Does gender diversity in Dutch hospital boards influence the
performance of Dutch hospitals?
Master Thesis V.M. Duimelaar 0516546
Date final version: 23/08/2015
Supervisor: ms. dr. S. Dominguez Martinez
University of Amsterdam
Faculty: Economics and Business (FEB) Specialization: Organization Economics
Contents
1. Introduction 3
2. Related literature 6
2.1 Theoretical background of the influence of diversity on performance 6
2.2 Results of various empirical studies 8
2.3 Hypotheses 11
3. Data and methodology 12
4. Results 16
4.1 Descriptive statistics 17
4.2 OLS regression 18
4.3 Fixed effects regression 20
5. Discussion 23
6. Conclusion 27
References 29
1. Introduction
In most companies, a part of the work is working in teams. People work in teams to achieve company goals and to contribute to better performance. At the present time, a large number of organizational decisions are assigned to teams instead of
individuals (Hoogendoorn, Oosterbeek and Van Praag, 2013). Teams are controlled by the management. One of the main tasks of a board is to monitor and control the management (Jensen and Meckling, 1976) and to guide the firm in the strategic direction that is described in the vision, mission, values and goals of the firm.
According to Hoogendoorn et al. (2013, p. 1514), the diversity of a team is “one of the potential determinants of the effectiveness of a team”. Hoogendoorn et al. (2013) claim that the results of their investigation could provide useful information about the consequences of the degree of gender diversity in teams that perform in similar settings, such as corporate boards for instance. In their study for a general collective intelligence factor in the performance of a team, Woolley, Chabris,
Pentland, Hashmi and Malone (2010) found that a greater proportion of women in a team contributes to the team’s performance, due to a higher level of the social sensitivity measure of the concerning team members. Moreover, Hoogendoorn et al. (2013) state that, according to Desvaux, Devillard-Hoellinger and Baumgarten
(2007), one of the general reasons mentioned in the public media supporting higher gender diversity in corporate boards is an expansion of the pool from which talent is drawn. Other reasons mentioned are for example complementarities and the
improvement of mutual learning. According to Campbell and Mínguez-Vera (2008) it is argued that diversity has a positive effect on innovation and creativity, because these characteristics have a tendency to vary consistently with demographic variables like gender, instead of being randomly distributed in the population. Furthermore, gender diversity can improve problem solving because of a broader view that occurs from a board with higher gender diversity. Because of the diversity in perspectives, the board will better understand the complexity of the business surroundings and therefore enable a better decision-making process (Campbell and Mínguez-Vera, 2008).
The past decades, gender diversity has become more important in boards and executive positions. So far, some research has been done on the diversity of board of directors; most of that research took place in the United States. In the existing
literature, the majority of the research focused on the relationship between gender diversity in boards and firm performance. Furthermore, most of the research focused on publicly listed firms. However, to my knowledge the gender diversity of Dutch hospital boards has not been studied yet. Therefore, this study will investigate if gender diversity in Dutch hospital boards influences the performance of Dutch hospitals. Focusing on the Dutch hospitals brings a new insight on the already existing literature. Below this will be discussed more in detail.
Eeckloo, Delesie and Vleugels (2007) state in their article that corporate governance stands for the system by which private corporations are directed and controlled. They explain agency theory on the basis of the shareholders of the corporation and the management of the corporation. The shareholders of the
corporation, who are mainly interested in making profit, represent the principal in the principal-agent model. The management of the corporation, who should try to
maximize profit, is the agent in the principal-agent model. The problem is that the shareholders do not know if the management made the right decisions and control the operations of the company effectively (Eeckloo et al., 2007). Besides that,
management can have an advantage in their knowledge, because they are closer to the process and they could use it for their own benefit. According to Eeckloo et al. (2007), the solution to this problem lies in the ‘checks and balances’, wherein the board of directors of a corporation plays an important role. The board has to make sure that the interest of the company and its shareholders are pursued by the management. Aside from the control function, the board is also concerned with strategy development, counseling and networking (Eeckloo et al., 2007).
With regard to hospital governance, there are some important differences. According to Eeckloo et al. (2007) hospital governance consists of two parts: the set of structures and processes that define the strategic direction for the hospital and the means by which resources are assembled and allocated to achieve the strategic direction. One of the differences with corporate governance is that most hospitals have no shareholders, since they are public or non-profit private organizations. The main focus will be on the interest of the stakeholders instead of the shareholders. This makes the decision making process more complicated, as the maximization principle is not the main interest anymore. Furthermore, because hospitals are more complex than most other institutions, they have less transparent transactions that are also more difficult to value (Eeckloo et al., 2007). A third difference is that the
organization of a hospital consists of ‘four worlds’ that each have their own decision-making process and difference in management. The four worlds within a hospital are: cure (medical community), care (nurses), administrative hierarchy (management of the hospital) and trustees (formal board). Because of the different ways of organizing, the different activities and the different mindsets in these four worlds, in the end the hospital does not consist of one organization, but four (Eeckloo et al., 2007).
Recently there have been several studies that investigate the relationship between gender diversity and firm performance. While some studies show a positive relationship between gender diversity and firm performance, other studies did not find a significant relationship.
Erhardt, Werbel and Shrader (2003) found in their study that firms wherein women are part of the board of directors have a better firm financial performance than firms with exclusively men as member of the board of directors. Campbell and Mínguez-Vera (2008) came to a similar conclusion, where there is a positive
relationship between gender diversity of the board and firm value.
On the other hand, Carter et al. (2010) and Rose (2007) did not find a
significant relationship between gender diversity and firm performance. Adams and Ferreira (2009) found a negative relationship between gender diversity of board of directors and firm performance.
In this paper I would like to investigate if gender diversity in Dutch hospital boards influences the performance of Dutch hospitals. Studies on board of director diversity and firm performance will be applied on the hospital sector. Whereas boards of directors of firms only decide while having the shareholder’s interest in mind, the boards of directors of hospitals also have to take the interest of the stakeholders into account. I expect that higher gender diversity in Dutch hospital boards will be related to better hospital performance. The expectation is based on the legitimate arguments of amongst others Hoogendoorn et al. (2013), Woolley et al. (2010), Desvaux et al. (2007) and Campbell and Mínguez-Vera (2008). Also, the increased importance of the topic contributes to the motivation for this investigation, because board diversity has gained popularity in the academic literature, as well as in government policies and organizational strategy. Furthermore, I believe that the added value to the decision-making process is higher for boards that are more diverse, because of the higher level of social sensitivity in groups that consist of a larger percentage of women (Woolley et al., 2010).
