Skip to main content
.
Share with LinkedInShare with TwitterShare with facebook

Highlights

What GAO Found

The overall proportion of racial or ethnic minorities in the Department of State’s (State) full-time, permanent, career workforce increased from 28 to 32 percent from fiscal year 2002 to fiscal year 2018. The direction of change for specific groups varied. For instance, the proportion of Hispanics increased from 5 percent to 7 percent, while the proportion of African Americans decreased from 17 to 15 percent. Also, the proportion of racial or ethnic minorities and women was lowest at management and executive levels.

Diversity in State Department Workforce in Fiscal Years 2002 and 2018

GAO’s analyses of State data for fiscal years 2002 through 2018 found differences in promotion outcomes for racial or ethnic minorities and whites. These differences existed in both descriptive analyses, which calculated simple averages, and in adjusted analyses, which controlled for certain individual and occupational factors that could influence promotion. Compared with the descriptive analyses, the adjusted analyses found smaller percentage differences between promotion outcomes for racial or ethnic minorities and whites in State’s Civil Service and some evidence of smaller percentage differences in State’s Foreign Service. GAO found generally lower promotion rates for racial or ethnic minorities than for whites. For example, controlling for factors such as education, years of service, and occupation, racial or ethnic minorities in the Civil Service had lower rates and odds of promotion than whites at each rank from early career through senior management. Also, both types of analysis found promotion outcomes for women compared with men were lower in the Civil Service and generally higher in the Foreign Service. For example, women in the Foreign Service were more likely than men to be promoted in early to mid career.

State has identified some barriers to equal opportunity but should consider other issues that could indicate potential barriers to diversity. In its annual reports to the Equal Employment Opportunity Commission (EEOC), State has identified issues such as underrepresentation of Hispanic employees. However, other State analysis, GAO’s analysis, and GAO’s interviews with State employee groups highlighted additional issues that could indicate barriers. For example, State’s reports have not identified discrepancies in midcareer promotion of racial or ethnic minorities relative to whites, which GAO found in its analysis. Taking additional steps to identify diversity issues could enhance State’s ability to detect and remove barriers to equal participation in its workforce.

Why GAO Did This Study

State has expressed a commitment to maintaining a workforce that reflects the diverse composition of the United States and has undertaken efforts to increase representation of diverse groups in its Civil and Foreign Services. EEOC requires some federal agencies, including State, to systematically identify, examine, and remove barriers to equal participation at all levels of their workforce and to report on such barriers annually.

GAO was asked to review issues related to the diversity of State’s workforce. This report examines (1) the demographic composition of State's workforce in fiscal years 2002 through 2018; (2) any differences in promotion outcomes for various demographic groups in State’s workforce; and (3) the extent to which State has identified barriers to diversity in its workforce. GAO analyzed State’s personnel data for its full-time, permanent, career workforce for fiscal years 2002 through 2018. GAO analyzed the number of years until promotion from early career ranks to the executive rank in both the Civil and Foreign Services. GAO’s analyses do not completely explain the reasons for differences in promotion outcomes, which may result from various unobservable factors. Thus, GAO’s analyses do not establish a causal relationship between demographic characteristics and promotion outcomes. GAO also reviewed State documents and interviewed State officials and members of 11 employee groups.

What GAO Recommends

GAO recommended that State take additional steps to identify diversity issues that could indicate potential barriers to equal opportunity in its workforce. State concurred with this recommendation.

Recommendations

Recommendations for Executive Action

NumberAgencyRecommendation
1Department of State

The Secretary of State should take additional steps to identify diversity issues that could indicate potential barriers to equal opportunity in its workforce. For example, State could conduct additional analyses of workforce data and of employee groups’ feedback. (Recommendation 1)

View recommendation(s) status

Introduction 


The Department of State (State) has expressed a commitment to maintaining a workforce that reflects the diverse composition of the United States. In addition, State has worked to increase representation of diverse groups in its Civil and Foreign Services. However, concerns about the demographic composition of State’s workforce are longstanding. For example, in 1989, we reported on the underrepresentation of minorities and women at middle and senior levels of the Foreign Service.[1] More recently, in a hearing before the Senate Committee on Foreign Relations, State’s Director General of the Foreign Service and Director of Human Resources stated that the diversity of the agency’s workforce remains a concern and that the agency is working hard to increase diversity.[2] As of the end of fiscal year 2018, State had 22,806 full-time, permanent, career employees, including 9,546 in its Civil Service and 13,260 in its Foreign Service.

We were asked to review issues related to the diversity of State’s workforce. This report examines (1) the demographic composition of State's workforce in fiscal years 2002 through 2018, (2) any differences in promotion outcomes for various demographic groups in State’s workforce, and (3) the extent to which State has identified any barriers to diversity in its workforce.

To examine the demographic composition of State’s workforce over time, we analyzed State’s Global Employment Management System (GEMS) data on the department’s full-time, permanent, career workforce[3] for fiscal years 2002 through 2018.[4] We were unable to analyze the numbers and percentages based on sexual orientation, because federal personnel records do not include these data. For each year, we calculated the demographic composition of the workforce by racial or ethnic group and by gender for State overall and for State’s Civil and Foreign Services.[5] In addition, we compared the demographics of State’s workforce in fiscal year 2018 with the most recent available data on demographics of (1) the federal workforce, as reported by the Office of Personnel Management (OPM), and (2) the relevant civilian labor force, from the Census Bureau’s Equal Employment Opportunity (EEO) tabulation.[6] For both the Civil and Foreign Services, we examined workforce composition by racial or ethnic group and by gender across ranks for fiscal year 2018.[7] Through a review of documentation, electronic testing, and interviews with knowledgeable agency officials, we determined that these data were sufficiently reliable for our purposes.

To examine promotion outcomes for various racial or ethnic minorities[8] and women in State’s workforce, we conducted two types of analyses using State’s GEMS data on its full-time, permanent, career workforce for fiscal years 2002 through 2018.[9]

  • We conducted descriptive analyses of State’s data, calculating simple averages to compare promotion rates for racial or ethnic minorities and whites and for women and men.[10]
  • We conducted adjusted analyses using a multivariate statistical method (i.e., duration analysis)[11] that accounted for certain individual and occupational factors other than racial or ethnic minority status and gender that could influence promotion, including the length of time it takes to be promoted. Specifically, we used a discrete-time multivariate statistical logit model to analyze the number of yearly cycles it took to be promoted up to the executive level from General Schedule (GS) grade 11 in the Civil Service and from Class 4 in the Foreign Service.[12] We examined the statistical relationship between promotion and racial or ethnic minority status and gender,[13] incorporating various individual and position-specific characteristics[14] in the models to control for differences in promotion outcomes.[15] Our analyses do not completely explain the reasons for differences in promotion outcomes, which may result from various unobservable factors. Thus, our analyses do not establish a causal relationship between demographic characteristics and promotion outcomes.

To examine the extent to which State has identified barriers to diversity, we reviewed annual reports on workforce diversity that State submitted to the U.S. Equal Employment Opportunity Commission (EEOC) for fiscal years 2009 through 2018 as well as workforce analyses that State conducted. We also met with officials from State’s Office of Civil Rights and Bureau of Human Resources. In addition, we conducted structured interviews with representatives of employee groups in State’s Civil and Foreign Services. Appendix I provides additional details of our scope and methodology.[16]

We conducted this performance audit from April 2018 to January 2020 in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives.

Background 


Diversity in the Federal Workforce

Title VII of the Civil Rights Act of 1964 and Section 501 of the Rehabilitation Act of 1973 mandate that all federal personnel decisions be made without discrimination on the basis of race, color, religion, sex, national origin, reprisal, or disability and require that agencies establish a program of equal employment opportunity for all federal employees and applicants.[17] The EEOC has oversight responsibility for those programs, and EEOC regulations direct agencies to maintain a continuing affirmative program to promote equal opportunity and to identify and eliminate discriminatory practices and policies.[18] In support of those programs, agencies are to, among other things, conduct a continuing campaign to eradicate discrimination from the agency’s personnel policies, practices, and working conditions.[19]

EEOC’s Management Directive 715 (MD-715) provides policy guidance and standards for establishing and maintaining effective affirmative programs of equal employment opportunity. Through MD-715, EEOC provides that, as a part of a model EEO program to prevent unlawful discrimination, federal agencies are to regularly evaluate their employment practices to identify barriers to EEO in the workplace, take measures to eliminate identified barriers, and report annually on these efforts to EEOC.[20]

EEOC’s MD-715 guidance lays out a four-step process for federal agencies to identify and eliminate barriers to their workforce diversity (see fig. 1).

Figure 1: U.S. Equal Employment Opportunity Commission’s (EEOC) Management Directive 715 Process for Identifying and Eliminating Barriers to Federal Agency Workforce Diversity

Note: EEOC defines a trigger as a trend, disparity, or anomaly that suggests the need for further inquiry into a particular policy, practice, procedure, or condition. It defines a barrier as an agency policy, principle, or practice that limits or tends to limit employment opportunities for members of a particular gender, race, ethnic background, or disability status.

As figure 1 shows, the first step of the process calls for agencies to analyze various sources of workforce data to look for trends, disparities, or anomalies—which this report collectively refers to as diversity issues—that suggest the need for further inquiry into a particular policy, practice, procedure, or condition.[21] Diversity issues can be identified on the basis of various sources of information, such as workforce statistics or employee surveys.

The second step of the process calls for agencies to conduct an investigation to pinpoint barriers that could be the causes of any diversity issues. EEOC reporting requirements state that a barrier is an agency policy, procedure, practice, or condition that limits, or tends to limit, employment opportunities for members of a particular group on the basis of their sex, race, ethnic background, or disability status. According to EEOC’s instructions for identifying and eliminating barriers, agencies are required to move beyond treating the symptom (e.g., anomalies found in workforce demographics) to eliminate the underlying barrier, or cause of the symptom (e.g., lack of career development opportunities).

The third step of the process is to devise a plan to eliminate any barriers that are not job related and consistent with business necessity and to report annually to EEOC. The fourth step is to assess the success of plans to eliminate barriers. In this report, we focus on the first two steps of the process.

State Initiatives to Increase Workforce Diversity

State’s Office of Civil Rights and Bureau of Human Resources jointly oversee initiatives to increase State’s workforce diversity.

  • Office of Civil Rights. According to State, the Office of Civil Right’s mission is to propagate fairness, equity, and inclusion at the agency. The Secretary has delegated to the Director of the Office of Civil Rights the tasks of advancing diversity within the department and ensuring equal opportunity to all employees. The office develops State’s annual MD-715 report and submits it to EEOC. In addition, the office works with relevant bureaus to gather and analyze necessary data and information to complete the MD-715 report.
  • Bureau of Human Resources. According to State, the Bureau of Human Resources coordinates inter- and intra-agency efforts to advance diversity and inclusion. The bureau’s Senior Advisor for Diversity, Inclusion, and Outreach provides guidance and support to the bureaus in implementing initiatives to recruit a diverse workforce and improve career development. This official also leads the development of State’s Diversity and Inclusion Strategic Plan in coordination with agency stakeholders. According to Bureau of Human Resources officials, this process includes receiving feedback from employee representative groups. Additionally, according to State officials, the bureau conducts workforce data analysis at the request of relevant internal and external stakeholders.

Office of Civil Rights officials told us that, to guide the office’s and the Bureau of Human Resources’ diversity efforts, they regularly meet with employee groups that represent various demographic populations in the agency.[22] According to State officials, employee groups serve as a link between diverse employee constituencies and State’s senior management, Office of Civil Rights staff, and Bureau of Human Resources staff.

State’s most recent Diversity and Inclusion Strategic Plan, issued in 2016, describes the department’s recruitment, career development, and bureau-level diversity initiatives, among others.[23] According to the plan, having a workforce that reflects the composition of U.S. citizenry is a long-standing commitment that continues to be a department priority “today and into the future.”

Recruitment. According to Bureau of Human Resources officials, State’s recruitment efforts for the Civil Service and the Foreign Service[24] include the following efforts to enhance diversity:

  • Diplomats in Residence. According to State officials, as of January 2019, this initiative has assigned 16 Foreign Service Officers and specialists to university campuses throughout the United States, as well as 10 recruiters based in Washington, D.C. The officers visit historically black colleges and universities and Hispanic-serving institutions as well as institutions with significant minority enrollment.
  • Thomas R. Pickering Foreign Affairs Fellowship Program and Charles B. Rangel International Affairs Program. These two programs recruit candidates for the Foreign Service by providing graduate fellowships to college seniors and college graduates. Both programs seek to attract highly talented and qualified individuals who represent ethnic, racial, gender, social, and geographic diversity.
  • Outreach initiatives. According to State officials, recruiters for the department participate in career fairs, present on panels, and host information sessions at the conferences of partners with a focus on diversity and inclusion, such as the Hispanic Association of Colleges and Universities and the Congressional Black Caucus Foundation.

Career development. According to State, the agency views mentoring and career development as vital elements to enhance leadership skills, retain employees, and develop an agile workforce. The officials stated that mentoring and career development initiatives help employees develop the skills needed to reach the senior ranks. State maintains mentoring programs for both the Civil and Foreign Services. For example, State reported that the agency had initiated a Senior Executive Service career development program in which, as of October 2018, more than 70 percent of the selected participants were women and 50 percent identified as a minority.

Bureau-level initiatives. Some regional and functional bureaus lead efforts to increase diversity and inclusion. According to State’s Senior Advisor for Diversity, Inclusion, and Outreach, bureau leaders set the tone, and provide support, for bureau-level initiatives. Examples of these initiatives include the following:

  • Bureau of European and Eurasian Affairs’ Driving Diversity, Growth, and Excellence. Through this initiative, senior leaders mentor midlevel employees and introduce employees to the bureau’s expectations regarding diversity and inclusion.
  • Bureau of Diplomatic Security’s Diversity Working Group. The purpose of this initiative is to support and advise bureau leadership on initiatives to promote diversity and inclusion throughout the bureau and all of its directorates.[25]

Characteristics of State’s Civil and Foreign Services

As figure 2 shows, State had 22,806 full-time, permanent, career employees at the end of fiscal year 2018—an increase of more than 38 percent since 2002.[26] From fiscal year 2002 to fiscal year 2018, the number of full-time, permanent, career Civil Service employees increased from 6,831 to 9,546, or by nearly 40 percent. During the same period, the number of full-time, permanent, career Foreign Service employees increased from 9,739 to 13,260, or by 36 percent.

Figure 2: Numbers of Full-Time Permanent Career Employees in Department of State’s Civil and Foreign Services, Fiscal Years 2002-2018

Note: The data shown reflect numbers of full-time, permanent, career employees at the end of each fiscal year.

Civil Service

State’s Civil Service made up 42 percent of the department’s full-time, permanent, career workforce at the end of fiscal year 2018. Civil Service employees serve alongside Foreign Service employees in professional, technical, administrative, and clerical positions; help formulate and implement U.S. foreign policy; provide strategic and logistical support to U.S. diplomatic missions; and issue passports and travel warnings, among other functions. Civil Service employees are on the GS classification system, which has 15 ranks, ranging from GS-1 (lowest) to GS-15 (highest), followed by the executive rank.

Table 1 shows the number and percentage of employees in each Civil Service rank. Additionally, table 1 provides information on the promotion rates from ranks below executive for promotions effective in fiscal year 2018. Each year, State promotes different numbers of its Civil Service employees. Promotion generally becomes more competitive for the higher ranks. For example, 15 percent of GS-11 employees in fiscal year 2017 were promoted to GS-12 in fiscal year 2018, while 1 percent of GS-15 employees in fiscal year 2017 were promoted to the executive rank.

Table 1: Number and Percentage of State’s Civil Service Employees in Each Rank and Rates of Promotion to Higher Ranks, Fiscal Year 2018
RankNumber of employees in rankPercentage of employees in rankRate of promotion to next-higher rank, %a
Executive1572N/A
GS-15992101
GS-141,841194
GS-132,885306
GS-121,3871517
GS-111,4141515
GS-10 and lower870926
Legend: GS = General Schedule.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: The data shown reflect numbers and percentages of Civil Service employees at the end of fiscal year 2018.

aWe calculated the promotion rate as the number of newly elevated employees in the next-higher rank in fiscal year 2018 divided by the number of employees in the given rank in fiscal year 2017. For example, 15 percent of GS-11 employees in fiscal year 2017 were promoted to GS-12 in fiscal year 2018. For GS-10 and lower, we calculated the promotion rate as the average of the number of newly elevated employees in each of the next-higher ranks in fiscal year 2018 divided by the number of employees in the given ranks in fiscal year 2017.

Foreign Service

State’s Foreign Service made up 58 percent of the agency’s full-time, permanent, career workforce at the end of fiscal year 2018. Foreign Service employees serve as either generalists or specialists. Foreign Service generalists help formulate and implement U.S. foreign policy and are assigned to work in one of five career tracks: consular, economic, management, political, or public diplomacy. Foreign Service specialists support and maintain the functioning of overseas posts and serve in 25 different skill groups, filling positions such as security officer or information management.

Foreign Service Officers enter at Class 4, 5, or 6, depending on education and experience. Officers can be promoted up to Class 1, after which they can apply for the executive rank. Table 2 shows the number of employees at each of these ranks.[27] Table 2 also shows promotion rates across ranks for promotions effective in fiscal year 2018. Each year, State promotes different numbers of its Foreign Service employees. Promotion generally becomes more competitive for the higher ranks. For example, 15 percent of Class 4 employees in fiscal year 2017 were promoted to Class 3 in fiscal year 2018, while 4 percent of Class 1 employees in fiscal year 2017 were promoted to the executive rank.

Table 2: Number and Percentage of State’s Foreign Service Employees in Each Rank and Rate of Promotion to Higher Rank, Fiscal Year 2018
RankNumber of employees in rankPercentage of employees in rankRate of promotion to next-higher rank, %a
Executive9337N/A
Class 11,605124
Class 22,846217
Class 33,644279
Class 43,1732415
Class 5645546
Class 6 and lower414334
Legend: N/A = not applicable.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: The data shown reflect numbers and percentages of Foreign Service employees at the end of fiscal year 2018.

aWe calculated the promotion rate as the number of newly elevated employees in the next-higher rank in fiscal year 2018 divided by the number of employees in the given rank in fiscal year 2017. For example, 15 percent of Class 4 employees in fiscal year 2017 were promoted to Class 3 in fiscal year 2018. Competitive promotion for Foreign Service generalists starts at Class 4. For Class 6 and lower, we calculated the promotion rate as the average of the number of newly elevated employees in each of the next-higher ranks in fiscal year 2018 divided by the number of employees in the given ranks in fiscal year 2017.

Major Findings 

Overall Proportion of Racial or Ethnic Minorities at State Has Grown, but Proportions of African Americans and Women Have Declined

Proportion of Racial or Ethnic Minorities at State Increased, While Proportion of African Americans Decreased

Proportion of Racial or Ethnic Minorities at State Increased Overall

From fiscal year 2002 to fiscal year 2018, the proportion of racial or ethnic minorities[28] among State’s full-time, permanent, career employees[29] increased from 28 percent to 32 percent, as figure 3 shows.[30] This increase in the proportion of racial or ethnic minorities at State overall was driven solely by an increase in the proportion of racial or ethnic minorities in the Foreign Service. During this period,

  • the proportion of racial or ethnic minorities in the Civil Service decreased slightly from 44 percent to 43 percent[31] and
  • the proportion of racial or ethnic minorities in the Foreign Service increased from 17 percent to 24 percent.[32]

Figure 3: Proportions of White and Racial or Ethnic Minority Employees at Department of State Overall and in Civil Service and Foreign Service, Fiscal Years 2002 and 2018

Notes: The data shown reflect numbers and proportions of white and racial or ethnic minority employees at the end of fiscal years 2002 and 2018. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” includes individuals whose race or ethnicity is not identified. Percentages may not sum to 100 because of rounding. For instances where the racial or ethnic group changed over time for an employee record, we assigned the most recent value to all available years.

Proportion of Racial or Ethnic Minorities at State Was Lower Than in Federal Workforce but Higher Than in Relevant Civilian Labor Force

We compared State’s proportion of ethnic and racial minorities to proportions in the federal workforce and relevant civilian labor force (RCLF). Our comparison of State workforce data with data from OPM’s Federal Equal Opportunity Recruitment Program (FEORP) report for the federal government (last produced in fiscal year 2016) found the following:

  • The proportion of racial or ethnic minorities at State in fiscal year 2018 (32 percent) was lower than the proportion in the federal workforce in fiscal year 2016 (36 percent). For more details, see appendix III.[33]
  • The proportion of racial or ethnic minorities at State increased from 28 percent in fiscal year 2002 to 32 percent in fiscal year 2018. In contrast, the proportion of racial or ethnic minorities in the federal workforce increased from 31 percent in fiscal year 2002 to 36 percent in fiscal year 2016.

Our comparison of State workforce data from fiscal year 2018 with data for the RCLF[34] (from 2006 through 2010—the most recent available data) found that the proportions of racial or ethnic minorities at State were higher[35] than the proportions in the RCLF for three occupational groups: officials and managers, professional workers, and administrative support workers.[36] For more details, see appendix III.

Proportion of African Americans at State Decreased, While Proportions of Hispanics, Asians, and Other Racial or Ethnic Minorities Increased

Although the overall proportion of racial or ethnic minorities increased at State from fiscal year 2002 to fiscal year 2018, the direction of change for specific racial or ethnic minority groups varied—the proportion of African Americans fell, while the proportions of Hispanics, Asians, and other racial or ethnic minorities rose.[37] The overall number of employees at State increased from 16,570 to 22,806; however, the proportion of African Americans fell from 17 percent to 15 percent of all employees, as shown in figure 3.[38] Our analysis found that the overall decline in the proportion of African Americans at State reflects a substantial decline in the proportion of African Americans in State’s Civil Service.

