Gender diversity and firm value: an extension of mean-variance portfolio theory.

ABSTRACT

This study sought to innovatively extend the financial concept of mean-variance portfolio theory into the areas of organizational management and firm financial performance. More specifically, the primary empirical analysis focused on female advancement within organizations and the associated financial performance of the firms. Based on these initial results, a subsequent analysis examined the relationship between board size and firm value. Organizational data were collected from two distinct time periods (1996: "buff" market, economic expansion and 2001: "bear" market, economic recession). The results from both the 1996 and 2001 samples indicated: 1) a positive relationship between the percentage of female board members (directors) and Market-Value-Added (MVA), 2) a negative relationship between the number of female board members (directors) and MVA, and 3) a negative correlation between board size and MVA The results, with regard to board composition and size, have implications for not only firm financial performance but also corporate governance structure.

Keywords: Gender Diversity, Firm Value, Portfolio Theory, Board Size, Market-Value-Added.

1. INTRODUCTION

Efforts made to eliminate the "glass ceiling" (i.e. the observed phenomenon of females not advancing to the highest levels of organizational management in proportional numbers) have been grounded in perspectives ranging from legal/moral/ethical (e.g. "it's discriminatory and wrong") to social/census data and trends (e.g. "women will continue to comprise an increasing proportion of the qualified workforce") to strategic/management (e.g. "it's just good business"). Yet, as current news reports and studies indicate, the "glass ceiling" continues to persist and the progress for women has been slow. Discussion and evaluation of this issue are complicated by the inconclusive empirical work to date.

With this as a backdrop, this study has two purposes: 1) to propose an innovative framework (based on an extension of a financial-diversification theory) through which to analyze the "glass ceiling" phenomenon and 2) to provide additional insight into the gender diversity--firm financial performance relationship through an empirical analysis.

First, a review of both the theoretical and empirical literature in this area will be presented. More specifically, the literature review will discuss the theoretical background and framework for the study and present findings from previous gender diversity--firm performance analyses. This study's empirical analysis (study design, data and analysis, and results) will then be presented. This will be followed by the summary and discussion. Finally, study limitations and future research opportunities will be discussed.

2. LITERATURE REVIEW

2.1 Theoretical

This empirical study is based on a theoretical framework established by Markowitz (1952). Markowitz developed the concept of "mean-variance" portfolio theory. It states that an "opportunity set" of individual investments (assets) with particular risk/return characteristics can be combined in various proportions to produce a "feasible set" of portfolios. Of this feasible set, some portfolios are better or more "efficient" than the others (i.e. maximum return for a given amount of risk or minimum risk for a given level of return). An investor can then choose the portfolio that has the risk/return combination that best meets his or her needs.

This study seeks to apply Markowitz's investment framework based on individual assets, feasible portfolios, and maximum return to the efficient use of human resources. More specifically, this analysis will seek to identify and describe the observed relationship(s) between the degree of female advancement within a sample of firms (i.e. a proxy for the "glass ceiling") and firm financial performance. In the past, the original financial framework has been extended from portfolios of securities to include portfolios of tangible assets and projects in capital budgeting analysis. It is an extension of this type that serves as a basis for this study's theoretical foundation. The paradigm is as follows:

Markowitz created portfolios by combining individual, financial assets (investments). This study proposes that a firm contains a portfolio of individual, human assets (i.e. employees). With Markowitz's theory, if the opportunity set of individual assets is reduced (constrained), the resulting feasible region is reduced thereby reducing the potential returns. If a firm artificially constrains the opportunity set of human assets (employees) either intentionally or by accident, the potential feasible region of returns for the firm is reduced. In other words, if a firm is "missing" productive, valuable human assets (female employees) that could improve its expected return, the result should be (relatively) diminished firm financial performance and value.