The data used for this research will be collected from the annual reports of Dutch hospitals for the years 2011-2014. The sample consists of a balanced panel data set for all of the STZ-hospitals (cooperating top clinical teaching hospitals) and all of university medical centers in the Netherlands. The sample includes 144 annual observations for 36 Dutch teaching hospitals.
The structure of this paper is as follows: in section 2 I will discuss the theoretical background of the influence of diversity on performance and the results of various empirical studies. My hypotheses are based on the literature and will also be
discussed in section 2. Section 3 will explain the data and methodology used for this research. The results will be discussed in section 4. Finally, section 5 concludes with a discussion and conclusion.
2. Related literature
In this section, the theoretical background of the influence of diversity on
performance and the results of various empirical studies will be discussed, followed by the hypotheses that are based on the literature.
2.1 Theoretical background of the influence of diversity on performance
The theoretical background for board diversity and firm financial performance is principally based on four theories: resource dependence theory, human capital theory, agency theory and social psychological theory.
Pfeffer and Salancik (1978) formulated resource dependence theory by the link between diversity of the board and external dependencies. Higher diversity of the board gives the board the opportunity to address more external resources. In addition to this theory, Hillman, Cannella and Paetzold (2000) claim that different types of board members will provide different types of advantageous resources to the organization. Thus a higher diversity of the board will supply more profitable
resources and therefore better financial performance. Carter et al. (2010) state that the information that is provided by the board to the managers will be improved because of the unique information kept by the diverse board members. Due to the differences in gender and ethnicity, unique information sets are captured that are available to the board, which will improve decision-making (Carter et al., 2010).
According to Terjesen, Sealy and Singh (2009) the human capital theory originates from Becker (1964). This theory investigates the role of a man/woman’s stock of education, skills and experience in improving observational and fruitful competences that can be utilized to the advantage of this person and the organization. From historical perspective, the female sex invested less in educational development and labor experience than their male counterparts; therefore women would have a lower salary and less promotion opportunities (Terjesen et al., 2009). In addition, Terjesen et al. (2009) state that gender diversity causes board members to have unique human capital. Carter et al. (2010) summarize that because of the diverse and unique human capital, boards will be influenced by the diversity in boards. However, Carter et al. (2010) state, without further explanation, that the result could have a positive or negative effect on financial performance.
Another theory that takes board diversity into account is the agency theory. One of the main tasks of a board is to monitor and control management, according to Jensen and Meckling (1976) this function of the board is fundamental to the concept of agency theory. Carter, Simkins and Simpson (2003) argue that boards that are more diverse are better capable of monitoring management, since board diversity will lead to higher board independence. This in turn could lead to better financial
performance. However, Carter et al. (2003) claim that the prediction of the link between more diverse boards and firm performance is not evident from the agency theory. In accordance with the perspective of the agency theory, it is less plausible that board members are obligated to managers. In addition, Carter et al. (2010) state that it is possible that some factors, like the degree of ownership of the organization, could have a stronger effect on monitoring by the board than on board
independence. Carter et al. (2010) conclude that the support for a positive effect on financial performance caused by the diversity in the board is more provided by the resource dependence theory than by the agency theory, however the agency theory does not exclude the advantageous effects that diversity in the board could have.
The last theory to be discussed refers to social psychology. Westphal and Milton (2000) claim that group decision-making will be less efficient if demographic minorities are part of the board. Social cohesion between groups will be lower when there are differences in demography and because of the social barriers it is possible that the influence on group decisions by minority points of view is less likely
could have a very large influence on group decision-making (Westphal and Milton, 2000). According to Carter et al. (2010) it is therefore possible that diverse board members do not have an influence on the board because of the internal group dynamics of the board. Furthermore, according to Lau and Murnighan (1998) decision-making could be more time-consuming and less effective because greater diversity between board members could cause more different opinions and critical thoughts. Carter et al. (2010) summarize that according to the social psychological theory, diversity in the board could influence the financial performance of a firm both positively and negatively.
2.2 Results of various empirical studies
Recently there have been several studies that investigate the relationship between gender diversity and firm performance. While some studies show a positive
relationship between gender diversity and firm performance, other studies did not find a significant relationship.
Erhardt, Werbel and Shrader (2003) found in their study that firms wherein women are part of the board of directors have a better firm financial performance than firms with exclusively men as member of the board of directors. They used data from 127 large companies in the United States between the years 1993-1998. In the study of Erhardt et al. (2003) return on assets and return on investment were used as financial performance data. Diversity was measured as the percentage of women and minorities on the board of directors. The data were investigated by correlation and regression analyses. The results of the research confirmed the hypothesis that more diverse boards are positively related to organizational performance.
Campbell and Mínguez-Vera (2008) came to a similar conclusion, where there is a positive relationship between gender diversity and firm value. They examined the relationship between gender diversity of Spanish company boards and firm financial performance. The sample for the panel data analysis consisted of 68 non-financial companies and 408 observations during the period from January 1995 to December 2000. Gender diversity was measured by the percentage of women on the board, as well as a dummy variable with the value of one when there is at least one female board member, and zero if there is no woman part of the board. An approximation of Tobin’s Q was used as measure of firm value. Furthermore, a number of control variables were added, such as the size of the firm, the return on assets and the debt
level. To control for potential edogeneity of the variables, Campbell and Mí nguez-Vera (2008) used a two-stage least squares panel data regression. This method is used when the error terms of the dependent variable are correlated with the
independent variables. In the first stage, the included endogenous variables are regressed against the included exogenous variables. In the second stage the dependent variable is regressed against the included exogenous variables and the predicted endogenous variables from the first regression (Stock and Watson, 2012). The panel data methodology is used to clear out any unobservable heterogeneity that could exist between the firms used in the sample (Campbell and Mínguez-Vera, 2008). When unobservable heterogeneity is correlated with the independent
variables, which could bias the acquired coefficients, the fixed effects method is used. On the other hand, when the effects are not correlated with explanatory
variables, the random effects model is applied (Campbell and Mínguez-Vera, 2008). The results of Campbell and Mínguez-Vera (2008) showed that a woman as member of the board does not, in itself, influence firm value. At the same time, they found that board diversity has a positive effect on firm financial performance.
