2.1. Female Executives and Firm Performance in Family Firms
The incorporation of women into the corporate governance structures has become a relevant topic in recent years, especially since national and international regulations appeared. It is a fact that women remain consistently underrepresented in the upper management echelons of firms. The progressive incorporation of women into managerial positions has been one of the relevant aspects of good corporate governance. To prove the benefit for the firms of having women in management positions has been a trending topic in academic research. Especially, the study of the impact of female leadership on firm level performance is a key issue.
Unfortunately, there is no coherent theoretical framework that comprises all aspects of gender and how it could affect firm performance, although the resource-based theory has claimed for the outperform of women in managerial positions. According to Barney [
21], resource-based theory argues that human capital resources are key to competitive advantage. Employee and management capabilities are firm-level resources that are difficult for competitors to imitate. Katzenbach et al. [
22] concluded that many companies have underutilized human resources by not including women and minorities.
This theory suggests that firms employing more women managers have probably better succeeded at recruiting capable managers from the total available talent pool. Nevertheless, literature addressing the links between firm performance and female presence in governance is inconclusive, especially in the context of female presence on boards of directors. In this latter case, literature is mainly based on large and listed companies, and according to Carter et al. [
23], the net effect of gender diversity can be either positive or negative from a financial performance perspective depending on the different internal and external circumstances of the firm. The existing literature seems to confirm this idea.
Numerous studies find a positive relationship between women on boards and firm performance [
2,
3,
4]. More recently, Terjesen et al. [
24] conducted a study on 3876 public firms in 47 countries and found that firms with female directors have better financial performance. On the contrary, Shrader et al. [
5] analyzed the participation of women on the boards of 200 large U.S. firms found a negative impact on performance. Adams and Ferreira [
25] for a sample of U.S. firms found that the average effect of gender diversity on firm performance was negative. He and Huang [
26] also found a negative link for 530 U.S. manufacturing firms.
In Europe, the Scandinavian countries have been a pioneer in gender issues and the different studies show, as well, no consensus in their results. Smith et al. [
6] and Nielsen and Huse [
27] found positive effects, while Randoy et al. [
28] and Rose [
29] did not find a significant diversity effect on a firm’s performance.
More recently, Masud et al. [
30] examined a sample of companies in South Asian countries and did not find any significant relationship between female directors and the Environmental Sustainability Reporting Performance (ESRP). This result was expected, as female participation in SA countries’ board decisions is very limited compared to developed countries.
In Spain, Campbell and Minguez-Vera [
31] and Martin-Ugedo and Minguez-Vera [
32] obtained a positive effect between gender diversity. Nevertheless, Minguez-Vera and Martin [
33] analyzed the presence of women on the board of directors in a sample of SMEs, finding a negative impact on the return on equity (ROE).
Although the board of directors is considered part of the TMT, some members are not directly involved in the management of the company and, given the small proportion of women reaching senior positions, it seemed relevant to study the relationship between executive women and firm performance, especially considering that not-public firms sometimes have an informal board of directors. Women in top management and its impact on firm performance has been much less studied.
Women bring a number of strengths and skills to the management teams. Several studies analyze different management styles between men and women [
5,
34,
35].
Based on surveys and interviews with female leaders, Rosener [
36] found that women show an interactive leadership style. They encourage participation and share power and information with their subordinates. Sandberg [
37] revealed that women create a more flexible environment fostering exchange of ideas and knowledge. This style should produce positive effects on firm performance and the existing literature seems to support this idea and offers more conclusive results than those studies based on board of directors. Thus, the vast majority of the limited studies that analyzes the relationship between the presence of executive women and firm performance have found a positive influence on financial return [
5,
6,
7,
8,
9]. So, according to this latter literature, based on executive women, we expect similar results in our sample.
Hypothesis 1. Female executives exert a positive effect on firm’s return on assets (ROA).
This topic presents additional interest in the specific case of family firms, due to the fact that the number of women that reach managerial positions is higher in the family businesses [
38,
39]. According to Martinez-Jiménez [
40], family-owned and family-controlled businesses have higher percentages of women in the c-suite, as well as in top management positions and on the board, than other types of companies.
