The Global Women’s Leadership Initiative Index is a global index that measures national and subnational levels, and looks at leadership across the five sectors of government: executive, legislative, judiciary, civil service and national security. It does not rank but rather allows for a comparison of countries over time to analyze the indicators that impact women’s access to, participation and power they hold as leaders across government sectors globally.
The Women in Public Service Project (WPSP) is dedicated to closing the gap in leadership and to this end, designed the Global Women’s Leadership Initiative Index (Leadership Index) to get a systems-level view of the state of women’s leadership across governments and across the globe. With an eye towards action-oriented analysis and moving the needle on gender parity, the Leadership Index goes beyond simple percentages of women as the measure of parity and addresses the ecosystem at large to capture the nuanced reality of parity in the public sector. It does this by measuring the pathways, positions and power women hold.
The Pathways pillar addresses women’s access to positions of leadership and the factors that enable or hinder women from entering these positions. As a sub index, pathways measures the relative strength of structural factors and takes stock of the policies, practices, institutions, and dynamics that shape women’s access to leadership positions. This approach will help us understand and identify what obstacles present the most persistent barriers, and recognize the successful policies, skills and practices that enable women to attain leadership positions.
Skills attainment (Note: For the indicators in the Skills Attainment cluster, we have chosen to measure gender gaps rather than absolute levels to understand in context how women’s experience differs from men.)
Access to labor markets
Access to public sector
The positions pillar measures the descriptive representation of women in leadership positions. The notion of descriptive representation refers to the extent to which a representative resembles those being represented.  Given that 51% of the global population is female, then the notion of fair descriptive representation in public leadership should reflect that reality. The Leadership Index examines women’s representation in the executive, legislative, judiciary, public administration, and national security across national and subnational levels.
The National level refers to the central government at the highest level of an individual country; it does not include international organizations or supra-national bodies. The Subnational level refers to the intermediate and local levels of government when applicable, as not all governments include an intermediate level. The subnational level data for this index is limited to the executive branch and legislative branch as this is the data that is currently available. We included data on regional executives and mayors where data was available as well as regional legislatures and municipal councils.
In defining subnational governments, there are competing definitions of what a subnational governments entails as well as a lack of data at this level. UN agencies are currently working on these definitions while developing metrics for the Sustainable Development Goals. For these reasons, the Leadership Index chose to rely upon the existing European Commission methodology from the Gender Balance in Decision-Making project for defining intermediate and local government. We found that this methodology was comprehensive enough for us to use, provided a good working definition in the absence of a standard, was relevant to the index, and the data collection methods are high quality.
As governments are organized differently from country to country, state to state and city to city, creating a consistent comparable list of what positions qualify as leadership and which do not is nearly impossible at an aggregate level in a way we can apply consistently across governments. Therefore, in keeping with a similar approach to the European Commission’s methodology and with organizational management theorist K. Provan’s  characterization, we define leadership positions as positions that fulfill one of two criteria:
While these definitions are not identical, both hinge on the powers granted by the position either within an organization or as part of a decision-making body that may or may not be larger than a single organization or department. These definitions do not limit decision-making to set of titles, but rather takes into consideration the different structures across government sectors and government systems and are sensitive to context.
Clusters within this pillar were organized according to the five sectors of government and include metrics at the national and subnational levels.
It is not enough to achieve parity if women hold half of the positions but effectively wield less power. This pillar measures both the formal powers granted to positions and enacted power of women holding leadership positions across sectors of government to understand at a newer, deeper level the gaps that exist between men and women holding positions of leadership. We are adopting Katz and Kahn’s (1966) simple definition of power as the capacity to exert influence. Therefore we look at two distinct dynamics:
This pillar is designed to look at the challenges to achieving the goals that women face once they have reached a level of leadership. If men consistently hold positions that wield more power, have larger budgets, or are closer to focal points of power, then we cannot say that gender parity is achieved.
Relevant to the study of power and its gendered dimensions, Oxley and Fox in their 2004 paper “Women in Executive Office: Variation across American States” in Political Research Quarterly use the FPI as a tool demonstrating the FPI’s applicability in the 3 P’s to Parity framework. Oxley and Fox examine in American states the experiences of women seeking office varies from state to state using FPI in their analysis. While the FPI only reflects the power of the governor, Oxley and Fox use the FPI as a proxy for the institutional power of the executive offices in a given state. This work sets a precedent for using the FPI as a valid way of measuring power from a gendered lens which we have now integrated into the 3 P’s to Parity framework.