On the other hand, Carter et al. (2010) did not find a significant relationship between gender diversity and firm performance. They used firms of the Standard & Poor’s 500 index for the years 1998-2002. The sample consisted of an unbalanced panel of 641 firms and 2,563 firm years. Tobin’s Q and return on assets were used as financial performance data. The number of women on the board of directors was used as a diversity measure. To estimate the relationship between firm performance and the number of female board members, a fixed effects regression with lagged variables of all of the independent variables was applied. Furthermore, a three-stage least squares regression was performed because of the presence of edogeneity in the fixed effects method indicated by the Hausman tests (Carter et al., 2010). Carter et al. (2010) explain that they use a three-stage least squares method instead of the two-stage least squares method, because the three-stage least squares method takes the cross-equation correlation that could be present in their data into account. The results of Carter et al. (2010) demonstrate that there is no significant relationship between gender diversity and firm financial performance.
Rose (2007) also did not find a significant relationship between firm
performance and gender diversity of the board. The sample in the paper consisted of all Danish listed firms over the years 1998-2001, resulting in 443 firm-time
observations. Tobin’s Q was used as a measure of firm performance. The proportion of female board members was used as measure for gender diversity. A
cross-sectional analysis was performed of which the coefficients were estimated by
ordinary least squares. The results showed no significant link between board gender diversity and firm performance. According to Rose (2007), one of the reasons for this outcome may be that unconventional members of the board could possibly have taken over the behaviors and norms of the conservative board members. An explanation for adopting the norms and behaviors is that it may be the only option seen from the point of view of top decision makers to be competent for the position of a member of the board of directors. Therefore, it is possible that the benefits from having women on the board will not be included and displayed in any measure of performance that is chosen (Rose, 2007).
Adams and Ferreira (2009) found a negative connection between gender diversity of board of directors and firm performance. They used firms of the Standard & Poor’s (S&P) 500, S&P MidCaps and S&P SmallCap firms gathered by the
Investor Responsibility Research Center (IRRC) for the years 1996-2003. The sample consisted of an unbalanced panel of 1,939 firms and 8,253 firm years. A proxy for Tobin’s Q and return on assets were used as financial performance data. The fraction of women on the board of directors was used as a diversity measure. To control for endogeneity problems resulting from omitted variables that affect both board diversity and firm performance, Adams and Ferreira (2009) used an ordinary least squares model with firm fixed effects and year dummies. Furthermore, to control for reverse causality they used an instrumental variables method. Without controlling for edogeneity of gender diversity, Adams and Ferreira (2009) found a positive relationship between gender diversity and firm performance at first. However, this result was not robust. The results of the instrumental variables regression
showed that the relationship between gender diversity and firm performance appears to be negative. On the other hand, more diverse board could positively influence firm performance in organizations with weak governance. The value of the firm decreases when gender quotas in boardrooms are enforced in organizations with strong
governance, because of the possibility of overmonitoring in these firms (Adams and Ferreira, 2009). In addition, the results also demonstrated some new evidence that female board members behave differently than their male counterparts. Women have for example better attendance records than male board members; if there is more
gender diversity in a board, men will have fewer attendance problems; and the stock returns of organizations with women on their boards are less volatile.
Hoogendoorn, Oosterbeek and Van Praag (2013) did their research in a slightly different way. They performed a field experiment to measure the effect of the share of female members in business teams on the financial performance of the teams. The business teams existed of around 12 first-year students per team who start up and run an enterprise as part of their education program. Students were randomly allocated to a team, dependent on their gender. The annual reports of 43 teams were used to obtain information about the financial performance of the
business teams. A team’s financial performance was measured by sales and profits. Hoogendoorn et al. (2013) included a quadratic specification for measuring the share of women. The majority of the teams had a share of women between 0.2 and 0.6. One of the main findings is the inverse U-shaped relationship between sales and the share of women. When the share of women increases, sales also increase for a share of women between 0.2 and 0.5. When the share of women increases further (i.e. exceeding 0.5), sales have the tendency to decrease (Hoogendoorn et al., 2013). Likewise, when the share of women is below 0.5, profits are increasing in the share of women. When the share of women increases further, the relationship between profits and the share of women is flat (Hoogendoorn et al., 2013).
Hoogendoorn et al. (2013) used a standard ordinary least squares model to present the results. In addition, the results from the median and the robust regression were also presented, since the results from the OLS-regression could be sensitive to outliers. The results of the study of Hoogendoorn et al. (2013) showed that teams with an equal distribution of gender have a better financial performance than teams of which the majority consists of men.
2.3 Hypotheses
This study investigates if gender diversity in Dutch hospital boards influences the performance of Dutch hospitals. Studies on board of director diversity and firm performance will be applied on the hospital sector.
The first hypothesis focuses on the aspect of gender diversity in Dutch hospital boards and is based on the results of the studies of Erhardt, Werbel and Shrader (2003) and Campbell and Mínguez-Vera (2008). The second hypothesis takes the female board membership into account (Campbell and Mínguez-Vera, 2008). While
some studies show a positive relationship between gender diversity and firm
performance, other studies did not find a significant or even a negative relationship. I expect that higher gender diversity in Dutch hospital boards will be related to better hospital performance. The expectation is based on the legitimate arguments of amongst others Hoogendoorn et al. (2013), Woolley et al. (2010), Desvaux et al. (2007) and Campbell and Mínguez-Vera (2008). Also, the increased importance of the topic contributes to the motivation for this investigation, because board diversity has gained popularity in the academic literature, as well as in government policies and organizational strategy. Furthermore, I believe that the added value to the decision-making process is higher for boards that are more diverse, because of the higher level of social sensitivity in groups that consist of a larger percentage of women (Woolley et al., 2010).