  • The proportion of African Americans in State’s Civil Service decreased from 34 percent to 26 percent in fiscal years 2002 through 2018.[39]
  • The proportion of African Americans in State’s Foreign Service increased from 6 percent to 7 percent over the same period.[40]

In contrast to the proportions of African Americans, the proportions of Hispanics, Asians, and other racial or ethnic minorities in State’s Civil Service and Foreign Services increased from fiscal year 2002 to fiscal year 2018.[41]

Proportion of Racial or Ethnic Minorities in Civil and Foreign Services Was Generally Much Smaller in Higher Ranks

Our analysis of State data for fiscal year 2018 found that the proportions of racial or ethnic minorities were lower than the proportions of whites at GS-11, GS-13, and higher ranks in the Civil Service and at all ranks in the Foreign Service, as figure 4 shows. For example, at GS-13, the proportion of whites (58 percent) exceeded the proportion of racial or ethnic minorities (42 percent). State data for fiscal year 2002 showed a similar difference between the two proportions.[42]

Figure 4: Proportions of White and Racial or Ethnic Minority Employees in the Department of State’s Civil and Foreign Services across Ranks, Fiscal Year (FY) 2018

Notes: The data shown reflect proportions of white and racial or ethnic minority employees at the end of fiscal year 2018. Racial or ethnic minorities exclude non-Hispanic whites and non-Hispanic employees whose race was unspecified. “Unspecified” includes individuals whose race or ethnicity is not identified. For instances where the racial or ethnic group changed over time for an employee record, we assigned the most recently reported group to all available years.

Additionally, as figure 4 shows, our analysis found that the proportions of racial or ethnic minorities in fiscal year 2018 were progressively lower in each rank above GS-12 in the Civil Service and above Class 5 in the Foreign Service.[43] Similarly, our analysis of State’s data for fiscal year 2002 found that the proportions of racial or ethnic minorities were generally progressively lower in higher ranks in both the Civil Service and the Foreign Service.

According to State officials, hiring diverse classes at the lower ranks of the Foreign Service improves representation at higher ranks over time.[44] Specifically, State officials noted that because rising from Class 4 to the Senior Foreign Service takes approximately 20 years, the diversity of the senior ranks should improve.

Proportion of Women at State Decreased Over Time

Decline in Proportion of Women in Civil Service Exceeded Increase in Proportion of Women in Foreign Service

From fiscal year 2002 to fiscal year 2018, the proportion of women at State decreased slightly, from 44 percent to 43 percent, as figure 5 shows.[45] Our analysis found that the overall decline in the proportion of women at State reflects a decline in the proportion of women in State’s Civil Service. Specifically:

  • The proportion of women in the Civil Service decreased from 61 percent to 54 percent.[46]
  • The proportion of women in the Foreign Service increased from 33 percent to 35 percent.[47]

Figure 5: Proportions of Men and Women in the Department of State and Its Civil Service and Foreign Service, Fiscal Years 2002 and 2018

Notes: The data shown reflect proportions of male and female employees at the end of fiscal years 2002 and 2018. For instances where the gender changed over time for an employee record, we assigned the most recent value to all available years.

In addition, the proportion of African American women at State decreased from 13 percent in fiscal year 2002 to 9 percent in fiscal year 2018. See the text box for additional details.

Proportion of Minority and Gender Groupings at the Department of State, Fiscal Years 2002-2018

The proportion of African American women at the Department of State (State) decreased from 13 percent in fiscal year 2002 to 9 percent in fiscal year 2018, contributing to overall decreases in the proportions of African Americans and women at State.a Our analysis found that the overall decline in the proportion of African American women at State reflects a decline in the proportion of African American women in State’s Civil Service.

  • Civil Service. The proportion of African American women in the Civil Service decreased from 27 percent to 17 percent.
  • Foreign Service. The proportion of African American women in the Foreign Service increased from 2 percent to 3 percent.

In contrast, the proportions of the following demographic groups increased at State overall, in the Civil Service, and in the Foreign Service:

  • African American men
  • Hispanic men
  • Hispanic women
  • Asian men
  • Asian women
  • Other racial or ethnic minority men
  • Other racial or ethnic minority womenb

The proportions of White men and White women decreased at State overall and in the Foreign Service but increased in the Civil Service.

Source: GAO analysis of Department of State data. | GAO-20-237

aOffice of Personnel Management data for fiscal years 2002 and 2016 show that the proportion of African American women in the federal workforce remained around 11 percent.

b"Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial.

Proportion of Women at State Was Similar to Federal Workforce but Mixed in Comparison with Relevant Civilian Labor Force

We compared the proportion of women at State with the proportions of women in the federal workforce and RCLF. Our comparison of State workforce data with federal government workforce data from OPM’s FEORP report (last published in fiscal year 2016) found the following:

  • The proportion of women was 43 percent both at State in fiscal year 2018 and in the federal workforce in fiscal year 2016.[48]
  • The proportion of women decreased slightly from 44 percent to 43 percent both at State in fiscal years 2002 through 2018 and in the federal workforce in fiscal years 2002 and 2016.

Our comparison of State workforce data from fiscal year 2018 with data from the RCLF[49] (from 2006 through 2010—the most recent available data) found that the proportion of women was lower[50] at State than in the RCLF for two occupational groups: (1) officials and managers and (2) professional workers. However, the proportion of women was higher at State than in the RCLF for administrative support workers.[51] For more details, see appendix III.

Proportions of Women in Civil and Foreign Services Were Generally Smaller in Higher Ranks

State data for fiscal year 2018 show that the proportions of women were lower than the proportions of men at GS-14 and higher ranks in the Civil Service and at Class 4 and higher ranks in the Foreign Service, as figure 6 shows. For example, in fiscal year 2018, the proportion of men at Class 4 (64 percent) exceeded the proportion of women (36 percent). In fiscal year 2002, a similar difference existed.[52]

Figure 6: Proportions of Women and Men in the Department of State’s Civil and Foreign Services across Ranks, Fiscal Year (FY) 2018

Notes: The data shown reflect proportions of male and female employees at the end of fiscal year 2018. For instances where the gender changed over time for an employee record, we assigned the most recent value to all available years.

As figure 6 shows, our analysis of State data for fiscal year 2018 for the Civil Service found progressively smaller proportions of women with every increase in rank, except between GS-11 and GS-12 and between GS-14 and GS-15.[53] Specifically, the proportion of women in the Civil Service in fiscal year 2018 was 69 percent at GS-10 or below, 58 percent at GS-11, 52 percent at GS-13, 46 percent at GS-14, and 38 percent at the executive level.

Similarly, our analysis of State data for fiscal year 2018 for the Foreign Service found progressively smaller proportions of women from Class 6 to Class 3 and from Class 1 to executive.[54] For example, the proportion of women in the Foreign Service in fiscal year 2018 was 68 percent at Class 6 or lower, 56 percent at Class 5, 36 percent at Class 4, and 29 percent at Class 3. Our analysis of State data for fiscal year 2002 similarly found progressively smaller proportions of women across all ranks, from Class 5 to executive, in the Foreign Service.

Promotion Outcomes Were Generally Lower for Racial or Ethnic Minorities Than for Whites and Differed for Women Relative to Men

Our analyses of State data for fiscal years 2002 through 2018 found differences between promotion outcomes for racial or ethnic minorities relative to whites and for women relative to men. We found these differences when conducting descriptive analyses, which calculated simple averages, as well as adjusted analyses, which controlled for certain individual and occupational factors other than racial or ethnic minority status and gender that could influence promotion. In particular, we found generally lower promotion rates for racial or ethnic minorities than for whites in both the Civil Service and the Foreign Service. For women relative to men, we found generally lower promotion rates in the Civil Service and generally higher promotion rates in the Foreign Service. However, our analyses do not completely explain the reasons for differences in promotion outcomes, which may result from various unobservable factors. Thus, our analyses do not establish a causal relationship between demographic characteristics and promotion outcomes.

We Conducted Descriptive and Adjusted Analyses of Promotion Outcomes

To examine promotion outcomes for various demographic groups in State’s Civil and Foreign Services, we conducted two types of analysis—descriptive and adjusted. We considered promotion to be an increase in rank between fiscal years. Our analyses include all individuals in the original rank and do not distinguish between individuals who did or did not apply or between those who were eligible or ineligible for promotion.[55] For the Foreign Service, we focused on promotions starting from Class 4, as these corresponded to the competitive promotion process for most Foreign Service employees. For the Civil Service, we focused on promotions starting from GS-11, which corresponds to Class 4 in the Foreign Service.

Descriptive Analysis

We conducted descriptive analyses of State data, calculating simple averages to compare actual promotion rates for racial or ethnic minorities and whites and for women and men in State’s Civil and Foreign Services. For each fiscal year, we calculated the rate of promotion from each rank as the number of newly elevated employees in the next-higher rank in the following fiscal year divided by the number of employees in the given rank in the current year. We used the promotion rates to compute absolute percentage differences and relative percentage differences in the promotion rates for the various groups.

While our descriptive analyses provide helpful context on promotion at State, they do not account for the variety of factors besides racial or ethnic minority status or gender that may affect promotion outcomes, nor do they show whether systematic delays in promotion exist.

Adjusted Analysis

To examine the statistical relationship between racial or ethnic minority status, gender, and promotion in State’s Civil and Foreign Services, we conducted adjusted analyses of State data, using a multivariate statistical method.[56] This method accounted for certain individual and occupational factors other than racial or ethnic minority status and gender that could influence promotion, including the length of time it takes to be promoted. Accounting for these factors, the regression results from the adjusted analyses produced adjusted promotion rates, odds ratios, and percentage differences in relative odds of promotion. See the text box for the specific control variables we used in our adjusted analyses.

Control Variables Used in Adjusted Analysis of Department of State Data

In addition to controlling for racial or ethnic minority status and gender, our adjusted analyses of State data for the Department of State’s (State) Civil and Foreign Services controlled for a number of factors that could influence promotion.

  • For both the Civil and Foreign Services, we controlled for length of time in each grade or class prior to promotion; years of federal service; age when hired at State; veteran’s status; graduation from a college or university considered Ivy League or located in the District of Columbia,a Virginia, or Maryland; use of long-term leave in the prior year; change between service types; occupation; and fiscal years.
  • For the Foreign Service, we also controlled for service in a hardship assignment in the prior year; overseas service in the prior year; and proficiency in a hard language.

Including these factors allowed us to estimate differences, if any, in the odds of promotion even if the racial or ethnic minority employees and White employees were hired at the same age; had the same years of federal experience; worked at State in the same fiscal years; and had the same gender, time spent in each grade or class prior to promotion, veteran’s status, education background, use of long-term leave, service change record, occupation, hardship or overseas assignment, and language proficiency. However, our analyses do not completely explain the reasons for differences in promotion outcomes, which may result from various unobservable factors. Thus, our analyses do not establish a causal relationship between demographic characteristics and promotion outcomes.

Source: GAO. | GAO-20-237

aWe included these variables because there may be a perception that graduates from a college or university considered Ivy league would be high-quality applicants to State and because some of the colleges or universities located in the District of Columbia, Virginia, or Maryland have highly respected programs related to foreign service that may provide networking opportunities.

By accounting for a variety of individual and occupational factors, the multivariate statistical method we used for our adjusted analysis can estimate the extent to which racial or ethnic minority status and gender are related to promotion outcomes. The objective of this method is not to establish that racial or ethnic minority status and gender are key causal factors in promotion outcomes. However, this method can provide insights into whether differences between promotion outcomes for racial or ethnic minority groups and genders persist after certain individual and occupational factors, such as length of service, have been accounted for.

Because our adjusted analyses used a multivariate statistical method, they have certain limitations. First, our adjusted analyses do not completely explain the reasons for differences in odds of promotion. While we included numerous control variables relevant to promotion, various unobservable factors for which our adjusted analyses did not explicitly account—for example, employees’ skills, motivation, performance, or abilities[57]—may have caused differences in promotion outcomes. [58] Second, the presence of institutional budget constraints could affect the number of available promotion slots across State, which may help explain some of the observed differences in promotion outcomes.[59] Thus, our analyses do not establish a causal relationship between demographic characteristics and promotion outcomes.

Promotion Outcomes Were Generally Lower for Racial or Ethnic Minorities Than for Whites at All Ranks Except Executive

Descriptive and Adjusted Promotion Rates and Adjusted Promotion Odds in Civil Service Were Generally Lower for Racial or Ethnic Minorities Than for Whites

Both our descriptive analysis and adjusted analysis of data for State’s Civil Service found that the promotion rate[60] was generally lower for racial or ethnic minorities than for whites (see table 3).[61] In addition, our adjusted analysis found that racial or ethnic minorities in State’s Civil Service generally had lower odds of promotion than their white counterparts.

Table 3: Promotion Outcomes for Whites and Racial or Ethnic Minorities in the Department of State’s Civil Service, Fiscal Years 2002-2018
GS-11 to GS-12GS-12 to GS-13GS-13 to GS-14GS-14 to GS-15GS-15 to exec.
Descriptive analysis
Promotion rate for whites, %24.825.18.25.31.6
Promotion rate for racial or ethnic minorities, %18.414.66.13.91.3
Percentage point difference between promotion rate for racial or ethnic minorities and promotion rate for whites-6.4-10.6-2.1-1.4-0.3
Percentage difference between promotion rate for racial or ethnic minorities and promotion rate for whites, %-25.7-42.0-25.8-26.6-16.1
Adjusted analysis

Promotion rate for whites, %

95 percent confidence interval, %

23.2

[22.1, 24.4]

21.8

[20.7, 22.9]

7.9

[7.5, 8.4]

5.2

[4.8, 5.6]

1.6

[1.3, 1.9]

Promotion rate for racial or ethnic minorities, %

95 percent confidence interval, %

19.8

[18.8, 20.9]

17.4

[16.3, 18.6]

6.5

[6.0, 7.1]

4.1

[3.6, 4.7]

1.5

[0.9, 2.1]

Percentage point difference between promotion rate for racial or ethnic minorities and promotion rate for whites-3.4-4.4-1.4-1.1-0.1

Odds ratio for promotion for racial or ethnic minorities relative to whites

95 percent confidence interval, %

0.738**

[0.657, 0.829]

0.707**

[0.641, 0.780]

0.806**

[0.729, 0.891]

0.782**

[0.672, 0.910]

0.957

[0.611, 1.496]

Percentage difference between promotion odds for racial or ethnic minorities and promotion odds for whites, %

95 percent confidence interval, %

-26.2**

[-34.3, -17.1]

-29.3**

[-35.9, -22.0]

-19.4**

[-27.1, -10.9]

-21.8**

[-32.8, -9.0]

-4.3

[-38.9, 49.6]

Legend: GS = General Schedule, exec. = executive, ** = statistical significance at p-value < 0.01; * = statistical significance at p-value < 0.05.
Source: GAO analysis of State data. | GAO-20-237

Notes: The p-value represents the smallest level of significance for which our estimate results in a rejection of the hypothesis of there being no difference in the odds of promotion. For each rank, the promotion rates based on the descriptive analyses for whites and for racial or ethnic minorities represent an average of the number of newly elevated whites or racial or ethnic minorities in the next-higher rank in the following year, divided by the number of whites or racial or ethnic minorities in the given rank in the current year. Given this methodology, we were not able to calculate promotion rates for fiscal year 2018, because the State data we analyzed ended in fiscal year 2018. For the descriptive analysis, we calculated the percentage point difference and percentage difference for racial or ethnic minorities relative to whites on the basis of unrounded promotion rates; thus, differences are due to rounding. We calculated the percentage difference for racial or ethnic minorities relative to whites as the unrounded percentage point difference divided by the unrounded promotion rate for whites. For the adjusted analysis of State data, we conducted discrete-time duration analysis using a logit model that controlled for a variety of factors relevant to promotion, and we analyzed the time duration (number of years) to be promoted. The adjusted analysis does not completely explain the reasons for differences in odds of promotion. While various independent variables capture and control for many different characteristics across different demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

As table 3 shows, our descriptive analysis of the data for State’s Civil Service found that the average percentage of racial or ethnic minorities promoted from GS-11 and higher ranks was lower than the average percentage of whites promoted from the same ranks. For example, our descriptive analysis found that in fiscal years 2002 through 2017, an average of 18.4 percent of racial or ethnic minorities was promoted from GS-11 to GS-12, compared with an average of 24.8 percent of whites.[62] This negative 6.4 percentage point difference indicates that the average rate of promotion from GS-11 to GS-12 was 25.7 percent lower for racial or ethnic minorities than for whites.[63] However, our descriptive analysis does not account for the variety of factors besides racial or ethnic minority status that may affect promotion rates, including the length of time it takes to be promoted.[64]

Our adjusted analysis of the data for State’s Civil Service, controlling for factors other than racial or ethnic minority status that could influence promotion, found that racial or ethnic minorities had lower adjusted rates of promotion and lower odds of promotion from each rank, from early career (GS-11) through senior manager level (GS-15), than their white counterparts.[65] Specifically, our adjusted analysis of State data for fiscal years 2002 through 2018 found the following:

  • The average adjusted rate of promotion from GS-11 to GS-12 for racial or ethnic minorities was 19.8 percent, compared with an average of 23.2 percent of whites. This statistically significant difference[66] indicates that the odds of promotion from GS-11 to GS-12 in the Civil Service were 26.2 percent lower for racial or ethnic minorities than for whites.[67]
  • Our estimates of the adjusted rates and odds of promotion from GS-12 to GS-13, from GS-13 to GS-14, and from GS-14 to GS-15 were also statistically significantly lower for racial or ethnic minorities than for whites.
  • While the adjusted rate of promotion from GS-15 to executive was lower for racial or ethnic minorities than for whites, there was no statistically significant difference in the odds of promotion from GS-15 to executive for racial or ethnic minorities relative to whites in the Civil Service.[68]
  • Compared with the descriptive analysis, the adjusted analysis generally found a smaller percentage difference in promotion outcomes for racial or ethnic minorities relative to whites in the Civil Service.

Figure 7 shows key results of our descriptive and adjusted analyses of State data for racial or ethnic minorities and whites in State’s Civil Service.

Figure 7: Relative Differences in Promotion Rates and Adjusted Promotion Odds for Racial or Ethnic Minorities Compared with Whites in the Department of State’s Civil Service, Fiscal Years 2002-2018

Notes: For each rank, the promotion rates based on the descriptive analyses for whites and for racial or ethnic minorities represent an average of the number of newly elevated whites or racial or ethnic minorities in the next-higher rank in the following year, divided by the number of whites or racial or ethnic minorities in the rank in the current year. Given this methodology, we were not able to calculate promotion rates for fiscal year 2018, because the State data we analyzed ended in fiscal year 2018. For the descriptive analysis, we calculated the percentage difference for racial or ethnic minorities relative to whites as the difference between the unrounded promotion rates for racial or ethnic minorities and whites divided by the unrounded promotion rate for whites; thus, differences are due to rounding. For the adjusted analysis of State data, we conducted discrete-time duration analysis using a logit model that controlled for a variety of factors relevant to promotion, and we analyzed the time duration (number of years) to be promoted. The adjusted analysis does not completely explain the reasons for differences in odds of promotion. While various independent variables capture and control for many different characteristics across different demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Descriptive Promotion Rates Were Generally Lower for Racial or Ethnic Minorities Than Whites in Foreign Service, but Differences in Adjusted Promotion Rates and Odds Were Generally Not Statistically Significant

Our descriptive analysis of data for State’s Foreign Service found that the rate of promotion[69] was generally lower for racial or ethnic minorities than for whites.[70] In addition, our adjusted analysis found differences in the promotion rates for racial or ethnic minorities and whites. However, the only statistically significant difference between promotion outcomes for the two groups in the Foreign Service was for promotion from Class 4 to Class 3, where both the rate and the odds of promotion were lower for racial or ethnic minorities than for whites. See table 4 for more details.

Table 4: Promotion Outcomes for Whites and Racial or Ethnic Minorities in the Department of State’s Foreign Service, Fiscal Years 2002-2018
Class 4 to Class 3Class 3 to Class 2Class 2 to Class 1Class 1 to executive
Descriptive analysis
Promotion rate for whites, %17.014.39.67.7
Promotion rate for racial or ethnic minorities, %16.212.68.07.9
Percentage point difference between promotion rate for racial or ethnic minorities and promotion rate for whites-0.9-1.7-1.50.2
Percentage difference between promotion rate for racial or ethnic minorities and promotion rate for whites, %-5.0-12.2-15.82.7
Adjusted analysis

Promotion rate for whites, %

95 percent confidence interval, %

17.2

[16.9, 17.5]

14.0

[13.7, 14.4]

9.4

[9.1, 9.7]

7.6

[7.2, 8.0]

Promotion rate for racial or ethnic minorities, %

95 percent confidence interval, %

15.7

[15.1, 16.3]

13.5

[12.9, 14.2]

8.8

[8.1, 9.5]

8.2

[7.3, 9.2]

Percentage point difference between promotion rate for racial or ethnic minorities and promotion rate for whites-1.5-0.5-0.60.6

Odds ratio for promotion for racial or ethnic minorities relative to whites

95 percent confidence interval, %

0.872**

[0.820, 0.927]

0.951

[0.880, 1.027]

0.925

[0.835, 1.026]

1.097

[0.944, 1.275]

Percentage difference between promotion odds for racial or ethnic minorities and promotion odds for whites, %

95 percent confidence interval, %

-12.8**

[-18.0, -7.3]

-4.9

[-12.0, 2.7]

-7.5

[-16.5, 2.6]

9.7

[-5.6, 27.5]

Legend: ** = statistical significance at p-value < 0.01; * = statistical significance at p<0.05.
Source: GAO analysis of State data. | GAO-20-237

Notes: The p-value represents the smallest level of significance for which our estimate results in a rejection of the hypothesis of there being no difference in the odds of promotion. For each rank, the promotion rates based on the descriptive analyses for whites and for racial or ethnic minorities represent an average of the number of newly elevated whites or racial or ethnic minorities in the next-higher rank in the following year, divided by the number of whites or racial or ethnic minorities in the given rank in the current year. Given this methodology, we were not able to calculate promotion rates for fiscal year 2018, because the State data we analyzed ended in fiscal year 2018. For the descriptive analysis, we calculated the percentage point difference and percentage difference for racial or ethnic minorities relative to whites on the basis of unrounded promotion rates; thus, differences are due to rounding. We calculated the percentage difference for racial or ethnic minorities relative to whites as the unrounded percentage point difference divided by the unrounded promotion rate for whites. For the adjusted analysis of State data, we conducted discrete-time duration analysis using a logit model that controlled for a variety of factors relevant to promotion, and we analyzed the time duration (number of years) to be promoted. The adjusted analysis does not completely explain the reasons for differences in odds of promotion. While various independent variables capture and control for many different characteristics across different demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

As table 4 shows, our descriptive analysis of the data for State’s Foreign Service found that for Class 4 and higher ranks, a lower average percentage of racial or ethnic minorities than of whites was promoted from each rank except Class 1. For example, our descriptive analysis found that in fiscal years 2002 through 2017, an average of 16.2 percent of racial or ethnic minorities were promoted from Class 4 to Class 3, compared with an average of 17.0 percent of whites.[71] This negative 0.9 percentage point difference indicates that the average rate of promotion from Class 4 to Class 3 was 5.0 percent lower for racial or ethnic minorities than for whites.[72] However, this descriptive analysis does not account for the variety of factors besides racial or ethnic minority status that may affect promotion rates, including the length of time it takes to be promoted.[73]

Our adjusted analysis of the data for State’s Foreign Service, controlling for factors other than racial or ethnic minority status that could influence promotion, found that racial or ethnic minorities had lower adjusted rates and odds of promotion from Class 4 to Class 3 than their white counterparts. [74] Specifically, our adjusted analysis of State data for fiscal years 2002 through 2018 found the following:

  • On average, the adjusted rate of promotion from Class 4 to Class 3 for racial or ethnic minorities was 15.7 percent, compared with 17.2 percent of whites. This statistically significant difference indicates that the odds of promotion from Class 4 to Class 3 in the Foreign Service were 12.8 percent lower for racial or ethnic minorities than for whites.
  • The adjusted rates and odds of promotion from Class 3 to Class 2 and from Class 2 to Class 1 were lower for racial or ethnic minorities than for whites, and the adjusted rates and odds of promotion for Class 1 to executive were higher for racial or ethnic minorities than for whites. However, we did not find any statistically significant differences in the odds of promotion from Class 3 to Class 2, from Class 2 to Class 1, and from Class 1 to executive for racial or ethnic minorities relative to whites in the Foreign Service. That is, we could not conclude that there was a statistical relationship between racial or ethnic minority status and promotion at these class levels.
  • Compared with the descriptive analysis, the adjusted analysis found a larger percentage difference in promotion outcomes from Class 4 to Class 3 for racial or ethnic minorities relative to whites.