The greatest challenge to measuring the state of women’s leadership in the public sector is the lack of data availability, which WPSP is addressing by forming a number of data partnerships but the gaps still exist. (See “How we Manage Missing Data”)
Due to the data gaps, we include, where possible, proxies for indicators that are not yet being tracked in a standardized way on a global scale. These proxies are chosen by their demonstrated ability to predict the missing indicator or with its proven relationship with evidence provided by impact evaluation or some other experimental design.
The GWLII measures the 3 P’s to Parity by equally weighting the three pillars: Pathways, Positions, and Power. Within each of the three pillars, indicators are organized into clusters and sub-clusters (when necessary). Clusters are a statistical grouping of indicators that are normalized, scaled, and weighted to create standardized values that make possible meaningful analysis of the index. They combine and capture information from several indicators to illuminate a particular aspect of women’s leadership in the public sphere.
First, the indicators are scaled to normalize the various types of data for comparability and to determine relative high and low scores. The scores are scaled to values between 0 and 5 (5 best) using this normalization formula:
Scaled value = 5 x ( indicator value - minimum value/ maximum value - minimum value)
The maximum value for position indicators is set at 50% to reflect the 50% representation by 2050 goal of WPSP. Maximum values and minimum values for all other indicators are set by the observed maximum of the data set (which includes more countries than the 50 in this iteration of the index) or a normative maximum determined by social science reasoning. Any value higher than the maximum is capped.
The scaled indicator values are then combined using a weighted arithmetic average to form clusters. Those cluster scores are then, in turn, combined with a weighted arithmetic average to form the pillar scores. The overall index score is an average of the three pillar scores, acknowledging that each of the 3 P’s is of equal importance in the mission for parity. That average is then scaled by a factor of 20 to reach a final score between 0-100. In a perfect world, a score of 100 would mean that not only has a country reached the goal of 50% of leadership positions held by women, but that structural barriers to entry of those offices have been removed AND that women wield the same power as their male colleagues once in office. It is a theoretical maximum that we should strive for representing a deeper and more nuanced interpretation of gender parity in the public sector.
As previously mentioned, much of the data needed to fully assess the state of women’s public leadership is missing. The WPSP team seeks to address gaps through expanding data partnerships and extensive research. Where gaps are left, the index has to adapt. Rather than estimating or modeling missing data where often baseline measurements do not exist, the index calculated scores by inputting a null value for the missing point and removing the weight of the missing indicator when calculating cluster scores. This was done on a country by country basis to make sure as much data could but included without punishing countries for missing data. The only exception here is education data, where we did model missing data using multiple imputation methods. It is the case that education statistics are not always reported year on year, particularly in developed nations as they are assumed to be fairly static (e.g. once a country has achieved nearly 100% literacy, it tends to remain near 100%). Given that reliable baselines exist for these indicators and missing data could be modeled with relative accuracy, we made this decision to ensure that the Skills Attainment cluster reflects the realities of education attainment and included those estimates.
The GWLII is a global Index and as such, hopes to span all countries throughout its lifetime. With this first iteration, we settled on 50 countries - the 35 OECD countries plus 15 non-OECD countries chosen based on population, regional representation and data availability. The decision to include the OECD countries was simple given the amount and quality of data available for them. As for the other 15 countries, by choosing countries with large populations, we hope to capture as many women as possible in this first iteration while also choosing a sample representative of regional diversity.
 Derived from Pitkin’s work on representation and its definitions
 Provan, K. G. (1980). Recognizing, measuring, and interpreting the potential/enacted power distinction in organizational research. Academy of Management Review, 5(4), 549-559.
 Katz, D., & Kahn, R. L. (1966). The social psychology of organizations. New York: Wiley.
 Beyle, T. (1968). The Governor's Formal Powers: A View from the Governor's Chair. Public Administration Review, 28(6), 540-545. doi:10.2307/973331
 Oxley, Z., & Fox, R. (2004). Women in Executive Office: Variation across American States. Political Research Quarterly, 57(1), 113-120. doi:10.2307/3219838