Hypothesis 1:
Higher gender diversity in Dutch hospital boards will be related to better hospital performance
Hypothesis 2:
Hospitals with a female board member have a better performance
3. Data and methodology
The data used for this research are collected from the annual reports of Dutch
hospitals for the years 2011-2014. The sample consists of a balanced panel data set for all of the STZ-hospitals (cooperating top clinical teaching hospitals) and all of university medical centers in the Netherlands. The panel data methodology is used to clear out any unobservable heterogeneity that could exist between the firms used in the sample (Campbell and Mínguez-Vera, 2008). With regard to panel data, Stock and Watson (2012, p. 389) explain that “by studying changes in the dependent variable over time, it is possible to eliminate the effect of omitted variables that differ across entities but are constant over time”.
The sample includes 144 annual observations for 36 Dutch teaching hospitals. The sample of the 36 hospitals is divided into all of the 28 STZ-hospitals and all of the 8 university medical centers. The names of the hospitals are provided in the appendix.
The Netherlands has approximately 90 hospitals, which can be divided into three levels: academic hospitals (university medical centers), top clinical hospitals and general or regional hospitals. All of the hospitals provide basic care, the routine treatments. If a treatment becomes more complicated, a patient will be referred to an academic or top clinical hospital for the treatment (depending on the disease). In an academic hospital (also called an university medical center) specialists practice treatments that are so rare or complicated that they can almost exclusively be performed in academic hospitals. The academic hospitals are affiliated to an
university that offers the programme in Medicine. Top clinical hospitals are hospitals that perform complex care in addition to basic care. These hospitals are specialized in one or more care areas and carry out many innovative research to stay ahead in their specialty. Top clinical hospitals are usually larger than general hospitals and receive patients from a wider area than the smaller general hospitals. General hospitals treat patients with diseases, which are more common. These hospitals are usually smaller and are mainly visited by patients from the region.
To measure the relationship between hospital board’s gender diversity and the hospital’s performance, I will use the following formula:
Hospital performance = + 1GenderDiversity + nControl + ε
Different measures can be considered when investigating the financial position of a hospital. In this study, hospital performance will be measured by the net income of the hospital. The annual reports of hospitals show the net income on their income statement, which represents how much money remains after all expenses, interests and taxes have been deducted. The profitability of a company over a certain period of time is represented by the net income of that company; therefore it is an important financial measure. Different studies regarding hospital financial performance use net income as a financial measure (Molinari, Alexander, Morlock and Lyles, 1995; Burns, Gimm and Nicholson, 2004; Kaissi, Begun and Welson, 2008).
I will use the percentage of women on the board as a measure for gender diversity (Erhardt et al., 2003; Campbell and Mínguez-Vera, 2008; Adams and Ferreira, 2009). A gender diversity dummy will be used as a second measure for gender diversity (Campbell and Mínguez-Vera, 2008). The dummy will be 0 if there are no female board members and 1 if one or more women are part of the board of directors. From the outset it is not evident how women on the board of directors influence the
performance of a hospital. It could generate a positive sign of the coefficient if a higher level of gender diversity is linked with company strategies that are more creative (Campbell and Mínguez-Vera, 2008). At the same time, a negative sign of the coefficient could be possible due to the fact that women being part of the board of directors could lead to more conflict on the board (Campbell and Mínguez-Vera, 2008). When it turns out that gender diversity has no influence on the performance of a firm, then the relationship is expected to be insignificant (Campbell and Mí nguez-Vera, 2008).
For the first measure of gender diversity (percentage of women on the board of directors), the quadratic specification for the percentage of women will be included in one of the models, because of a possible inverse U-shaped relationship between the financial performance measure and the share of women (Hoogendoorn et al., 2013).
When using the percentage of women as member of the board of directors instead of the number of women as a gender diversity measure, it should be considered that diversity increases for 0-50%, but decreases again for 50-100%. However, the latter is not the case in this study, since the highest percentage of women as member of the board of directors is 50%.
Because it is more likely that larger hospitals also generate higher net incomes, the number of beds available will be added as a control variable to control for the hospital size (Menachemi, Burkhardt, Shewchuk, Burke, and Brooks, 2006; Campbell and Mínguez-Vera, 2008). Another control variable that will be used is the number of board members as board size (Erhardt et al., 2003; Carter et al., 2010). According to Yermack (1996), there is a negative relationship between board size and firm
performance. The results of his study show that smaller boards of directors are more efficient.
An ordinary least squares (OLS) regression will be used in order to test if higher gender diversity in Dutch hospital boards is related to better hospital performance. Initially, I was planning to use a fixed effects model on hospital and year to control for omitted variables and to address changes that are unobservable over time (Carter et al., 2010). These unobservable characteristics could influence the performance of a hospital and could bias the OLS estimates (Stock and Watson, 2012). In the study of Campbell and Mínguez-Vera (2008) the fixed effects method is used when
unobservable heterogeneity is correlated with the independent variables, which could bias the acquired coefficients. On the other hand, when the effects are not correlated with explanatory variables, Campbell and Mínguez-Vera (2008) used the random effects model.
When performing the Hausman test to identify which specification of the model should be used, it turns out that the random effects model should be used. The Hausman test tests whether unobservable heterogeneity is correlated with the independent variables (Campbell and Mínguez-Vera, 2008). Therefore, I will use the random effects model instead of the fixed effects model to control for omitted
variables and unobservable characteristics over time (Campbell and Mínguez-Vera, 2008).
Different models are tested to see what the effect is when the control variables are added to the model and how this influences the coefficients. The following models are tested for both the OLS regression model and the random effect regression model:
(1) Income =
+ 1Percentage of Women + ε
(2) Income =
+ 1Percentage of Women + 2(Percentage of Women)
2 + ε (3) Income = + 1Percentage of Women + 2Beds + 3Board Size +4UMC + ε
(4) Income = + 1Female Dummy + ε
(5) Income =
+ 1Female Dummy + 2Beds + 3Board Size +4UMC + ε
The OLS regression models as well as the random effects regression models use robust standard errors. Before performing all of the regressions, a test for
variable (income) and the control variables (beds and board size). The variance inflation factor showed that there was no sign of multicollinearity.
