Report No: . Toward Developing a Mobility and Gender Index . January 2024 Transport Global Practice . . 1 © 2024 The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Rights and Permissions Disclaimer: This work is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions: The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Attribution: Please cite the work as follows: “Nato Kurshitashvili, Anne Laure Humbert and Laila Ait Bihi Ouali. World Bank. 2024. “Toward Developing a Mobility and Gender Index�. ©World Bank.� All queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202- 522-2625; e-mail: pubrights@worldbank.org. 2 Table of Contents ACKNOWLEDGMENTS ...................................................................................................................... 4 ABBREVIATIONS AND ACRONYMS ................................................................................................ 6 ABSTRACT ............................................................................................................................................ 7 1 INTRODUCTION .......................................................................................................................... 8 1.1 The value added of this research ............................................................................................. 9 1.2 Limitations .............................................................................................................................. 9 2 CONCEPTUAL FRAMEWORK ................................................................................................... 9 2.1 Dimensions ........................................................................................................................... 11 Dimension 1: Gender and mobility patterns: modes, locations, timings ...................................... 11 Dimension 2: Gendered barriers to mobility................................................................................. 12 Dimension 3: Gender and mobility infrastructure ........................................................................ 14 Dimension 4: Gendered perceptions of transport.......................................................................... 15 Dimension 5: Women’s representation in the transport sector ...................................................... 15 Dimension 6: Legal and policy frameworks ................................................................................. 15 2.2 Data pre-processing and short-list of indicators.................................................................... 16 3 MEASUREMENT FRAMEWORK ............................................................................................. 19 3.1 Computation of scores .......................................................................................................... 19 3.1.1. Gap vs levels ....................................................................................................................... 19 3.1.2. Metric for the Mobility and Gender Index .......................................................................... 20 3.2 Checking for outliers and data treatment .............................................................................. 21 3.3 Aggregation and presentation of the scores .......................................................................... 21 4 VISUALIZATIONS – PRELIMINARY FINDINGS ................................................................... 22 4.1 Dimension 2: Barriers ........................................................................................................... 22 4.2 Dimension 4: Gendered perceptions of mobility .................................................................. 23 4.3 Dimension 5. Women’s representation in the transport sector. ............................................. 23 5 POLICY IMPLICATIONS............................................................................................................ 25 6 CONCLUDING REMARKS AND NEXT STEPS ...................................................................... 26 7 REFERENCES ............................................................................................................................. 27 8 APPENDIX ................................................................................................................................... 30 3 ACKNOWLEDGMENTS This exploratory research was conceived within the World Bank’s Transport Global Practice, and prepared by Nato Kurshitashvili (Senior Gender Specialist, World Bank), Anne Laure Humbert ( Co- Director of the Centre for Diversity Policy Research and Practice, Oxford Brookes University) and Laila Ait Bihi-Ouali (Assistant Professor, University of Southampton) under the guidance of Karla Gonzalez Carvajal and Binyam Reja (World Bank Practice Managers) in 2022-23. The team is grateful to Divyanshi Wadhwa (World Bank), Eduardo Zambrano (California Polytechnic State University), and Karla Dominguez Gonzalez (World Bank) for their contributions as team members, and Ronak Sunil Shah - for the visualizations. The authors would like to thank the peer reviewers Arturo Ardila Gomez, Carolina Monsalve, Matias Herrera Dappe, Rita Ramalho, and Tea Trumbic from the World Bank, and Gaëlle Ferrant (OECD) and Inés Sánchez de Madariaga (Universidad Politécnica de Madrid, UNESCO) for their insightful comments. 4 LIST OF TABLES AND HEATMAPS Heatmaps Heatmap 1 - Dimension 2: Gendered barriers to mobility ................................................................... 22 Heatmap 2 - Dimension 4: Gendered perceptions of mobility. ............................................................. 23 Heatmap 3 - Dimension 5: Women’s representation in the transport sector ......................................... 24 Tables Table 1 - Examples of scores associated with different metrics ........................................................... 20 Table 2 - Summary of steps applied to compute metrics ...................................................................... 21 Table 3 - Selection process summary for eligible variables potentially included in the MGI .............. 30 Table 4 - Conceptual framework, measurement framework and indicators ......................................... 34 Table 5 - Examples of scores for sex-disaggregated indicators using the modified minmax approach 35 Table 6 - Scores for indicators (after treatment) ................................................................................... 36 Table 7 - Scores for dimensions ............................................................................................................ 40 5 ABBREVIATIONS AND ACRONYMS BRT Bus Rapid Transit EIGE European Institute for Gender Equality GDP Gross Domestic product ILO International Labour Organization MENA The Middle East and North Africa MGI Mobility and Gender Index OECD Organization for Economic Co-Operation and Development UNDP United Nations Development Programme UNECE United Nations Economic Commission for Europe UN Women The United Nations Entity for Gender Equality and the Empowerment of Women WBL Women, Business, and the Law WEF World Economic Forum 6 ABSTRACT Although the benefits of a gender-inclusive This paper summarizes the exploratory research approach to mobility for transport conducted by the World Bank’s Transport Global decarbonization, access to jobs, and human Practice in 2022-23 to construct a Mobility and capital advancement have been increasingly Gender Index (MGI). The report presents a six- recognized globally, this topic has not received dimension theoretical framework, outlines data sufficient attention. The lack of attention to pre-processing and indicator selection gendered mobility barriers is partly due to a procedures, and describes the technical steps limited understanding of the wider benefits of taken to develop the measurement framework for inclusive transport services for development, the index. Various approaches for measuring which is caused by the absence of sex- gender gaps and levels are explored, with disaggregated mobility data highlighting gender potential aggregation within dimensions and the inequalities. presentation of scores. One of the obstacles to this gender-inclusive The report includes visualizations of some of the approach is the absence of a global gender dimension scores as heatmaps and highlights key indicator to track gender-based inequalities in findings. Finally, it acknowledges data gaps and mobility across countries and time. The lack of outlines the necessary next steps to develop the such an index (i) hinders policymakers and MGI. development agencies from prioritizing this issue, setting project, program, and policy priorities, and monitoring performance, and (ii) discourages efforts to improve the quantity and quality of sex-disaggregated data related to mobility. 7 1 INTRODUCTION Addressing mobility challenges that both women and men but especially women (and girls) experience has both intrinsic and instrumental value: intrinsic value, as the ability to move around represents a basic freedom for everyone, and instrumental, since mobility is critical to accumulate human capital by, for example, accessing health and education, and to get desired jobs. There is also an environmental imperative for addressing gendered mobility barriers. Globally, women rely more heavily on public transport and walking than men. However, women’s mobility patterns are often not a matter of preference but necessity. Care responsibilities, reduced access to a car, and less disposable income shape women’s transport choices and have an unintended, albeit the environmentally desirable result of a lower carbon footprint than men. This means that without interventions to make transportation more amenable for all, and especially for women, an increase of women in the paid workforce could see their use of cars converge with men’s use over time. In addition, while women’s lower carbon footprint may be desirable environmentally, their current travel patterns, which are localized, are also barriers to their economic independence and their full participation in economic life. In this context, adopting a people- centered, inclusive approach to mobility is vital not only for women’s and girls’ social and economic empowerment but for a just transition to the decarbonization of transport (Munoz-Raskin et al, 2022). Although the benefits of a people-centered approach to mobility for transport decarbonization, jobs, and human capital advancement have increasingly been recognized within the World Bank and globally in the development discourse, this topic has not received the utmost attention it deserves. The lack of adequate attention is partly due to a lack of a fuller understanding of the wider benefits that gender- responsive transport service brings to development, which itself is caused by the lack of sex- disaggregated mobility data, which would highlight gender gaps in mobility. One of the stumbling blocks is the absence of a global gender indicator that would track gender-based mobility issues across countries