Despite these facts, the study of female leadership and its impact in firm performance of family firms has not received much attention. The study of Bjuggren [
41] is an exception where differences in firm performance due to female leadership in family and non-family firms was analyzed and the results indicated that female leadership makes much more of a positive difference on performance in family firms while this effect is negative in non-family firms.
Campopiano et al. [
10] provide a recent review of research on women’s involvement in family firms and identified three main aspects addressed by literature: the entry of women in family firms through entrepreneurship and succession, women’s career development in family firms, and women’s static presence in family firms.
In our study, we consider that women’s career development is the aspect that most affects women as executives. Traditionally, women have had preference to work in family firms. According to Dumas and Lyman et al. [
42,
43], family firms offer women many advantages such as flexible schedules (which help them combine their professional responsibilities with childcare), job security, more opportunities for personal growth, and greater chance to access managerial positions. Family businesses offer women opportunities that other businesses do not. For example, Salganicoff [
44] reported better positions, higher income, more flexibility in work schedules, and more job security for women who work in family businesses than for women who work in non-family businesses.
Nevertheless, women also face several problems in reaching top positions even though they are now as academically qualified as male counterparts. The invisibility problems and discrimination still exist in the base of many companies, which are essentially patriarchal. This problem is even harder in a country like Spain, in which there is still a certain belief in a distribution of roles between sexes and women are supposed to stay at home taking care of the family. In this sense, there is a research line focused on the gender pay gap (GPG) among managers. Based on a sample of Spanish managers, Scicchitano [
45] showed that women earned significantly less than men across wage distributions.
Based on the above, we consider that women who have access to leadership positions have had to overcome many obstacles that have made them better leaders, so we expect them to have a positive effect on firm performance. This is also consistent with the resource-based theory, which implies that human capital is the key factor for firms and it is also important in family firms. Therefore, we propose the following hypothesis:
Hypothesis 2. Female executives exert a positive effect on ROA in family firms.
2.2. Female Executives, Family Status, and Firm Performance
In this family firm context, the family status of the managers represents a key issue. Nevertheless, it is the object of conceptual controversy as different interpretations can be given according to the theoretical approach. Agency theory is double edged, on one hand it has been used to suggest that firms run by family executives benefit from lower agency costs because there is an alignment between owners and managers [
46,
47] and on the other, family executives are said to use their superior positions to benefit themselves at the expense of the company [
48,
49].
Furthermore, stewardship theory also applies to family firms and family leaders. It considers that family executives are especially attached and loyal to their firm. These managers are highly motivated and manage for the long-term.
Among the main issues, the most important research lines have been focused on whether the CEO is or is not a family member [
11,
12,
13,
14] and participation of family members in the TMT. The results of these studies are inconclusive.
Nevertheless, only a few studies have related the family status of family firm managers with gender. Cromie and Sullivan [
17] compared career experiences of executive women who were both family and non-family members. To do so, they interviewed 80 family women executives and 62 non-family executive women. They showed on their study that family members had advantages over non-family members in developing careers in family firms. In their study, the authors demonstrated that female managers that were family members had a long association with the family business because, generally, they had been working in the business for longer than their non-family counterparts. Family executives also experienced a sense of achievement and enjoyed job security and flexibility.
These results are consistent with stewardship theory assumptions, therefore we propose a third hypothesis:
Hypothesis 3. Executive women that are family members have a positive impact on firm performance.
2.5. Descriptive Statistics
The main feature we found regarding female executives (Femaleexecutives) was their expected fairly low participation. Executive women accounted for 12.67 per cent of the cases analyzed, but the median value was 0, meaning the likelihood was to find no executive women at all. Family firms prevailed, representing 60.95 per cent of the sample; non-family firms represented the remaining 39.05 per cent. This data was unsurprising, as over two-thirds of all Spanish companies are family owned.