Figure 8 shows key results of our descriptive and adjusted analyses of State data for racial or ethnic minorities and whites in the Foreign Service.

Figure 8: Relative Differences in Promotion Rates and Adjusted Promotion Odds for Racial or Ethnic Minorities Compared with Whites in the Department of State’s Foreign Service, Fiscal Years 2002-2018

Notes: For each rank, the promotion rates based on the descriptive analyses for whites and for racial or ethnic minorities represent an average of the number of newly elevated whites or racial or ethnic minorities in the next-higher rank in the following year, divided by the number of whites or racial or ethnic minorities in the rank in the current year. Given this methodology, we were not able to calculate promotion rates for fiscal year 2018, because the State data we analyzed ended in fiscal year 2018. For the descriptive analysis, we calculated the percentage difference for racial or ethnic minorities relative to whites as the difference between the unrounded promotion rates for racial or ethnic minorities and whites divided by the unrounded promotion rate for whites; thus, differences are due to rounding. For the adjusted analysis of State data, we conducted discrete-time duration analysis using a logit model that controlled for a variety of factors relevant to promotion, and we analyzed the time duration (number of years) to be promoted. The adjusted analysis does not completely explain the reasons for differences in odds of promotion. While various independent variables capture and control for many different characteristics across different demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Promotion Outcomes for Women Relative to Men Varied for Civil and Foreign Services

Descriptive Promotion Rates Were Generally Lower for Women Than Men in Civil Service, but Differences in Adjusted Promotion Rates and Odds Were Not Statistically Significant

Our descriptive analysis of State data for fiscal years 2002 through 2018 found that the rate of promotion[75] in State’s Civil Service was generally lower for women than for men.[76] However, our adjusted analysis did not find any statistically significant differences in the promotion rates or odds of promotion for women relative to men in the Civil Service. See table 5 for more details.

Table 5: Promotion Outcomes for Women and Men in the Department of State’s Civil Service, Fiscal Years 2002-2018
GS-11 to GS-12GS-12 to GS-13GS-13 to GS-14GS-14 to GS-15GS-15 to exec.
Descriptive analysis
Promotion rate for men, %23.220.67.54.71.5
Promotion rate for women, %20.519.17.45.21.5
Percentage point difference between promotion rate for women and promotion rate for men-2.7-1.5-0.10.50.0
Percentage difference between promotion rate for women and promotion rate for men, %-11.6-7.1-1.49.8-0.7
Adjusted analysis

Promotion rate for men, %

95 percent confidence interval, %

21.7

[20.4, 22.9]

20.1

[18.9, 21.2]

7.6

[7.0, 8.1]

4.9

[4.4, 5.4]

1.6

[1.3, 1.9]

Promotion rate for women, %

95 percent confidence interval, %

21.4

[20.4, 22.4]

19.5

[18.4, 20.5]

7.3

[6.8, 7.8]

4.9

[4.5, 5.4]

1.5

[1.1, 1.9]

Percentage point difference between promotion rate for women and promotion rate for men-0.3-0.6-0.20.0-0.1

Odds ratio for promotion for women relative to men

95 percent confidence interval, %

0.977

[0.862, 1.108]

0.953

[0.865, 1.049]

0.966

[0.880, 1.061]

1.006

[0.879, 1.150]

0.951

[0.708, 1.277]

Percentage difference between promotion odds for women and promotion odds for men, %

95 percent confidence interval, %

-2.3

[-13.8, 10.8]

-4.7

[-13.5, 4.9]

-3.4

[-12.0, 6.1]

0.6

[-12.1, 15.0]

-4.9

[-29.2, 27.7]

Legend: GS = General Schedule, exec. = executive, ** = statistical significance at p-value < 0.01; * = statistical significance at p<0.05.
Source: GAO analysis of State data. | GAO-20-237

Notes: The p-value represents the smallest level of significance for which our estimate results in a rejection of the hypothesis of there being no difference in the odds of promotion. For each rank, the promotion rates based on the descriptive analyses for men and for women represent an average of the number of newly elevated men or women in the next-higher rank in the following year, divided by the number of men or women in the given rank in the current year. Given this methodology, we were not able to calculate promotion rates for fiscal year 2018, because the State data we analyzed ended in fiscal year 2018. For the descriptive analysis, we calculated the percentage point difference and percentage difference for women relative to men on the basis of unrounded promotion rates; thus, differences are due to rounding. We calculated the percentage difference for women relative to men as the unrounded percentage point difference divided by the unrounded promotion rate for men. For the adjusted analysis of State data, we conducted discrete-time duration analysis using a logit model that controlled for a variety of factors relevant to promotion, and we analyzed the time duration (number of years) to be promoted. The adjusted analysis does not completely explain the reasons for differences in odds of promotion. While various independent variables capture and control for many different characteristics across different demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

As table 5 shows, our descriptive analysis of the data for State’s Civil Service found that the average percentage of women promoted from GS-11 through GS-13 was lower than the average percentage of men. For example, our descriptive analysis found that in fiscal years 2002 through 2017, an average of 20.5 percent of women were promoted from GS-11 to GS-12, compared with an average of 23.2 percent of men.[77] This negative 2.7 percentage point difference indicates that the average rate of promotion from GS-11 to GS-12 was 11.6 percent lower for women than for men.[78] However, this descriptive analysis does not account for the variety of factors besides gender that may affect promotion rates, including the length of time it takes to be promoted.[79]

Our adjusted analysis of the State data, controlling for factors other than gender that could influence promotion, found no statistically significant differences in the rates or odds of promotion for women compared with men in the Civil Service. Specifically, the adjusted analysis for fiscal years 2002 to 2018 found the following:

  • The adjusted rates and odds of promotion from GS-11 to GS-12, from GS-12 to GS-13, from GS-13 to GS-14, and from GS-15 to executive were lower for women than for men.
  • Our estimates of the odds of promotion from GS-14 to GS-15 were higher for women than for men.
  • In all cases we did not find any statistically significant differences in the odds of promotion at any rank for women relative to men in the Civil Service. That is, we could not conclude that there was a statistical relationship between gender and promotion at these ranks.[80]

Figure 9 shows key results of our descriptive and adjusted analyses of State data for men and women in State’s Civil Service.

Figure 9: Relative Differences in Promotion Rates and Adjusted Promotion Odds for Women Compared with Men in the Department of State’s Civil Service, Fiscal Years 2002-2018

Notes: For each rank, the promotion rates based on the descriptive analyses for men and for women represent an average of the number of newly elevated men or women in the next-higher rank in the following year, divided by the number of men or women in the rank in the current year. Given this methodology, we were not able to calculate promotion rates for fiscal year 2018, because the State data we analyzed ended in fiscal year 2018. For the descriptive analysis, we calculated the percentage difference for women relative to men as the difference between the unrounded promotion rates for women and men divided by the unrounded promotion rate for men; thus, differences are due to rounding. For the adjusted analysis of State data, we conducted discrete-time duration analysis using a logit model that controlled for a variety of factors relevant to promotion, and we analyzed the time duration (number of years) to be promoted. The adjusted analysis does not completely explain the reasons for differences in odds of promotion. While various independent variables capture and control for many different characteristics across different demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

While we found no statistically significant difference in odds of promotion for all women in the Civil Service relative to men, we found generally lower odds of promotion for racial or ethnic minority women and men than for white men. See the text box for additional details.

Adjusted Analysis Found Racial or Ethnic Minority Women and Men in Civil Service Generally Had Lower Odds of Promotion Than White Men

In addition to examining differences in odds of promotion based on Department of State employees’ racial or ethnic minority status and gender, we adjusted for multiple comparisons by separately examining the intersection of these demographic characteristics. We conducted this analysis for the following four groups:

  • White men (2,907 employees in fiscal year 2018)
  • White women (2,559 employees in fiscal year 2018)
  • Racial or ethnic minority men (1,500 employees in fiscal year 2018)
  • Racial or ethnic minority women (2,576 employees in fiscal year 2018)

We used a discrete-time multivariate statistical logit model to control for the time duration (number of years) prior to promotion to each General Schedule (GS) rank from GS-11 to executive. After controlling for factors that could influence promotion,a our analysis found the following:

  • There was no statistically significant difference in the odds of promotion for White women relative to White men in the Civil Service.b Thus, we could not conclude that there was a statistical relationship between promotion and being a White woman.
  • Racial or ethnic minority men in the Civil Service had statistically significantly lower odds of promotion than White men from each rank from GS-11 through GS-15.
  • Racial or ethnic minority women in the Civil Service had statistically significantly lower odds of promotion than White men from each rank from GS-11 through GS-14.

Our analyses do not completely explain the reasons why differences may exist in the odds of promotion across demographic groups. Various unobservable factors could be present that may account for differences in odds of promotion. Thus, our analyses do not establish a causal relationship between demographic characteristics and promotion outcomes.

Source: GAO analysis of Department of State data. | GAO-20-237.

aWe controlled for employees’ time in each rank prior to promotion; racial or ethnic minority status; years of federal experience; age when entering State; veteran’s status; taking long-term leave; graduating from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland; changing between the Foreign and Civil Services; occupation; and fiscal years.

bWe express our confidence in the precision of our estimates as statistically significant differences, which refers to the likelihood of an observed difference being due to chance. We consider differences in our estimates to be statistically significant if they were statistically significant at the 95 percent level. In contrast, “practical significance” refers to the magnitude of an observed difference.

Descriptive and Adjusted Promotion Rates and Adjusted Promotion Odds in Foreign Service Were Generally Higher for Women Than Men in Early to Midcareer

Our descriptive and adjusted analyses of data for fiscal years 2002 through 2018 for State’s Foreign Service both found that the rate[81] and odds of promotion were generally higher for women than for men, as table 6 shows.[82]

Table 6: Promotion Outcomes for Men and Women in the Department of State’s Foreign Service, Fiscal Years 2002-2018
Class 4 to Class 3Class 3 to Class 2Class 2 to Class 1Class 1 to executive
Descriptive analysis
Promotion rate for men, %17.113.08.97.7
Promotion rate for women, %16.316.110.27.7
Percentage point difference between promotion rate for women and promotion rate for men-0.83.11.30.0
Percentage difference between promotion rate for women and promotion rate for men, %-4.724.314.7-0.3
Adjusted analysis

Promotion rate for men, %

95 percent confidence interval, %

16.5

[16.2, 16.9]

13.5

[13.2, 13.9]

9.1

[8.8, 9.5]

7.5

[7.1, 8.0]

Promotion rate for women, %

95 percent confidence interval, %

17.6

[17.0, 18.1]

14.7

[14.1, 15.3]

9.7

[9.1, 10.2]

8.0

[7.4, 8.7]

Percentage point difference between promotion rate for women and promotion rate for men1.01.20.60.5

Odds ratio for promotion for women relative to men

95 percent confidence interval, %

1.094**

[1.031, 1.162]

1.127**

[1.051, 1.209]

1.075

[0.989, 1.169]

1.079

[0.958, 1.215]

Percentage difference between promotion odds for women and promotion odds for men, %

95 percent confidence interval, %

9.4**

[3.1, 16.2]

12.7**

[5.1, 20.9]

7.5

[-1.1, 16.9]

7.9

[-4.2, 21.5]

Legend:** = statistical significance at p-value < 0.01; * = statistical significance at p-value < 0.05.
Source: GAO analysis of State data. | GAO-20-237

Notes: The p-value represents the smallest level of significance for which our estimate results in a rejection of the hypothesis of there being no difference in the odds of promotion. For each rank, the promotion rates based on the descriptive analyses for men and for women represent an average of the number of newly elevated men or women in the next-higher rank in the following year, divided by the number of men or women in the given rank in the current year. Given this methodology, we were not able to calculate promotion rates for fiscal year 2018, because the State data we analyzed ended in fiscal year 2018. For the descriptive analysis, we calculated the percentage point difference and percentage difference for women relative to men on the basis of unrounded promotion rates; thus, differences are due to rounding. We calculated the percentage difference for women relative to men as the unrounded percentage point difference divided by the unrounded promotion rate for men. For the adjusted analysis of State data, we conducted discrete-time duration analysis using a logit model that controlled for a variety of factors relevant to promotion, and we analyzed the time duration (number of years) to be promoted. The adjusted analysis does not completely explain the reasons for differences in odds of promotion. While various independent variables capture and control for many different characteristics across different demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

As table 6 shows, our descriptive analysis of the data for State’s Foreign Service found that the average percentage of women promoted from Classes 3 and 2 was higher than the average percentage of men.[83] For example, our descriptive analysis found that in fiscal years 2002 through 2017, an average of 16.1 percent of women were promoted from Class 3 to Class 2, compared with an average of 13.0 percent of men.[84] This 3.1 percentage point difference indicates that the average rate of promotion from Class 3 to Class 2 was 24.3 percent higher for women than for men.[85] However, this descriptive analysis does not account for the variety of factors besides gender that may affect promotion rates, including the length of time it takes to be promoted.[86]

Our adjusted analysis of the data for State’s Foreign Service, controlling for factors other than gender that could influence promotion, found that women in the Foreign Service had higher adjusted rates of promotion and higher odds of promotion than men in early to midcareer.[87] Specifically, our adjusted analysis of the data for fiscal years 2002 through 2018 found the following:

  • On average, the adjusted rate of promotion from Class 4 to Class 3 for women in the Foreign Service was 17.6 percent, compared with 16.5 percent of men. This statistically significant difference indicates that the odds of promotion from Class 4 to Class 3 were 9.4 percent higher for women than for men.
  • Our estimates for the adjusted promotion rates and odds of promotion from Class 3 to Class 2 were also statistically significantly higher for women than for men.[88]
  • While the adjusted rates of promotion from Class 2 to Class 1 and Class 1 to executive were higher for women than for men, there was no statistically significant difference in the odds of promotion at these ranks for women relative to men in the Foreign Service. That is, we could not conclude that there was a statistical relationship between gender and promotion at these ranks.
  • Compared with the descriptive analysis, our adjusted analysis found a smaller percentage difference in promotion outcomes from Class 3 to Class 2 for women relative to men. Our adjusted analysis also found positive, rather than negative, percentage difference in promotion outcomes from Class 4 to Class 3 for women relative to men.

Figure 10 displays key results of our descriptive analysis and adjusted analysis of State data.

Figure 10: Relative Differences in Promotion Rates and Adjusted Promotion Odds for Women Compared with Men in the Department of State’s Foreign Service, Fiscal Years 2002-2018

Notes: For each rank, the promotion rates based on the descriptive analyses for men and for women represent an average of the number of newly elevated men or women in the next-higher rank in the following year, divided by the number of men or women in the rank in the current year. Given this methodology, we were not able to calculate promotion rates for fiscal year 2018, because the State data we analyzed ended in fiscal year 2018. For the descriptive analysis, we calculated the percentage difference for women relative to men as the difference between the unrounded promotion rates for women and men divided by the unrounded promotion rate for men; thus, differences are due to rounding. For the adjusted analysis of State data, we conducted discrete-time duration analysis using a logit model that controlled for a variety of factors relevant to promotion, and we analyzed the time duration (number of years) to be promoted. The adjusted analysis does not completely explain the reasons for differences in odds of promotion. While various independent variables capture and control for many different characteristics across different demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

In addition, the higher odds of promotion from Class 4 to Class 3 and from Class 3 to Class 2 for women in the Foreign Service reflect higher odds of promotion for white women than for white men. See the text box for additional details.

Adjusted Analysis Found White Women in Foreign Service Had Higher Odds of Promotion Than White Men in Early to Midcareer

In addition to examining differences in odds of promotion related to Department of State employees’ racial or ethnic minority status and gender, we adjusted for multiple comparisons by separately examining the intersection of these demographic characteristics. We conducted this analysis for the following four groups:

  • White men (6,611 employees in fiscal year 2018)
  • White women (3,368 employees in fiscal year 2018)
  • Racial or ethnic minority men (1,949 employees in fiscal year 2018)
  • Racial or ethnic minority women (1,320 employees in fiscal year 2018)

We used a discrete-time multivariate statistical logit model to control for the time duration (number of years) prior to promotion to each rank from Class 4 to executive. After controlling for factors that could influence promotion,a our analysis found the following:

  • White women in the Foreign Service had statistically significantly higher odds of promotion from Class 4 to Class 3 and from Class 3 to Class 2 than white men.b
  • Racial or ethnic minority men in the Foreign Service had statistically significantly lower odds of promotion from Class 4 to Class 3 than white men.
  • There was no statistically significant difference in the odds of promotion for racial or ethnic minority women relative to white men in the Foreign Service. That is, we could not conclude that there was a statistical relationship between being a racial or ethnic minority woman and promotion.

Our analyses do not completely explain the reasons for differences in the odds of promotion across demographic groups. Various unobservable factors could be present that may account for differences in odds of promotion; thus, our analyses do not establish a causal relationship between demographic characteristics and promotion outcomes.

Source: GAO analysis of Department of State data. | GAO-20-237.

aWe controlled for employees’ time in each rank prior to promotion; racial or ethnic minority status; years of federal experience; age when entering State; veteran’s status; taking long-term leave; graduating from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland; changing between the Foreign and Civil Services; occupation; having a hardship assignment in the prior year; having an overseas post in the prior year; proficiency in a hard language; and fiscal years.

bWe express our confidence in the precision of our estimates as statistically significant differences, which refers to the likelihood of an observed difference being due to chance. We consider differences in our estimates to be statistically significant if they were statistically significant at the 95 percent level. In contrast, “practical significance” refers to the magnitude of an observed difference.

State Has Identified Some Diversity Issues but Should Consider Other Issues That Could Indicate Potential Barriers

While State has identified some diversity issues in its reports to EEOC, State’s internal workforce analyses, its employee groups, and our analyses have identified additional issues that could indicate potential barriers on which State has not reported. EEOC’s MD-715 calls for federal agencies to regularly evaluate their employment practices to identify barriers to equal opportunity in the workplace, take measures to eliminate identified barriers, and report annually on these efforts to EEOC.

In fiscal years 2009 through 2018, State’s annual MD-715 reports identified and analyzed a total of 11 diversity issues related to participation of racial or ethnic minorities and women: (1) underrepresentation of Asian Americans in the senior ranks (reported 1 year); (2) underrepresentation of women in the senior ranks (reported 2 years); (3) underrepresentation of African Americans in the senior ranks (reported 2 years); (4) underrepresentation of Native American/Pacific Islander/Alaskan Natives (reported 2 years); (5) underrepresentation of women in the Foreign Service (reported 3 years); (6) underrepresentation of African Americans in the Foreign Service (reported 3 years); (7) underrepresentation of minorities in the senior ranks (reported 4 years); (8) underrepresentation of Hispanics (reported 6 years); (9) underrepresentation of individuals with disabilities (reported 8 years); (10) higher attrition of women in a particular bureau (reported 1 year); and (11) higher attrition of minorities in a particular bureau (reported 1 year), as seen in table 7.[89]

Table 7: Numbers and Types of Diversity Issues Identified by the Department of State, Fiscal Years 2009-2018
Fiscal yearTotal
2009201020112012201320142015201620172018
Underrepresentation of Asian Americans in the senior ranks1
Underrepresentation of Women in the senior ranksa2
Underrepresentation of African Americans in the senior ranks2
Underrepresentation of Native American/Pacific Islander/Alaskan Natives2
Underrepresentation of women in Foreign Service3
Underrepresentation of African Americans in the Foreign Service3
Underrepresentation of minorities in the senior ranks4
Underrepresentation of Hispanics6
Underrepresentation of individuals with disabilities8
Higher attrition of women in a particular bureau1
Higher attrition of minorities in a particular bureau1
Total555623221233
Legend: ✓ = identified in Management Directive 715 (MD-715) report for the fiscal year, — = not identified in MD-715 report for the fiscal year.
Source: GAO analysis of Department of State MD-715 reports for fiscal years 2009 through 2019. | GAO-20-237

aIn 2009, State’s MD-715 report specifically cited underrepresentation of African American and Hispanic females in the senior ranks.

However, State’s workforce analysis, State employee groups, and our analysis have identified additional diversity issues, which State should consider when evaluating employment practices to identify barriers to equal opportunity in the workplace.

State’s analysis. Analysis of State workforce data conducted by State’s Bureau of Human Resources has revealed diversity issues that could indicate potential barriers. For example, in May 2018, the bureau produced a demographic trend analysis of full-time permanent employees that showed some growth in minority representation, including an overall increase in racial or ethnic minorities in both services. However, the analysis also showed that the proportion of African Americans in the Civil Service had declined from 33 percent in 2000 to 28 percent in 2017 and that the proportion of women in the agency overall had declined slightly, from 45 percent to 44 percent.