4. Results
Figure 1 shows the relationship between the percentage of women and the financial performance measure. It turns out that there is no inverse U-shaped relationship between the financial performance measure and the share of women, which was the case in the study of Hoogendoorn et al. (2013). The relationship between income and the percentage of women seems to be decreasing when the percentage of women increases. Since there is no inverse U-shaped relationship between the percentage of women in the board of directors and the financial performance measure
(measured by income), the quadratic specification for the percentage of women is not included in model (3), in which the control variables were added.
4.1 Descriptive statistics
Table 1 shows the descriptive statistics for the sample used in this research. The financial performance measures point out that the hospitals in the sample were on average economically successful over the time period investigated.
Table 1: Descriptive statistics. The table provides summary statistics for the entire sample. The sample consists of a balanced panel data set for all of the STZ-hospitals (cooperating top clinical teaching hospitals) and all of the university medical centers in the Netherlands. The sample includes 144 annual observations of 36 hospitals for the period 2011-2014. The data used for this research are collected from the annual reports of Dutch hospitals.
Variable Mean Standard dev. Min Max
Firm characteristic Hospital 18.5 10.43 1 36 Year 2012.5 1.12 2011 2014 Income (x1000) 424,850 28,700 151,000 1,310,000 Income/beds 550,305 239,527 205,568 1,116,394 Profit (x1000) 8,723 7,670 -14,147 36,900 Beds 737.51 221.71 347 1320 UMC 0.22 0.42 0 1 Board characteristic Board Size 3.25 1.09 2 5 Percentage Women 9.01 15.5 0 50 Percentage Women2 319.6 639.65 0 2500 Female Dummy 0.28 0.45 0 1 Female CEO 0.02 0.14 0 1 Firm characteristics
Net income has a mean of €424,850,094 with a minimum of €151,000,000 and a maximum of €1,310,000,000. When net income is adjusted for hospital size (income divided by the number of beds available), it has a mean of €550,305. Profit has a mean of €8,723,321 with a minimum of €-14,147,000 and a maximum of
investigated in this research is 738. The minimum number of beds available for the hospitals investigated in this research is 347, whereas the maximum number of beds available is 1320. 22% of the hospitals in this research are Dutch university medical centers (academic hospitals).
Board characteristics
On average 28% of the Dutch hospitals investigated in this research have one or more female board members. Of these boards, on average 9% consists of the female sex. Only 2% of the female board members are CEO. The mean of the board size variable is 3.25 for the Dutch hospitals investigated in this research. The smallest board consists of 2 members and 5 board members are the maximum in this research.
4.2 OLS regression
Table 2 provides the results from the OLS regression for the models (1) to (5). The coefficient for the first gender diversity measure, the percentage of women as member of the board of directors, is negative in both models (1) and (3). Whereas the coefficient for the gender diversity is significant in model (1), adding the control variables to the regression in model (3) causes the coefficient of the gender diversity measure to be insignificant. However, the sign of the coefficient does not change.
In model (2) the squared term of the percentage of women as board member is included in the regression of model (1). Both the percentage of women and the squared term of the percentage of women turn out to be significant. However, whereas the percentage of women has a positive sign, the coefficient of squared term of the percentage of women is negative.
As a second measure for gender diversity a female dummy was used in model (4). The sign of the coefficient of the dummy is negative, however the coefficient is not significant. When adding the control variables to the regression in model (5), the coefficient of the dummy becomes positive, yet the coefficient is still not significant.
Accordingly, when the control variables are added to the different models, the coefficients of the gender diversity measures appear to be insignificant. The results present no significant evidence of the relationship between female board members and hospital performance. The presence of a woman on the board seems to have no effect on the financial performance of the Dutch hospitals investigated in this
research. However, the presence of female board members could possibly have an impact in other ways that the financial performance of an organization. In addition, Rose (2007) explains a possible process of socialization where women that are part of the board have taken over the behavior and norms of their male counterparts.
The control variable for the number of beds available is significant both in model (3) and model (5) for a p-value less than 0.01. This means that the financial performance of a hospital will increase with €364,117 in model (3) and €366,530 in model (5) when increasing the number of beds available by one.
The control variable for board size is also significant both in model (3) and model (5) for a p-value less than 0.01. This means that the financial performance of a hospital will increase with €21,100,000 in model (3) and €20,600,000 in model (5) when increasing the board by one member. This result is in contrast to the findings of Yermack (1996). Yermack (1996) provides evidence that firms with smaller board of directors have a better financial performance.
The coefficient for UMC (university medical centers) is significant at the 1% level. This result indicates that UMC’s perform significantly better than non-UMC’s.
Table 2: OLS regression. This table reports the OLS regression results for models (1) to (5) that estimates the relationship between hospital performance (measured by income) and gender diversity. The sample consists of a balanced panel data set for all of the
STZ-hospitals (cooperating top clinical teaching STZ-hospitals) and all of the university medical centers in the Netherlands. The sample includes 144 annual observations of 36 hospitals for the period 2011-2014. The data used for this research are collected from the annual reports of Dutch hospitals. Standard errors are adjusted for potential heteroskedasticity. T-statistics of robust standard errors are in parentheses. Asterisks indicate significance at 0.01 (***), 0.05 (**) and 0.10 (*) levels.
Dependent variable: Income
Independent variables: (1) (2) (3) (4) (5) Percentage Women -2,845,350** 8,901,240* -49,260 (-2.35) (1.74) (-0.13) Percentage Women2 -296,765*** (-2.78) Female Dummy (x1000) -53,300 2,872 (-1.02) (0.19) Beds 364,117*** 366,530*** (6.02) (6.19) Board Size (x1000) 21,100*** 20,600*** (3.22) (3.06) UMC (x1000) 508,000*** 508,000*** (14.86) (14.93) Constant (x1000) 450,000*** 440,000*** -24,600 440,000*** -26,100 (16.05) (15.35) (-0.50) (15.38) (-0.54) R-squared 0.02 0.06 0.91 0.01 0.91 Adj. R-squared - - - - - F-statistic 5.51** 28.08*** 174.59*** 1.05 181.75*** P-value (F-stat) 0.02 0.00 0.00 0.31 0.00
4.3 Random effects regression
Table 3 shows the results from the random effects regression for models (1) to (5). The coefficient for the first gender diversity measure, the percentage of women as member of the board of directors, is positive in both models (1) and (3). This is
different from the results of the OLS regression models (1) and (3), in which the coefficient of the gender diversity measure is negative. Whereas the coefficient for the gender diversity is significant in the OLS regression model (1), the coefficient in the random effects regression model (1) turns out to be insignificant, just as the coefficient of the random effects model (3).