Expanding upon this theoretical premise from a strategic perspective, Rosener (1995) states that women have a unique management style and are underutilized by organizations (as indicated by economic data). While men tend to use a command and control leadership approach (characterized by hierarchical lines of authority, the amassing of power and information, a win-lose decision process, and the value of sameness), women employ an interactive leadership style. This interactive approach focuses on consensus building, comfort with ambiguity, and the sharing of power and information. These "interactive" attributes are believed to be most effective in today's fast-changing, service-oriented, and entrepreneurial organizations. Ultimately, Rosener argues, organizations need to realize that utilization of the distinctive leadership abilities of women will have a positive impact on their bottom lines (financial performance). This thesis seems to be gaining some acceptance in the corporate world. For instance, Leonard Schaeffer, chairman and chief executive of Wellpoint Health Networks, and his board of directors compiled a list of the skills and expertise desired in their directors. He indicated that since women make up half the talent pool, it would not be prudent to exclude them. In the end, Mr. Schaeffer believes that a diverse board makes good business sense (Hymowitz, 2003).

2.2 Empirical

While significant research has been conducted in the general areas of diversity, corporate boards, and firm performance; this portion of the literature review will focus on the relatively limited number of gender-diversity studies as related to officers (management), corporate board composition, and firm financial performance. Given this set of variables, the following empirical review will be presented in two distinct sections: 1) gender diversity at the corporate officer (management) level and firm financial performance and 2) gender diversity at the board of director level and firm financial performance.

2.2.1 Corporate Officer/Management

Williams and O'Reilly (1998) produced a comprehensive synthesis of the empirical research done in the general area of "diversity" over the 40-year period prior to 1998. Of the 62 "field" (as opposed to "lab") studies cited, only two even tangentially analyzed gender diversity as it related to some sort of firm (financial) performance measure.

First, Pelled, Eisenhardt, and Xin (1997) examined the relationship between gender diversity and performance. In particular, they evaluated a total of 45 corporate work teams involved in cognitive tasks from three major firms. Their results found no significant correlation between gender and group performance.

Second, Riordan and Shore (1997) studied a gender-productivity relationship by evaluating 98 work groups (over 1500 employees) at a life insurance company. No statistically significant correlation was found between gender and perceived productivity.

Recently, however, Dwyer, Richard, and Chadwick (2003) specifically evaluated the gender diversity-firm financial performance relationship by analyzing a sample of banks. Employee productivity and return on equity (ROE) were used as measures of firm performance. Taken as an independent factor, gender diversity was positively correlated with both productivity and ROE but was not statistically significant in either case. When a gender diversity-organizational culture interaction effect was analyzed, the statistical relationships were significant at times, but the direction of the correlations became mixed.

2.2.2 Board of Directors

Zahra and Stanton (1988) examined the relationship between gender diversity and firm financial performance. The sample consisted of 100 Fortune 500 firms and performance variables included return on equity, profit margin on sales, earnings per share, dividends per share, and profit. They did not find a statistically significant relationship between board gender-diversity and firm financial performance.

Shrader, Blackburn, and lies (1997) analyzed the relationship between firm financial performance (as measured by return on assets and return on equity) and the percentage of female board members. The sample consisted of approximately 200 Fortune 500 firms. Their results indicated a statistically significant negative relationship.

Carter, Simkins, and Simpson (2003) evaluated the relationship between board gender-diversity and firm value. The study analyzed the Fortune 1000 firms for the year 1997. Using Tobin's Q as a measure of firm value, they found statistically significant positive relationships for both the "presence of women" and "percentage of women" factors.

A recent study by the Conference Board of Canada indicated that a gender-diverse board appeared to be associated with efforts to improve corporate governance/performance. The research found that 94% of boards in North America with three or more women insisted on conflict-of-interest guidelines. This was in contrast to 58% of all-male boards. In addition, 72% of boards with two or more women conducted formal board-performance evaluations as opposed to 49% of all-male boards. One significant factor seemed to be the presence of several women on a board. It appeared more likely that a lone female would end up accepting the view of the dominant group (Hymowitz, 2003).

3. STUDY DESIGN

Given the very limited empirical work and the inconclusive findings in these areas, it is clear that additional analysis is needed. This study attempts to improve upon and add to the previous work in this area by:

1) Using as its theoretical framework an innovative extension of mean-variance portfolio theory.

2) Evaluating not only the number and percentage of women on the board but also as corporate officers (management).

3) Using Market Value Added (a market measure of firm financial performance that isolates the equity component of firm value) in addition to traditional accounting measures (ROE, ROA).

4) Evaluating two distinctly different time periods (1996: economic expansion, "up" stock market and 2001: economic recession, "down" stock market) rather than just a single year or consistent economic period.

5) Incorporating firm size as measured by market capitalization.