State employee groups. During our structured interviews with 11 employee groups, representatives of the groups discussed a variety of issues related to diversity at State.[90] Examples include the following:

  • Employee group representatives expressed concern about representation of minorities in the higher ranks of both the Civil and Foreign Services. For example, representatives told us that for some minority groups, it is difficult to be promoted above the GS-13 level.
  • Employee group representatives voiced perceptions that it takes longer for women and racial or ethnic minorities to be promoted. For example, representatives of one group told us that it takes longer for employees with diverse backgrounds to reach GS-13 in the Civil Service and Class 2 in the Foreign Service and that very few of these employees are promoted beyond those levels.

Our analysis. Our analysis identified additional diversity issues that may indicate potential barriers, such as persistently lower representation of minorities in the higher ranks of the Civil and Foreign Services. Additionally, while the adjusted percentage difference in odds of promotion to the executive rank was 53 percent higher for African Americans than for whites between fiscal years 2002 and 2018, this tendency disappeared when we analyzed the odds of promotion for African Americans to the executive rank for promotions after fiscal year 2011.[91] In addition, our analysis showed discrepancies in promotion outcomes for racial or ethnic minorities in early and midcareer career relative to whites.

State has reported on some of the issues it has identified, but its workforce data, our interviews with employee groups, and our analysis indicate that there are other issues that State should consider to gain a full understanding of the diversity issues, such as discrepancies in early and midcareer promotion outcomes, that could indicate potential barriers in its workforce. Until State takes steps to explore these issues, State could be missing opportunities to investigate, identify, and remove barriers that impede members of some demographic groups from realizing their full potential.

Conclusions 


According to State, to represent the United States to the world, the agency must have a workforce that reflects the rich composition of its citizenry. State has implemented several plans, activities, and initiatives to improve diversity and representation throughout the ranks of its workforce. However, longstanding diversity issues persist at the agency, such as underrepresentation of racial or ethnic minorities and women in the senior ranks.

EEOC MD-715 states that equality of opportunity is essential to attracting, developing and retaining the most qualified workforce to support the agency’s achievement of its strategic mission. To achieve this goal, EEOC calls for each federal agency to, among other things, identify and eliminate barriers that impair the ability of individuals to compete in the workplace because of race, national origin, sex or disability. While State has identified some diversity issues, additional issues may exist that it has not highlighted in its MD-715 reports. For example, our analysis showed that women in the Civil Service and racial or ethnic minorities in both services were less likely to be promoted through the midcareer ranks, yet State’s MD-715 reports have focused only on the underrepresentation of women, including racial or ethnic minority women, in the senior ranks. As a result, State may have an incomplete picture of issues affecting diversity in its workforce. Taking additional steps to identify diversity issues could help State properly direct its resources to investigate, identify, and remove barriers to a diverse workforce.

Agency Comments 


We provided a draft of this product to State, EEOC, and OPM for comment. In its comments, reproduced in appendix XIV, State concurred with the recommendation and stated that the agency will continue to work on initiatives to recruit, retain, develop, and empower a diverse, capable workforce. EEOC and OPM did not provide comments.

As agreed with your offices, unless you publicly announce the contents of this report earlier, we plan no further distribution until 30 days from the report date. At that time, we will send copies to We are sending copies of this report to the appropriate congressional committees, the Secretary of State, the Chair of the EEOC, and the Director of OPM. In addition, the report will be available at no charge on the GAO website at https://gao.gov.

If you or your staff have any questions about this report, please contact me at (202) 512-6881 or bairj@gao.gov. Contact points for our Offices of Congressional Relations and Public Affairs may be found on the last page of this report. GAO staff who made significant contributions to this report are listed in appendix XV.

Jason Bair

Director, International Affairs and Trade

Congressional Addressees

The Honorable Robert Menendez
Ranking Member
Committee on Foreign Relations
United States Senate

The Honorable Gary C. Peters
Ranking Member
Committee on Homeland Security and Governmental Affairs
United States Senate

The Honorable Eliot L. Engel
Chairman
Committee on Foreign Affairs
House of Representatives

The Honorable Joaquin Castro
Chairman
Subcommittee on Oversight and Investigations
House of Representatives

The Honorable Cory A. Booker
United States Senate

The Honorable Benjamin L. Cardin
United States Senate

The Honorable Christopher Coons
United States Senate

The Honorable Marco Rubio
United States Senate

The Honorable Brian Schatz
United States Senate

The Honorable Jeanne Shaheen
United States Senate

The Honorable Karen Bass
House of Representatives

The Honorable Ami Bera, M.D.
House of Representatives

Appendixes 


Appendix I: Objectives, Scope, and Methodology

Our objectives were to examine (1) the demographic composition of the Department of State's (State) workforce in fiscal years 2002 through 2018, (2) any differences in promotion outcomes for various demographic groups in State’s workforce, and (3) the extent to which State has identified any barriers to diversity in its workforce.

Data

To examine the demographic composition of State’s workforce and any differences in promotion outcomes for various demographic groups, we analyzed State’s personnel data from its Global Employment Management System database for its full-time, permanent, career workforce for fiscal years 2002 through 2018. Our main data request, which we tailored on the basis of conversations with agency officials and our own consideration of the availability of data, included two types of data and covered 42,473 unique employees.

For each fiscal year, we analyzed record-level status data that reflected State’s employees as of September 30 (the end of the fiscal year). We requested and received yearly snapshots with record-level data on all full-time, permanent, career State employees in fiscal years 2002 through 2018. Specifically, we requested demographic and administrative data, including race, ethnicity, gender, grade or class, age or date of birth, date of entry to State, years of service, veteran’s status, occupation, location or duty station, and each employee’s unique identifier. We additionally analyzed record-level dynamic data that included personnel actions, such as promotions or separations, and each employee’s unique identifier.

In addition to requesting the personnel data, we requested other data from State and obtained publicly available data. We merged these additional data with State’s personnel data. We also requested State leave data as well as education, language, and historical assignment data. We obtained “Post (Hardship) Differential Percentage of Basic Compensation” data from State’s website for fiscal years 2002 through 2018. Following guidance from the U.S. Equal Employment Opportunity Commission, we used data on nine federal job categories and their correspondence to specific occupation codes to match federal job categories to the occupations of State’s employees.

We analyzed the State data at our audit site at State’s headquarters in Washington, D.C.; we did not hold or analyze any record-level data at our headquarters. We assessed the reliability of all data sets and of the data elements that were critical to our analyses and determined that they were sufficiently reliable for our analyses. Specifically, we reviewed documentation on the general design and structure of the data sets, interviewed State officials who were knowledgeable about the data, and completed our own electronic testing to assess the accuracy and completeness of the data used in our analyses.

Demographic Composition

To examine the demographic composition of State's workforce in fiscal years 2002 through 2018, we analyzed State data to determine summary statistics on State’s full-time, permanent, career workforce.[92] For State overall, the Civil Service, and the Foreign Service, we analyzed the numbers and percentages of racial or ethnic minorities, in total and by gender, and of women for each year from fiscal year 2002 through fiscal year 2018. In addition, we analyzed these numbers and percentages by occupation and rank, including General Service (GS) grade for the Civil Service, salary class for the Foreign Service, and executive rank (i.e., Senior Executive Service or Senior Foreign Service).[93]

For the purpose of this report, racial or ethnic minority status corresponds to instances where the racial or ethnic group is neither non-Hispanic white nor unspecified. The Hispanic group included Hispanics of all races. The remaining non-Hispanic racial or ethnic groups included white, African American, Asian, other, and unspecified. Our analysis for the category we report as “other” included non-Hispanic members of the American Indian or Alaskan Native, Native Hawaiian or other Pacific Islander, and two or more races. For instances where an employee’s reported racial or ethnic category changed, we assigned the most recent value to all available years.[94]

In addition, we compared the demographics of State’s workforce in fiscal year 2018 with (1) demographics of the federal workforce[95] in fiscal year 2016 as reported by the Office of Personnel Management (OPM) and (2) demographics of the relevant civilian labor force[96] in 2006 through 2010 from the Census Bureau’s Equal Employment Opportunity tabulation.[97]

Promotion Analyses

We considered promotion to be an increase in rank between fiscal years.[98] We include in our analyses all individuals in the original rank and do not distinguish between individuals who did or did not apply or between those who were eligible or ineligible for promotion.[99] To examine promotion for various demographic groups in State’s workforce, we analyzed State personnel data, and we conducted two types of analyses.

Descriptive Analysis

We compared the annual promotion rates for racial or ethnic minorities with the annual promotion rates for whites and compared the promotion rates for women with the promotion rates for men. For each fiscal year and rank, we calculated these rates as the number of newly elevated employees in the next-higher rank in the following fiscal year divided by the number of employees in the given rank in the current year.

Adjusted Analysis

We examined adjusted promotion rates, odds ratios, and the percentage difference[100] in relative odds of promotion, from our adjusted analysis of State data.[101] We conducted adjusted analyses using a multivariate statistical method, (specifically duration analysis), which accounted for certain individual and occupational factors other than racial or ethnic minority status and gender that could influence promotion, including the length of time it takes to be promoted.[102] Specifically, we used a discrete-time multivariate statistical logit model for each rank to analyze the number of yearly cycles it took for promotion up to the executive level from Civil Service rank GS-11 and from Foreign Service rank Class 4. That is, we conducted adjusted analysis from each rank, from GS-11 through GS-15 up to executive for the Civil Service and from Class 4 through Class 1 up to executive for the Foreign Service.[103]

A variety of factors besides racial or ethnic minority status and gender may affect promotion outcomes; therefore, in our regression models, we controlled for factors that are likely to be associated with promotion, which helped us to examine the statistical relationship between racial or ethnic minority status and gender and promotion.[104] These factors can be time consistent (e.g., racial or ethnic minority status, gender) or time varying (e.g., having long-term leave in the prior year, having worked overseas in the prior year). We conducted adjusted analyses in which we incorporated the time-consistent and time-varying factors separately and together. In addition to incorporating racial or ethnic minority status and gender in the regressions, we incorporated various employee- and position-specific characteristics, such as an employee’s (1) time in each rank before promotion; (2) years of prior federal government experience; (3) age at the time of entering State; (4) receipt of veterans’ preference points; (5) having long-term leave in the previous year (i.e., having taken more than 2 weeks of consecutive leave more than twice in the previous year); (6) having graduated from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland;[105] (7) having transferred between the Civil and Foreign Services; (8) having worked overseas in the previous year (for the Foreign Service); (9) having worked in at a location where the hardship differential was 20 percent or more (Foreign Service only) in the previous year;[106] (10) proficiency in a hard language (Foreign Service only);[107] (11) occupation;[108] and (12) fiscal years.[109] For Civil Service, we also clustered the standard errors on a code for the organizational structure (i.e., division of State smaller than the bureau level) of State.[110] We identified these attributes as being relevant to promotion by reviewing relevant literature and interviewing agency officials.

Our primary model is a pooled model that includes all employees who we used to determine summary statistics on the department’s full-time, permanent, career workforce for fiscal years 2002 through 2018.[111] In addition, we conducted adjusted analysis before and after fiscal year 2011, when “Executive Order 13583—Establishing a Coordinated Government-wide Initiative to Promote Diversity and Inclusion in the Federal Workforce” was signed.

We also reviewed other potential methodologies, and after taking into consideration the strengths and limitations of these other methodologies, we relied on the multivariate statistical method to examine how, if at all, promotion outcomes differed across demographic groups in State’s workforce. In addition, we received feedback on our methodology from three academic experts with relevant expertise, the Office of Personnel Management, the U.S. Equal Employment Opportunity Commission, and the Department of State.[112] The experts and agencies reviewed our methodology, assessed its strengths and limitations, and provided comments. We incorporated their comments as appropriate and disclosed additional limitations as necessary. In addition, we conducted a number of sensitivity analyses, such as examining the robustness of our models to the inclusion of different sets of control variables (see app. XI), applying the multivariate statistical method for different permutations of racial or ethnic minority status (see app. XII), and applying the multivariate statistical method only for individuals who entered their rank at State in 2003 or later.

Limitations and Other Considerations

Our estimates of the differences in promotion rates resulting from our descriptive analyses of State data for racial or ethnic minorities relative to whites and for women relative to men, respectively, may be limited by several factors and should be interpreted with caution. This analysis does not account for any factors besides racial or ethnic minority status and gender that may affect promotion rates. Hypothetically, if an employee’s occupation relates to promotion opportunities, and if racial or ethnic minorities happen to be in occupations that have more limited promotion opportunities, examining promotion rates without accounting for occupation may suggest that promotion rates for racial or ethnic minorities are lower than promotion rates for whites. Additionally, possible variability in promotion rates across years makes it challenging to examine promotion patterns.[113]

Our estimates from the adjusted analysis of State data may be limited by the following factors and should be interpreted with caution.

  • Unobservable factors. Our adjusted analyses took into account a variety of factors that may help explain some of the differences in odds of promotion, such as characteristics of the individual employees (e.g., employees' time in each rank before promotion), occupation, and fiscal years. However, we may not have taken into account all possible factors, including various unobservable factors that may cause differences in odds of promotion. For example, some unobservable factors that our analyses may not have captured include employees’ skills, motivation, performance, or abilities.[114] The effects of these unobservable factors could decrease or increase our estimates of odds of promotion. Because our analyses do not completely explain the reasons for differences in promotion outcomes, which may result from various unobservable factors, our analyses do not establish a causal relationship between demographic characteristics and promotion outcomes.[115]
  • Occupation segmentation. We controlled for occupation to help estimate the statistical relationship between promotion outcomes and racial or ethnic minority status and gender that exists beyond any statistical relationship between occupation and promotion outcomes. In other words, by controlling for occupation, we accounted for whether certain occupations have more limited promotion potential. However, controlling for occupation may have prevented us from considering any differences in promotion outcomes due to systematic differences in occupation distribution or segmentation across different racial or ethnic groups and gender. If racial or ethnic minorities or women tend to be segmented in occupations that have relatively limited promotion potential, we might have observed lower odds of promotion for those groups compared with whites or men, respectively, if we had not controlled for occupation. See appendix XI for the results of a model that controlled for other characteristics of the individual employees relevant to promotion (model 3) but did not control for occupation.
  • Differences in attrition. While our adjusted analysis accounted for several factors that may be related to an employee’s prospects for promotion, there may still be some residual differences in promotion prospects for employees who left State relative to those who stayed. However, because we controlled for a variety of factors that may affect the odds of promotion, any residual differences between employees who left and those who stayed would be unrelated to these factors. In particular, we controlled for racial or ethnic minority status and gender, so residual differences between employees who left and those who stayed would be unrelated to these characteristics. Behavioral motivations and outcomes related to attrition may influence racial or ethnic minorities and women differently than whites and men, respectively. The potential existence of differential trends related to attrition could be one explanation for differences in odds of promotion.
  • Types of promotion. By controlling for occupation, we controlled for situations where some occupations may be more likely to have career-ladder (i.e., noncompetitive) than competitive promotions.[116] In addition, by analyzing promotions separately by rank level while controlling for occupation, we controlled for situations where the promotion structure may have changed from noncompetitive to competitive. However, our estimates do not explicitly differentiate between noncompetitive and competitive promotions. Promotions within an employees' career ladder tend to be more likely than competitive promotions, and we are not accounting for this difference. The effect of the promotion types could decrease or increase our estimates of odds of promotion.
  • Promotion applicants and eligibility. We accounted for the time all employees spent in each rank before promotion. However, we did not account for whether the employee was actively applying for promotion or was eligible for promotion. Thus, our estimates are based on the individuals in the original rank, not on the applicants for promotion or on those eligible for promotion. In addition, data regarding employees who applied for promotion were not available. Employees’ eligibility for promotion may differ across the Civil and Foreign Services, occupations, and job series. These nuances make it impractical to distinguish eligibility for each employee on the basis of the available data. The effect of applicant and eligibility status could decrease or increase our estimates of odds of promotion.
  • Budget constraints. The specific number of promotion slots available each year may vary by annual budget constraints. We controlled for some aspects of this budget constraint by including control variables for each fiscal year, which would be relevant if promotion opportunities were affected by budget constraints that varied across fiscal years. However, because of data availability constraints, our estimates may not capture the specific number of promotion slots available each year. In addition, our estimates may not capture the extent to which fiscal year budget constraints affect promotion opportunities differently across occupations or bureaus. The effect of these budget constraints could decrease or increase our estimates of odds of promotion.

State’s Identification of Diversity Issues

To examine the extent to which State has identified barriers to diversity in its workforce, we reviewed State’s Management Directive 715 reports to the U.S. Equal Employment Opportunity Commission for fiscal years 2009 through 2018. We also reviewed State’s workforce analyses, such as a workforce composition analysis developed by State’s Bureau of Human Resources in May 2018. We met with officials from State’s Office of Civil Rights and Bureau of Human Resources. In addition, we conducted structured interviews with representatives from 11 of 13 employee groups representing current employees in the Civil and Foreign Services.[117] In these interviews, we asked a set of structured interview questions, which we had developed with a methodologist, to identify employees’ perspectives about diversity at the agency and their perceptions of State’s diversity efforts. We met with the following affinity groups: Arab-Americans in Foreign Affairs Agencies, the Asian American Foreign Affairs Association, the Carl T. Rowan Chapter of Blacks in Government, the Council for Career Entry Professionals, the Disability Action Group, Executive Women at State, Gays and Lesbians in Foreign Affairs Agencies, the Hispanic Employees Council of Foreign Affairs Agencies, the South Asian–American Employee Association, the Thursday Luncheon Group, and Veterans at State.[118]

We conducted this performance audit from April 2018 to January 2020 in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives.


Appendix II: Analysis of Department of State Workforce Data, Fiscal Years 2002-2018

The following figures and tables present numbers and proportions of employees in racial, ethnic, and gender groups in the Department of State (State) overall and in State’s Civil and Foreign Services in fiscal years 2002 through 2018.

Figure 11: Percentages of White Employees and Racial or Ethnic Minority Employees in the Department of State, Fiscal Years 2002-2018

Notes: The data shown reflect percentages of white and racial or ethnic employees at the end of each fiscal year. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates that race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.
Table 8: Numbers and Percentages of White Employees and Racial or Ethnic Minority Employees in the Department of State, Fiscal Years (FY) 2002-2018
WhiteAfrican AmericanHispanicAsianOtherUnspecifiedTotal
FY 2002Number11,6352,89879965230927716,570
Percentage70175422100
FY 2003Number12,2543,01389074433818817,427
Percentage70175421100
FY 2004Number12,6393,03792582137810817,908
Percentage71175521100
FY 2005Number12,9823,0739728634115318,354
Percentage71175520100
FY 2006Number13,1783,0841,027909433118,632
Percentage71176520100
FY 2007Number13,4873,1501,1119634661619,193
Percentage70166520100
FY 2008Number13,8533,2071,1881,0085112619,793
Percentage70166530100
FY 2009Number14,3893,2791,2481,0845662820,594
Percentage70166530100
FY 2010Number15,1493,3661,3231,1876523521,712
Percentage70166530100
FY 2011Number15,7443,4561,3971,2977292822,651
Percentage70156630100
FY 2012Number15,9033,4761,4351,365801622,986
Percentage69156630100
FY 2013Number16,1693,5241,4891,374876623,438
Percentage69156640100
FY 2014Number16,2033,5091,5501,400935623,603
Percentage69157640100
FY 2015Number16,1043,5101,5841,396965523,564
Percentage68157640100
FY 2016Number16,1973,5321,6221,4331,011823,803
Percentage68157640100
FY 2017Number15,9143,4631,6401,4171,004623,444
Percentage68157640100
FY 2018Number15,4453,3221,6321,4069851622,806
Percentage68157640100
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: The data shown reflect numbers and percentages of white and racial or ethnic employees at the end of each fiscal year. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates that race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.

Figure 12: Percentages of White Employees and Racial or Ethnic Minority Employees in the Department of State’s Civil Service, Fiscal Years 2002-2018

Notes: The data shown reflect percentages of white and racial or ethnic employees at the end of each fiscal year. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates that race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.
Table 9: Numbers and Percentages of White Employees and Racial or Ethnic Minority Employees in the Department of State’s Civil Service, Fiscal Years (FY) 2002-2018
WhiteAfrican AmericanHispanicAsianOtherUnspecifiedTotal
FY 2002Number3,7002,337278265161906,831
Percentage54344421100
FY 2003Number3,9132,415307302166557,158
Percentage55344421100
FY 2004Number4,0032,407313319179317,252
Percentage55334420100
FY 2005Number4,1182,416332335190157,406
Percentage56334530100
FY 2006Number4,2482,42236534920017,585
Percentage56325530100
FY 2007Number4,5302,479431391224158,070
Percentage56315530100
FY 2008Number4,7732,533492416251258,490
Percentage56306530100
FY 2009Number4,8672,560515413270268,651
Percentage56306530100
FY 2010Number5,0992,580544458286339,000
Percentage57296530100
FY 2011Number5,4302,640568500323279,488
Percentage57286530100
FY 2012Number5,5652,62658052335149,649
Percentage58276540100
FY 2013Number5,8132,665600532398410,012
Percentage58276540100
FY 2014Number5,7472,63362954541649,974
Percentage58266540100
FY 2015Number5,7722,631660543423310,032
Percentage58267540100
FY 2016Number5,9102,657674579452710,279
Percentage58267640100
FY 2017Number5,7592,577666574441510,022
Percentage57267640100
FY 2018Number5,4662,44664556641949,546
Percentage57267640100
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: The data shown reflect numbers and percentages of white and racial or ethnic employees at the end of each fiscal year. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates that race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.

Figure 13: Percentages of White Employees and Racial or Ethnic Minority Employees in the Department of State’s Foreign Service, Fiscal Years 2002-2018

Notes: The data shown reflect percentages of white and racial or ethnic employees at the end of each fiscal year. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates that race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.
Table 10: Numbers and Percentages of White Employees and Racial or Ethnic Minority Employees in the Department of State’s Foreign Service, Fiscal Years (FY) 2002-2018
WhiteAfrican AmericanHispanicAsianOtherUnspecifiedTotal
FY 2002Number7,9355615213871481879,739
Percentage8165422100
FY 2003Number8,34159858344217213310,269
Percentage8166421100
FY 2004Number8,6366306125021997710,656
Percentage8166521100
FY 2005Number8,8646576405282213810,948
Percentage8166520100
FY 2006Number8,930662662560233011,047
Percentage8166520100
FY 2007Number8,957671680572242111,123
Percentage8166520100
FY 2008Number9,080674696592260111,303
Percentage8066520100
FY 2009Number9,522719733671296211,943
Percentage8066620100
FY 2010Number10,050786779729366212,712
Percentage7966630100
FY 2011Number10,314816829797406113,163
Percentage7866630100
FY 2012Number10,338850855842450213,337
Percentage7866630100
FY 2013Number10,356859889842478213,426
Percentage7767640100
FY 2014Number10,456876921855519213,629
Percentage7767640100
FY 2015Number10,332879924853542213,532
Percentage7677640100
FY 2016Number10,287875948854559113,524
Percentage7667640100
FY 2017Number10,155886974843563113,422
Percentage7677640100
FY 2018Number9,9798769878405661213,260
Percentage7577640100
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: The data shown reflect numbers and percentages of white and racial or ethnic employees at the end of each fiscal year. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates that race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.

