be combined with propensity score matching techniques that permit the inclusion of several dimensions in the selection of control groups, reinforcing the comparability of the selected sample. In addition, the association techniques suffer from some limitations when different aspects need to be combined into a single analysis. In this regard, regression techniques could offer deeper insights into some of the results obtained in terms of job stability, where individual characteristics of workers might explain some of the diversity observed both in the SE and in the control group.
In terms of analytical issues, future work should address some of the limitations of the present paper. Firstly, it would be interesting to conduct a deeper analysis of the quality of employment generated by the SE for some vulnerable groups, such as some age and/or qualification groups. Another perspective that should be addressed is the contribution of the SE to territorial cohesion, and in particular the analysis of its impact in rural areas. Some previous works (see [37] for example) highlight the key role of SE in those areas, but further research is required. Finally, it would be interesting to dig deeper into labor stability issues and particularities, such as gender differences in working lives for specific age groups, a key aspect related to maternity, work-life balance and gender gaps in the labor market. Moreover, the dilemma between labor stability and retributions is addressed in the SE literature, where retribution adjustment is used as a mechanism to achieve higher stability, especially during crises. In that sense, the capacity of the social economy to react to the effects of the economic crisis on employment and wage adjustment has been acknowledged [46]. In fact, the analysis of SE retributions, including gender gaps (and the degree to which that gap that might be attributed to discrimination) is a crucial aspect in the study of the contribution of the SE to the SDGs. Some research considers SE entities as society's response to a state of wellbeing, social justice and inclusion, which differentiates them from purely mercantilism-based businesses [60].