Model:

FirmFinancialPerformance=[B.sub.0]+[B.sub.1]Beta+[B.sub.2]Size+[B.sub.3] FemaleOff+[B.sub.4]%FemaleOff+[B.sub.5]FemaleBrdMem+[B.sub.6]% FemaleBrdMem

Where:

Dependent Variables:

Firm Financial Performance Measures include:

ROE: Return on Equity

ROA: Return on Assets

MVA: Market Value Added (mkt. value of equity--book value of equity)

Independent Variables:

Control Variables:

Beta: Beta of the stock (systematic risk) (Sharpe, 1964)

Size: Size of the company as measured by market capitalization (Fama and French, 1992)

Study Variables:

FemaleOff: Number of female officers

%FemaleOff: Percentage of female officers

FemaleBrdMem: Number of female board members

%FemaleBrdMem: Percentage of female board members

4. DATA AND ANALYSIS

1) Data were collected for approximately 100 of the largest firms as determined by revenue for both 1996 (n=105) and 2001 (n=102).

2) Data for the Independent Female variables were obtained from Catalyst for 1996 and 2001.

3) Data for the Dependent and Control variables were obtained from the S&P Stock Market Encyclopedia for 1996 and 2001.

Regression analysis was conducted on the complete data sets for both 1996 (an "up" year, middle of bull market, economic expansion) and 2001 (a "down" year, middle of bear market, economic recession). This was done in order to control for the "economic" and "market" effects.

5. RESULTS

Results from the 1996 total sample analysis indicate the following (Table #1):

1) A statistically significant, positive relationship between the percentage of female officers and ROE.

2) A statistically significant, positive relationship between the percentage of female board members and MVA.

3) A statistically significant, negative relationship between the number of female board members and MVA.

Results from the 2001 total sample analysis indicate the following (Table #2):

1) A statistically significant, positive relationship between the percentage of female board members and MVA.

2) A statistically significant, negative relationship between the number of female board members and MVA

Most interestingly, in both the 1996 and 2001 complete samples, a positive relationship existed between the percentage of female board members (directors) and MVA while a negative relationship occurred between the number of female board members (directors) and MVA. These consistent, individual, MVA-related results imply that MVA is increased when (holding all else constant): 1) the percentage of female board members is increased and 2) the number of female board members is reduced. Indirectly, this combination seems to support the proposition that boards should be smaller rather than larger. Thus, a follow-up analysis was conducted using the following model:

Firm Financial Performance = [B.sub.0]+[B.sub.1]Beta+[B.sub.2]Size+ [B.sub.3]TotalBrdMem

Where:

Dependent Variables:

Firm Financial Performance Measures include:

ROE: Return on Equity

ROA: Return on Assets

MVA: Market Value Added (mkt. value of equity--book value of equity)

Independent Variables:

Control Variables:

Beta: Beta of the stock (systematic risk) (Sharpe, 1964)

Size: Size of the company as measured by market capitalization (Fama and French, 1992)

Study Variables:

TotalBrdMem: Total number of board members (directors)

In both the 1996 and 2001 complete samples, a statistically significant, negative relationship was observed between the total number of board members and MVA. This confirmed the previous individual results (Tables #1 and #2). Although not intended to be a formal part of the study, this ancillary finding of negative correlation between board size and MVA contributes one more piece of information to the prior work in this area. This observation compares to previous findings as follows:

Chaganti, Mahajan, and Sharma (1985) used bankruptcy as a firm financial performance indicator. In their sample of 21 pairs of failed and non-failing retail firms, they found that smaller boards were associated with higher rates of bankruptcy.

Zahra and Stanton (1988) examined the relationship between board size and firm financial performance. The sample consisted of 100 Fortune 500 firms and performance variables included return on equity, profit margin on sales, earnings per share, dividends per share, and profit. They did not find a statistically significant relationship between board size and firm financial performance.

Yermack (1996) carried out an extensive analysis of the board size-firm financial performance relationship. In particular, he found 1) a negative correlation between board size and firm value as measured by Tobin's Q, 2) a negative correlation between board size and firm financial performance as measured by return on assets, return on sales, and sales-to-assets, and 3) abnormal, positive stock returns around the announcement of a significant board-size reduction (and the opposite for an announcement of an increase in board size).