Table 11: Numbers and Percentages of Men and Women in the Department of State, Fiscal Years (FY) 2002-2018
MenWomenTotal
FY 2002Number9,2317,33916,570
Percentage5644100
FY 2003Number9,7577,67017,427
Percentage5644100
FY 2004Number10,0487,86017,908
Percentage5644100

FY 2005

Number10,2998,05518,354
Percentage5644100
FY 2006Number10,4158,21718,632
Percentage5644100
FY 2007Number10,6528,54119,193
Percentage5645100
FY 2008Number10,9938,80019,793
Percentage5644100
FY 2009Number11,4929,10220,594
Percentage5644100
FY 2010Number12,1339,57921,712
Percentage5644100
FY 2011Number12,7359,91622,651
Percentage5644100
FY 2012Number12,93110,05522,986
Percentage5644100
FY 2013Number13,23310,20523,438
Percentage5644100
FY 2014Number13,37510,22823,603
Percentage5743100
FY 2015Number13,40310,16123,564
Percentage5743100
FY 2016Number13,51610,28723,803
Percentage5743100
FY 2017Number13,31610,12823,444
Percentage5743100
FY 2018Number12,9759,83122,806
Percentage5743100
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: The data shown reflect numbers and percentages of men and women at the end of each fiscal year. If an employee’s recorded gender changed during the period shown, we assigned the most recently recorded value to all years for that employee.

Table 12: Numbers and Percentages of Men and Women in the Department of State’s Civil Service, Fiscal Years (FY) 2002-2018
MenWomenTotal
FY 2002Number2,6924,1396,831
Percentage3961100
FY 2003Number2,8704,2887,158
Percentage4060100
FY 2004Number2,9494,3037,252
Percentage4159100
FY 2005Number3,0564,3507,406
Percentage4159100
FY 2006Number3,1434,4427,585
Percentage4159100
FY 2007Number3,3584,7128,070
Percentage4258100
FY 2008Number3,5684,9228,490
Percentage4258100
FY 2009Number3,6744,9778,651
Percentage4258100
FY 2010Number3,8525,1489,000
Percentage4357100
FY 2011Number4,1345,3549,488
Percentage4456100
FY 2012Number4,2515,3989,649
Percentage4456100
FY 2013Number4,4805,53210,012
Percentage4555100
FY 2014Number4,4815,4939,974
Percentage4555100
FY 2015Number4,5655,46710,032
Percentage4655100
FY 2016Number4,7215,55810,279
Percentage4654100
FY 2017Number4,6115,41110,022
Percentage4654100
FY 2018Number4,4095,1379,546
Percentage4654100
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: The data shown reflect numbers and percentages of men and women at the end of each fiscal year. If an employee’s recorded gender changed during the period shown, we assigned the most recently recorded value to all years for that employee.

Table 13: Numbers and Percentages of Men and Women in the Department of State’s Foreign Service, Fiscal Years (FY) 2002-2018
Men WomenTotal
FY 2002Number6,5393,2009,739
Percentage6733100
FY 2003Number6,8873,38210,269
Percentage6733100
FY 2004Number7,0993,55710,656
Percentage6733100
FY 2005Number7,2433,70510,948
Percentage6634100
FY 2006Number7,2723,77511,047
Percentage6634100
FY 2007Number7,2943,82911,123
Percentage6634100
FY 2008Number7,4253,87811,303
Percentage6634100
FY 2009Number7,8184,12511,943
Percentage6535100
FY 2010Number8,2814,43112,712
Percentage6535100
FY 2011Number8,6014,56213,163
Percentage6535100
FY 2012Number8,6804,65713,337
Percentage6535100
FY 2013Number8,7534,67313,426
Percentage6535100
FY 2014Number8,8944,73513,629
Percentage6535100
FY 2015Number8,8384,69413,532
Percentage6535100
FY 2016Number8,7954,72913,524
Percentage6535100
FY 2017Number8,7054,71713,422
Percentage6535100
FY 2018Number8,5664,69413,260
Percentage6535100
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: The data shown reflect numbers and percentages of men and women at the end of each fiscal year. If an employee’s recorded gender changed during the period shown, we assigned the most recently recorded value to all years for that employee.

Appendix III: Comparison of State Department Workforce with Federal Government and Relevant Civilian Labor Force

We compared summary statistics for the Department of State’s (State) workforce overall with summary statistics for the federal government and relevant civilian labor force (RCLF).

Comparison of State and Federal Workforce

We compared summary statistics calculated from State personnel data for fiscal year 2018 with summary statistics for the federal government for fiscal year 2016 published in the Federal Equal Opportunity Recruitment Program (FEORP) report.[119]

Our comparison of State personnel data with data from the Office of Personnel Management’s FEORP report for the federal government found differences between the proportions of racial or ethnic minorities at State and those in the federal workforce.[120] In particular, the proportions of minorities in general and of African Americans and Hispanics in particular were lower at State in fiscal year 2018 than in the federal workforce in fiscal year 2016. However, proportions of women at State and in the federal workforce were similar (see table 14).

Table 14: Percentages of Employees across Demographic Groups in the Department of State in Fiscal Year (FY) 2018 and the Federal Workforce in FY 2016
State, FY 2018Federal workforce, FY 2016
Racial or ethnic group
White6864
Racial or ethnic minority3236
African American1518
Hispanic79
Asian66
Other44
Gender
Men5757
Women4343
Source: GAO analysis of Department of State (State) and Office of Personnel Management data. | GAO-20-237

Notes: The data shown reflect percentages of employees at the end of each fiscal year. “Racial or ethnic minority” includes employees whose recorded race or ethnicity is neither non-Hispanic white nor unspecified. “Other” includes employees whose recorded race or ethnicity is Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. If an employee’s recorded racial or ethnic group or gender changed during the period shown, we assigned the most recently recorded value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.

Comparison of State’s Workforce with RCLF across Equal Employment Opportunity Commission Groupings

We compared workforce summary statistics from State with RCLF summary statistics from the Census Bureau’s Equal Employment Opportunity tabulation for three of the Equal Employment Opportunity Commission (EEOC) occupational classification system’s nine categories.[121] Using an EEOC table that cross-classifies OPM occupation codes and federal sector occupational categories, we classified each State employee into one of the nine categories. We compared State and RCLF statistics for the following three categories, corresponding to 99 percent of State’s full-time, permanent employees in fiscal year 2018: Officials and Managers, Professional Workers, and Administrative Support Workers.[122]

Our comparison of State workforce data with RCLF data found larger proportions of racial or ethnic minorities at State than in the RCLF for Officials and Managers, Professional Workers, and Administrative Support Workers (see tables 15 through 19).[123] The proportions of women were lower at State than in the RCLF for Officials and Managers and Professional Workers but were higher for Administrative Support Workers (see tables 15 through 17).[124]

Table 15: Percentages of Officials and Managers across Demographic Groups at the Department of State in Fiscal Year (FY) 2018 and in Relevant Civilian Labor Force (RCLF) in 2006-2010
State, FY 2018RCLF, 2006-2010
Racial or ethnic group
White6578
Racial or ethnic minority3521
African American189
Hispanic87
Asian54
Other42
Gender
Men5956
Women4144
Source: GAO analysis of Department of State (State) and Equal Employment Opportunity Commission data. | GAO-20-237

Notes: The data shown for fiscal year 2018 reflect percentages of employees at the end of the fiscal year. Officials and managers represented 11,353 of State's full-time, permanent workforce in fiscal year 2018. “Racial or ethnic minority” includes employees whose recorded race or ethnicity is neither non-Hispanic white nor unspecified. “Other” includes employees whose recorded race or ethnicity was Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. If an employee’s recorded racial or ethnic group or gender changed during the period shown, we assigned the most recent value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.

Table 16: Percentages of Professional Workers across Demographic Groups at the Department of State in Fiscal Year (FY) 2018 and in Relevant Civilian Labor Force (RCLF) in 2006-2010
State, FY 2018RCLF, 2006-2010
Racial or ethnic group
White7377
Racial or ethnic minority2722
African American98
Hispanic66
Asian86
Other41
Gender
Men5945
Women4155
Source: GAO analysis of Department of State (State) and Equal Employment Opportunity Commission data. | GAO-20-237

Notes: The data shown for fiscal year 2018 reflect percentages of professional workers at the end of the fiscal year. Professional workers represented 9,890 of State's full-time, permanent workforce in fiscal year 2018. “Racial or ethnic minority” includes employees whose recorded race or ethnicity is neither non-Hispanic white nor unspecified. “Other” includes employees whose recorded race or ethnicity was Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. If an employee’s recorded racial or ethnic group or gender changed during the period shown, we assigned the most recent value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.

Table 17: Percentages of Administrative Support Workers across Demographic Groups at the Department of State in Fiscal Year (FY) 2018 and in Relevant Civilian Labor Force (RCLF) in 2006-2010
State, FY 2018RCLF, 2006-2010
Racial or ethnic group
White5472
Racial or ethnic minority4527
African American2512
Hispanic811
Asian63
Other62
Gender
Men2125
Women7975
Source: GAO analysis of Department of State (State) and Equal Employment Opportunity Commission data. | GAO-20-237

Notes: The data shown for fiscal year 2018 reflect percentages of administrative support workers at the end of the fiscal year. Administrative support workers represented 1,265 of State's full-time, permanent workforce in fiscal year 2018. “Racial or ethnic minority” includes employees whose recorded race or ethnicity is neither non-Hispanic white nor unspecified. “Other” includes employees whose recorded race or ethnicity was Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. If an employee’s recorded racial or ethnic group or gender changed during the period shown, we assigned the most recent value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.

Appendix IV: Analysis of Demographic Data on Executives in the Department of State

Figures 14 and 15 present our analysis of data on executive employees, by racial or ethnic group and by gender, at the Department of State (State) overall and in State’s Civil Service and Foreign Service in fiscal years 2002 and 2018.[125] As figure 14 shows, the percentage of racial or ethnic minority executives increased from 12 percent to 13 percent, driven by an increase in the number of racial or ethnic minority executives in the Civil Service.[126] In addition, the percentage of African American executives at State declined from 6 percent to 3 percent, driven by a decline in the number of African American executives in the Foreign Service.

Figure 14: Percentages of White Executives and Racial or Ethnic Minority Executives in the Department of State and Its Civil and Foreign Services, Fiscal Years (FY) 2002 and 2018

Notes: The data shown reflect percentages of white and racial or ethnic minority executives at the end of each fiscal year. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates that race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all years for that employee. Because of rounding, percentages may not sum precisely to 100 and percentages of racial or ethnic minority employees may not sum precisely to the totals shown.

As figure 15 shows, the percentage of executive women at State rose from 26 percent in fiscal year 2002 to 33 percent in fiscal year 2018 and increased in both the Civil and Foreign Services.

Figure 15: Percentages of Executive Men and Women in the Department of State and Its Civil and Foreign Services, Fiscal Years (FY) 2002 and 2018

Notes: The data shown reflect percentages of executive men and women at the end of each fiscal year. If an employee’s recorded gender changed during the period shown, we assigned the most recently recorded value to all years for that employee.

To compare State and federal government workforce data, we contrasted summary statistics on executive employees calculated from State personnel data for fiscal year 2018 with summary statistics on executives from federal government workforce data for fiscal year 2016 that were published in the Office of Personnel Management’s Federal Equal Opportunity Recruitment Program (FEORP) report.[127] As table 18 shows, our comparison of State workforce data with the FEORP data found a higher proportion of white executives and a lower proportion of African American executives at State than in the federal workforce overall.[128]

Table 18: Percentages of Executives in Demographic Groups at the Department of State in Fiscal Year (FY) 2018 and in the Federal Workforce in FY 2016
State, FY 2018Federal government, FY 2016
Racial or ethnic group
White8779
Minority1321
African American311
Hispanic55
Asian34
Other22
Gender
Men6765
Women3335
Source: GAO analysis of Department of State (State) and Office of Personnel Management data. | GAO-20-237

Notes: The data shown reflect percentage of executives at the end of each fiscal year. We analyzed data for those listed as EX/AD/ES in State personnel data for its Civil Service, those listed as CM/MC/OC in State personnel data for its Foreign Service, and those listed as Senior Executive Service for the federal government. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates individuals whose race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all available years for that employee. Percentages may not sum precisely to 100 because of rounding.

Appendix V: Analysis of Data on Veterans at the Department of State

We analyzed Department of State (State) data on employees hired with veterans’ preference in fiscal years 2002 through 2018. The following tables present the numbers and percentages of employees hired with or without veterans’ preference in State’s workforce overall and in State’s Civil and Foreign Services during that period.

Table 19: Numbers and Percentages of Employees Hired with or without Veterans’ Preference in the Department of State, Fiscal Years (FY) 2002-2018
Hired with veterans’ preference Not hired with veterans’ preference Total
FY 2002 Number 2,399 14,171 16,570
Percentage 14 86 100
FY 2003 Number 2,510 14,917 17,427
Percentage 14 86 100
FY 2004 Number 2,563 15,345 17,908
Percentage 14 86 100
FY 2005 Number 2,569 15,785 18,354
Percentage 14 86 100
FY 2006 Number 2,556 16,076 18,632
Percentage 14 86 100
FY 2007 Number 2,613 16,580 19,193
Percentage 14 86 100
FY 2008 Number 2,738 17,055 19,793
Percentage 14 86 100
FY 2009 Number 2,899 17,695 20,594
Percentage 14 86 100
FY 2010 Number 3,080 18,632 21,712
Percentage 14 86 100
FY 2011 Number 3,391 19,260 22,651
Percentage 15 85 100
FY 2012 Number 3,529 19,457 22,986
Percentage 15 85 100
FY 2013 Number 3,759 19,679 23,438
Percentage 16 84 100
FY 2014 Number 3,877 19,726 23,603
Percentage 16 84 100
FY 2015 Number 4,026 19,538 23,564
Percentage 17 83 100
FY 2016 Number 4,152 19,651 23,803
Percentage 17 83 100
FY 2017 Number 4,139 19,305 23,444
Percentage 18 82 100
FY 2018 Number 4,021 18,785 22,806
Percentage 18 82 100
Source: GAO analysis of Department of State data. | GAO-20-237

Note: The data shown reflect numbers and percentages of employees at the end of each fiscal year.

Table 20: Numbers and Percentages of Employees Hired with or without Veterans’ Preference in the Department of State’s Civil Service, Fiscal Years (FY) 2002-2018
Hired with veterans’ preference Not hired with veterans’ preference Total
FY 2002 Number 925 5,906 6,831
Percentage 14 86 100
FY 2003 Number 1,003 6,155 7,158
Percentage 14 86 100
FY 2004 Number 1,030 6,222 7,252
Percentage 14 86 100
FY 2005 Number 1,054 6,352 7,406
Percentage 14 86 100
FY 2006 Number 1,100 6,485 7,585
Percentage 15 86 100
FY 2007 Number 1,144 6,926 8,070
Percentage 14 86 100
FY 2008 Number 1,263 7,227 8,490
Percentage 15 85 100
FY 2009 Number 1,331 7,320 8,651
Percentage 15 85 100
FY 2010 Number 1,430 7,570 9,000
Percentage 16 84 100
FY 2011 Number 1,633 7,855 9,488
Percentage 17 83 100
FY 2012 Number 1,723 7,926 9,649
Percentage 18 82 100
FY 2013 Number 1,888 8,124 10,012
Percentage 19 81 100
FY 2014 Number 1,917 8,057 9,974
Percentage 19 81 100
FY 2015 Number 2,017 8,015 10,032
Percentage 20 80 100
FY 2016 Number 2,109 8,170 10,279
Percentage 21 79 100
FY 2017 Number 2,055 7,967 10,022
Percentage 21 80 100
FY 2018 Number 1,933 7,613 9,546
Percentage 20 80 100
Source: GAO analysis of Department of State data. | GAO-20-237

Note: The data shown reflect numbers and percentages of employees at the end of each fiscal year.

Table 21: Numbers and Percentage of Employees Hired with or without Veterans’ Preference in the Department of State’s Foreign Service, Fiscal Years (FY) 2002-2018
Hired with veterans’ preference Not hired with veterans’ preference Total
FY 2002 Number 1,474 8,265 9,739
Percentage 15 85 100
FY 2003 Number 1,507 8,762 10,269
Percentage 15 85 100
FY 2004 Number 1,533 9,123 10,656
Percentage 14 86 100
FY 2005 Number 1,515 9,433 10,948
Percentage 14 86 100
FY 2006 Number 1,456 9,591 11,047
Percentage 13 87 100
FY 2007 Number 1,469 9,654 11,123
Percentage 13 87 100
FY 2008 Number 1,475 9,828 11,303
Percentage 13 87 100
FY 2009 Number 1,568 10,375 11,943
Percentage 13 87 100
FY 2010 Number 1,650 11,062 12,712
Percentage 13 87 100
FY 2011 Number 1,758 11,405 13,163
Percentage 13 87 100
FY 2012 Number 1,806 11,531 13,337
Percentage 14 86 100
FY 2013 Number 1,871 11,555 13,426
Percentage 14 86 100
FY 2014 Number 1,960 11,669 13,629
Percentage 14 86 100
FY 2015 Number 2,009 11,523 13,532
Percentage 15 85 100
FY 2016 Number 2,043 11,481 13,524
Percentage 15 85 100
FY 2017 Number 2,084 11,338 13,422
Percentage 16 84 100
FY 2018 Number 2,088 11,172 13,260
Percentage 16 84 100
Source: GAO analysis of Department of State data. | GAO-20-237

Note: The data shown reflect numbers and percentages of employees at the end of each fiscal year.

Appendix VI: Data on Individuals with Disabilities at the Department of State

Table 22 shows the proportions of permanent employees with a disability in the Civil Service and Foreign Service in fiscal years 2005 through 2017, using summary statistics from the Department of State’s (State) Management Directive 715 (MD-715) reports to the Equal Employment Opportunity Commission.[129] As the table shows, the proportion of permanent employees with disabilities increased in the Civil Service and decreased in the Foreign Service during this period.

Table 22: Percentages of Permanent Employees with a Disability in the Department of State’s Civil and Foreign Services, Fiscal Years 2005-2017
Fiscal year Civil Service Foreign Service
2005 6 7
2006 6 7
2007 6 7
2008 6 7
2009 5 6
2010 5 5
2011 5 5
2012 5 5
2013 7 5
2014 7 4
2015 6 4
2016 7 4
2017 8 4
Source: Department of State. | GAO-20-237

Notes: The percentages shown are those cited in the Department of State’s Management Directive-715 reports for fiscal years 2005 through 2017.

Appendix VII: Analysis of Data on New Employees at the Department of State, Fiscal Years 2003-2018

In addition to analyzing the demographic composition of the Department of State’s (State) workforce, we analyzed State personnel data to determine summary statistics on employees hired in fiscal years 2003 through 2018. We considered an employee to have been hired in a given fiscal year if the employee first appeared in State’s personnel data for that year.[130] Because the State data we reviewed began in fiscal year 2002, we were unable to identify employees who were hired in that fiscal year; thus, fiscal year 2003 is the first for which we were able to identify newly hired employees.

The following figures present proportions of newly hired employees in racial, ethnic, and gender groups in State overall and State’s Civil Service and Foreign Service in fiscal years 2003 through 2018.

Figure 16: Percentages of Newly Hired White and Racial or Ethnic Minority Employees in the Department of State and Its Civil and Foreign Services, Fiscal Years (FY) 2003 and 2018

Notes: The data shown reflect percentages of newly hired white and racial or ethnic minority employees at the end of each fiscal year. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates that race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all years for that employee. Percentages may not sum precisely to 100 because of rounding.

Figure 17: Percentages of Newly Hired Men and Women in the Department of State and Its Civil and Foreign Services, Fiscal Years (FY) 2003 and 2018

Notes: The data shown reflect percentages of newly hired men and women at the end of each fiscal year. If an employee’s recorded gender changed during the period shown, we assigned the most recently recorded value to all years for that employee.

Appendix VIII: Analysis of Data on Attrition at the Department of State, Fiscal Years 2003-2018

In addition to analyzing the demographic composition of the Department of State’s (State) workforce, we analyzed State personnel data to determine summary statistics for employees who left State for reasons other than retirement or death in fiscal years 2003 through 2018. Figures 18 and 19 present the percentages of such employees in various racial, ethnic, and gender groups at State overall and in State’s Civil Service and Foreign Service in fiscal years 2003 and 2018.

Figure 18: Percentages of White and Racial or Ethnic Minority Employees Who Left the Department of State and Its Civil and Foreign Services, Fiscal Years 2003 and 2018

Notes: The data shown reflect percentages of white and racial or ethnic minority employees who had left the department as of the end of each fiscal year. Our analysis includes data only for employees who left for reasons other than retirement or death. “Other” includes Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and non-Hispanic multiracial. “Unspecified” indicates that race or ethnicity was not recorded. If an employee’s recorded race or ethnicity changed during the period shown, we assigned the most recently recorded value to all years for that employee. Because of rounding, percentages may not sum precisely to 100 and percentages of racial or ethnic minority employees may not sum precisely to the totals shown.

Figure 19: Percentages of Men and Women Who Left the Department of State and Its Civil and Foreign Services, Fiscal Years 2003 and 2018

Notes: The data shown reflect percentages of men and women who had left the department as of the end of each fiscal year. Our analysis includes data only for employees who left for reasons other than retirement or death. If an employee’s recorded gender changed during the period shown, we assigned the most recently recorded value to all years for that employee.

Table 23 presents attrition rates for white and racial or ethnic minority employees who left the Department of State in fiscal years 2003 through 2018 for reasons other than retirement or death.