In model (2) the squared term of the percentage of women as board member is included in the regression of model (1). Both the percentage of women and the squared term of the percentage of women turn out to be significant. However, whereas the percentage of women has a positive sign, the coefficient of squared term of the percentage of women is negative. These outcomes are similar to the outcomes of the OLS regression model (2).
As a second measure for gender diversity a female dummy was used in model (4). The sign of the coefficient of the dummy is positive, moreover the coefficient turns out to be significant. This is different from the result of the OLS regression model (4), in which the coefficient for the female dummy is negative and insignificant. When adding the control variables to the regression in model (5), the coefficient of the dummy is still positive, yet not significant.
Again, the results of the random effects regression present no significant evidence of the relationship between female board members and hospital
performance when the control variables are added to the different regression models. The presence of a woman on the board seems to have no effect on the financial performance of the Dutch hospitals investigated in this research, similar to the results of the OLS regression. However, the presence of female board members could possibly have an impact in other ways that the financial performance of an
organization. In addition, Rose (2007) explains a possible process of socialization where women that are part of the board have taken over the behavior and norms of their male counterparts.
For the random effects regression models, both the control variable for the number of beds available and the control variable for board size are not significant. These results do not correspond with the results of the OLS regression models, in which both the control variable for the number of beds available and the control variable for board size are significant at the 1% level.
The coefficient for UMC (university medical centers) is significant at the 1% level, similar to the results of the OLS regression models. This result indicates that UMC’s perform significantly better than non-UMC’s.
Table 3: Random effects regression. This table reports the random effects regression results for models (1) to (5) that estimates the relationship between hospital performance (measured by income) and gender diversity. The sample consists of a balanced panel data set for all of the STZ-hospitals (cooperating top clinical teaching hospitals) and all of the university medical centers in the Netherlands. The sample includes 144 annual observations of 36 hospitals for the period 2011-2014. The data used for this research are collected from the annual reports of Dutch hospitals. Standard errors are adjusted for potential
heteroskedasticity. Z-statistics of robust standard errors are in parentheses. Asterisks indicate significance at 0.01 (***), 0.05 (**) and 0.10 (*) levels.
Dependent variable: Income
Independent variables: (1) (2) (3) (4) (5) Percentage Women 516,245 1,863,296** 267,171 (1.64) (2.05) (0.86) Percentage Women2 -35,504* (-1.94) Female Dummy (x1000) 18,800* 12,600 (1.91) (1.26) Beds 27,610 27,228 (0.75) (0.73) Board Size (x1000) 27,000 26,100 (1.61) (1.55) UMC (x1000) 597,000*** 598,000*** (8.74) (8.74) Constant (x1000) 420,000*** 420,000*** 182,000*** 420,000*** 184,000*** (8.87) (8.84) (2.85) (8.88) (2.88) R-squared 0.02 0.00 0.85 0.01 0.85 Adj. R-squared - - - - - Wald Chi2 2.69 4.29 109.34*** 3.64* 113.19***
Overall, the results show that there is no significant relationship between gender diversity in Dutch hospital boards and the performance of Dutch hospitals
investigated in this research. This is similar to the results of Carter et al. (2010) and Rose (2007), who also did not find a significant relationship between gender diversity and firm performance. According to Rose (2007), one of the reasons for this outcome may be that unconventional members of the board could possibly have taken over the behaviors and norms of the conservative board members. An explanation for adopting the norms and behaviors is that it may be the only option seen from the point of view of top decision makers to be competent for the position of a member of the board of directors. Therefore, it is possible that the benefits from having women on the board will not be included and displayed in any measure of performance that is chosen (Rose, 2007).
However, this research found that when increasing board size by one member, the financial performance of a hospital also increases. This result is in contrast to the findings of Yermack (1996). Furthermore, it turns out that UMC’s perform significantly better than non-UMC’s.
5. Discussion
To check if the obtained results are robust, the OLS regressions are performed for models (1) to (5) with profit, instead of income, as financial performance measure. Table 4 shows the results from the OLS regression for models (1) to (5) with profit as dependent variable. The coefficient for the first gender diversity measure, the
percentage of women as member of the board of directors, is negative in models (1). This is similar to the result of the OLS regression for model (1) with income as
dependent variable, however the coefficient in table 4 is not significant, as opposed to the coefficient for the gender diversity measure in table 2. When adding the control variables to the regression in model (3), the coefficient for the percentage of women as member of the board becomes positive, yet still insignificant. The coefficient of the gender diversity measure in model (3) with income as financial performance measure is also insignificant, however the sign of this coefficient is negative.
In model (2) the squared term of the percentage of women as board member is included in the regression of model (1). Both the percentage of women and the squared term of the percentage of women turn out to be significant. This is similar to
the results of the OLS regression for model (2) with income as dependent variable. However, the coefficients of both terms of the gender diversity measure are negative, whereas the coefficients of model (2) in table 2 are positive for the percentage of women and negative for the squared term of the percentage of women.
Table 4: OLS regression. This table reports the OLS regression results for models (1) to (5) that estimates the relationship between hospital performance (measured by profit) and gender diversity. The sample consists of a balanced panel data set for all of the
STZ-hospitals (cooperating top clinical teaching STZ-hospitals) and all of the university medical centers in the Netherlands. The sample includes 144 annual observations of 36 hospitals for the period 2011-2014. The data used for this research are collected from the annual reports of Dutch hospitals. Standard errors are adjusted for potential heteroskedasticity. T-statistics of robust standard errors are in parentheses. Asterisks indicate significance at 0.01 (***), 0.05 (**) and 0.10 (*) levels.