and time. The lack of such a composite index (i) prevents policymakers and development agencies from drawing adequate attention to this issue, setting project, program, and policy priorities, and benchmarking and monitoring performance across countries and times, and (ii) hampers enthusiasm to improve the quantity and quality of data collected in the area of gender and mobility. Despite the dearth of sex-disaggregated mobility data, in 2022-23, the World Bank’s Transport Global Practice (GP) undertook analytical work to explore the feasibility of creating a global Mobility and Gender Index (MGI), which would track progress in reducing gender-based inequalities in mobility across countries and time. The work resulted in a conceptual framework with six broad analytical categories (or “dimensions�) that could be used to underpin the empirical measurement of mobility from a gender perspective and provided an overview of a shortlist of indicators that could be included in a corresponding measurement framework to support the proposed six dimensions and to build the desired index. The conceptual framework is summarized in the chapter 2 of this Paper. The study concluded that due to data limitations at the time of research, an index measuring gender inequalities in mobility cannot be computed with the available open-source data, either for the world or for any specific region with more advanced data availability. The work determined that there are not enough data for most countries and for most of the indicators of interest. Instead, the team (i) constructed scores for five out of the six dimensions using the publicly available limited data as well as the Gallup dataset, which the team purchased since the latter includes a number of relevant indicators in the area of gender and mobility, and (ii) is currently developing a dashboard to visualize these scores. The chapter 3 of the paper provides the measurement framework detailing steps from the computation of dimension scores to their aggregation. The chapter 4 provides some of the preliminary visualizations of the dimension scores, and the chapter 5 concludes with the proposed next steps. 8 1.1 The value added of this research This report presents an initial attempt to construct a Mobility and Gender Index (MGI), though these efforts have been hindered by significant (gender) data gaps. Nevertheless, the report: • Pioneers a robust cross-sectional measurement framework of gender and mobility that has been subjected to rigorous theoretical and empirical validation for the first time. • Provides a platform for various stakeholders, ranging from policymakers to development agencies, to identify the areas where (gender) data gaps exist. • Advocates for more consistent and higher quality data collation. Given these characteristics, this innovative instrument can serve a dual role as a diagnostic mechanism as well as an instrument for advocacy. 1.2 Limitations This research faced several limitations. Firstly, our work was unable to go beyond a binary understanding of gender, as the data were only available for the sex categories of female and male. We recognize the inherent complexity and range of gender identities and are conscious of the limitations imposed by this binary perspective. Secondly, we were unable to integrate an intersectional perspective into our analysis. We are fully aware of the systematic disadvantages women encounter in society, which are created and perpetuated by power inequalities and that intersect with gender – aspects that the MGI does not currently capture. Despite this intersectional data gap, measuring inequalities based on sex provides useful insights that cut across other forms of disadvantage, helping improve mobility for minoritized and marginalized groups, even where data are unavailable. Our study also incorporates Gallup data, which are not publicly accessible and requires an annual subscription. While the cost of this resource is not considerable, securing annual funding represents a potential challenge for the sustained progression of this project. Also, as it currently stands, this work does not allow comparisons over time. However, this is a feature to consider in future iterations, once a more robust measurement framework is established and, importantly, when more indicators become available. Finally, our study does not currently employ ‘big data’. We examined traditional data sources, such as surveys, censuses, and administrative records, along with alternative ‘big data’ sources, like Facebook API, satellite data, and individual mobile phone data. While these sources offer information on specific movement patterns of individuals, they present data at the micro-level, limiting our ability to aggregate it at the national level without more advanced mapping analysis tools. Therefore, big data was excluded from the current research. However, as we secure more data and analytical tools and as our work on the index advances, the potential for incorporating big data can be further explored. 2 CONCEPTUAL FRAMEWORK Research on transport and mobility has historically been gender-blind in that it has failed to consider how women and men (and other gender groups outside this binary) might have different experiences (Law, 1999). However, the new perspectives for understanding the role of mobility in recent years have brought greater concerns for existing gender inequalities in the sector, which encouraged quite a lot of research in this area (Alam et al, 2022). Yet, the studies considering a gender perspective typically focus on women, thereby conflating gender dynamics with the experiences of women alone. Men (and other gender groups) also have a gender, and as such, have a gendered experience of transport and mobility. Also, most studies examine topics related to the productive sphere – i.e., commuting from home to a paid workplace – and less so other topics related to caring activities, which disproportionally fall to women, and thereby not representing women’s mobility and transport perspectives adequately. 9 A feminist perspective on gender and mobility is needed to better unpack what might be the gendered social and cultural geographies of mobility. Sánchez de Madariaga (2013) calls for looking into the ‘mobility of care,’ that is, giving a more balanced focus on paid and unpaid work when researching or developing policies on transport and mobility. Doing so would not only benefit women – though their experiences would be less invisible as a result – but would benefit all, as better care within the family or the community will provide better societal outcomes for all, irrespective of their sex or gender. Examining social and cultural geographies of mobility cannot be done without also looking into the different needs and patterns across industrialized and developing countries, as well as intersectionalities between sex, gender, and other factors such as age, ethnicity, disability, race, location, sexual orientation, or income. For instance, the needs of a woman with a disability living in a rural part of a country with poor public transportation will be qualitatively different from those of a non-disabled woman in a city with at least some options for urban transportation. In a similar vein, an elderly low- income man with limited mobility would have more acutely pronounced mobility needs than an able- bodied young man with more resources to afford private transportation (Kurshitashvili et al., 2022). It is worth repeating that women’s and men’s daily lives can differ significantly. Differences apply globally, with variations as to their intensity rather than any discernible differences in how they play out in different countries, as illustrated by various global measures of gender inequalities, e.g., World Economic Forum’s (WEF) Global Gender Gap Index, the Gender Inequality Index of the United Nations Development Programme (UNDP), and the Gender Equality Index of the European Institute for Gender Equality (EIGE). Gender inequalities can be considered as outcomes – e.g., employment, education – that result from social norms and institutions present in different countries and that contribute to gender inequalities. Social institutions influencing women’s and men’s mobility often include but are not limited to household division of labor, norms about what is permissible for women and men to do, or laws that discriminate on the basis of sex or gender, e.g., the Social Indicators and Gender Index of the Organization for Economic Co-Operation and Development (OECD). One of the starkest gender inequalities that affects mobility is the division of time and responsibilities for paid and unpaid work, as it means that women and men have different needs and that mobility options respond unequally to these different needs. Gender inequalities are best understood as influenced by factors that play out at the micro- (e.g., education, income), meso- (e.g., family as an institution, rural/urban location), and macro-levels (e.g., national policies and legal frameworks). The topic of gender and mobility is no exception, and it is therefore useful to examine it as such (for an ecological framework on how transportation systems interact with micro-, meso- and macro-level factors, see Alam et al., 2022). These factors shape women’s and men’s mobility needs, and in turn affect equal access to employment, education, health, or leisure. To measure gender inequalities in relation to transport and mobility, this paper first proposes a six- dimension conceptual framework that explains how gender affects mobility across these micro-, meso- and macro-levels, and that is subsequently used to inform the development of a measurement framework for the MGI. The framework examines how gender is a strong determinant of mobility patterns (modes, location, timings), the barriers women and men face in relation to mobility, then discusses gender in relation to the mobility infrastructure, as well as women’s representation in the transport sector before outlining how these are framed by norms, policies, and perceptions at the societal level. 