Table 23: Attrition Rates for White and Racial or Ethnic Minority Employees Who Left the Department of State and Its Civil or Foreign Service in Fiscal Years 2003-2018
Department of StateCivil ServiceForeign Service
Fiscal year Whites, %Racial or ethnic minorities, %Whites, %Racial or ethnic minorities, %Whites, %Racial or ethnic minorities, %
20031.91.82.82.01.51.4
20042.32.84.03.21.52.0
20052.73.24.43.91.92.0
20063.02.74.43.42.31.5
20073.02.64.13.02.42.0
20083.13.64.54.22.42.6
20092.82.23.92.62.21.7
20102.32.23.32.41.91.9
20112.52.33.53.01.91.3
20122.51.93.22.32.11.4
20132.52.43.43.02.01.5
20142.62.44.02.91.91.7
20153.23.54.84.42.32.2
20163.03.14.63.92.12.2
20173.03.04.63.92.11.8
20183.13.15.13.92.12.1
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: Our analysis includes data only for employees who left for reasons other than retirement or death. “Racial or ethnic minorities” includes employees whose recorded race or ethnicity is neither non-Hispanic white nor unspecified.

Table 24 presents rates of attrition for men and women who left the Department of State in fiscal years 2003 through 2018 for reasons other than retirement or death.

Table 24: Attrition Rates for Men and Women Who Left the Department of State and Its Civil or Foreign Service in Fiscal Years 2003-2018
Department of StateCivil ServiceForeign Service
Fiscal yearMen, %Women, %Men, %Women, %Men, %Women, %
20032.32.42.82.92.11.8
20042.73.04.13.92.22.0
20053.03.24.74.12.32.1
20063.03.44.14.12.52.5
20072.83.04.03.42.32.4
20083.13.74.44.62.42.6
20092.52.83.33.42.12.0
20102.22.52.83.22.01.7
20112.22.83.33.51.72.0
20122.32.53.32.81.92.2
20132.32.72.93.41.91.9
20142.42.73.53.51.81.9
20153.33.25.14.32.42.1
20163.03.24.54.12.22.0
20172.83.34.04.52.21.9
20182.83.54.54.72.02.2
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Our analysis includes data only for employees who left for reasons other than retirement or death.

Appendix IX: Analysis of Data on Promotion Rates at the Department of State, Fiscal Years 2013-2017

As table 25 shows, our analysis of yearly promotion rates for fiscal years 2013 through 2017 at the Department of State (State) found that promotion rates for white employees exceeded the rates for racial or ethnic minority employees for

  • promotions from GS-11 and every higher rank for every year, except from GS-15 to executive in 2 years in State’s Civil Service, and
  • promotions from Class 4 and higher ranks for 16 of the 20 possible year-rank combinations in State’s Foreign Service.
Table 25: Years When Promotion Rates for White Employees Exceeded Promotion Rates for Racial or Ethnic Minority Employees in the Department of State’s Civil and Foreign Services, Fiscal Years 2013-2017
Promotion typeFY 2013FY 2014FY 2015FY 2016FY 2017Total for FYs 2013-2017
Civil Service
GS-15 to executive3 of 5 years
GS-14 to GS-155 of 5 years
GS-13 to GS-145 of 5 years
GS-12 to GS-135 of 5 years
GS-11 to GS-125 of 5 years
Foreign Service
Class 1 to executive5 of 5 years
Class 2 to Class 14 of 5 years
Class 3 to Class 23 of 5 years
Class 4 to Class 34 of 5 years
Legend: FY = fiscal year, GS = General Schedule, ✓ = higher promotion rate for whites than racial or ethnic minorities, – = higher promotion rate for racial or ethnic minorities than whites.
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: For each promotion type and fiscal year, we calculated the promotion rates for white and racial or ethnic minority employees, respectively, as the number of newly elevated white or racial or ethnic minority employees in the next-higher rank in the following fiscal year divided by the number of whites or racial or ethnic minorities in the given rank in the current year. We examined differences in the unrounded promoted rates.

Table 26 shows the promotion rates for white employees and racial or ethnic minority employees in State’s Civil and Foreign Services in fiscal years 2013 through 2017.

Table 26: Promotion Rates for White Employees and Racial or Ethnic Minority Employees in the Department of State’s Civil and Foreign Services, Fiscal Years 2013-2017
Percentages
Promotion typeFY 2013FY 2014FY 2015FY 2016FY 2017
Civil Service: Whites
GS-15 to executive0.71.82.01.10.6
GS-14 to GS-154.05.54.52.64.3
GS-13 to GS-146.68.78.44.35.9
GS-12 to GS-1327.923.724.518.222.9
GS-11 to GS-1222.718.320.519.616.3
Civil Service: Racial or ethnic minorities
GS-15 to executive2.10.52.40.50.0
GS-14 to GS-153.32.72.62.33.5
GS-13 to GS-144.26.15.13.85.8
GS-12 to GS-1314.415.216.511.712.0
GS-11 to GS-1216.215.217.315.213.9
Foreign Service: Whites
Class 1 to executive8.57.98.77.34.1
Class 2 to Class 19.28.38.68.16.8
Class 3 to Class 213.312.712.211.38.7
Class 4 to Class 315.517.017.717.115.2
Foreign Service: Racial or ethnic minorities
Class 1 to executive5.47.87.27.03.3
Class 2 to Class 16.17.86.38.86.4
Class 3 to Class 214.310.411.311.88.3
Class 4 to Class 316.214.817.415.415.0
Legend: FY = fiscal year, GS = General Schedule.
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: For each promotion type and fiscal year, we calculated the promotion rates for white and racial or ethnic minority employees, respectively, as the number of newly elevated white or racial or ethnic minority employees in the next-higher rank in the following fiscal year divided by the number of whites or racial or ethnic minorities in the given rank in the current year. This analysis does not take into account the variety of factors besides racial or ethnic minority status that may affect promotion rates, including the length of time it takes to be promoted.

As table 27 shows, our analysis of yearly promotion rates for fiscal years 2013 through 2017 showed that men were promoted at a higher rate than women

  • from GS-11 and higher ranks for 15 of the 25 possible year-rank combinations and
  • from Class 4 and higher ranks for four of the 20 possible year-rank combinations in the Foreign Service.
Table 27: Years When Promotion Rates for Men Exceeded Promotion Rates for Women in the Department of State’s Civil and Foreign Services, Fiscal Years 2013-2017
Promotion typeFY 2013FY 2014FY 2015FY 2016FY 2017Total for FYs 2013-2017
Civil Service
GS-15 to executive2 of 5 years
GS-14 to GS-152 of 5 years
GS-13 to GS-144 of 5 years
GS-12 to GS-134 of 5 years
GS-11 to GS-123 of 5 years
Foreign Service
Class 1 to executive1 of 5 years
Class 2 to Class 10 of 5 years
Class 3 to Class 20 of 5 years
Class 4 to Class 33 of 5 years
Legend: FY = fiscal year, GS = General Schedule, ✓ = higher promotion rate for men than women, – = higher promotion rate for women than men.
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: For each promotion type and fiscal year, we calculated the promotion rates for white and racial or ethnic minority employees, respectively, as the number of newly elevated white or racial or ethnic minority employees in the next-higher rank in the following fiscal year divided by the number of whites or racial or ethnic minorities in the given rank in the current year. We examined differences in the unrounded promoted rates.

Table 28 shows the promotion rates for men and women in State’s Civil and Foreign Services in fiscal years 2013 through 2017.

Table 28: Promotion Rates for Men and Women in the Department of State’s Civil and Foreign Services, Fiscal Years 2013-2017
Percentages
Promotion typeFY 2013FY 2014FY 2015FY 2016FY 2017
Civil Service: Men
GS-15 to executive0.61.51.91.60.4
GS-14 to GS-153.44.24.12.63.8
GS-13 to GS-146.27.87.04.16.1
GS-12 to GS-1322.020.319.215.117.6
GS-11 to GS-1220.819.319.917.414.9
Civil Service Women
GS-15 to executive1.61.52.40.20.7
GS-14 to GS-154.25.33.62.34.3
GS-13 to GS-145.27.67.24.05.6
GS-12 to GS-1320.118.521.114.616.9
GS-11 to GS-1218.915.418.417.515.2
Foreign Service: Men
Class 1 to executive8.07.78.36.74.3
Class 2 to Class 18.37.87.47.85.8
Class 3 to Class 212.611.010.79.87.5
Class 4 to Class 315.417.117.517.016.6
Foreign Service: Women
Class 1 to executive8.28.28.88.43.4
Class 2 to Class 19.19.19.59.28.6
Class 3 to Class 215.614.514.715.011.1
Class 4 to Class 316.415.117.916.012.3
Legend: FY = fiscal year, GS = General Schedule.
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: For each promotion type and fiscal year, we calculated the promotion rates for white and racial or ethnic minority employees, respectively, as the number of newly elevated white or racial or ethnic minority employees in the next-higher rank in the following fiscal year divided by the number of whites or racial or ethnic minorities in the given rank in the current year. This analysis does not take into account the variety of factors besides racial or ethnic minority status that may affect promotion rates, including the length of time it takes to be promoted.

Appendix X: Analysis of Data on Employee Years in Rank at the Department of State, Fiscal Years 2002-2018

Our analysis found that racial or ethnic minorities at the Department of State generally spent more years in each rank than whites did in both the Civil Service and the Foreign Service (see table 29).

Table 29: Average Years in Rank for Whites and Racial or Ethnic Minorities in the Department of State’s Civil and Foreign Services, Fiscal Years 2002-2018
All employeesPromoted employees
RankWhitesRacial or ethnic minoritiesWhitesRacial or ethnic minorities
Civil Service
Executive6.45.0N/AN/A
GS-156.15.85.34.2
GS-145.25.34.14.2
GS-134.55.13.33.6
GS-122.73.91.82.7
GS-112.73.51.62.1
Foreign Service
Executive6.16.5N/AN/A
Class 14.94.75.24.9
Class 25.15.15.05.0
Class 34.04.34.04.3
Class 43.73.93.84.0
Legend: GS = General Schedule, N/A = not available.
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: Average years in rank for promoted executives is not available because we did not examine promotion above GS-15 or Class 1.

In addition, our analysis found that women generally spent fewer years in each rank than men did in both the Civil Service and the Foreign Service (see table 30).

Table 30: Average Years in Rank for Men and Women in the Department of State’s Civil and Foreign Services, Fiscal Years 2002-2018
All employeesPromoted employees
RankMenWomenMenWomen
Civil Service
Executive6.46.0N/AN/A
GS-156.16.05.25.1
GS-145.55.04.33.9
GS-134.84.73.43.5
GS-123.23.22.02.2
GS-112.93.21.71.9
Foreign Service
Executive6.26.2N/AN/A
Class 15.04.85.15.2
Class 25.24.75.14.9
Class 34.23.84.23.9
Class 43.83.73.93.8
Legend: GS = General Schedule, N/A = not available.
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: Average years in rank for promoted executives is not available because we did not examine promotion above GS-15 or Class 1.

Appendix XI: Full Promotion Regression Results

Tables 31, 32, 38, and 39 provide summaries of the multivariate statistical regression results (specifically, duration regression results) for our estimates of the percentage differences in odds of promotion for racial or ethnic minorities compared with whites and for women compared with men in the Civil and Foreign Services. Our analyses do not completely explain the reasons for differences in promotion outcomes, which may result from various unobservable factors. Thus, our analyses do not establish a causal relationship between demographic characteristics and promotion outcomes.

These summary tables present estimates from six regression models, one of which, model 5, is also presented in the body of this report. All models controlled for the time employees spent in each rank—that is, in each General Schedule (GS) grade for the Civil Service or salary class for the Foreign Service—prior to promotion.

  • Model 1a controlled only for racial or ethnic minority status when estimating the percentage differences in odds of promotion for racial or ethnic minorities compared with whites. Model 1b controlled only for gender when estimating the percentage difference in odds of promotion for women compared with men.
  • Model 2 controlled for both racial or ethnic minority status and gender.
  • Model 3 controlled for racial or ethnic minority status, gender, and the following additional individual control variables that may be positively or negatively related to promotion outcomes.
    • Model 3 controlled for the following variables that may be positively related to promotion outcomes: graduating from a college or university considered Ivy League, because there may be a perception that graduates from these colleges or universities would be high quality applicants to State; graduating from a college or university located in the District of Columbia, Virginia, or Maryland, because some of these colleges or universities have highly respected programs related to foreign service that may provide networking opportunities; having a hardship assignment in the prior year (Foreign Service only); and having proficiency in a hard language (Foreign Service only).
    • Model 3 controlled for the use of long-term leave in the prior year, a variable that may be negatively related to promotion outcomes.[131]
    • Model 3 controlled for the following variables that may be positively or negatively related to promotion outcomes: years of federal government experience, age when entering State, veteran’s status, changing between the Foreign and Civil Services, and having an overseas post in the prior year (Foreign Service only). [132]
  • Model 4 controlled for the same variables as model 3 as well as for occupation, because occupations may vary in their statistical relationship to promotion outcomes. That is, certain occupations may be either positively or negatively related to promotion outcomes.
  • Model 5, presented in the body of the report, controlled for the same variables as model 4 as well as for fiscal year fixed effects (indicator variables representing the fiscal year), because available promotion slots (and resulting promotion outcomes) may be related to budget constraints that vary across fiscal years. In addition, model 5 clustered the standard errors on organization for the Civil Service, because available promotion slots (and resulting promotion outcomes) may be related to specific organizations in the Civil Service.[133]
  • Model 6 controlled for the same variables as model 5 but used data for fiscal years 2011 through 2018 only.

In addition, tables 33 through 37 and tables 40 through 43 provide the full regression results for the first five models by all promotion stages that we analyzed in the Civil and Foreign Services, respectively. While tables 31, 32, 38, and 39 present the full regression results as estimates of percentage differences, tables 33 through 37 and tables 40 through 43 present the regression results as odds ratios. Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic were less likely to be promoted. Odds ratios that are statistically significant and greater than 1.00 indicate that individuals with the given characteristic were more likely to be promoted. To convert the values in tables 33 through 37 and tables 40 through 43 to the values in tables 31, 32, 38, and 39, we linearly transformed the estimates. That is, the values for the estimates in tables 31, 32, 38, and 39 are equal to the values in tables 33 through 37 and tables 40 through 43 multiplied by 100 minus 100. The values for the standard errors in tables 31, 32, 38, and 39 are equal to the values in tables 33 through 37 and tables 40 through 43 multiplied by 100. For example, in table 33, the estimate for model 1a is 0.818; we arrived at the percentage difference of negative 18 percent in table 31 by 0.818*100-100. Additionally, in table 33, the estimate for the standard error for model 1a is (0.0305); we arrived at the converted standard error of (3) in table 31 by (.0305)*100.

Table 31 presents the summary of the regression results for our estimates of the percentage differences in odds of promotion for racial or ethnic minorities compared with whites in the Civil Service. We observed that the statistically significantly lower odds of promotion for racial or ethnic minorities from GS-11 through GS-15 were consistent across all of our models, which examined different combinations of factors that could influence promotion (i.e., models 1a through 5). In addition, our statistically significant results were consistent when we examined the more recent time period fiscal years 2011 through 2018 (see model 6).

Table 31: Percentage Differences in Promotion Odds for Racial or Ethnic Minorities Compared with Whites in Department of State’s Civil Service, Fiscal Years 2002-2018 and 2011-2018
Percentage difference (standard error)
Control variablesGS-11 to GS-12GS-12 to GS-13GS-13 to GS-14GS-14 to GS-15GS-15 to exec.
Fiscal years 2002-2018
Model 1a: Racial or ethnic minority status-18***-40***-27***-29***-18
(3)(2)(3)(5)(17)
Model 2: Racial or ethnic minority status and gender-18***-40***-27***-30***-18
(3)(2)(3)(5)(17)
Model 3: Racial or ethnic minority status, gender, and additional individual-level control variables-15***-33***-24***-30***-20
(3)(3)(3)(5)(17)
Model 4: Racial or ethnic minority status, gender, additional individual-level control variables, and occupation-25***-29***-23***-27***-15
(3)(3)(4)(6)(18)
Model 5: Racial or ethnic minority status, gender, additional individual-level control variables, and occupation (with fiscal year fixed effects and standard errors clustered on organization)-26***-29***-19***-22***-4
(4)(4)(4)(6)(22)
Fiscal years 2011-2018
Model 6: Racial or ethnic minority status, gender, additional individual-level control variables, and occupation (with fiscal year fixed effects and standard errors clustered on organization)-25***-26***-20***-23**-28
(7)(5)(6)(9)(23)
Legend: GS = General Schedule, exec. = executive, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State (State) data. | GAO-20-237

Notes: We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each GS level shown. In all models, we controlled for the time that employees spent in each grade before promotion. Additional individual-level control variables include employees’ years of government service; age when entering State; veteran’s status; taking long-term leave; graduating from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland; and changing between the Civil and Foreign Services. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 32 presents the summary of the regression results for our estimates of the percentage difference in odds of promotion for women compared with men in the Civil Service. We observed that the statistical insignificance of our estimates was generally consistent across all of our models, which examined different combinations of factors that could influence promotion (i.e., model 1b–model 5). In addition, the statistical insignificance of our estimates was consistent for the more recent time period fiscal years 2011 through 2018 (see model 6).

Table 32: Percentage Differences in Promotion Odds for Women Compared with Men in the Department of State’s Civil Service, Fiscal Years 2002-2018 and 2011-2018
Percentage difference (standard error)
Control variablesGS-11 to GS-12GS-12 to GS-13GS-13 to GS-14GS-14 to GS-15GS-15 to exec.
Fiscal years 2002-2018
Model 1b: Gender-5-9**-47-1
(4)(3)(4)(7)(15)
Model 2: Gender and racial or ethnic minority-3-30111
(4)(4)(4)(7)(15)
Model 3: Gender, racial or ethnic minority, and additional individual-level control variables-2-4-7*-4-7
(4)(4)(4)(6)(14)
Model 4: Gender, racial or ethnic minority, additional individual-level control variables, and occupation-1-6-6-7-8
(5)(4)(4)(6)(14)
Model 5: Gender, racial or ethnic minority, additional individual-level control variables, and occupation (with fiscal year fixed effects and standard errors clustered on organization)-2-5-31-5
(6)(5)(5)(7)(14)
Fiscal years 2011-2018
Model 6: Gender, racial or ethnic minority, additional individual-level control variables, and occupation (with fiscal year fixed effects and standard errors clustered on organization)-6-8-4-47
(8)(6)(7)(10)(23)
Legend: GS = General Schedule, exec. = executive, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State (State) data. | GAO-20-237

