SOCIAL TOPICS (Archive): WORKPLACE

Published, July 1997

       What factors influence the advancement of women into senior management and the representation of women on boards of directors? Our clients have long been interested in these questions. For seven years, we have collected data on the diversity of boards and senior management teams, and have voted against non-diverse boards on proxy ballots.

       On behalf of our client, the Women’s Equity Mutual Fund, we analyzed our 1995 data, examining the relationship between gender diversity on boards and the “glass ceiling” issue of gender diversity in senior management. To our surprise, we found no statistically significant relationship at the aggregate level. Boards with an above average proportion of women might or might not be found in conjunction with above average representation of women in senior management.

       This year, we took other possible factors into account. We looked for relationships in our 1996 data on 165 companies between gender inclusivity for boards and senior management and company characteristics such as size, degree of insider or institutional ownership, and economic sector.1 We found that the factors most important in determining gender representation on boards differed from those determining representation in senior management (officers).

       For gender inclusivity on boards, the composition of senior management is the most significant factor. Companies with above average numbers of female senior managers are half again as likely to have diverse boards as those with less than an average number of female managers. For these companies, no other explanatory variable is statistically significant. In companies without women in senior management, the company’s industry is an important factor. Healthcare, consumer product, communication, industrial materials, consumer cyclical, and utility companies were over four times more likely to have an average or above representation of women on the board than companies in the producer products, energy, financial services, and technology industries. Among the 40 companies in higher representation industries, 62% had 10% or greater representation of women on the board, with an overall average of 11.4% women directors. Among the 45 companies in the lower representation group, women achieved a 10% or greater representation on the board in only 7 of the 45 companies, or 16%; the overall average representation of women directors was only 4.9% for these 45 companies.

       Company size is the most important factor determining the representation of women in senior management. Smaller companies are half again as likely to achieve an average or above average representation of women officers as larger companies. The degree of insider ownership is also important, but complex. Small companies with either very high or very low insider ownership are much more likely to have women in senior management. Small companies in our sample with low insider ownership include utilities, gas pipelines, a few industrial companies, and healthcare related companies. Those companies with high insider ownership are generally recently founded, in rapidly changing fields. In several cases, women were involved in the founding, either as member of a team of scientists, or in partnership with a spouse.

       Another interesting finding concerned the companies selected for our Innovations Fund, which seeks companies with positive contributions to society in either product, or process, or management, but not specifically for gender diversity in management. Although these are small companies with medium levels of insider ownership, they had more than twice the representation of women in senior management as similar companies not selected for this Fund, and, as a group, the Innovations Fund companies had an above average representation of women.

       For companies above $1.5 billion in market capitalization, less than 10% insider ownership leads to an approximately average representation of women among senior management. But, companies above $1.5 billion in size, with insider ownership of more than 10%, have only a one in ten chance of an average level of representation in senior management. This is even more pronounced in the companies above $1.5 but below $15 billion — in these 16 companies, not one of them achieved even an average level of representation of women in senior management; only 5 had any women in senior management, and the average level of representation of women in senior management among these companies was only 1.7%. Among the four largest companies with high insider ownership, representation of women in senior management was just slightly less than average, at 9.7% of total managers.

       Thus, the profile of the company with the most resistant glass ceiling appears to be a medium size company, between $1.5 billion and $15 billion in market capitalization, with greater than 10% insider ownership. One quarter of these companies were in financial services. The most permeable glass ceilings are found at small companies, less than $1.5 billion in market capitalization, with either very little or very high insider ownership, in either heavily regulated industries such as utilities, or rapidly evolving fields, such as healthcare and environmental technology. For companies with high insider ownership, it also helps if women were there at the beginning — or have founded the company.

       Neither the representation of women on the board as we defined it nor the degree of institutional ownership were highly significant in explaining the presence of women in senior management. For the last seven years, we have been voting against non-diverse boards, in part as a strategy to improve the diversity of senior management. This analysis suggests that, while board gender diversity may be a small factor in affecting senior management, company size, degree of control by insiders, and the overall progressivity of the firm (as captured by inclusion in the Innovations Fund) are more important factors. Gender inclusivity in senior management may need to be addressed more directly.

1 We used a statistical technique called CHAID, or Chi-square Analysis of Integer Data. First, the data are grouped into categories, such as sectors of the economy, and high, medium, or low amounts of institutional or insider ownership. Then we defined two “success” variables, one for women in senior management and one for women on boards of directors, so that a company was “successful” if it had 10% or more women in senior management or on the board — approximately the average levels. The technique then examines the strength of the relationship between each of the possible explanatory variables and “success,” based on the probability level of the associated chi-square, ranking them by the most explanatory, then the next, then the next, etc. At each stage of explanation, the cases are divided into groups based on the probability level of the explanatory variable. Then this smaller group is examined to see if there are further explanations of variations in “success” within this group.