10 Figure 1: Dimensions informing the conceptual and measurement frameworks. Legal and policy frameworks Women’s representation in the transport sector Barriers Perceptions Infrastructure Modes, locations, timings Source: authors’ work. 2.1 Dimensions Dimension 1: Gender and mobility patterns: modes, locations, timings Gender is a strong determinant of mobility choice and constraint, particularly as it has a strong effect on travel patterns expressed in transport modes women and men use, locations they go to, and timings of their travel. While the report refers to women and men’s differences in a general sense, there are certainly heterogeneous patterns within each group, and indeed other gender groups besides women and men. • Modes: On average, based on the limited available data, women make a higher proportion of trips using public transportation and/or on foot. Percentages may differ depending on the level of private car and motorcycle use in the country, but this general pattern remains consistent in most of the countries where the data are available. By contrast, men make more trips by car, motorcycle, and bicycle. Women are more likely to use cheaper forms of public transport (buses may be cheaper than trains for example) (Duchène, 2011). However, to remain mobile, some women end up paying more than men for alternative forms of transport such as taxis and ride- sharing options, even though men on average have higher salaries and higher employment rates. On the other hand, men are more likely to be motorized and get access to private vehicles in households where their number is limited. Men are also more likely to use bicycles, as caring responsibilities prevent women with children from using cycling as a transportation mode (Shaw et al., 2020) in addition to cultural constraints as well as affordability and safety concerns that hold them back from cycling. • Locations: Due to greater care responsibilities, women are more likely to make shorter trips, often for multiple purposes mirroring their multiple roles, for example, dropping children at school before travelling to a workplace, shopping, or health-related trips (Song et al., 2019; 11 Zhao et al., 2015), often referred to as “trip chaining.� Trip chaining can be problematic in that it means reliance on a transport network that may not be able to respond to women’s complex travel needs to access scattered locations leaving them spatially entrapped and time poorer and ultimately limiting their employment options. It also means that fare structures need to be able to cater to the need for women to make multiple stops without additional expense (Booth et al., 2000). Because of greater reliance on public transport, women are also more constrained in relation to their spatial reach, not only in the places they can access but how far they are able to travel. • Timings: Timings when women and men tend to use transport are affected by care responsibilities, but also by concerns over safety and security and transport availability. Women are likely to avoid the use of public transport at night, or more likely to require transport in early or mid-afternoon for school pick-ups. Women are also more likely to avoid peak-times travel, where there are fewer services limiting their job opportunities and curtailing their freedom of movement. Dimension 2: Gendered barriers to mobility The transport related barriers to mobility from a gender lens have been summarized using the five broad (and sometimes overlapping) categories outlined in Box 1: availability, affordability, acceptability (social and cultural), accessibility (physical) and security and safety of public and private transport (Alam et al., 2022). Box 1: gendered barriers to mobility a. Availability: This refers to connectivity and coverage of the transportation system. In rural areas availability of transport infrastructure may be more salient than availability of transport services, while in urban areas transport services may be more salient. For people who own vehicles, availability of public transport may be less salient than people who don’t own vehicles and rely on public transport to move around. b. Affordability: This refers to travel costs and the extent to which people can afford to travel when and where they want. It includes both the direct (financial) cost, as well as, the opportunity cost of potential consumption that is foregone in exchange for mandatory trips. The same travel cost may be affordable for some people but not others. c. (Social and Cultural) Acceptability: This refers to the quality of transportation infrastructure and user comfort and reliability. It also includes differing judgments, attitudes and behavioral reactions to women and men traveling and using various modes of transport. People travelling with children, or the elderly may view the comfort and reliability of the same transportation system differently than those travelling alone. a. (Physical) Accessibility: This refers to the ease with which an individual can access opportunities (e.g., employment, health care, education, etc.). Able bodied people and people with disabilities may view the accessibility of the same transportation system differently. b. Safety and Security: This refers to safety from crime and perception of security when using transportation systems. Source: Alam et al. (2022, p. 15) Availability: Availability can be gendered in that public transport systems may be under-developed though they are in greater demand from women than from men, overall. The locations at which transport is available may also not cater equally to women and men, due to their different mobility patterns in relation to locations and timings. Here whether transport options have adequate first- and last-mile connectivity is of particular importance affecting women and girls’ mobility more than that of men and 12 boys. For example, often public transport either does not serve any internal routes in the communities, is only available at certain times – only during the daytime, for example – or the service is excessively infrequent. This results in women walking long distances or taking informal modes of transport to cover internal routes, which can be expensive and unsafe. A lack of evening transport also means that many women working in the services sector are less likely to take night-shift jobs (Kurshitashvili et al., 2021). Also, and importantly, limited availability of formal transport can push women towards using informal services, which tend to be inconsistent in their tariffs and schedules, bringing additional risk factors to women’s safety and complicating their already complex travel (Dominguez Gonzalez et al., 2020). Affordability: Women’s access to economic resources is more limited than that of men, constraining women’s affordability of transport. Furthermore, gendered mobility patterns can mean that women pay more such as when they trip chain or opt for different modes or timings to ensure their safety from gender-based violence. Because of the lower availability of reliable or safe transport, women can be left with having to pay for taxis or other forms of private transport. This has been termed the ‘pink transport tax’, which is a gender-based price discrimination – an extra amount that women as well as other categories of users such as older people or people with disabilities pay for transportation (Dandapat & Maitra, 2020; Mejía-Dorantes & Soto Villagrán, 2020). The 2018 study conducted by the New York University’s Rudin Center for Transportation (Kaufman et al., 2018) found that in New York City the median monthly extra travel costs for women incurred for caretaking reasons reached up to US$50 more than their regular travel expenses, while men incurred zero additional costs. The same study also identified that safety concerns led women to modify their travel behaviors, which added another US$50 to their travel costs. Women are particularly affected by transport systems that charge flat rates per line or per journey and do not offer integrated fare systems. This is because as women travel multiple destinations within one trip, they often pay numerous single fare tickets during a chained trip. Acceptability (Social and Cultural): Acceptability is related to the social and cultural barriers that women encounter when traveling. Societal attitudes – often shaped by religious or cultural norms – can constrain the acceptability of women’s mobility. For example, family members, and in particular men, can discourage the use of transport options such a walking, cycling or public transport. This can be related to whether mobility is seen as appropriate and/or safe. In some countries where women’s access to public transport or walking is low, driving or being driven become important prerequisites to participation to work, education, leisure, or access to health services (Kerzhner et al., 2018), which not only limits women’s and girls’ freedom of movement but hampers the efforts to decarbonize transport and accelerate energy transition. Where women increase their participation in work, education or leisure in the public sphere, mobility itself can be understood as a defiance to traditional gender norms (Neupane & Chesney-Lind, 2014), and punished as such. Physical Accessibility: Women’s transport options are hindered by accessibility, including factors such as long walking distances to stations or other terminals, which exacerbates existing concerns about safety and security. Access to transport can be hindered by gendered factors such as travelling with children and dealing with a pushchair where there is no step-free access, where transport is overcrowded. Safety and personal security: Globally, real and perceived threats of gender-based violence – that is physical (including sexual) and verbal harassment – in transport and broadly in public space represent one of the biggest mobility barriers affecting women and girls, disproportionately more than it affects men and boys (Ball & Wesson, 2017; Freedman, 2002; Hickey, 2014; Horii & Burgess, 2012; Mungai & David, 2006). The most common forms of violence against women and girls on public transport range from leering, winking, and offensive gestures to unwanted touching/groping and pressing against women and girls as well as indecent exposure and assault. These transgressions are pervasive in different contexts. In some countries, sexual harassment is so widespread that women report being 13 accustomed to ‘routine groping’, though they remain concerned about possible escalation to more serious assaults (Mansoor & Hasan, 2016). Safety concerns disproportionally impact women who have lower incomes. A qualitative study (Dominguez Gonzalez et al., 2020) on women’s mobility barriers conducted in three Latin American cities, Rio de Janeiro, Buenos Aires, and Lima, revealed that low-income women prioritized safety over affordability. If they could afford the costs, women were willing to pay more for transport modes that provided a greater sense of safety. In other focus groups in the same study, women who were economically worse off were not able to exert this choice even if it was their preference. Women tend to adapt and/or constrain their mobility patterns to navigate lower safety and security whether these are real or perceived (Stark & Meschik, 2018). Factors that come into consideration are waiting times, transfer in isolated locations, avoiding journeys at certain times (anxious about safety at night) or levels of occupancy (Chowdhury, 2019). Women weigh up safety concerns with timings frequently, prioritizing the former over the latter (Shirgaokar, 2019). They are also more likely to use strategies such as traveling in groups, or taking more expensive modes of travel, as reported earlier, out of concerns for safety or even avoid travel altogether. Timings and locations create various degrees of exposure to violence, e.g., sexual harassment most prevalent in crowded rush hour (Hutson & Krueger, 2018; Lewis et al., 2021) as it offers perpetrators proximity and anonymity with little risk to be caught (Neupane & Chesney-Lind, 2014), but