Notes: We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each GS level shown. In all models, we controlled for the time that employees spent in each grade before promotion. Additional individual-level control variables include employees’ years of government service; age when entering State; veteran’s status; taking long-term leave; graduating from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland; and changing between the Civil and Foreign Services. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 33: Odds Ratios for Promotion from GS-11 to GS-12 in the Department of State’s Civil Service, Fiscal Years 2002-2018
Odds ratio (standard error)
Control variableModel 1aModel 1bModel 2Model 3Model 4Model 5
Woman0.9540.9750.9820.9860.977
(0.0365)(0.0375)(0.0413)(0.0459)(0.0626)
Racial or ethnic minority0.818***0.820***0.850***0.750***0.738***
(0.0305)(0.0307)(0.0338)(0.0330)(0.0436)
Veteran's preference1.0721.0981.162*
(0.0615)(0.0678)(0.0939)
Ivy League college or university2.315***1.1191.093
(0.314)(0.165)(0.175)
Washington, D.C.–area college or university2.382***1.207***1.238***
(0.110)(0.0624)(0.0775)
Age at entry1.0070.9971.002
(0.0142)(0.0147)(0.0193)
Age at entry, squared0.999***1.0001.000
(0.00019)(0.00020)(0.00027)
Years of government service0.9960.969***0.969***
(0.00714)(0.00733)(0.0109)
Years of government service, squared0.999***1.000*1.000
(0.00027)(0.00027)(0.00038)
Long leave in prior year0.9430.9070.928
(0.154)(0.160)(0.172)
Changed service0.9760.8100.865
(0.345)(0.302)(0.277)
Occupation
Security administration 1.905***1.784*
(0.372)(0.563)
Foreign affairs1.946***2.123***
(0.162)(0.283)
Human resources management2.664***2.921***
(0.299)(0.457)
Miscellaneous administration0.712***0.686***
(0.0474)(0.0915)
Management program analysis1.555***1.799***
(0.134)(0.282)
Financial administration and program0.622***0.616
(0.0715)(0.225)
Budget analysis2.204***2.356***
(0.294)(0.463)
Passport, visa examining0.145***0.151***
(0.00908)(0.0166)
Information technology management0.9200.900
(0.0971)(0.176)
Duration controls
Fiscal year controls
Constant3.283***3.136***3.328***3.258***4.407***3.160***
(0.321)(0.311)(0.333)(0.870)(1.254)(1.286)
Observations21,48721,48721,48721,48721,48720,089
Legend: GS = General Schedule, ✓ = controls applied, — = not applicable, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic are less likely to be promoted, while odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to be promoted. For example, the estimated odds ratio for racial or ethnic minority employees for promotion from GS-11 to GS-12 is 0.738 (model 5), which means that the odds of promotion for racial or ethnic minority employees are about 74 percent of the odds for white employees. We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each GS grade shown. In all models, we controlled for the time that employees spent in each grade before promotion. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 34: Odds Ratios for Promotion from GS-12 to GS-13 in the Department of State’s Civil Service, Fiscal Years 2002-2018
Odds ratio (standard error)
Control variableModel 1aModel 1bModel 2Model 3Model 4Model 5
Woman0.911**0.9750.9590.9450.953
(0.034)(0.037)(0.0394)(0.0410)(0.0468)
Racial or ethnic minority0.599***0.601***0.669***0.708***0.707***
(0.0223)(0.023)(0.0263)(0.0294)(0.0355)
Veteran's preference0.895**0.9731.015
(0.0495)(0.0558)(0.0719)
Ivy League college or university2.545***1.470***1.419**
(0.319)(0.203)(0.233)
Washington, D.C.–area college or university1.395***1.157***1.161***
(0.0573)(0.0503)(0.0575)
Age at entry1.0211.0191.021
(0.0142)(0.0144)(0.0163)
Age at entry, squared0.999***0.999***0.999***
(0.00019)(0.00019)(0.000218)
Years of government service0.962***0.971***0.968***
(0.00652)(0.00680)(0.00886)
Years of government service, squared1.0001.0001.000
(0.00025)(0.00026)(0.000289)
Long leave in prior year0.8850.9440.939
(0.142)(0.154)(0.164)
Changed service1.361**1.451**1.355
(0.214)(0.228)(0.291)
Occupation
Security administration0.550***0.540***
(0.0648)(0.0830)
Foreign affairs4.161***4.361***
(0.283)(0.513)
Human resources management1.227**1.209
(0.0992)(0.184)
Miscellaneous administration0.9280.893
(0.0642)(0.134)
Management program analysis1.581***1.595***
(0.102)(0.193)
Financial administration and program0.799*0.784
(0.106)(0.319)
Budget analysis1.447***1.437**
(0.138)(0.220)
Passport, visa examining0.490***0.468***
(0.0387)(0.0542)
Information technology management1.245***1.173
(0.104)(0.212)
Duration controls
Fiscal year controls
Constant2.049***1.827***2.076***2.083***1.1541.095
(0.193)(0.174)(0.199)(0.547)(0.312)(0.371)
Observations22,09722,09722,09722,09422,09420,729
Legend: GS = General Schedule, ✓ = controls applied, — = not applicable, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic are less likely to be promoted, while odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to be promoted. For example, the estimated odds ratio for racial or ethnic minority employees for promotion from GS-12 to GS-13 is 0.707 (model 5), which means that the odds of promotion for racial or ethnic minority employees are about 71 percent of the odds for white employees. We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each GS grade shown. In all models, we controlled for the time that employees spent in each grade before promotion. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 35: Odds Ratios for Promotion from GS-13 to GS-14 in the Department of State’s Civil Service, Fiscal Years 2002-2018
Odds ratio (standard error)
Control variableModel 1aModel 1bModel 2Model 3Model 4Model 5
Woman0.9601.0000.929*0.9360.966
(0.0395)(0.0416)(0.0411)(0.0423)(0.0461)
Racial or ethnic minority0.734***0.734***0.757***0.766***0.806***
(0.0324)(0.0327)(0.0343)(0.0361)(0.0412)
Veteran's preference0.852***0.9231.046
(0.0521)(0.0578)(0.0677)
Ivy League college or university1.622***1.473***1.419***
(0.171)(0.158)(0.148)
Washington, D.C.–area college or university1.221***1.157***1.181***
(0.0535)(0.0512)(0.0547)
Age at entry1.035**1.027*1.023
(0.0156)(0.0155)(0.0160)
Age at entry, squared0.999***0.999***0.999***
(0.00020)(0.00020)(0.00021)
Years of government service1.0041.0030.980**
(0.00745)(0.00755)(0.00809)
Years of government, service squared0.999***0.999***0.999**
(0.00028)(0.00028)(0.00030)
Long leave in prior year0.8070.8000.809
(0.127)(0.126)(0.124)
Changed service0.602***0.555***0.599***
(0.0861)(0.0803)(0.0996)
Occupation
Security administration0.493***0.438***
(0.0635)(0.0571)
Foreign affairs1.0231.060
(0.0609)(0.0982)
Human resources management1.0191.120
(0.106)(0.158)
Miscellaneous administration0.482***0.503***
(0.0489)(0.0569)
Management program analysis0.763***0.844*
(0.0540)(0.0817)
Financial administration and program0.8740.955
(0.151)(0.145)
Budget analysis0.9340.983
(0.114)(0.149)
Passport, visa examining0.424***0.450***
(0.0617)(0.0751)
Information technology management0.883*0.831*
(0.0650)(0.0821)
Duration controls
Fiscal year controls
Constant0.0465***0.0431***0.0465***0.0353***0.0472***0.0912***
(0.00572)(0.00535)(0.00580)(0.0105)(0.0143)(0.0310)
Observations37,25437,25437,25437,25437,25434,395
Legend: GS = General Schedule, ✓ = controls applied, — = not applicable, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic are less likely to be promoted, while odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to be promoted. For example, the estimated odds ratio for racial or ethnic minority employees for promotion from GS-13 to GS-14 is 0.806 (model 5), which means that the odds of promotion for racial or ethnic minority employees are about 81 percent of the odds for white employees. We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each GS grade shown. In all models, we controlled for the time that employees spent in each grade before promotion. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 36: Odds Ratios for Promotion from GS-14 to GS-15 in the Department of State’s Civil Service, Fiscal Years 2002-2018
Odds ratio (standard error)
Control variableModel 1aModel 1bModel 2Model 3Model 4Model 5
Woman1.0711.1070.9650.9331.006
(0.0675)(0.0702)(0.0642)(0.0627)(0.0688)
Racial or ethnic minority0.713***0.704***0.698***0.727***0.782***
(0.0536)(0.0532)(0.0533)(0.0571)(0.0605)
Veteran's preference0.849*0.8841.028
(0.0810)(0.0855)(0.106)
Ivy League college or university1.631***1.509***1.452***
(0.208)(0.195)(0.204)
Washington, D.C.–area college or university1.0441.0401.066
(0.0693)(0.0692)(0.0739)
Age at entry1.060**1.062**1.056**
(0.0255)(0.0257)(0.0243)
Age at entry, squared0.999***0.999***0.999***
(0.00032)(0.00033)(0.00030)
Years of government service1.019*1.021*0.979*
(0.0110)(0.0111)(0.0114)
Years of government service, squared0.999**0.999**1.000
(0.00038)(0.00038)(0.00039)
Long leave in prior year0.6570.633*0.651*
(0.170)(0.164)(0.168)
Changed service0.707*0.7420.924
(0.140)(0.148)(0.191)
Occupation
Security administration0.8570.813
(0.208)(0.227)
Foreign affairs1.339***1.377***
(0.114)(0.148)
Human resources management1.1031.401
(0.227)(0.323)
Miscellaneous administration1.457***1.463***
(0.197)(0.210)
Management program analysis1.380***1.462***
(0.151)(0.204)
Financial administration and program1.0471.103
(0.242)(0.305)
Budget analysis1.403*1.504*
(0.244)(0.320)
Passport, visa examining0.9781.003
(0.506)(0.450)
Information technology management0.9440.964
(0.112)(0.128)
Duration controls
Fiscal year controls
Constant0.0270***0.0239***0.0258***0.0171***0.0133***0.0328***
(0.00508)(0.00454)(0.00491)(0.00808)(0.00636)(0.0169)
Observations23,62323,62323,62323,62323,62321,792
Legend: GS = General Schedule, ✓ = controls applied, — = not applicable, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic are less likely to be promoted, while odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to be promoted. For example, the estimated odds ratio for racial or ethnic minority employees for promotion from GS-14 to GS-15 is 0.782 (model 5), which means that the odds of promotion for racial or ethnic minority employees are about 78 percent of the odds for white employees. We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each GS grade shown. In all models, we controlled for the time that employees spent in each grade before promotion. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 37: Odds Ratios for Promotion from GS-15 to Executive in the Department of State’s Civil Service, Fiscal Years 2002-2018
Odds ratio (standard error)
Control variableModel 1aModel 1bModel 2Model 3Model 4Model 5
Woman0.9931.0130.9260.9210.951
(0.147)(0.152)(0.143)(0.144)(0.143)
Racial or ethnic minority0.8230.8210.8050.8550.957
(0.167)(0.168)(0.166)(0.181)(0.218)
Veteran's preference0.7160.7750.864
(0.192)(0.209)(0.252)
Ivy League college or university1.428*1.3351.203
(0.296)(0.280)(0.255)
Washington, D.C.–area college or university1.0611.0581.076
(0.166)(0.166)(0.172)
Age at entry1.174**1.184***1.178***
(0.0750)(0.0759)(0.0738)
Age at entry, squared0.998***0.998***0.998***
(0.00086)(0.00087)(0.00085)
Years of government service1.053**1.052*1.017
(0.0277)(0.0279)(0.0268)
Years of government service, squared0.999*0.999*0.999
(0.00081)(0.00082)(0.00076)
Long leave in prior year1.0040.9991.146
(0.596)(0.593)(0.731)
Changed service0.7190.7331.104
(0.369)(0.378)(0.597)
Occupation
Security administration
Foreign affairs0.9860.977
(0.169)(0.208)
Human resources management1.0061.229
(0.530)(0.659)
Miscellaneous administration0.7890.815
(0.225)(0.268)
Management program analysis0.5470.593
(0.218)(0.215)
Financial administration and program0.7260.727
(0.376)(0.372)
Budget analysis1.5921.917
(0.749)(0.889)
Passport, visa examining6.249***6.307**
(3.986)(5.080)
Information technology management0.414**0.455*
(0.177)(0.200)
Duration controls
Fiscal year controls
Constant0.0129***0.0125***0.0129***0.0008***0.0008***0.0018***
(0.00597)(0.00583)(0.00599)(0.00105)(0.00097)(0.00224)
Observations13,32513,32513,32513,32513,15812,183
Legend: GS = General Schedule, ✓ = controls applied, — = not applicable, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic are less likely to be promoted, while odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to be promoted. We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each GS grade shown. In all models, we controlled for the time that employees spent in each grade before promotion. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

.

Table 38 summarizes the regression results for our estimates of the percentage differences in odds of promotion for racial or ethnic minorities compared with whites in the Foreign Service. We observed that, while the lower odds of promotion for racial or ethnic minorities from Class 4 to Class 3 were consistently statistically significant across our models, we also found statistically significantly lower odds of promotion in fiscal years 2002 through 2018 for racial or ethnic minorities from Class 3 to Class 2 and from Class 2 to Class 1 (see models 1a through 4) when controlling for various subsets of factors. Similarly, when we examined the more recent period fiscal years 2011 through 2018, we found that racial or ethnic minorities in the Foreign Service were statistically significantly less likely than whites to be promoted from Class 3 to Class 2 (see model 6).[134]

Table 38: Percentage Differences in Promotion Odds for Minorities Compared with Whites in the Department of State’s Foreign Service, Fiscal Years 2002-2018 and 2011-2018
Percentage difference (standard error)
Control variablesClass 4 to Class 3Class 3 to Class 2Class 2 to Class 1Class 1 to executive
Fiscal years 2002-2018
Model 1a: Racial or ethnic minority status-9***-18***-20***2
(3)(3)(4)(8)
Model 2: Racial or ethnic minority status and gender-9***-19***-21***2
(3)(3)(4)(8)
Model 3: Racial or ethnic minority status, gender, and additional individual-level control variables-14***-20***-16***6
(3)(3)(4)(8)
Model 4: Racial or ethnic minority status, gender, additional individual-level control variables, and occupation-17***-13***-13***6
(3)(3)(4)(8)
Model 5: Racial or ethnic minority status, gender, additional individual-level control variables, and occupation (with fiscal year fixed effects)-13***-5-810
(3)(4)(5)(8)
Fiscal years 2011-2018
Model 6: Racial or ethnic minority status, gender, additional individual-level control variables, and occupation (with fiscal year fixed effects)-11**-10**-12*-7
(4)(5)(6)(11)
Legend: *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State (State) data. | GAO-20-237

Notes: We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each salary class shown. In all models, we controlled for the time that employees spent in each salary class before promotion. Additional individual-level control variables include employees’ years of government service; age when entering State; veteran’s status; taking long-term leave; graduating from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland; changing between the Civil and Foreign Services; having a hardship assignment in the prior year; having an overseas post in the prior year; and proficiency in a hard language. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 39 presents the summary of the regression results for our estimates of the percentage differences in odds of promotion for women compared with men in the Foreign Service. Before controlling for occupation and fiscal years, we found that, in general, women in the Foreign Service were statistically significantly less likely than men to be promoted from Class 4 to Class 3 (models 1b through 3). However, this effect was driven by data for one Foreign Service occupation, secretary; in fiscal year 2018, 91 percent of the 766 Foreign Service secretaries were women.[135] When we also controlled for occupation (i.e., secretary and others) and fiscal years, we found that women in the Foreign Service were statistically significantly more likely than men to be promoted from Class 4 to Class 3, even in the more recent period fiscal years 2011 through 2018 (see models 5 and 6). The statistically significantly higher odds of promotion for women from Class 3 to Class 2 were generally consistent across all of our models, which examined various combinations of factors that could influence promotion (i.e., models 1b through 5). However, using data from fiscal years 2011 through 2018, we found no statistically significant difference in odds of promotion from Class 3 to Class 2 for women relative to men. As a result, we could not conclude that there was a statistical relationship between gender and promotion from Class 3 to Class 2 in that more recent period (see model 6). In fiscal years 2011 through 2018, women had statistically significantly higher odds of promotion from Class 2 to Class 1 and from Class 1 to executive (see model 6).

Table 39: Percentage Differences in Promotion Odds for Women Compared with Men in the Department of State’s Foreign Service, Fiscal Years 2002-2018 and 2011-2018
Percentage difference (standard error)
Control variablesClass 4 to Class 3Class 3 to Class 2Class 2 to Class 1Class 1 to executive
Fiscal years 2002-2018
Model 1b: Gender-5**38***23***0
(2)(4)(5)(6)
Model 2: Gender and racial or ethnic minority-5*38***24***0
(2)(4)(5)(6)
Model 3: Gender, racial or ethnic minority, and additional individual-level control variables-11***25***26***6
(2)(4)(5)(6)
Model 4: Gender, racial or ethnic minority, additional individual-level control variables, and occupation6*8**76
(3)(4)(4)(6)
Model 5: Gender, racial or ethnic minority, additional individual-level control variables, and occupation (with fiscal year fixed effects)9***13***8*8
(3)(4)(5)(7)
Fiscal years 2011-2018
Model 6: Gender, racial or ethnic minority, additional individual-level control variables, and occupation (with fiscal year fixed effects)11**513**24**
(5)(5)(7)(11)
Legend: *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State (State) data. | GAO-20-237

Notes: We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each salary class shown. In all models, we controlled for the time that employees spent in each salary class before promotion. Additional individual-level control variables include employees’ years of government service; age when entering State; veteran’s status; taking long-term leave; graduating from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland; changing between the Civil and Foreign Services; having a hardship assignment in the prior year; having an overseas post in the prior year; and proficiency in a hard language. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 40: Odds Ratios for Promotion from Class 4 to Class 3 in the Department of State’s Foreign Service, Fiscal Years 2002-2018
Odds ratio (standard error)
Control variableModel 1aModel 1bModel 2Model 3Model 4Model 5
Woman0.948**0.951*0.889***1.055*1.094***
(0.0244)(0.0246)(0.0244)(0.0309)(0.0333)
Racial or ethnic minority0.906***0.908***0.865***0.832***0.872***
(0.0260)(0.0261)(0.0255)(0.0251)(0.0273)
Veteran's preference1.0330.763***0.995
(0.0398)(0.0310)(0.0421)
Hard language1.301***1.329***1.261***
(0.0753)(0.0782)(0.0769)
Ivy League college or university0.9981.179***1.225***
(0.0454)(0.0557)(0.0601)
Washington, D.C.–area college or university0.947*1.0101.029
(0.0271)(0.0298)(0.0315)
Age at entry0.792***0.808***0.799***
(0.00964)(0.0101)(0.0104)
Age at entry, squared1.002***1.002***1.002***
(0.00016)(0.00017)(0.00017)
Years of government service1.051***1.061***1.017**
(0.00677)(0.00688)(0.00686)
Years of government service, squared0.997***0.998***0.998***
(0.0003)(0.0003)(0.0003)
Overseas in prior year1.328***1.601***1.949***
(0.0425)(0.0543)(0.0729)
Hardship in prior year1.0201.0151.000
(0.155)(0.157)(0.161)
Long leave in prior year0.818*0.8350.959
(0.0904)(0.0933)(0.111)
Changed service0.525***0.499***0.631***
(0.0363)(0.0350)(0.0463)
Occupation
Economist1.254***1.188***
(0.0607)(0.0594)
Administrative officer1.602***1.502***
(0.0827)(0.0802)
Passport, visa examiner1.280***1.323***
(0.0635)(0.0679)
Public affairs specialist1.395***1.408***
(0.0692)(0.0724)
Security administrator2.714***2.799***
(0.127)(0.137)
Secretary0.285***0.307***
(0.0324)(0.0363)
Computer specialists1.0230.795***
(0.0580)(0.0468)
Information systems management
Office service management supervisor3.031***2.748***
(0.271)(0.253)
Other3.713***3.557***
(0.235)(0.235)
Duration controls
Fiscal year controls
Constant0.0060***0.0060***0.0061***0.7600.354***1.522
(0.00067)(0.00067)(0.00068)(0.182)(0.0867)(0.394)
Observations55,55455,55455,55455,55455,55252,384
Legend: ✓ = controls applied, — = not applicable, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic are less likely to be promoted, while odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to be promoted. For example, the estimated odds ratio for women for promotion from Class 4 to Class 3 is 1.094 (model 5), which means that the odds of promotion for women are about 109 percent of the odds for men. We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each salary class shown. In all models, we controlled for the time that employees spent in each grade before promotion. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 41: Odds Ratios for Promotion from Class 3 to Class 2 in the Department of State’s Foreign Service, Fiscal Years 2002-2018
Odds ratio (standard error)
Control variableModel 1aModel 1bModel 2Model 3Model 4Model 5
Woman1.379***1.382***1.248***1.078**1.127***
(0.0424)(0.0425)(0.0401)(0.0366)(0.0403)
Racial or ethnic minority0.818***0.815***0.796***0.871***0.951
(0.0292)(0.0291)(0.0290)(0.0328)(0.0375)
Veteran's preference0.462***0.534***0.760***
(0.0225)(0.0272)(0.0403)
Hard language1.291***0.9600.807***
(0.0661)(0.0509)(0.0451)
Ivy League college or university1.750***1.201***1.126**
(0.0914)(0.0652)(0.0649)
Washington, D.C.–area college or university1.259***1.101***1.100***
(0.0416)(0.0378)(0.0395)
Age at entry0.9810.9821.026
(0.0150)(0.0155)(0.0171)
Age at entry, squared1.0001.0000.999***
(0.000210)(0.000218)(0.000229)
Years of government service1.102***1.133***0.996
(0.00727)(0.00780)(0.00775)
Years of government service, squared0.997***0.997***1.000
(0.00030)(0.00031)(0.000304)
Overseas in prior year1.158***1.211***1.624***
(0.0373)(0.0404)(0.0598)
Hardship in prior year1.685***1.757***2.010***
(0.236)(0.254)(0.313)
Long leave in prior year0.679***0.639***0.782**
(0.0714)(0.0687)(0.0883)
Changed service0.9080.742***1.220**
(0.0787)(0.0657)(0.118)
Occupation
Economist0.713***0.641***
(0.0410)(0.0390)
Administrative officer1.177***1.213***
(0.0737)(0.0818)
Passport, visa examiner0.571***0.519***
(0.0336)(0.0322)
Public affairs specialist0.9751.099
(0.0583)(0.0713)
Security administrator0.251***0.229***
(0.0146)(0.0138)
Secretary0.00409***0.00665***
(0.00291)(0.00475)
Computer specialist0.177***0.158***
(0.0131)(0.0120)
Information systems management
Office service management supervisor0.278***0.262***
(0.0293)(0.0281)
Other0.579***0.447***
(0.0382)(0.0309)
Duration controls
Fiscal year controls
Constant0.0013***0.0012***0.0012***0.0014***0.0025***0.00051***
(0.00020)(0.00017)(0.00018)(0.00043)(0.00078)(0.000187)
Observations45,68545,68545,68545,68445,67942,036
Legend: ✓ = controls applied, — = not applicable, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic are less likely to be promoted, while odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to be promoted. For example, the estimated odds ratio for women for promotion from Class 3 to Class 2 is 1.127 (model 5), which means that the odds of promotion for women are about 113 percent of the odds for men. We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each salary class shown. In all models, we controlled for the time that employees spent in each grade before promotion. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 42: Odds Ratios for Promotion from Class 2 to Class 1 in the Department of State’s Foreign Service, Fiscal Years 2002-2018
Odds ratio (standard error)
Control variableModel 1aModel 1bModel 2Model 3Model 4Model 5
Woman1.232***1.240***1.260***1.0661.075*
(0.0478)(0.0482)(0.0508)(0.0446)(0.0457)
Racial or ethnic minority0.799***0.792***0.841***0.869***0.925
(0.0398)(0.0395)(0.0425)(0.0446)(0.0485)
Veteran's preference0.566***0.765***0.983
(0.0412)(0.0576)(0.0770)
Hard language1.177**1.0470.979
(0.0764)(0.0691)(0.0661)
Ivy League college or university1.306***1.182***1.125*
(0.0742)(0.0701)(0.0680)
Washington, D.C.–area college or university1.157***1.110**1.114***
(0.0461)(0.0453)(0.0462)
Age at entry0.934***0.926***0.978
(0.0201)(0.0203)(0.0218)
Age at entry, squared1.001**1.001*1.000
(0.00031)(0.00032)(0.00032)
Years of government service1.130***1.158***1.009
(0.00881)(0.00920)(0.0111)
Years of government service, squared0.996***0.996***0.999***
(0.00032)(0.00032)(0.00036)
Overseas in prior year1.258***1.229***1.468***
(0.0505)(0.0503)(0.0640)
Hardship in prior year0.9391.0071.132
(0.185)(0.200)(0.230)
Long leave in prior year0.597***0.608***0.668**
(0.101)(0.103)(0.116)
Changed service0.8800.644***1.116
(0.121)(0.0902)(0.165)
Occupation
Economist0.897*0.918
(0.0584)(0.0604)
Administrative officer1.937***2.033***
(0.134)(0.143)
Passport, visa examiner1.0131.070
(0.0697)(0.0746)
Public affairs specialist2.390***2.781***
(0.168)(0.204)
Security administrator0.599***0.663***
(0.0424)(0.0474)
Secretary
Computer specialist
Information systems management0.333***0.446***
(0.0345)(0.0471)
Office service management supervisor1.0831.409**
(0.170)(0.226)
Other1.658***1.591***
(0.138)(0.136)
Duration controls
Fiscal year controls
Constant0.0087***0.0078***0.0081***0.0149***0.0202***0.0033***
(0.00145)(0.00130)(0.00135)(0.00594)(0.00822)(0.00161)
Observations39,58739,58739,58739,58739,56936,727
Legend: ✓ = controls applied, — = not applicable, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic are less likely to be promoted, while odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to be promoted. We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each salary class shown. In all models, we controlled for the time that employees spent in each grade before promotion. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Table 43: Odds Ratios for Promotion from Class 1 to Executive in the Department of State’s Foreign Service, Fiscal Years 2002-2018
Odds ratio (standard error)
Control variableModel 1aModel 1bModel 2Model 3Model 4Model 5
Woman1.0011.0001.0611.0591.079
(0.0566)(0.0566)(0.0620)(0.0630)(0.0653)
Racial or ethnic minority1.0231.0231.0551.0641.097
(0.0757)(0.0757)(0.0791)(0.0801)(0.0840)
Veteran's preference0.787**0.781**0.983
(0.0891)(0.0911)(0.118)
Hard language1.255**1.253**1.197*
(0.115)(0.116)(0.114)
Ivy League college or university1.149*1.206**1.111
(0.0824)(0.0904)(0.0846)
Washington, D.C.–area college or university1.0651.0731.093
(0.0600)(0.0609)(0.0633)
Age at entry0.891***0.885***0.923***
(0.0239)(0.0240)(0.0251)
Age at entry, squared1.001***1.002***1.001**
(0.00038)(0.00038)(0.00038)
Years of government service1.142***1.133***0.995
(0.0164)(0.0165)(0.0175)
Years of government service, squared0.997***0.997***0.999**
(0.00047)(0.00048)(0.00053)
Overseas in prior year1.316***1.333***1.566***
(0.0742)(0.0755)(0.0948)
Hardship in prior year0.7790.8120.826
(0.261)(0.274)(0.282)
Long leave in prior year0.417***0.411***0.482**
(0.125)(0.123)(0.146)
Changed service0.6850.580**0.766
(0.171)(0.147)(0.201)
Occupation
Economist0.865*0.877
(0.0757)(0.0778)
Administrative officer1.314***1.400***
(0.125)(0.136)
Passport, visa examiner1.1591.151
(0.112)(0.113)
Public affairs specialist1.0191.053
(0.105)(0.111)
Security administrator1.274**1.552***
(0.134)(0.168)
Computer specialist
Information systems management0.9651.233
(0.150)(0.196)
Office service management supervisor0.216***0.259***
(0.111)(0.134)
Other0.727**0.579***
(0.0909)(0.0771)
Duration controls
Fiscal year controls
Constant0.0094***0.0094***0.0094***0.0213***0.0225***0.0036***
(0.00226)(0.00227)(0.00226)(0.0111)(0.0119)(0.00269)
Observations22,80022,80022,80022,80022,79821,195
Legend: ✓ = controls applied, — = not applicable, *** = statistically significant at p-value < 0.01, ** = statistically significant at p-value < 0.05, * = statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Note: Odds ratios that are statistically significant and lower than 1.00 indicate that individuals with the given characteristic are less likely to be promoted, while odds ratios that are statistically significant and greater than 1.00 indicate that individuals with that characteristic are more likely to be promoted. We conducted discrete-time duration analysis using logit models to analyze the time duration (number of years) before promotion from each salary class shown. In all models, we controlled for the time that employees spent in each grade before promotion. The overall baseline population for the duration analysis represents individuals who possessed none of the characteristics indicated by the list of control variables. These analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes.

