Sex-based occupational segregation in U.S. state bureaucracies, 1987-97.(Statistical Data Included)

Introduction

The literature on the distribution of women and men in public-sector jobs and the integration of women into government managerial ranks is replete with evidence that women often face glass walls, especially in certain types of agencies (Lewis and Emmert 1986; Pfeifer and Davis-Blake 1987; Kellough 1989, 1990; Guy and Duerst-Lahti 1992; Bullard and Wright 1993; Cornwell and Kellough 1994; Guy 1994; Lewis and Nice 1994; Naif 1994; Newman 1994; Riccucci and Saidel 1997). The glass wall metaphor refers to occupational segregation attributed to barriers that restrict women's access to certain types of jobs (or agencies) or to factors that concentrate women within certain types of jobs (or agencies). Glass walls are likely to persist when (1) the agency and its clientele do not engage in efforts to remove impediments to change; and/or (2) skills necessary to perform jobs in a given agency are not highly valued outside the agency.

Large-scale studies of sex-based occupational segregation have been conducted on U.S. federal government (Rosenbloom 1977; Lewis and Emmert 1986; Kellough 1989, 1990) and municipal government workforces (Miller, Kerr, and Reid 1999), but the lack of access to comparative public employment data on states has resulted in a dearth of generalizable empirical studies on employment in state-level bureaucracies. Numerous studies employ summary data from the U.S. Equal Employment Opportunity Commission's (EEOC) Job Patterns for Minorities and Women In State and Local Government (Cayer and Sigelman 1980; Moore and Mazey 1986; Sigelman and Dometrius 1986; Lewis and Nice 1994; Dometrius and Sigelman 1997). (1) Many other studies are based on samples drawn from one state or just a few states (Bayes 1989; Rehfuss 1986; Hale, Kelly, and Burgess 1989; Kelly et al. 1991; Duerst-Lahti and Johnson 1992; Guy 1992; Newman 1994). The data employed by Bullard and Wright (1993) and Riccucci and Saidel (1997) are limited to agency heads from across the 50 states and gubernatorial appointees from nearly all 50 states, respectively. Previous research on the distribution of state jobs provides some useful conceptual and analytic frameworks, but findings from these studies provide little basis for generalizing about the employment patterns of career administrative and professional personnel.

In this article, we examine the distribution of women and men in state-level administrative and professional positions by agency type and over time in each of the 50 states to determine whether agency missions are associated with the extent and nature of glass walls. We are interested in the following questions: (1) what is the distribution of female and male administrators and professionals in various functional areas in state governments (police, corrections, natural resources/parks, highways, public welfare, etc.); and (2) is the underrepresentation and/or overrepresentation of female or male administrators and professionals in various functional areas related to the agency missions in those functional areas?

These questions are important for several reasons. Greater access to quality jobs, including public-sector managerial positions, promotes the economic, social, and political progress of women, and it may result in longterm benefits through altered socialization processes (Kanter 1977; MacManus 1981; Tolleson-Rinehart 1991; Guy and Duerst-Lahti 1992). Greater representation for women among managerial personnel is also likely to result in changes in management styles and leadership processes, perhaps making them more innovative and democratic (Tolleson-Rinehart 1991; Duerst-Lahti and Johnson 1992; Stivers 1993). The increased presence of women is also likely to have a distinctive impact on policy outputs (Mezey 1978; Stewart 1980; Stanwick and Kleeman 1983; Welch 1985; Gelb and Palley 1996; Carroll, Dodson, and Mandel 1991; Dodson and Carroll 1991; Tolleson-Rinehart 1991; Thomas 1994; but see Donahue 1997; Ford and Dolan 1999).

The focus on women in state bureaucracies is important because state governments function as distinct entities with their own constitutions, laws, and independently raised revenues (Kelly et al. 1991). Osborne (1988) and Van Horn (1989) argue that state governments have become more responsive, innovative, and effective; however, these efforts are by no means distributed uniformly across all states (Brudney, Hebert, and Wright 1999). Nonetheless, over the last three decades there has been a marked increase in state agency activism (Wright, Yoo, and Cohen 1991; Bullard and Wright 1993). Bullard and Wright (1993) argue that administrative activism in the states converged with women's activism to create expanded opportunities for women at the state level, often in the form of newly created agencies and programs. They also argue that, all else being equal, new agencies and programs are less likely than existing agencies and programs to be subject to constraints invidious to women (Bullard and Wright 1993). Such expanded opportunities provide reason to think that, in general, employment patterns should indicate evidence of growth in women's share of managerial positions in state government.