Dalton, Daily, Johnson, and Ellstrand (1999) conducted a meta-analysis of 131 samples and observed that board size tended to be positively associated with various measures of firm financial performance. This was found to be true whether the performance indicators were accounting-based or market-based.

It is quite obvious that the findings in this area are also far from conclusive and that additional research is needed.

6. SUMMARY AND DISCUSSION

In general, this analysis achieved its intended goals of improving upon and adding to the previous work in this area. First, it used as its theoretical framework an innovative extension of a fundamental financial theory. Second, the analysis utilized a unique model and drew upon data from two distinctive, economic/market time periods. Finally, the primary, statistically significant results were consistent throughout the study. Specifically, in both the 1996 and 2001 complete samples, a positive relationship existed between the percentage of female board members (directors) and MVA while a negative relationship occurred between the number of female board members (directors) and MVA. Indirectly, this combination of MVA-related results also seemed to support the proposition that boards should be smaller rather than larger. Subsequent analysis confirmed this contention. In particular, a statistically significant negative relationship was observed between the total number of board members and MVA in the 1996 as well as the 2001 complete samples.

Overall, the statistical results indicated that enhanced financial performance (as measured by MVA) was associated with 1) a smaller board of directors, 2) a board comprised of a relatively high proportion of women, and 3) a board consisting of a relatively low number of women. Less bureaucracy and improved operational efficiency associated with a smaller board (and thus any corresponding decrease in the number of female directors as a consequence of a board-size reduction) may be possible explanations for the observed negative correlations. Meanwhile, the increased influence and utilization of female leadership and managerial skills may explain the positive correlation between the percentage of female directors and MVA. In terms of firm financial performance and value maximization, these results definitely have implications for those involved in corporate governance and management as well as members of the general investment community.

7. LIMITATIONS AND FUTURE OPPORTUNITIES

The data set for each year of the study consisted of approximately 100 of the largest corporations based on revenue. In the future, the composition of the sample could be determined by using different criteria for size (e.g. market capitalization, total assets, etc.). Also, the sample could be altered to include or focus on medium and/or small firms. As has been noted elsewhere, "supply" related factors may play a role in the apparent lack of women at these organizational levels. In other words, for positions in particular industries, the proportion of qualified or interested females may be less than the proportion of females in the aggregate population. Thus, a follow-up study focusing on individual industries would be in order. In addition, different variables could be used to evaluate firm financial performance (e.g. excess value, economic value added, cash flow, Tobin's Q, etc.). More years or an extended time period could also be considered for evaluation.

Interestingly, this list of potential improvements also reflects the vast research opportunities that are available in this relatively under-investigated area. The importance and value of continuing to examine this topic is best reflected in the following quotes:

"Given the resources many of the firms are pouring into efforts to diversify their work forces and given the large number of women on Wall Street, the lack of progress is surprising. It's gotten much better, but there is still a glass ceiling with less reason for a glass ceiling. Corporations are now losing many of the best and the brightest."--D. Soder (in an article by R. Abelson), The New York Times, July 26, 2001.

"This call for change is hardly a new one. For years, we've known that by 2010, 70 percent of new entrants into the workforce will be women and/or people of color. We've seen a steady increase in the presence of female corporate officers, but the pace has been incremental at best. We're still talking less than 20 percent. How can that be, when women are getting more than 50 percent of BAs and MAs, 43 percent of PhDs, nearly 50 percent of law degrees and medical degrees, and slightly more than 33 percent of business degrees? There has never been a greater need for fresh air in business--and if that's what we're really after, then we have to get fresh people."--S. Wellington, American Way, Feb. 15, 2003.

REFERENCES

Abelson, R., "A Survey of Wall Street Finds Women Disheartened", The New York Times, July 26, 2001, C1.

Carter D., Simkins, B., and Simpson, W., "Corporate Governance, Board Diversity, and Firm Value", Financial Review, Vol. 38 (1), 2003, 33-53.

Chaganti, R., Mahajan, V., and Sharma, S., "Corporate Board Size, Composition and Corporate Failures in Retailing Industry", Journal of Management Studies, Vol. 22, 1985, 400-416.

Dalton, D., Daily, C., Johnson, J., and Ellstrand, A., "Number of Directors and Financial Performance: A Meta-Analysis", Academy of Management Journal, Vol. 42 (6), 1999, 674-686.