Appendix XII: Promotion Regression Results for Various Demographic Groups

Tables 44 and 45 summarize the multivariate statistical regression results (specifically, duration regression results) for our estimates of the percentage differences in odds of promotion for two groupings of racial or ethnic minorities in the Civil and Foreign Services.[136]

  • We examined odds of promotion for African Americans and non–African American racial or ethnic minorities compared with whites.
  • We examined odds of promotion for the individual racial or ethnic groups—African Americans, Hispanics, Asians, and other racial or ethnic minorities—compared with whites.

Our analyses do not completely explain the reasons for differences in promotion outcomes, which may result from various unobservable factors. Thus, our analyses do not establish a causal relationship between demographic characteristics and promotion outcomes.

In addition to presenting the estimates for the two groupings of racial or ethnic minorities, tables 45 and 46 present estimates from three regression models. All models controlled for the time that employees spent in each grade prior to promotion.

  • The first model controlled only for the racial or ethnic minority variables relevant for the grouping of racial or ethnic minorities.
    • For the first grouping, the model controlled for whether the employee was African American or a non–African American racial or ethnic minority.
    • For the second grouping, the model controlled for whether the employee was African American, Hispanic, Asian, or another racial or ethnic minority.
  • The second model clustered the standard errors on organization (for the Civil Service only) and controlled for the racial or ethnic minority variables in the first model, gender, and the following additional control variables:
    • Years of federal government experience
    • Age when entering State
    • Veteran’s status
    • Use of long-term leave
    • Graduating from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland
    • Transferring between the Foreign and Civil Services
    • Having a hardship assignment in the prior year (Foreign Service only)
    • Having an overseas post in the prior year (Foreign Service only)
    • Proficiency in a hard language (Foreign Service only)
    • Occupation
    • Fiscal year fixed effects (indicator variables representing the fiscal year)
  • The third model controlled for everything in the second model, but the data were limited to fiscal years 2011 through 2018.

Table 44 presents the summary of the regression results for our estimates of the percentage differences in odds of promotion for the two groupings of racial or ethnic minorities compared with whites in the Civil Service.

  • For the first grouping, we found statistically significantly lower odds of promotion from GS-11 through GS-15 for African Americans than for whites in fiscal years 2002 through 2018 (model 2).[137] The odds of promotion from GS-11 to GS-12, GS-12 to GS-13, and GS-14 to GS-15 were also statistically significantly lower for non–African American racial or ethnic minorities during the same period.[138]
  • For the second grouping, we found statistically significantly lower odds of promotion from GS-11 to GS-12 and from GS-14 to GS-15 for Asians than for whites in fiscal years 2002 through 2018.[139]
Table 44: Percentage Differences in Odds of Promotion for Groupings of Racial or Ethnic Minorities Compared with Whites in Department of State’s Civil Service, Fiscal Years 2002-2018 and 2011-2018
GS-11 to GS-12 GS-12 to GS-13 GS-13 to GS-14 GS-14 to GS-15 GS-15 to exec.
Model 1: Estimate for racial or ethnic minorities relative to whites for fiscal years 2002-2018
African American -19*** -50*** -33*** -34*** -32
(4) (2) (4) (6) (21)
African American and non-African American racial or ethnic minorities compared with whites
Non-African American racial or ethnic minority (Hispanic, Asian, or other racial or ethnic minority) -17*** -14*** -15*** -22** -4
(4) (5) (5) (8) (25)
African American, Hispanic, Asian, and other racial or ethnic minorities compared with whites
Hispanic -28*** -15** -20** 3 29
(5) (7) (8) (16) (45)
Asian -11 -5 -1 -40*** -1
(7) (8) (9) (10) (38)
Other racial or ethnic minority -2 -24*** -27*** -17 -72
(10) (7) (9) (17) (28)
Model 2: Estimate for racial or ethnic minorities relative to whites, controlling for gender, additional individual-level control variables, and occupation (with fiscal year fixed effects and standard errors clustered on organization), for fiscal years 2002-2018
African American -32*** -38*** -25*** -24*** -14
(5) (4) (5) (8) (30)
African American and non-African American racial or ethnic minorities compared with whites
Non-African American racial or ethnic minority (Hispanic, Asian, or other racial or ethnic minority) -17** -12** -11* -19** 3
(6) (5) (6) (8) (27)
African American, Hispanic, Asian, and other racial or ethnic minorities compared with whites
Hispanic -16* -9 -11 -1 38
(8) (9) (9) (15) (48)
Asian -25** -9 -6 -35*** -1
(10) (9) (9) (11) (42)
Other racial or ethnic minority -8 -19* -21* -16 -66
(13) (9) (10) (17) (34)
Model 3: Estimate for racial or ethnic minorities relative to whites, controlling for gender, additional individual-level control variables, and occupation (with fiscal year fixed effects and standard errors clustered on organization), for fiscal years 2011-2018
African American -34*** -39*** -24*** -35*** -47
(7) (5) (7) (11) (26)
African American and non-African American racial or ethnic minorities compared with whites
Non-African American racial or ethnic minority (Hispanic, Asian, or other racial or ethnic minority) -14 -3 -15* -11 -11
(10) (8) (8) (13) (32)
African American, Hispanic, Asian, and other racial or ethnic minorities compared with whites
Hispanic -10 -4 -18 20 72
(14) (12) (12) (24) (71)
Asian -26* -4 -6 -32 -54
(12) (13) (13) (16) (35)
Other racial or ethnic minority -4 -1 -26* -17 -58
(20) (15) (12) (24) (43)
Legend: exec. = executive, GS = General Schedule, *** statistically significant at p-value < 0.01, ** statistically significant at p-value < 0.05, * statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: We conducted discrete-time duration analysis using logit models to analyze time duration (number of years) before promotion from each GS level shown. For each model, we considered two groupings of racial or ethnic minorities. For the first grouping, we examined odds of promotion for African Americans and non–African American racial or ethnic minorities relative to whites. For the second grouping, we examined odds of promotion for African Americans, Hispanics, Asians, and other racial or ethnic minorities relative to whites. In all models, we controlled for the time that employees spent in each grade before promotion. In models 2 and 3 we also controlled for the additional individual-level variables, including years of government service; age when entering State; veteran’s status; taking long-term leave; graduating from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland; and transferring between the Foreign and Civil Services. The overall baseline population for the duration analysis comprises individuals who possessed none of the characteristics indicated by the control variables. Our analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes. Standard errors are shown in parentheses.

Table 45 presents the summary of the regression results for our estimates of the percentage differences in odds of promotion for the two groupings of racial or ethnic minorities compared with whites in the Foreign Service.

  • For the first grouping, we found statistically significantly lower odds of promotion from Class 4 to Class 3 and statistically significantly higher odds of promotion from Class 1 to executive for African Americans than for whites in fiscal years 2002 through 2018 (model 2).[140] We also found statistically significantly lower odds of promotion from Class 4 to Class 3 for non–African American racial or ethnic minorities than for whites during the same period (model 2).[141]
  • In the second grouping, we found statistically significantly lower odds of promotion from Class 4 to Class 3 for Hispanics than for whites.[142]
Table 45: Percentage Differences in Odds of Promotion for Groupings of Racial or Ethnic Minorities Compared with Whites in Department of State’s Foreign Service, Fiscal Years 2002-2018 and 2011-2018
Class 4 to Class 3 Class 3 to Class 2 Class 2 to Class 1 Class 1 to executive
Model 1: Estimates for racial or ethnic minorities relative to whites for fiscal years 2002-2018
African American -12** -25*** -20** 42***
(5) (5) (7) (17)
African American and non-African American racial or ethnic minorities compared with whites
Non–African American racial or ethnic minority (Hispanic, Asian, or other racial or ethnic minority) -9*** -15*** -20*** -11
(3) (3) (5) (8)
African American, Hispanic, Asian, and other racial or ethnic minorities compared with whites
Hispanic -5 -25*** -18** -12
(5) (5) (7) (11)
Asian -6 -3 -20** -10
(5) (6) (7) (13)
Other racial or ethnic minority -20*** -16** -26** -12
(5) (7) (10) (19)
Model 2: Estimates for racial or ethnic minorities relative to whites, controlling for gender, additional individual-level variables, and occupation (with fiscal year fixed effects), for fiscal years 2002-2018
African American -23*** -10 -8 53***
(4) (6) (8) (19)
African American and non-African American racial or ethnic minorities compared with whites
Non–African American racial or ethnic minority (Hispanic, Asian, or other racial or ethnic minority) -9** -3 -7 -5
(3) (4) (6) (9)
African American, Hispanic, Asian, and other racial or ethnic minorities compared with whites
Hispanic -11** -9 -10 -12
(5) (6) (8) (11)
Asian -9* 1 -8 0
(5) (7) (9) (15)
Other racial or ethnic minority -4 4 3 9
(7) (10) (14) (24)
Model 3: Estimates for racial or ethnic minorities relative to whites, controlling for gender, additional individual-level variables, and occupation (with fiscal year fixed effects), for fiscal years 2011-2018
African American -19*** -15* -17 -29
(6) (8) (10) (17)
African American and non-African American racial or ethnic minorities compared with whites
Non–African American racial or ethnic minority (Hispanic, Asian, or other racial or ethnic minority) -8* -9 -10 0
(4) (5) (8) (13)
African American, Hispanic, Asian, and other racial or ethnic minorities compared with whites
Hispanic -9 -15* -4 -2
(7) (8) (11) (18)
Asian -4 -4 -12 15
(7) (8) (11) (23)
Other racial or ethnic minority -11 -5 -19 -18
(8) (12) (15) (24)
Legend: *** statistically significant at p-value < 0.01, ** statistically significant at p-value < 0.05, * statistically significant at p-value < 0.1.
Source: GAO analysis of Department of State data. | GAO-20-237

Notes: We conducted discrete-time duration analysis using logit models to analyze time duration (number of years) before promotion from each salary class shown. For each model, we considered two groupings of racial or ethnic minorities. For the first grouping, we examined odds of promotion for African Americans and non–African American racial or ethnic minorities relative to whites. For the second grouping, we examined odds of promotion for African Americans, Hispanics, Asians, and other racial or ethnic minorities relative to whites. In all models, we controlled for the time that employees spent in each grade before promotion. In models 2 and 3 we also controlled for the additional individual-level variables, including years of government service; age when entering State; veteran’s status; taking long-term leave; graduating from a college or university considered Ivy League or located in the District of Columbia, Virginia, or Maryland; transferring between the Foreign and Civil Services; having a hardship assignment in the prior year; having an overseas post in the prior year; and having proficiency in a hard language. The overall baseline population for the duration analysis comprises individuals who possessed none of the characteristics indicated by the control variables. Our analyses do not completely explain why differences in odds of promotion exist. While various independent variables capture and control for many characteristics across demographic groups, unobservable factors may account for differences in odds of promotion; thus, our regression results do not establish a causal relationship between demographic characteristics and promotion outcomes. Standard errors are shown in parentheses.

Appendix XIII: Department of State Diversity Initiatives

The Department of State (State) provided us with examples of its agency-wide diversity initiatives and the intended purposes of the initiatives. In addition, State’s Diversity and Inclusion Strategic Plan for 2016 included both agency-wide and bureau-led initiatives to promote diversity. See table 46 for more details.

Table 46: Summary of Department of State’s Agency-Wide Recruitment and Career Development Diversity Initiatives and Bureau-Led Diversity Initiatives
Initiative Purpose
Recruitment
Diplomats in Residence Program

To assign 16 Foreign Service officers and specialists to university campuses throughout the United States and 10 additional recruiters based in Washington, D.C., to recruit diverse candidates for the Foreign Service.

The program works with historically black colleges and universities and Hispanic-serving institutions as well as institutions with significant minority enrollment.

Thomas P. Pickering Foreign Affairs Fellowship

To provide graduate fellowships to college seniors and college graduates, including mentoring and professional development for individuals from minority groups historically underrepresented in the Department of State (State), women, and those with financial need.

Fellows become Foreign Service officers upon completing their degrees and fulfilling Foreign Service entry requirements.

Charles B. Rangel International Affairs Program

To provide graduate fellowships to college seniors and graduate programs, including mentoring and professional development for individuals from minority groups historically underrepresented in State, women, and those with financial need.

Fellows become Foreign Service officers upon completing their degrees and fulfilling Foreign Service entry requirements.

Foreign Affairs Information Technology Fellowship ProgramTo hire diverse, qualified individuals who will receive tuition assistance of up to $37,500 annually for an information technology-related degree. Those who successfully complete the program and the Foreign Service entry requirements receive an appointment as a Foreign Service information technology specialist.
Consular Fellows ProgramTo provide fellowships to speakers of target languages to serve overseas in limited non-career appointments of up to 5 years. State plans to hire 150 to 200 consular fellows annually through 2020.
U.S. Foreign Service Internship ProgramTo provide paid internships for selected students from underrepresented groups for two summers.
Workforce Recruitment ProgramTo hire college students and recent graduates with disabilities for summer or permanent positions. The agency typically funds 8 to 10 hires each year.
Selective Placement ProgramTo recruit and retain individuals with disabilities for careers at State.
Veterans Innovation PartnershipTo promote foreign affairs career opportunities for veterans. State provides 12-month, full-time appointments and developmental opportunities to participants.
Career Development
International Career Advancement ProgramTo increase midlevel opportunities for professional development through a professional leadership development program for highly promising midcareer Civil Service and Foreign Service employees.
SES Career Development

Agency-managed Senior Executive Service Candidate Development Program

As of October 2019, more than 70 percent of the selected participants were women and 50 percent identified as a minority.

Bureau-Led Initiatives
Bureau of African Affairs Cultivating ExcellenceTo provide professional development, training, and networking opportunities.
Bureau of Administration “ABCs of A” programTo bring new hires together with senior leaders throughout the bureau on a quarterly basis for “brown-bag” type sessions.
Bureau of Diplomatic Security’s Diversity Working GroupTo support and advise the bureau’s leadership on initiatives to promote diversity and inclusion throughout the bureau and all of its directorates.
Bureau of European and Eurasian Affairs’ Driving Diversity, Growth, and Excellence ProgramFor senior leaders to mentor midlevel employees and introduce employees to the bureau’s expectations regarding diversity and inclusion.

Bureau of Human Resources

  • Work Life Wellness Council
  • Leadership Council
  • Diversity and Inclusion Council
  • To develop policies, programs, and activities that advance the goals of creating a safe workplace while increasing employee productivity, morale, and retention.
  • To promote, guide, and instill a culture of leadership through the adoption and demonstration of State’s leadership and management principles.
  • To create a workforce that is diverse in all ways, including, but not limited to, age, gender identity, race, sexual orientation, physical or mental ability, and ethnicity.
Bureau of Information Resource Management’s“Eye on IRM” programTo promote a sense of belonging among new hires and to direct communication between employees and senior leaders through periodic brown-bag events.
Bureau of International Narcotics and Law Enforcement’s Diversity RoundtableTo promote dialogue and identify actionable steps to foster an inclusive work environment in the bureau through quarterly meetings.
Bureau of Near Eastern Affairs Diversity and Inclusion CouncilTo shape the bureau’s diversity and inclusion mission and vision statements as well as its action plan.
Bureau of Public Affairs Diversity InitiativeA diversity leadership council is implementing a sponsor and orientation program for new employees, monthly professional development sessions, and targeted outreach by diverse speakers to diverse domestic audiences. The bureau also reached out to employee affinity groups as part of the Foreign Service bidding process and launched a bureau diversity intranet Page.

Bureau of Western Hemisphere Affairs

  • Diversity Council
  • WHA Leads
  • To build a more diverse and inclusive bureau where differences are valued and all employees are confident that their contributions matter.
  • To develop a bureau-wide collaborative leadership culture by empowering and inspiring all employees serving in the bureau through regular engagement with State's leadership and management principles, mentoring, and professional development.
Source: GAO analysis of State data. | GAO-20-237

Appendix XIV: Comments from the Department of State

Appendix XV: GAO Contact and Staff Acknowledgments

GAO Contact

Jason Bair, (202) 512-6881 or bairj@gao.gov.

Staff Acknowledgments

In addition to the contact named above, Emil Friberg (Assistant Director), Julia Jebo Grant (Analyst-in-Charge), Nisha Rai, Moon Parks, Justin Fisher, Melinda Cordero, Courtney Lafountain, Kathleen McQueeney, Dae Park, K. Nicole Willems, Reid Lowe, Christopher Keblitis, and Camille Pease made key contributions to this report.

References

Figures

  1. Diversity in State Department Workforce in Fiscal Years 2002 and 2018
  2. Figure 1: U.S. Equal Employment Opportunity Commission’s (EEOC) Management Directive 715 Process for Identifying and Eliminating Barriers to Federal Agency Workforce Diversity
  3. Figure 2: Numbers of Full-Time Permanent Career Employees in Department of State’s Civil and Foreign Services, Fiscal Years 2002-2018
  4. Figure 3: Proportions of White and Racial or Ethnic Minority Employees at Department of State Overall and in Civil Service and Foreign Service, Fiscal Years 2002 and 2018
  5. Figure 4: Proportions of White and Racial or Ethnic Minority Employees in the Department of State’s Civil and Foreign Services across Ranks, Fiscal Year (FY) 2018
  6. Figure 5: Proportions of Men and Women in the Department of State and Its Civil Service and Foreign Service, Fiscal Years 2002 and 2018
  7. Figure 6: Proportions of Women and Men in the Department of State’s Civil and Foreign Services across Ranks, Fiscal Year (FY) 2018
  8. Figure 7: Relative Differences in Promotion Rates and Adjusted Promotion Odds for Racial or Ethnic Minorities Compared with Whites in the Department of State’s Civil Service, Fiscal Years 2002-2018
  9. Figure 8: Relative Differences in Promotion Rates and Adjusted Promotion Odds for Racial or Ethnic Minorities Compared with Whites in the Department of State’s Foreign Service, Fiscal Years 2002-2018
  10. Figure 9: Relative Differences in Promotion Rates and Adjusted Promotion Odds for Women Compared with Men in the Department of State’s Civil Service, Fiscal Years 2002-2018
  11. Figure 10: Relative Differences in Promotion Rates and Adjusted Promotion Odds for Women Compared with Men in the Department of State’s Foreign Service, Fiscal Years 2002-2018
  12. Figure 11: Percentages of White Employees and Racial or Ethnic Minority Employees in the Department of State, Fiscal Years 2002-2018
  13. Figure 12: Percentages of White Employees and Racial or Ethnic Minority Employees in the Department of State’s Civil Service, Fiscal Years 2002-2018
  14. Figure 13: Percentages of White Employees and Racial or Ethnic Minority Employees in the Department of State’s Foreign Service, Fiscal Years 2002-2018
  15. Figure 14: Percentages of White Executives and Racial or Ethnic Minority Executives in the Department of State and Its Civil and Foreign Services, Fiscal Years (FY) 2002 and 2018
  16. Figure 15: Percentages of Executive Men and Women in the Department of State and Its Civil and Foreign Services, Fiscal Years (FY) 2002 and 2018
  17. Figure 16: Percentages of Newly Hired White and Racial or Ethnic Minority Employees in the Department of State and Its Civil and Foreign Services, Fiscal Years (FY) 2003 and 2018
  18. Figure 17: Percentages of Newly Hired Men and Women in the Department of State and Its Civil and Foreign Services, Fiscal Years (FY) 2003 and 2018
  19. Figure 18: Percentages of White and Racial or Ethnic Minority Employees Who Left the Department of State and Its Civil and Foreign Services, Fiscal Years 2003 and 2018
  20. Figure 19: Percentages of Men and Women Who Left the Department of State and Its Civil and Foreign Services, Fiscal Years 2003 and 2018
  21.  

Tables

  1. Table 1: Number and Percentag