serious sexual assaults at times where women are more isolated. Dimension 3: Gender and mobility infrastructure Infrastructure: Women’s mobility challenges are often mistakenly understood to stem only from the design and operations of public transport service. In fact, they also concern the transport infrastructure and broader public space which shapes people’s decisions about if, when, how and when to travel affecting their travel patterns. For example, as women walk more and depend more on public transport, they are disproportionally impacted by the lack of broader public transport infrastructure, for instance, poor pedestrian pathways, a lack of street lighting or lighting at bus stops, or safe road crossings as well as generally inconvenient access to transport facilities. Consequently, this inadequate public transport infrastructure renders women’s daily journeys – which are already time-constrained in view of their multiple responsibilities – even more difficult. For many women and girls though, the act of venturing into a public space – such as passing through a market or walking down a crowded or deserted street – produces anxiety and affects their perception of safety. To alleviate this anxiety, considerations such as safety, accessibility, comfort, and a sense of belonging to the community are equally as important to public safety as affordable and reliable public transport service. Ensuring a safe environment, is an integral part of creating more amendable door to door mobility experience. Some of the factors that create safety and comfort include good street lighting, easy to read signage, general visibility of the area, clear and well-kept paths, and mixed used areas (Kurshitashvili et al., 2021; Soraganvi, 2017). Technology: Forms of on-demand transport such as bicycle sharing systems as well as demand- responsive public and private mobility services (picking up and dropping passengers in response to their needs without running a fixed schedule) have increased, enabled by technological developments. It is therefore both transport investment but also innovations that are shaping mobility. Transport infrastructure is shaped by the technological tools available to women and men, such as real time information or providing real-time bus information to not only improve service quality but also alleviate security concerns particularly among women. In recent years, mobile apps have played a key role in allowing users to determine the duration of waiting time. The use of artificial intelligence for transport services is also promising to respond to demand, and thus develop demand-responsive mobility services. However, it is important to consider the digital divide and that access to the internet and smart- 14 phone continues to be lower among women than men, constraining women’s access to these l innovations (Singh, 2020). Smart mobility is therefore contingent, amongst others, on equalizing women’s and men’s access to and use of economic and financial resources, including bank accounts and payment cards (Masikini & Baruah, 2020). Dimension 4: Gendered perceptions of transport Assessing solely the nature of the transport infrastructure and services available paints only a partial picture of women’s mobility: women’s perceptions of comfort, price, availability, safety, and enjoyment from departure to the end of their journey reflect the performance of a transport network ‘as a lived experience’. Perception issues can undermine any serious attempt to provide quality and accessible transportation infrastructure to women (Dominguez Gonzalez et al., 2020). Some studies show statistically significant differences in perception of safety of public transport and more broadly, public space among women and men significantly affecting mobility patterns and travel behaviours of women (Coppola & Silvestri, 2021; Polko & Kimic, 2021). Women’s perception of transport issues can fail to be captured in data, for reasons ranging from underreporting (Neupane & Chesney-Lind, 2014) to the lack of sex-disaggregated data. Nevertheless, a substantial strand of literature indicates significant gender differences: these studies mostly focus on urban spaces, and on gender perceptions of urban spaces and transport infrastructure (AitBihiOuali et al., 2019; Loukaitou-Sideris & Fink, 2008; Yavuz & Welch, 2010). They are helpful to understand the nature and extent of the barriers that impair women’s social and economic inclusion. Nevertheless, most studies rely only on local data, and there is a scarce number of studies focusing on several countries or several continents. Dimension 5: Women’s representation in the transport sector Gender can remain absent from transport policymaking especially when there is a low representation of women in the transport sector including as strategists, planners, or policymakers, but also as service providers. Women for example remain greatly under-represented in most of the customer front facing roles such as among bus and taxi drivers though women represent about half, if not more of their customer base. Though women’s representation and involvement does not guarantee that gender perspectives will be addressed, it is nevertheless more likely that this will lead to a greater recognition and reflection of their needs in transport policymaking. Employing more women in the sector at all levels from boardroom to platform can lead to more gender-responsive transport service development by bringing women’s perspectives to the decision-making table and by gender diversifying the transport service, which can influence the adoption of more gender-inclusive transport service and infrastructure planning, design, and operations. Of note is that many experts believe that only when women comprise 30 percent or more of an organization, they can start affecting change. Dimension 6: Legal and policy frameworks The decision of women to move for different purposes, including access to economic opportunities is influenced by gender norms and institutions that dictate if, when, where and how women and girls can or cannot move. These restrictive norms are often reinforced further by laws that discriminate based on gender, such as for example, laws in some countries that restrict women travelling in the same way as men. Transport policies can play a key role in combating gender inequalities by tackling discriminatory social norms and institutions that negatively impact on women’s mobility and instead emphasize the benefits to society that can be gained by improving mobility for both women and men (Akyelken, 2013). 15 2.2 Data pre-processing and short-list of indicators At the start of the data collection process, a review of existing data sources on mobility and gender has been undertaken. The full initial set of eligible data is presented in Table 3 in Appendix. This first set of eligible variables was then put through a list of selection criteria. The main principles underpinning the data selection process are listed below: 1. Ensuring sufficient data quality: The data potentially eligible to an inclusion into the final indicators was, for the most part, coming from open data sources published by reputable international institutions (e.g., OECD, World Bank, ILO). This choice was motivated by the necessity to ensure that the data collection would be peer-reviewed internally and documented externally and thus lead to variables meeting minimum data quality requirements. 2. Principle of replicability: The use of open-source data allows for a transparent replication of the measurements undertaken in this project. However, we also deemed necessary to allow for these indicators to be replicable also over time, and for more countries as wider data collections could be expected to be undertaken in future years. On the basis of these main principles, several filters have been put in place to further select the variables. - Filter 1: Geographic coverage: since the aim of this research is to document global gender and mobility patterns, data which were not available for a sufficiently large number of countries were discarded. In many cases, the variables were only available for Europe (and the European Union countries more specifically). In the final set of variables, selected data are available for at least two regions, if not globally – with the exception of Dimension 3. Dimension 3 features two indicators which are available only for Europe and the collection was discontinued at the time of writing. These indicators measure the “Average absolute accessibility in urban areas (distance: 15mn/4km)� for two travel modes (walking and public transport). Given the relevance of these two indicators, the decision was made to retain and visualize these indicators on the dashboard to encourage future data collection with wider geographic coverage. - Filter 2: Data periodicity and cut-off period: All the variables obtained from data collection before 2018 were deemed ineligible and discarded from the analysis. The underlying assumptions behind this five-year data cut-off period is (i) a rapid increase in investment in transport infrastructure and services in recent years globally, and (ii) older data may not depict the reality of transport systems and individual mobility patterns. In most cases, eligible data are collected at a yearly frequency. - Filter 3: Data availability: Ensuring the future replicability of results is a central objective of this research. Therefore, open datasets have been prioritized to ensure that the data were collected following a transparent and documented process and that this data remain freely available in the years to come. However, the team was facing the necessity to include sex- disaggregated data, which was not available worldwide through open data sources, and the Gallup World Poll data was used as it was meeting all the other requirements with its worldwide geographic coverage, its yearly data collection frequency, and the provision of sex- disaggregated data. Three variables are used for indicators present in dimension 2 (Barriers) and dimension 4 (Perceptions). - Filter 4: Redundant variables, the presence of alternatives and conceptual relevance: Since the first selection of eligible variables, all eligible data have been further screened to test 16 whether there were better data alternatives and/or whether some variables were redundant with others in the present list. o Redundant variables: The OECD variable measuring women’s feeling of safety when walking alone at night, collected from the Better Life Index, was eliminated from the list of eligible variables. The Gallup World Poll’s similar variable (“Do you feel safe walking alone at night in your community?�) superseded the OECD variable since the former has worldwide geographical coverage. o Conceptual relevance: The variable of the World Bank’s “Women, Business and the Law� index (“Is there legislation specifically addressing domestic violence?�) which the team considered for the dimension 6 (policy and legal frameworks) was discarded due to its conceptual irrelevance to the phenomenon being measured. Moreover, two other variables (Gallup World Poll’s “In the city or area where you live, do you have confidence in the local police force?