Dependent variable: Profit
Independent variables: (1) (2) (3) (4) (5) Percentage Women -14,444 -348,534*** 16,741 (-0.38) (2.74) (0.47) Percentage Women2 -9,170*** (-3.50) Female Dummy (x1000) 597 1,207 (0.43) (0.91) Beds 900 1,138 (0.30) (0.39) Board Size (x1000) 375 266 (0.73) (0.48) UMC (x1000) 9,783*** 9,843*** (4.48) (4.54) Constant (x1000) 8,853*** 8,516*** 4,516* 8,558*** 4,498* (11.80) (11.10) (1.84) (11.18) (1.89) R-squared 0.00 0.05 0.32 0.00 0.32 Adj. R-squared - - - - - F-statistic 0.14 11.99*** 8.85*** 0.18 9.44*** P-value (F-stat) 0.71 0.00 0.00 0.67 0.00
As a second measure for gender diversity a female dummy was used in model (4). The sign of the coefficient of the dummy is positive, however the coefficient is not significant. The sign of the coefficient is different from the result of the OLS
regression for model (4) with income as dependent variable, in which the coefficient for the female dummy is negative, yet also insignificant. When adding the control variables to the regression in model (5), the coefficient of the dummy is still positive, but also still insignificant. This result is similar to the result of the OLS regression for model (5) with income as dependent variable.
When the control variables are added to the different models, the coefficients of the gender diversity measures appear to be insignificant. The results present no significant evidence of the relationship between female board members and hospital performance. The presence of a woman on the board seems to have no effect on the financial performance of the Dutch hospitals investigated in this research.
For the OLS regression models with profit as dependent variable, both the control variable for the number of beds available and the control variable for board size are not significant. These results do not correspond with the results of the OLS regression models with income as dependent variable, in which both the control variable for the number of beds available and the control variable for board size are significant at the 1% level.
The coefficient for UMC (university medical centers) is significant at the 1% level, similar to the results of the OLS regression models in table 2. This result indicates that UMC’s perform significantly better than non-UMC’s.
When investigating the relationship between gender diversity and hospital financial performance, the presence of endogeneity should be considered. When it comes to gender diversity, it is possible that hospitals decide whether or not to let women be part of the board of directors. At the same time the female gender could decide on which hospitals they want to join. Hence, it is difficult to determine a causal
relationship when investigating hospital financial performance and gender diversity of the board of directors. It is possible that hospitals with female members on their boards achieve better financial performance because of the female representation on the board. On the other hand it is also possible that hospitals with better financial performance decide to engage female board members.
A way to deal with this reverse causality concern is to use the instrumental variables method. In order to be valid, an instrument has to be relevant (i.e. the instrument must be correlated with the included endogenous variable) and the instrument has to be exogenous (i.e. uncorrelated with the error term) (Stock and Watson, 2012). Adams and Ferreira (2009) performed an instrumental variables regression, in which they considered the share of male members of the board who have a connection with female board members in other boards as a valid instrument for the share of female board members as gender diversity variable.
Endogeneity could also emerge because of omitted variables, such as unobservable firm characteristics. These unobservable characteristics could
influence the performance of a hospital and could bias the OLS estimates (Stock and Watson, 2012). Campbell and Mínguez-Vera (2008), Adams and Ferreira (2009) and Carter et al. (2010) used among others the fixed effects model to control for omitted variables and to address changes that are unobservable over time. Whereas
Campbell and Mínguez-Vera (2008) used a two-stage least squares panel data regression, Carter et al. (2010) performed a three-stage least squares regression to address the issue of endogeneity.
Variables that could possibly be missing in this study are for instance the age of the hospital. Hospital age stands for the amount of years that a hospital is active and provides care to the patients. It is possible that the age of a hospital influences the performance of the hospital, due to the fact that hospitals that are active for more years also are more experienced. This could result in a better organizational
structure of the hospital and therefore a better performance.
Another variable that could be missing is the expertise on the board. Members of the board of directors that are skilled in the financial or medical field could have an influence on the decision-making process regarding the hospital’s financial
performance and the quality of care that is provided to the patient.
Furthermore, the independency of the board could also be considered as a missing variable. Outside directors (members of the board of directors that are not an employee or stakeholder in the hospital) are more likely to form unbiased judgments and presumably have very little or no conflict of interest with the organization. Their presence on the board can help prevent circumstances in which inside directors are allured to increase their influence on the organization, which in turn could be
Lastly, it is possible that the results of this study depend on the limited variation of female representation in the board of directors of the Dutch hospitals investigated in this research. There are only a few female board members representing the Dutch hospitals that are part of the sample in this study. This could possibly depend on the type of hospital, since I only investigated the Dutch teaching hospitals, but it could also be the case for all of the hospitals in the Netherlands.
6. Conclusion
This paper investigated if gender diversity in Dutch hospital boards influences the performance of Dutch hospitals. One of the main tasks of a board is to monitor and control the management (Jensen and Meckling, 1976) and to guide the firm in the strategic direction that is described in the vision, mission, values and goals of the firm. Whereas boards of directors of firms only decide while having the shareholder’s interest in mind, the boards of directors of hospitals also have to take the interest of the stakeholders into account.
Overall, the results of this research indicate that there is no significant
relationship between gender diversity in Dutch hospital boards and the performance of Dutch hospitals investigated in this paper. This is similar to the results of Carter et al. (2010) and Rose (2007), who also did not find a significant relationship between gender diversity and firm performance. According to Rose (2007), one of the reasons for this outcome may be that unconventional members of the board could possibly have taken over the behaviors and norms of the conservative board members. An explanation for adopting the norms and behaviors is that it may be the only option seen from the point of view of top decision makers to be competent for the position of a member of the board of directors. Therefore, it is possible that the benefits from having women on the board will not be included and displayed in any measure of performance that is chosen (Rose, 2007).
However, this research found that when increasing board size by one member, the financial performance of a hospital also increases. This result is in contrast to the findings of Yermack (1996). Furthermore, it turns out that UMC’s perform significantly better than non-UMC’s.
The dataset for this research only contains Dutch teaching hospitals, so the results are unilateral to the Netherlands and only apply to the teaching hospital sector. The outcomes could be different for other countries and other types of hospitals.
Furthermore, in order to obtain more insight in the different diversity measures and the effect on hospital financial performance, more extensive research is required. It is also recommended to use a larger dataset to have a better validation of the results.