Dwyer, S., Richard, O., and Chadwick, K., "Gender Diversity in Management and Firm Performance: The Influence of Growth Orientation and Organizational Culture", Journal of Business Research Vol. 56 (12), 2003, 1009-1019.

Fama, E. and French, K., "The Cross-Section of Expected Stock Returns", Journal of Finance, Vol. 47, 1992, 427-465.

Hymowitz, C., "In the U.S., What Will It Take to Create Diverse Boardrooms?", The Wall Street Journal, July 8, 2003, B1.

Markowitz, H., "Portfolio Selection", Journal of Finance, March, 1952, 77-91.

Pelled, L., Eisenhardt, K., and Xin, K., "Exploring the Black Box: An Analysis of Work Group Diversity, Conflict, and Performance", Administrative Science Quarterly, Vol. 44 (1), 1999, 1-28.

Riordan, C. and Shore, L., "Demographic Diversity and Employee Attitudes: An Empirical Examination of Relational Demography within Work Units", Journal of Applied Psychology, Vol. 82 (3), 1997, 342-358.

Rosener, J., America's Competitive Secret: Utilizing Women as a Management Strategy, Oxford University Press, New York, 1995.

Sharpe, W., "Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk", Journal of Finance, September, 1964, 425-442.

Shrader C., Blackburn, V., and lles, P., "Women in Management and Firm Financial Value: An Exploratory Study", Journal of Managerial Issues, Vol. 9 (3), 1997, 355-372.

Wellington, S., "Ideafest", American Way, February 15, 2003, 59.

Williams, K. and O'Reilly, C., "Demography and Diversity in Organizations: A Review of 40 Years of Research", In: Staw, B. and Cummings, L., editors. Research in Organizational Behavior, JAI Press, Greenwich, CT, 1998.

Yermack, D., "Higher Market Valuation of Companies with a Small Board of Directors", Journal of Financial Economics, Vol. 40 (2), 1996, 185-211.

Zahra, S. and Stanton, W., "The Implications of Board of Directors' Composition for Corporate Strategy and Performance", International Journal of Management, Vol. 5 (2), 1988, 229-236.

Jeff Heinfeldt, Ohio Northern University, Ada, Ohio, USA

Dr. Jeff Heinfeldt earned his Ph.D. at Kent State University in 1995. Currently he is an associate professor of finance at Ohio Northern University.

TABLE 1
STATISTICAL RESULTS FOR 1996 TOTAL SAMPLE (N=105)

INDEPENDENT
VARIABLES                                 DEPENDENT VARIABLES

                                     ROA         ROE          MVA

# of Female         Coefficient    -0.0793   -0.1216       -0.0447
Officers            p-value          .2227     .5444         .5212

% of Female         Coefficient     0.0602    0.6839         .0772
Officers            p-value          .4395     .0047 ***     .3562

# of Female Board   Coefficient     0.2898   -2.0061       -2.9836
Members             p-value          .8539     .6804         .0800 *

% of Female         Coefficient    -0.0419   -0.4182        0.3577
Board Members       p-value          .8256     .4785         .0829 *

Total Number of     Coefficient    -0.0561    0.0252       -0.3832
Board Members       p-value          .7216     .9608         .0237 **

* statistically significant at a .10 level

** statistically significant at a .05 level

*** statistically significant at a .01 level

TABLE 2
STATISTICAL RESULTS FOR 2001 TOTAL SAMPLE (N=102)

INDEPENDENT
VARIABLES                                 DEPENDENT VARIABLES

                                      ROA        ROE         MVA

# of Female          Coefficient    -0.1082    -0.1622     0.2119
Officers             p-value          .1979      .6471      .2279

% of Female          Coefficient     0.0698     0.2417    -0.1030
Officers             p-value          .5082      .5882      .6405

# of Female Board    Coefficient     0.3381     7.3508    -5.3554
Members              p-value          .8007      .1967      .0583 *

% of Female          Coefficient    -0.0333    -0.9922     1.0178
Board Members        p-value          .8721      .2585      .0203 **

Total Number of      Coefficient    -0.2050     1.0768    -0.9936
Board members        p-value          .3606      .2566      .0369 **

* statistically significant at a .10 level

** statistically significant at a .05 level