�, and International Transport Forum’s “Road Accident Fatalities�) were dismissed from the analysis since they were deemed redundant with their dimension. The Gallup variable on the confidence towards the police was considered less relevant when it comes to gender and mobility compared to the self-rated feeling of safety whilst walking at night. Moreover, the road accident facilities number was more of a safety indicator, thus better suited for the Barriers section – however, it did not meet the sex- disaggregation requirement of that section, and had to be discarded as a result. o Presence of alternative variables: The European Commission variable which contains information on the share of women employed in the transport sector (Dimension 5) was replaced with ILO data. The share of women employed in the transportation and storage sector (produced by ILO) superseded the European Commission variable with its worldwide geographic coverage and its yearly data collection frequency. - Filter 5: Compatibility with theoretical framework requirements: The conceptual framework has implications for the selection of relevant data, and whether or not dimensions require sex- disaggregated data. After a review, the following variables have been discarded: o Dimension 2: Gendered barriers to mobility: This dimension requires data to be sex disaggregated. The variable of the World Bank’s “Women, Business and the Law� index (“Can a woman travel outside her home in the same way as a man?�) is not sex disaggregated. Thus, it was eliminated, and priority was given to another indicator available for women and men (“Do you feel safe walking alone at night in your community?�). - Filter 6: Statistical tests: Further data selection was undertaken afterwards to ensure that the statistical structure of the data was fit for aggregation at the dimension level. The statistical structure tests can be divided into two main types: o Directionality tests: Directionality should be maintained within each dimension for variables to be kept – i.e., the correlations should remain positive between variables at the dimension level. o Correlation tests: Variables are eliminated when the correlations are too high (i.e., >0.9). - The main outcomes of the statistical tests led to the main eliminations for the reasons detailed below: o The average accessibility of driving variable (“Urban access – 15mn / 4km – Driving�) was deleted since this variable is negatively correlated to the average accessibility of 17 walking variable and biking variable (“Urban access – 15mn / 4km – Walking�, and (“Urban access – 15mn / 4km – Biking�). The average accessibility of driving variable is also uncorrelated to the average accessibility of public transport variable (“Urban access – 15mn / 4km – Public Transport�). The indicators focusing on walking and public transport have been prioritized over those related to the other transport modes due to their correlation and relevance to the sustainable mobility agenda. o The variable on average accessibility of biking (“Urban access – 15mn / 4km – Biking�) was dropped due to its high (>.9) correlation with the average accessibility of walking variable. 18 3 MEASUREMENT FRAMEWORK The integrated database was used to develop a measurement framework. This was done following an iterative process, whereby the correlation structure of indicators within and across domains was examined and informed decisions on possible sets of indicators based on both conceptual and statistical considerations. The conceptual framework, measurement framework and the associated indicators are presented in Table 4 in Appendix. The table shows the extent to which it is possible to provide an empirical measurement of the conceptual framework, and therefore where there are (gender) data gaps. 3.1 Computation of scores The construction of a composite indicator relies on the aggregation of different indicators. However, it is essential to ensure that these indicators are all interpreted in the same way (directionality) and are measured on the same scale (comparability). These two principles are used in the development of metrics, bound between 0 and 1, where the upper/lower bounds represent the best/worst situation relative to other countries for which data are available (though the arbitrary meaning given to them varies across existing indices, with some equating the best situation with 0 and other 1). These metrics allow for comparisons between different indicators and for further aggregation of these indicators into dimensions. The choice of metrics is not a simple one, and different choices have different implications. In this section, several approaches are reviewed – with examples from selected indices used as illustrations – to present the choices available for the Mobility and Gender Index. 3.1.1. Gap vs levels Metrics can focus on measuring gaps or levels. The gender gap approach is the one taken by the World Economic Forum’s Global Gender Gap Index. This index converts all indicators to gender ratios (𝑊/𝑀), as long as women’s achievements remain lower than that of men’s, and caps at 1 otherwise. The level approach can be seen in the construction of the UNDP’s Gender Development Index. This index relies on the creation of dimension indices, computed separately for women and men, on the basis 𝑊 – 𝑚𝑖𝑛𝑊 𝑀 – 𝑚𝑖𝑛𝑀 of the min-max procedure: minmax𝑊 = 𝑚𝑎𝑥 and minmax𝑀 = 𝑚𝑎𝑥 , with W and M the 𝑊 − 𝑚𝑖𝑛𝑊 𝑀 − 𝑚𝑖𝑛𝑀 mean values for women and men respectively. The minimum and maximum values are taken across countries in a given year. The min-max approach provides metrics that are much more easily interpretable, particularly for the purpose of policymaking, which relies on ensuring the higher possible level of achievement, while minimising gender gaps (Humbert & Hubert, 2021). The Europe Institute for Gender Equality’s Gender Equality Index adopts a more complex metric, which seeks to adjust gender gaps using the relative position of a country. However, this approach has been criticized for making it difficult to communicate the part of the score that is attributable to gender gaps and to levels of achievement (Permanyer, 2015), and for the arbitrary allocation of importance given to gender gaps and levels of achievement in the calculation of the index (Humbert & Hubert, 2021). As the Mobility and Gender Index aims to combine indicators that are sex-disaggregated with indicators that are not, it is not desirable to adopt a gender gap approach. Further, the gender gap approach is always problematic when used alone, as it fails to distinguish between a small gender gap where both women and men are doing badly from a small gender gap where women and men are performing well. Finally, whether a ratio or percentage point difference is adopted generates very different metrics, which are difficult to interpretate and communicate. This is illustrated in Table 1. 19 Table 1 - Examples of scores associated with different metrics Satisfaction with public 𝑊 /M 𝑀 − 𝑊 minmax𝑊 transport infrastructure 𝑊 10% Case 1 0.5 0.1 0.1 𝑀 20% 𝑊 40% Case 2 0.5 0.4 0.4 𝑀 80% 𝑊 70% Case 3 0.875 0.1 0.7 𝑀 80% 𝑊 1% Case 4 1 0 0.01 𝑀 1% 𝑊 99% Case 5 1 0 0.99 𝑀 99% 𝑊 – 𝑚𝑖𝑛𝑊 Note: Minmax calculated using theoretical min/max of 0/100, and minmax𝑊 = 𝑚𝑎𝑥𝑊 − 𝑚𝑖𝑛𝑊 3.1.2. Metric for the Mobility and Gender Index The metric used to compute scores for the Mobility and Gender Index is provided below. This metric focuses on levels rather than gaps and is based on the minmax approach. An advantage of this metric is that it can used and interpreted in similar ways for indicators that are sex-disaggregated and for those that are not, and makes it possible to aggregate indicators that are disaggregated by sex or not. For sex-disaggregated indicators, the harmonic mean is first computed between the levels for women and men. Second, the minimum and maximum across countries and wo/men are taken as the interval of reference, e.g., 𝑚𝑖𝑛(𝑊, 𝑀) and 𝑚𝑎𝑥 (𝑊, 𝑀). The use of the harmonic mean penalizes gender inequalities that may exist, and is therefore a way to account for gender gaps while keeping the emphasis on levels. For example, in Table 5 in Appendix, countries A and D both have an average of 15%, but country A achieves a lower overall score because it has a wider gender gap (13% for country A, compared with 15% for country D). The same mechanism applies to countries C and E, showing that the approach is invariant to the level themselves. The interval used for the minmax is calculated across women and men to represent the potential convergence to the highest achievement for the max (e.g., women should aspire to the same level of achievement as men), while the min captures the relative lower position of women in society generally. Finally, the scores obtained are rescaled to ensure they fall on the interval 0 to 1. For indicators that are not sex-disaggregated, the minmax is applied normally as: minmax𝑋 = 𝑋 – 𝑚𝑖𝑛𝑋 𝑚𝑎𝑥𝑋 − 𝑚𝑖𝑛𝑋 , with X the mean of the variable. The only exception is when an indicator represents a share, in which case a cap is applied at 50%, e.g., the parity point, before applying the minmax transformation in the usual way. The computations associated with each type of indicators are summarized in Table 2. 20 Table 2 - Summary of steps applied to compute metrics Non-disaggregated indicators Sex-disaggregated indicators Share of women D1.1 D2.1 D5.1 D3.1 D4.1 D3.2 D4.2 2 Obtain 𝑚𝑖𝑛 Compute 𝐻 = 1�𝑥 + 1�𝑥𝑀 Cap 𝑥𝑊 at 50% 𝑊 Obtain 𝑚𝑎𝑥 Obtain 𝑚𝑖𝑛𝑊,𝑀 Obtain 𝑚𝑖𝑛𝑊 𝑥−min Compute Obtain 𝑚𝑎𝑥𝑊,𝑀 Obtain 𝑚𝑎𝑥𝑊 𝑚𝑎𝑥−𝑚𝑖𝑛 𝐻−𝑚𝑖𝑛𝑊,𝑀 𝑥𝑊 −𝑚𝑖𝑛𝑊 Compute 𝑚𝑎𝑥 Compute 𝑊,𝑀 − 𝑚𝑖𝑛𝑊,𝑀 𝑚𝑎𝑥𝑊 − 𝑚𝑖𝑛𝑊 3.2 Checking for outliers and data treatment Applying these metrics provides the scores for each indicator, on an interval bound between 0 and 1. As per the recommendations of the European Commission’s Competence Centre on Composite Indicators and Scoreboards (COIN), the scores are screened for outliers (Nardo et al., 2008), on the basis of the following criteria: 1. skew < |2| 2. kurtosis < |3.5| Where necessary, indicators are winsorized, e.g., extreme values are replaced by the next nearest value, until a given indicator falls within the criteria outlined above. Two of the indicators were treated: D1.1 and D3.2. In the first case, the values for Japan and Switzerland were replaced with the value for France. In the second case, the values for Greece, Romania and Spain were replaced with the value for Bulgaria. After treatment, the table of scores for each of indicator is provided in Table 6 in Appendix. 3.3 Aggregation and presentation of the scores The aggregation within dimension is performed using the arithmetic mean. This means that full compensation between indicators is permitted. Different sets of indicators were considered, as per the shortlist of indicators. As part of an iterative process that involves checking the correlation structure for different solutions, the most appropriate set was retained. The scores for each dimension are provided in Table 7 in Appendix. 