References
Adams, R. B., & Ferreira, D. (2009). Women in the boardroom and their impact on governance and performance. Journal of financial economics, 94(2), 291-309. Becker, G. (1964). Human capital. Chicago, IL: University of Chicago Press. Burns, L. R., Gimm, G., & Nicholson, S. (2004). The financial performance of integrated health organizations. Journal of healthcare management/American
College of Healthcare Executives, 50(3), 191-211.
Campbell, K., & Mínguez-Vera, A. (2008). Gender diversity in the boardroom and firm financial performance. Journal of business ethics, 83(3), 435-451.
Carter, D. A., Simkins, B. J., & Simpson, W. G. (2003). Corporate governance, board diversity, and firm value. Financial Review, 38(1), 33-53.
Carter, D. A., D'Souza, F., Simkins, B. J., & Simpson, W. G. (2010). The gender and ethnic diversity of US boards and board committees and firm financial performance.
Corporate Governance: An International Review, 18(5), 396-414.
Desvaux, G., Devillard-Hoellinger, S., & Baumgarten, P. (2007). Women matter:
Gender diversity, a corporate performance driver. McKinsey.
Eeckloo, K., Delesie, L., & Vleugels, A. (2007). Where is the pilot? The changing shapes of governance in the European hospital sector. The journal of the Royal
Society for the promotion of health, 127(2), 78-86.
Eelzak, A. (2014). De financiële gezondheid van de ziekenhuizen. Amsterdam: Sint
Lucas Andreas Ziekenhuis.
Erhardt, N. L., Werbel, J. D., & Shrader, C. B. (2003). Board of director diversity and firm financial performance. Corporate governance: An international review, 11(2), 102-111.
Hillman, A. J., Cannella, A. A., & Paetzold, R. L. (2000). The resource dependence role of corporate directors: Strategic adaptation of board composition in response to environmental change. Journal of Management studies, 37(2), 235-256.
Hoogendoorn, S., Oosterbeek, H., & Van Praag, M. (2013). The impact of gender diversity on the performance of business teams: Evidence from a field experiment.
Management Science, 59(7), 1514-1528.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of financial economics, 3(4), 305-360. Kaissi, A. A., Begun, J. W., & Welson, T. (2008). Strategic planning processes and hospital financial performance. Journal of Healthcare Management, 53(3), 197.
Lau, D. C., & Murnighan, J. K. (1998). Demographic diversity and faultlines: The compositional dynamics of organizational groups. Academy of Management Review,
23(2), 325-340.
Menachemi, N., Burkhardt, J., Shewchuk, R., Burke, D., & Brooks, R. G. (2006). Hospital information technology and positive financial performance: a different
approach to finding an ROI. Journal of Healthcare Management/American College of
Healthcare Executives, 51(1), 40-58.
Molinari, C., Alexander, J., Morlock, L., & Lyles, C. A. (1995). Does the hospital board need a doctor? The influence of physician board participation on hospital financial performance. Medical care, 170-185.
Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations: A resource dependence approach. NY: Harper and Row Publishers.
Rose, C. (2007). Does female board representation influence firm performance? The Danish evidence. Corporate Governance: An International Review, 15(2), 404-413. Stock, J. H., & Watson, M. W. (2012). Introduction to Econometrics, Third Edition.
Pearson Education.
Terjesen, S., Sealy, R., & Singh, V. (2009). Women directors on corporate boards: A review and research agenda. Corporate governance: an international review, 17(3), 320-337.
Westphal, J. D., & Milton, L. P. (2000). How experience and network ties affect the influence of demographic minorities on corporate boards. Administrative Science
Quarterly, 45(2), 366-398.
Woolley, A. W., Chabris, C. F., Pentland, A., Hashmi, N., & Malone, T. W. (2010). Evidence for a collective intelligence factor in the performance of human groups.
Science, 330(6004), 686-688.
Yermack, D. (1996). Higher market valuation of companies with a small board of directors. Journal of financial economics, 40(2), 185-211.
www.jaarverslagenzorg.nl www.nvz-ziekenhuizen.nl www.stz.nl
Appendix
STZ-hospitals (cooperating top clinical teaching hospitals):
Name of the hospital Place
1. Amphia Ziekenhuis Breda
2. Atrium Medisch Centrum Heerlen
3. Canisius-Wilhelmina Ziekenhuis
Nijmegen
4. Catharina Ziekenhuis Eindhoven
5. Deventer Ziekenhuis Deventer
6. Gelre Ziekenhuizen Apeldoorn
7. HagaZiekenhuis Den Haag
8. Isala Klinieken Zwolle
9. Jeroen Bosch Ziekenhuis Den Bosch
10. Kennemer Gasthuis Haarlem
11. Maasstad Ziekenhuis Rotterdam
12. Martini Ziekenhuis Groningen
13. Maxima Medisch Centrum Eindhoven 14. Meander Medisch Centrum Amersfoort
15. Medisch Centrum Alkmaar Alkmaar 16. Medisch Centrum Haaglanden Den Haag 17. Medisch Centrum Leeuwarden Leeuwarden 18. Medisch Spectrum Twente Enschede
19. Onze Lieve Vrouwe Gasthuis
Amsterdam
20. Reinier de Graaf Groep Delft
21. Rijnstate Arnhem
22. Sint Franciscus Gasthuis Rotterdam 23. Sint Lucas Andreas
Ziekenhuis
Amsterdam
24. Spaarne Ziekenhuis Haarlem
25. St. Antonius Ziekenhuis Nieuwegein 26. St. Elisabeth Ziekenhuis Tilburg
27. VieCuri Venlo
28. Albert Schweitzer Ziekenhuis
University Medical Centers:
Name of the hospital Place 1. Academisch Medisch
Centrum
Amsterdam
2. Vrije Universiteit Medisch Centrum Amsterdam 3. Universitair Medisch Centrum Groningen Groningen 4. Leids Universitair Medisch Centrum Leiden 5. Academisch Ziekenhuis Maastricht Maastricht 6. Radboud Universitair Medisch Centrum Nijmegen 7. Erasmus Medisch Centrum Rotterdam 8. Universitair Medisch Centrum Utrecht Utrecht