Considering the small group of women in our sample, several key aspects are worthy of note. The main question was their representation in family firms, and we observed that most of these women were working in family businesses (69.30%), whereas 30.7% were working in non-family firms. The χ
2 coefficient (13.467 ***) in
Table 3 indicates that this difference was statistically significant, so we can state that women have better opportunities in family businesses than in their non-family counterparts. For a deeper understanding of these results, we have included a frequency distribution (
Table 4) that summarizes the relationship between our category variables (Familyfirm, Femaleexecutive, and Familyfemaleexecutive). Based on this table, it is possible to observe that although this could seem to be a good starting point, only 14.9% of family firms had some executive women. Even though this is a low value, it is above that found for non-family firms, where it was only 10.4%. The first implication is that further progress is necessary, both in family and non-family firms.
Finally, regarding women who work in family firms, their family relationship with the founding family is fairly close. We found that 7.37% of women in our sample were relatives of the main owner but this represented 86.16% of the total number of women working in family firms. This indicates that in our sample, family women had better access to management positions, a finding in line with that of other studies [
57,
58].
Table 5 shows mean statistics for the variables included in the data set, and
Table 6 shows the correlation matrix.
2.6. Methodology
The methodology was determined by taking into account two significant problems that arise when analyzing the impact of the governance structure on firm performance: endogeneity and unobservable heterogeneity. Regarding endogeneity, we found that our topic could be endogenous: because firm performs well, the family is more likely (or more easily) able to install a family female in the executive team. Therefore, we discussed an empirical specification that would solve two sources of endogeneity: omitted variables and reverse causality.
Omitted variables (both time-varying and fixed over time) may simultaneously affect firm performance and the female manager appointment process. For example, in the case of the firm’s attitude towards CSR issues, the stronger the sensitivity towards CSR, the higher the presence of female managers on the TMT. Furthermore, this attitude could affect firm performance through the reduced cost of capital [
59], higher valuation, and greater access to finance [
60,
61]. Moreover, the direction of causality between female managers and firm performance could be reversed. Women could prefer to work in family firms (or in firms with better performance), and at the same time family firms (or successfully directed firms) could prefer to establish a family female in the executive team (or to hire executive women).
To solve this problem, there exist several methods to use, basically through the use of instrumental variables; regardless, they are considered as fixed or random effects, as well as by the generalized method of moments (GMM). In this sense, we follow Li (2016) [
68] who reviews all the prevailing econometric remedies concluding that among all of them, GMM has the greatest correction effect on the bias. We then proposed a dynamic panel system generalized method of moments estimator (GMM) as a suitable estimator that takes into account whether female managers (regardless of their family link to the firm) have an impact on firm financial performance.
Finally, unobservable heterogeneity occurs when calculated estimators are biased by the tendency of every firm to act in a certain unobservable manner over time or randomly. Specifically, family businesses have many characteristics that make them different from other organizational structures [
69].
The generalized method of moments (GMM) not only controls for unobservable heterogeneity problems, it is also consistent with endogeneity through its use of instruments; in order to improve efficiency, the estimator developed by Arellano and Bond [
70] was used, which exploits all the moment restrictions specified by the model. Specifically, we lagged twice the explanatory variables as valid instruments. In this way, we solved the endogeneity problem caused by the fact that the explanatory variables of the models and firm financial performance could be determined simultaneously. Finally, the error term was split into three components (
εit =
dt +
ηi +
vit): the time effect,
dt, the individual effect,
ηi, and random disturbance,
vit.
Based on these considerations, we proposed the models below to test corresponding hypotheses:
In the first model we tested the influence of female executives on the performance of the firm, expecting a positive outcome, and taking into account the possible effect of all remaining controlling variables. The influence exerted by family firms was considered in the second model. We also measured the importance of family ties with the controlling shareholder in model 3 (see
Table 7). Since a stronger commitment of family female executives may in turn affect firm performance, we tested to what extent family ties of executive women were related to firm performance in model 3.
To check the potential mis-specification of the models, we used the Hansen J statistic of overidentifying restrictions to test the absence of correlation between the instruments and the error term. These statistics are shown together with the estimation results.