Theory and Hypotheses

When and where women enter public-sector management positions varies by agency type or mission (Saltzstein 1986; Stein 1986; Pfeifer and Davis-Blake 1987; Bullard and Wright 1993; Lewis and Nice 1994; Naif 1994; Newman 1994; Kearney and Sellers 1996; Miller, Kerr, and Reid 1999; Reid, Kerr, and Miller 2000). Lewis and Nice (1994) find that women employed in state and local governments are heavily overrepresented in public welfare, housing, and health agencies, while men are overrepresented in streets and highways, fire, and police departments. Riccucci and Saidel (1997) find that among state-level political appointees, women are severely underrepresented in utilities and transportation, police, fire, natural resources, and corrections agencies, but that women approach or surpass parity in public welfare, employment security, health, labor and human resources, and civil and human fights agencies. Bullard and Wright (1993) find that, as late as 1988, women were not represented at all or were extremely underrepresented at the highest administrative levels in law enforcement and corrections, agriculture, natural resources, commerce, attorneys general, economic development, labor, oil and gas, water pollution, and water resources agencies. The general picture that emerges is that women are poorly represented in regulatory and distributive agencies, but they are relatively well represented in redistributive agencies.

Some of the most interesting theoretical work on the relationship between agency missions and occupational segregation has been done by Lowi (1985) and Newman (1994). Lowi's (1985) general argument is that sex-based occupational segregation among managerial personnel varies depending on whether agency missions are, on balance, distributive, regulatory, or redistributive (see also Newman 1994; Miller, Kerr, and Reid 1999). Newman's (1994) analysis of Florida state bureaucracies suggests that most women work in redistributive and regulatory agencies. She argues that discrimination in the hiring and promotion of women is most severe in distributive agencies (Newman 1994). We follow the contributions of Lowi (1985) and Newman (1994) in order to test our theoretical arguments about sex-based occupational segregation in state bureaucracies for 1987-97 across all 50 states.

Typical distributive functions performed by state governments include (1) construction, repair, and administration of highways and bridges; (2) administration and management of forest and other state lands; (3) provision and operation of parks and recreational facilities; (4) historic preservation and beautification; and (5) development and management of water resources. The corresponding functions on the EEO-4 form (see appendix for a description of functions) are highways and streets, natural resources and parks and recreation, and community development (EEOC Form 164).

Newman (1994) and Lowi (1985) argue that mutually reinforcing relationships between agencies and their clientele make distributive processes resistant to change, thereby making it more difficult for women to enter these occupational areas. Distributive agencies are characterized by a reliance on professional and occupational norms, promotion of specialists (engineers, biologists, physical and social scientists, and the like) rather than generalists, limited due process requirements, relatively wide fields of discretion, and limited sensitivity to discriminatory practices (Corson and Paul 1966; Lowi 1985; Lewis and Emmert 1986; Newman 1994). Because of these factors, we believe males will be heavily overrepresented among managerial workforces in distributive agencies, and, over time, there will be little or no progress toward gender balance in these agencies.

Regulatory functions performed by state governments include (1) police protection; (2) operation of prisons, reformatories, and detention homes, including activities related to parole and probation; (3) fire protection; and (4) regulation of numerous business practices such as labor relations, securities, environmental conditions, banking, insurance, water transportation; utilities, energy, oil and gas, etc. The corresponding job categories on the EEO-4 are police, fire, corrections, and utilities and transportation (EEOC Form 164).

Regulatory policies impose obligations and sanctions that seek to control individual and collective behavior (Lowi 1964, 1972). Policies dealing with criminal behavior involving persons and property are clearly regulatory in nature.2 So are other policies, such as regulation of business, under which the likelihood of coercion is immediate and the applicability of coercion works through individual contact (Lowi 1972). We retain the regulatory policy label for convenience sake, but wish to point out that it may not simply be the regulatory nature of police and corrections that makes them gender segregated. Traditional policing agencies do not promote gender balance (Cayer and Sigelman 1980; Riccucci 1986; Warner, Steel, and Lovrich 1989), and these agencies have developed a host of practices that exclude women from high-level positions (Steel and Lovrich 1986; Stein 1986; Mladenka 1991; Newman 1993). Nor do we think that nonpolicing state regulatory agencies--those other than police, fire, and corrections--will hire large numbers of women into administrative and professional positions. However, we do expect the percentage of women in nonpolicing regulatory agencies-in this case, utilities and transportation--to be greater than percentages in traditional policing functions because (1) these agencies are likely to hire attorneys (see Newman 1994); and (2) women's enrollment in law schools has increased over the last two decades (Chronicle of Higher Education 1999).

Redistributive functions performed by state governments include (1) management of public welfare programs; (2) management of employment security; (3) mental health and retardation programs; (4) programs for the aging; (5) …