21 4 VISUALIZATIONS – PRELIMINARY FINDINGS Despite the dearth of relevant data, some interesting findings emerge when data are visualized. This section provides a few heatmaps visualizing some of the dimension scores for illustration. 4.1 Dimension 2: Barriers The “Barriers� dimension has severe data constraints with data available only for the Safety and Personal Security barrier (Gallup indicator “Individuals feeling safe walking alone at night in the city or area where they live�). Heatmap 1 below visualizes the scores for this dimension where 0 represents the lowest relative position of a country and 1 – the highest. Women globally face significant safety concerns in public space especially at night, albeit to different extent. Differences in country scores are evident, with Latin America, and the Caribbean and most of Sub-Saharan Africa performing worst. Better performers are in Europe and Central Asia, North America, China, the Middle East and North Africa. As with any global data, these scores need to be treated with caution when undertaking a more nuanced gender analysis at country level. For example, in some economies, social and cultural norms discourage or prohibit women from being in public space at night. Global data may not capture women’s concerns about safety of walking at night in these geographies. The country averages also often mask variations in safety between densely populated metropolitan areas and rural places. Nevertheless, many of these findings are aligned with individual country reports available in these some of these geographies. Heatmap 1 - Dimension 2: Gendered barriers to mobility Source: Gallup’s indicator “Individuals feeling safe walking alone at night in the city or area where they live�. 2022. Data available for 144 countries. The visualization was done by the team. 22 4.2 Dimension 4: Gendered perceptions of mobility Women’s and men’s perceptions of comfort, price, availability, safety, and enjoyment for the end-to- end journey can limit the use of an available transport network. Some studies show significant differences in the perception of safety of public transport and public space among women and men, shaping their mobility patterns and travel behaviours differently. Gallup’s two sex-disaggregated indicators measure satisfaction with roads and highways and with public transport systems in areas where people live, and the associated scores visualized in Heatmap 2 show countries relative positions in this context, where 0 represents the lowest relative position of a country and 1 – the highest. Heatmap 2 - Dimension 4: Gendered perceptions of mobility Source: Gallup Indicators “Individuals satisfied with roads and highways in the area where they live, and “Individuals satisfied with public transport systems in the area where they live by sex�. 2022. Data available for 145 countries. The visualization was done by the team. 4.3 Dimension 5. Women’s representation in the transport sector. Gender is often absent from transport policymaking, especially when there is a low representation of women in the transport sector in planning, policy, and operations. Women remain greatly under- represented in technical and managerial roles and in customer-facing roles such as bus and taxi drivers, though women represent at least half of the customer base. Though women’s representation and involvement at all levels do not guarantee that gender perspectives will be addressed, it opens the sector to recognizing and reflecting on women’s needs in transport policymaking and transport service development. The differences in women’s employment are significant in all 121 countries for which data from the International Labor Organization (ILO) are available, although there are wide variations among countries. Regions with the lowest levels of women’s employment in the transport and storage sector include the Middle East and North Africa (MENA), Sub-Saharan Africa, South Asia and the Latin America and the Caribbean. Heatmap 3 below presents relative scores for women’s representation in the transport sector worldwide, where 0 represents the lowest relative position of a country and 1 – the highest. . 23 Heatmap 3 - Dimension 5: Women’s representation in the transport sector Source: ILO indicator “Share of women employed in transport and storage sector�. 2022. Data available for 121 countries. The visualization was done by the team. 24 5 POLICY IMPLICATIONS This research explored the feasibility of creating a Mobility and Gender Index (MGI). Based on this work, a number of policy implications have emerged: • What does not get measured, does not get done. No global index tracks gender inequalities in mobility despite the growing evidence about development implications of gender inequalities in mobility. There is a wide range of global composite indicators that measure gender inequalities in other domains, such as health, education, economy, and civic participation (e.g., OECD’s Social Indicators and Gender Index, World Economic Forum’s Global Gender Gap Index, UNDP’s Gender Inequality Index and EIGE’s Gender Equality Index). • The chicken and egg situation: The lack of sex-disaggregated mobility data and lack of awareness of the importance of addressing these gaps reinforce each other in a vicious cycle: the former does not allow for creating a mobility and gender index whilst the latter does not create sense of urgency for investing in improved data collection. The lack of a global index prevents policy makers and development agencies from drawing attention to this issue, setting project, program and policy priorities, and benchmarking and monitoring performance. • Need for more and better sex-disaggregated mobility data: The data gaps signify that existing policies may not adequately address gender inequalities in mobility. Recognizing these gaps in current data, there should be a focus on improving both the quality and quantity of data collected. This may involve developing new survey instruments and/or modifying existing ones to better capture gendered experiences of mobility (within World Bank teams) as well as seeking external partnerships to collect data on gender differences in mobility. Some of the areas that the team will be exploring is how to use ‘Big data’, such as satellite data and how to amend household survey tools to fill in some of the data gaps in this area. • Potential for an index: Due to data limitations, an index measuring gender inequalities in mobility cannot be computed with the available open-source data. However, this research showed that it is possible to construct scores for select dimensions and to visualize some of the gender gaps in the area of mobility. This preliminary research can be used as a foundation to ignite a discussion towards data collection and to finalize the construction of the MGI. • Public Awareness, Advocacy and Collaborations: Policy initiatives should also include efforts to raise public awareness about the importance of gender equality in mobility and the benefits it brings in general, and about the importance of collecting more and better gender and mobility data, in particular. This might involve campaigns, educational programs, and various collaborations among donors, international and regional development organizations, policymakers, national statistical offices, non-profit organizations, and private sector players to share best practices, insights, and technical expertise in promoting gender-inclusive mobility. In conclusion, the development of the MGI will be an important step towards understanding and addressing gender inequalities in mobility. As the current product is hopefully further developed into an index, it will provide policymakers with a tool to help guide their actions and monitor progress creating more inclusive and sustainable mobility. However, further work is needed to improve data collection, increase resource allocation, and promote collaboration and awareness. 25 6 CONCLUDING REMARKS AND NEXT STEPS This technical report has presented the key findings of the exploratory research of the World Bank’s Transport GP to construct a Mobility and Gender Index (MGI). The chapter 2 emphasized a feminist perspective to understand gendered geographies of mobility and highlighted the importance of considering factors intersecting with gender; proposed a six-dimension theoretical framework for the index and outlined data pre-processing and indicator selection steps followed for the dimensions. Existing data sources were reviewed, and selection criteria ensured data quality and replicability. Variables were filtered based on the geographic coverage, periodicity, availability, redundancy, and compatibility. Statistical tests were used to ensure data integrity. Informed by the findings of the chapter 2, chapter 3 detailed technical steps followed in developing the measurement framework for the MGI. Different approaches for measuring gender gaps and levels were explored, and possible aggregation within dimensions and scores were presented. Lastly, the report provided visualizations –by translating some of the dimension scores into heatmaps, and shared key findings of the displayed visuals. The report acknowledges data gaps, uneven geographical coverage, and limitations, and outline the following next steps for the Mobility and Gender Index: 1. Data Enhancement: there is a need to address the significant gender data gaps to enable a more comprehensive measurement of gendered aspects of mobility through an index. This may involve collecting additional data and improving the geographical coverage in each dimension. 2. Refinement of Indicators: work focusing on continuously reviewing and updating the selected indicators is needed, together with attempting to identifying new indicators as they emerge or alternatively modifying existing ones to ensure they accurately capture gender and mobility-related aspects. 3. Methodological Improvements: Exploring alternative methods and approaches for computing scores and aggregating indicators to enhance the accuracy and robustness of the index. This may include refining the chosen metric, considering different weighting schemes, or incorporating additional statistical considerations. 4. Stakeholder Engagement: Engaging with relevant stakeholders, policymakers, and experts to gather feedback on the proposed index, discuss its implications, and explore potential applications for policymaking and gender-responsive interventions. 5. Publication and Dissemination: Sharing the findings, methodology, and scores of the Gender Index through reports, academic publications, or other communication channels to raise awareness, encourage discussions, and promote further research in the field. 26 7 REFERENCES AitBihiOuali, L., Graham, D. J., Barron, A., & Trompet, M. (2019). Gender differences in the perception of safety in public transport. Journal of the Royal Statistical Society: Series A (Statistics in Society), 183. Akyelken, N. (2013). Development and Gendered Mobilities: Narratives from the Women of Mardin, Turkey. Mobilities, 8(3), 424-439. https://doi.org/10.1080/17450101.2013.769725 Alam, M. M., Kurshitashvili, N., Dominguez Gonzalez, K., Gonzalez Carjaval, K., & Baruah, B. (2022). Is a Mile for One a Mile for All? A Knowledge Synthesis Report on Gender and Mobility (2000–20). https://openknowledge.worldbank.org/handle/10986/37354 Ball, K. S., & Wesson, C. J. (2017). Perceptions of unwanted sexual behaviour on public transport: exploring transport density and behaviour severity. Crime Prevention and Community Safety, 19(3), 199-210. https://doi.org/10.1057/s41300-017-0026-3 Booth, D., Hanmer, L., & Lovell, E. (2000). Transport and Poverty. Chowdhury, S. (2019). Role of Gender in the Ridership of Public Transport Routes Involving Transfers. Transportation Research Record, 2673(4), 855-863. https://doi.org/10.1177/0361198119837155 Coppola, P., & Silvestri, F. (2021). Gender Inequality in Safety and Security Perceptions in Railway Stations. Sustainability, 13(7), 4007. https://www.mdpi.com/2071-1050/13/7/4007 Dandapat, S., & Maitra, B. (2020). Preference heterogeneity in trip makers’ perception and policy Issues: A study with reference to bus services in Kolkata. Case Studies on Transport Policy, 8(4), 1504-1517. https://doi.org/https://doi.org/10.1016/j.cstp.2020.11.001 Dominguez Gonzalez, K., Machado, A. L., Bianchi Alves, B., Raffo, V., Guerrero, S., & Portabales, I. (2020). Why does she move? A Study of Women’s Mobility in Latin American Cities. Duchène, C. (2011). Gender and Transport (Discussion Paper, Issue. Freedman, A. (2002). Commuting Gazes: Schoolgirls, Salarymen, and Electric Trains in Tokyo. The Journal of Transport History, 23(1), 23-36. https://doi.org/10.7227/TJTH.23.1.4 Hickey, G. (2014). Anti-harassment campaigns for mass transit in the 21st century US: A critique from history. 5th International Conference on Women's Issues in TransportationFederation Internationale De L'Automobile (FIA) Institut Francais des Sciences et Technologies des Transports, de l'Aménagement et des Réseaux (IFSTTAR) Department of TransportationTransportation Research Board, Horii, M., & Burgess, A. (2012). Constructing sexual risk: ‘Chikan’, collapsing male authority and the emergence of women-only train carriages in Japan. Health, Risk & Society, 14(1), 41-55. https://doi.org/10.1080/13698575.2011.641523 Humbert, A. L., & Hubert, A. (2021). Gender Equality Index. In E. Ng, C. Stamper, & A. Klarsfeld (Eds.), Handbook on Diversity and Inclusion Indices: A Research Compendium. Edward Elgar. Hutson, A. C. F., & Krueger, J. C. (2018). The Harasser’s Toolbox: Investigating the Role of Mobility in Street Harassment. Violence Against Women, 25(7), 767-791. https://doi.org/10.1177/1077801218804100 Kaufman, S., Polack, C., & Campbell, G. (2018). The Pink Tax on Transportation: Women’s Challenges in Mobility. Kerzhner, T., Kaplan, S., & Silverman, E. (2018). Physical walls, invisible barriers: Palestinian women's mobility in Jerusalem [https://doi.org/10.1111/rsp3.12162]. Regional Science Policy & Practice, 10(4), 299-314. https://doi.org/https://doi.org/10.1111/rsp3.12162 Kurshitashvili, N., Carvajal, K., Pickup, L., C, M.-T., L., S., & L., F. (2021). Gender Equality in Transportation Course. https://olc.worldbank.org/content/gender-equality-transportation-self- paced Kurshitashvili, N., Carvajal, K., Saunders, K., & AitBihiOuali, L. (2022). Ways to Green Mobility- Perspectives on women and rail in Bosnia and Herzegovina and Serbia. https://openknowledge.worldbank.org/handle/10986/37324 Law, R. (1999). Beyond ‘women and transport’: towards new geographies of gender and daily mobility. Progress in Human Geography, 23(4), 567-588. https://doi.org/10.1191/030913299666161864 Lewis, S., Saukko, P., & Lumsden, K. (2021). Rhythms, sociabilities and transience of sexual harassment in transport: mobilities perspectives of the London underground. Gender, Place & Culture, 28(2), 277-298. https://doi.org/10.1080/0966369X.2020.1734540 27 Loukaitou-Sideris, A., & Fink, C. (2008). Addressing Women's Fear of Victimization in Transportation Settings: A Survey of U.S. Transit Agencies. Urban Affairs Review, 44(4), 554-587. https://doi.org/10.1177/1078087408322874 Mansoor, T., & Hasan, R. (2016). Gender Differences In The Fear Of Crime Victimization And Precautionary Behaviours. Pakistan Journal of Gender Studies, 12(1), 165-178. https://doi.org/10.46568/pjgs.v12i1.206 Masikini, N., & Baruah, B. (2020). Gender Equity in the “Sharing� Economy: Possibilities and Limitations. In L. A. Saraswati, B. Shaw, & H. Rellihan (Eds.), Introduction to Women’s, Gender & Sexuality Studies: Interdisciplinary and Intersectional Approaches (pp. 484-488). Oxford University Press. Mejía-Dorantes, L., & Soto Villagrán, P. (2020). A review on the influence of barriers on gender equality to access the city: A synthesis approach of Mexico City and its Metropolitan Area. Cities, 96, 102439. https://doi.org/https://doi.org/10.1016/j.cities.2019.102439 Mungai, M. w., & David, A. S. (2006). "No Mercy, No Remorse": Personal Experience Narratives about Public Passenger Transportation in Nairobi, Kenya. Africa Today, 52(3), 51-81. http://www.jstor.org/stable/4187721 Munoz-Raskin. R., Gonzalez Carvajal. K., Kurshitashvili. N. 2022. "The Climate Action Agenda: Why a people -centered approach to mobility matters". World Bank Blog. https://blogs.worldbank.org/transport/climate-action-agenda-why-people-centered-approach- mobility-matters Nardo, M., Saisana, M., Saltelli, A., Tarantola, S., Hoffman, A., & Giovannini, E. (2008). Handbook on constructing composite indicators. Methodology and user guide. OECD Publishing. https://doi.org/10.1787/533411815016 Neupane, G., & Chesney-Lind, M. (2014). Violence against women on public transport in Nepal: sexual harassment and the spatial expression of male privilege. International Journal of Comparative and Applied Criminal Justice, 38(1), 23-38. https://doi.org/10.1080/01924036.2013.794556 Permanyer, I. (2015). Why call it ‘equality’ when it should be ‘achievement’? A proposal to un-correct the ‘corrected gender gaps’ in the EU Gender Equality Index. Journal of European Social Policy, 25, 414-430. https://doi.org/10.1177/0958928715594545 Polko, P., & Kimic, K. (2021). Condition of urban park infrastructure in the context of perceived security of park users. IOP Conference Series: Earth and Environmental Science, 900. Sanchez de Madariaga, I. (2013). From Women in Transport to Gender in Transport: Challenging Conceptual Frameworks for Improved Policymaking. Journal of International Affairs, 67, 43. Shaw, C., Russell, M., Keall, M., MacBride-Stewart, S., Wild, K., Reeves, D., Bentley, R., & Woodward, A. (2020). Beyond the bicycle: Seeing the context of the gender gap in cycling. Journal of Transport & Health, 18, 100871. https://doi.org/https://doi.org/10.1016/j.jth.2020.100871 Shirgaokar, M. (2019). Operationalizing gendered transportation preferences: A psychological framework incorporating time constraints and risk aversion. Transport Policy, 75, 10-18. https://doi.org/https://doi.org/10.1016/j.tranpol.2018.12.010 Singh, Y. J. (2020). Is smart mobility also gender-smart? Journal of Gender studies, 29(7), 832-846. https://doi.org/10.1080/09589236.2019.1650728 Song, L., Kirschen, M., & Taylor, J. (2019). Women on wheels: Gender and cycling in Solo, Indonesia. Singapore Journal of Tropical Geography, 40(1), 140-157. https://doi.org/https://doi.org/10.1111/sjtg.12257 Soraganvi, A. S. (2017). Safe Public Places: Rethinking Design for Women Safety. International Journal on Emerging Technologies, 8(1), 304-308. https://www.researchtrend.net/ijet/pdf/59- %20109.pdf Stark, J., & Meschik, M. (2018). Women’s everyday mobility: Frightening situations and their impacts on travel behaviour. Transportation Research Part F: Traffic Psychology and Behaviour, 54, 311-323. https://doi.org/https://doi.org/10.1016/j.trf.2018.02.017 Yavuz, N., & Welch, E. W. (2010). Addressing Fear of Crime in Public Space: Gender Differences in Reaction to Safety Measures in Train Transit. Urban Studies, 47(12), 2491-2515. https://doi.org/10.1177/0042098009359033 28 Zhao, J., Wang, J., & Deng, W. (2015). Exploring bikesharing travel time and trip chain by gender and day of the week. Transportation Research Part C: Emerging Technologies, 58, 251-264. https://doi.org/https://doi.org/10.1016/j.trc.2015.01.030 29 8 APPENDIX Table 3 - Selection process summary for eligible variables potentially included in the MGI Dimension 1: Gender and mobility patterns: modes, locations, timings Latest year(s) Sex- Available for non- Source Variable Decision Post Screening available disaggregated European countries Modes Limited geographical Eurostat Total rail passengers 2004-2019 No Yes coverage Limited geographical Eurostat Rail passenger-km 2004-2019 No Yes coverage International Transport Rail passengers 2020 No Yes Kept in the final version Forum International Transport Road accident fatalities 2020 No Yes Relevance Forum Limited geographical OECD Total inland passenger transport (passenger-km) 2000-2019 No Yes coverage Daily public transport boardings (million) – WB Urban Transport Data discontinued and last Standard Bus / Regional Rail / Metro / LRT / 2014 No Yes Data Analysis Tool year available pre-cut-off date Tram / BRT Registered private vehicles and registered two- WB Urban Transport Data discontinued and last wheelers – Total number of registered motor 2014 No Yes Data Analysis Tool year available pre-cut-off date vehicles (Private, Total) WB Urban Transport Share of trips to work by car / by motorcycle / Data discontinued and last 2014 No Yes Data Analysis Tool bicycle/ by bus / by foot / by rail year available pre cut-off date Locations / Timings WB Urban Transport Data discontinued and last Average trip length by car (km) 2014 No Yes Data Analysis Tool year available pre cut-off date WB Urban Transport Data discontinued and last Average journey time (minutes) 2014 No Yes Data Analysis Tool year available pre cut-off date WB Urban Transport Data discontinued and last Average duration of journey to work (minutes) 2014 No Yes Data Analysis Tool year available pre cut-off date 30 Dimension 2: Gendered barriers to mobility Latest year(s) Sex- Available for non- Source Variable Decision Post Screening available disaggregated European countries Accessibility NO DATA Availability NO DATA Affordability Average monthly earnings of prime-age employees Did not pass the statistical ILO by sex, household type and presence of children 2021 Yes Yes tests (local currency) Women with account at financial institution or with Data discontinued and last UN Women mobile money-service provider (% of female 2017 Yes Yes year available pre cut-off population aged 15 and older) date Acceptability Compatibility issues: Data Can a woman travel outside her home in the same are not sex-disaggregated – WBL 2020 N/A Yes way as a man? yet this is required for this dimension Safety NO DATA Personal Security Redundant variable (Gallup data covers the same OECD Feeling Safe Walking alone at night 2019 Yes Yes question with wider geographical coverage) Do you feel safe walking alone at night in your Gallup Data 2022 Yes Yes Kept in the final version community? (WP113) In the city or area where you live, do you have Gallup Data 2022 Yes Yes Relevance confidence in the local police force? (WP112) 31 Dimension 3: Gender and mobility infrastructure Latest year(s) Sex- Available for non- Source Variable Decision Post Screening available disaggregated European countries OECD Urban access – 15mn / 4km – Driving 2018 N/A No Did not pass statistical tests OECD Urban access – 15mn / 4km – Walking 2018 N/A No Kept in the final version OECD Urban access – 15mn / 4km – Biking 2018 N/A No Did not pass statistical tests OECD Urban access – 15mn / 4km – Public Transport 2018 N/A No Kept in the final version Road Infrastructure at 31 December by Type of Limited geographical UNECE 1993-2019 N/A Yes Road, Country and Year coverage WB Urban Transport Size of network – Length of transit – Bus / Metro / Data discontinued and last 2014 N/A Yes Data Analysis Tool Tram / Light Rail - year available pre cut-off date WB Urban Transport Number of stations – Regional Rail/Commuter Rail Data discontinued and last 2014 N/A Yes Data Analysis Tool / Metro Rail / Light Rail / Tram / Trolley Bus / BRT year available pre cut-off date Peak Time fares for 10 km of journey by transport WB Urban Transport Data discontinued and last mode (Local currency) – Metro / Bus / Regional 2014 N/A Yes Data Analysis Tool year available pre cut-off date Rail / LRT / Tram / BRT (i) PT Fare as share of Income = Average fare for 10km*40*12/per capita GDP ; (ii) MT Fare as share WB Urban Transport Data discontinued and last of Income = Average fare for 10km*40*12/per 2014 N/A Yes Data Analysis Tool year available pre cut-off date capita GDP ; (iii) MT Fare for avg. trip length as share of minimum daily wage 32 Dimension 4: Gendered perceptions of transport Latest year(s) Sex- Available for non- Source Variable Decision Post Screening available disaggregated European countries Degree of satisfaction (city/area where you live): The public transportation systems : Very satisfied – Data discontinued and last year World Values Survey 2010-2014 Yes Yes Fairly satisfied – Fairly dissatisfied – Very available pre cut-off date dissatisfied In the city or area where you live, are you satisfied Gallup Data or dissatisfied with the roads and highways? 2022 Yes Yes Kept in the final version (WP91) In the city or area where you live, are you satisfied Gallup Data or dissatisfied with the public transportation 2022 Yes Yes Kept in the final version systems? (WP92) Dimension 5: Women’s representation in the sector Latest year(s) Sex- Available for non- Source Variable Decision Post Screening available disaggregated European countries Redundant variable – Replaced with ILO data with European Share of women employed in transport sector 2013-2017 N/A No better geographic coverage Commission and higher data collection frequency Dimension 6: Legal and policy frameworks Available Latest Sex- for non- Source Variable year(s) Decision Post Screening disaggregated European available countries Is there legislation specifically addressing domestic WBL 2020 N/A Yes Relevance violence? Data discontinued and last year SIGI Presence of legislation against harassment 2012 N/A Yes available pre-cut-off date 33 Table 4 - Conceptual framework, measurement framework and indicators Conceptual framework Measurement framework Domains Subdomains Domains Indicators Sex- disaggregation Modes, locations and timings Modes Passengers Carried (million Passenger-kms, rail) No Barriers to mobility Availability Affordability (Social and Cultural) Acceptability (Physical) Accessibility Safety and Security Safety and security Individuals feeling safe walking alone at night in the Yes city or area where they live by sex Infrastructure Infrastructure Average absolute accessibility in urban areas No (distance: 15mn/4km). Mode: Walking Average absolute accessibility in urban areas (distance: 15mn/4km). Mode: Public Transport Perceptions Perceptions Individuals satisfied with roads and highways in the Yes area where they live by sex Individuals satisfied with public transportation systems in the area where they live by sex Women’s representation in the Women’s Women’s representation in the transportation and Yes transport sector representation storage sector Legal and policy framework 34 Table 5 - Examples of scores for sex-disaggregated indicators using the modified minmax approach Min across countries Max across countries Minmax Country A Women 10% 5% 80% 0.07 Men 20% 15% 90% 0.07 Metric 13% 5% 90% 0.10 Country B Women 50% 5% 80% 0.60 Men 70% 15% 90% 0.73 Metric 58% 5% 90% 0.63 Country C Women 70% 5% 80% 0.87 Men 90% 15% 90% 1.00 Metric 79% 5% 90% 0.87 Country D Women 15% 5% 80% 0.13 Men 15% 15% 90% 0.00 Metric 15% 5% 90% 0.12 Country E Women 80% 5% 80% 1.00 Men 80% 15% 90% 0.87 Metric 80% 5% 90% 0.88 Country F Women 70% 5% 80% 0.87 Men 50% 15% 90% 0.47 Metric 58% 5% 90% 0.63 Country G Women 80% 5% 80% 1.00 Men 90% 15% 90% 1.00 Metric 85% 5% 90% 0.94 Country H Women 5% 5% 80% 0.00 Men 15% 15% 90% 0.00 Metric 8% 5% 90% 0.03 35 Table 6 - Scores for indicators (after treatment) D1.1 D2.1 D3.1 D3.2 D4.1 D4.2 D5.1 Afghanistan 0.00 0.25 0.27 0.00 Albania 0.00 0.62 0.53 0.49 0.22 Algeria 0.48 0.52 0.45 American Samoa Andorra Angola 0.09 Antigua and Barbuda Argentina 0.30 0.60 0.58 0.22 Armenia 0.91 0.52 0.49 0.23 Aruba Australia 0.55 0.62 0.67 Austria 0.83 0.37 1.00 0.75 0.76 0.49 Azerbaijan 0.01 0.84 0.63 0.66 Bahamas, The Bahrain 0.84 0.81 Bangladesh 0.63 0.85 0.79 0.08 Barbados 0.43 Belarus 0.57 0.63 0.63 0.47 Belgium 0.58 0.34 0.31 0.61 0.65 Belize 0.29 Benin 0.42 0.50 0.41 0.05 Bermuda Bhutan 0.11 Bolivia 0.33 0.68 0.65 0.09 Bosnia and Herzegovina 0.65 0.41 0.41 0.20 Botswana 0.16 0.58 0.64 0.37 Brazil 0.21 0.45 0.48 0.20 British Virgin Islands Brunei Darussalam 0.72 Bulgaria 0.18 0.54 1.00 0.61 0.60 Burkina Faso 0.45 0.47 0.43 0.09 Burundi 0.59 0.36 0.29 0.00 Cabo Verde 0.32 Cambodia 0.56 0.86 0.80 0.12 Cameroon 0.26 0.41 0.35 0.09 Canada 0.77 0.61 0.64 Cayman Islands Central African Republic Chad 0.27 0.42 0.32 0.03 Channel Islands Chile 0.03 0.29 0.57 0.61 0.33 China 0.94 0.86 0.86 Colombia 0.31 0.61 0.56 0.16 Comoros 0.65 0.56 0.46 0.10 Congo, Dem. Rep. 0.28 0.39 0.34 1.00 Congo, Rep. 36 Costa Rica 0.31 0.73 0.76 0.27 Côte d’Ivoire 0.33 0.42 0.36 0.02 Croatia 0.14 0.82 0.40 0.50 0.47 Cuba Curacao Cyprus 0.66 0.09 0.42 0.44 0.66 Czech Republic 0.66 0.76 0.34 0.32 0.86 0.87 0.51 Denmark 0.79 0.89 0.47 0.27 0.71 0.72 Djibouti Dominica Dominican Republic 0.21 0.65 0.70 0.08 Ecuador 0.28 0.71 0.72 0.13 Egypt 0.84 0.69 0.67 0.03 El Salvador 0.48 0.77 0.74 0.13 Equatorial Guinea Eritrea Estonia 0.22 0.81 0.46 0.22 0.85 0.85 0.52 Eswatini 0.26 0.64 0.63 0.49 Ethiopia 0.41 0.41 0.26 0.07 Faroe Islands Fiji 0.18 Finland 0.53 0.87 0.24 0.21 0.63 0.63 France 1.00 0.73 0.41 0.28 0.67 0.67 0.52 French Polynesia Gabon 0.13 0.21 0.18 Gambia 0.29 0.38 0.34 0.12 Georgia 0.79 0.70 0.70 0.20 Germany 0.69 0.71 0.27 0.18 0.69 0.71 Ghana 0.55 0.51 0.48 0.09 Gibraltar Greece 0.06 0.52 1.00 0.39 0.47 0.48 0.35 Greenland Grenada Guam Guatemala 0.42 0.73 0.70 0.12 Guinea 0.32 0.27 0.21 Guinea-Bissau 1.00 Guyana 0.28 Haiti 0.36 0.18 0.13 0.00 Honduras 0.45 0.71 0.70 0.05 Hong Kong 0.75 0.83 0.89 Hungary 0.57 0.68 0.58 0.83 0.71 0.73 0.49 Iceland 0.85 0.57 0.64 India 0.57 0.74 0.76 0.03 Indonesia 0.79 0.82 0.76 Iran 0.66 0.66 0.61 0.03 Iraq 0.58 0.60 0.51 0.00 Ireland 0.17 0.70 0.16 0.12 0.58 0.58 Isle of Man 37 Israel 0.75 0.63 0.63 0.36 Italy 0.47 0.61 0.72 0.35 0.35 0.41 Jamaica 0.60 0.67 0.72 0.30 Japan 1.00 0.77 0.57 0.61 0.46 Jordan 0.80 0.67 0.59 0.06 Kazakhstan 0.58 0.60 0.62 Kenya 0.41 0.48 0.50 0.38 Kiribati 0.53 Kosovo 0.75 0.66 0.59 0.17 Kuwait 0.91 0.76 0.80 Kyrgyzstan 0.57 0.70 0.70 0.07 Lao PDR 0.53 0.85 0.78 0.20 Latvia 0.19 0.67 0.53 0.18 0.78 0.77 Lebanon 0.39 0.34 0.27 0.09 Lesotho 0.14 0.53 0.54 0.09 Liberia 0.18 0.02 0.01 0.31 Libya 0.44 0.43 0.38 Liechtenstein Lithuania 0.12 0.69 0.54 0.20 0.67 0.70 Luxembourg 0.48 0.85 0.00 0.00 0.81 0.82 Macao SAR, China Madagascar 0.23 0.42 0.33 0.11 Malawi 0.29 0.35 0.32 Malaysia 0.46 0.62 0.64 Maldives 0.26 Mali 0.45 0.42 0.34 0.00 Malta 0.72 0.41 0.62 0.63 0.42 Marshall Islands 0.26 Mauritania 0.32 0.24 0.20 0.06 Mauritius 0.56 0.70 0.70 0.29 Mexico 0.00 0.27 0.60 0.60 0.18 Micronesia, Fed. Sts. Moldova 0.01 0.56 0.62 0.59 0.54 Monaco Mongolia 0.39 0.39 0.43 0.33 Montenegro 0.06 0.76 0.44 0.37 0.39 Morocco 0.55 0.51 0.50 Mozambique 0.34 0.56 0.46 0.10 Myanmar 0.52 0.79 0.72 0.07 Namibia 0.24 0.46 0.50 0.19 Nauru 0.00 Nepal 0.49 0.73 0.61 0.05 Netherlands 0.55 0.81 0.48 0.24 0.76 0.83 New Caledonia 0.60 New Zealand 0.55 0.53 0.59 Nicaragua 0.42 0.69 0.71 Niger 0.56 0.73 0.70 0.00 Nigeria 0.39 0.47 0.51 0.04 North Macedonia 0.01 0.66 0.50 0.47 0.24 38 Northern Mariana Islands Norway 0.30 0.95 0.24 0.32 0.65 0.65 Oman Pakistan 0.59 0.62 0.58 0.00 Palau 0.87 Palestine 0.62 0.75 0.75 Panama 0.40 0.67 0.69 0.22 Papua New Guinea Paraguay 0.32 0.42 0.40 Peru 0.28 0.44 0.46 0.13 Philippines 0.57 0.61 0.60 0.08 Poland 0.42 0.72 0.53 0.35 0.71 0.72 Portugal 0.28 0.85 0.46 0.43 0.58 0.62 Puerto Rico Qatar Romania 0.58 1.00 0.66 0.65 Russia 0.53 0.68 0.68 0.44 Rwanda 0.87 0.53 0.43 0.06 Samoa 0.20 San Marino São Tomé and Príncipe Saudi Arabia 0.84 0.84 0.71 Senegal 0.40 0.43 0.41 0.07 Serbia 0.03 0.77 0.54 0.52 0.40 Seychelles 0.57 Sierra Leone 0.37 0.19 0.12 0.00 Singapore 1.00 1.00 1.00 0.49 Sint Maarten (Dutch part) Slovak Republic Slovenia 0.26 0.92 0.14 0.10 0.65 0.66 0.32 Solomon Islands 0.13 Somalia 0.19 South Africa 0.13 0.65 0.65 South Korea 0.76 0.79 0.80 South Sudan Spain 0.34 0.80 1.00 0.71 0.64 0.67 Sri Lanka 0.58 0.70 0.63 0.07 St. Kitts and Nevis St. Lucia St. Martin (French part) St. Vincent and the Grenadines Sudan 0.04 Suriname 0.34 Sweden 0.78 0.77 0.33 0.14 0.75 0.74 Switzerland 1.00 0.88 0.49 0.30 0.91 0.92 0.52 Syrian Arab Republic Tajikistan 0.91 0.85 0.87 Tanzania 0.64 0.64 0.55 0.06 Thailand 0.56 0.74 0.69 0.38 39 Timor-Leste 1.00 Togo 0.40 0.22 0.19 0.06 Tonga 0.35 Trinidad and Tobago 0.24 Tunisia 0.49 0.34 0.29 0.20 Turkey 0.13 0.44 0.58 0.57 Turkmenistan 0.95 0.63 0.63 Turks and Caicos Islands Tuvalu 0.39 Uganda 0.34 0.45 0.39 0.02 Ukraine 0.28 0.47 0.60 0.58 United Arab Emirates 0.95 0.92 0.91 0.23 United Kingdom 0.74 0.27 0.09 0.70 0.74 United States 0.05 0.72 0.59 0.62 0.58 Uruguay 0.32 0.70 0.72 0.35 Uzbekistan 0.87 0.89 0.85 Vanuatu 0.17 Venezuela 0.05 0.00 0.00 Viet Nam 0.60 0.70 0.65 0.20 Virgin Islands (U.S.) Yemen 0.39 0.35 0.25 Zambia 0.27 0.43 0.43 0.08 Zimbabwe 0.28 0.37 0.33 0.17 Table 7 - Scores for dimensions D1 – D2 – D3 – D4 – D5 – Modes Barriers Infrastructure Perceptions Context Afghanistan 0.00 0.26 0.00 Albania 0.00 0.62 0.51 0.22 Algeria 0.48 0.49 American Samoa Andorra Angola 0.09 Antigua and Barbuda Argentina 0.30 0.59 0.22 Armenia 0.91 0.51 0.23 Aruba Australia 0.55 0.64 Austria 0.83 0.68 0.76 0.49 Azerbaijan 0.01 0.84 0.64 Bahamas, The Bahrain 0.83 Bangladesh 0.63 0.82 0.08 Barbados 0.43 Belarus 0.57 0.63 0.47 Belgium 0.58 0.33 0.63 Belize 0.29 40 Benin 0.42 0.46 0.05 Bermuda Bhutan 0.11 Bolivia 0.33 0.66 0.09 Bosnia and Herzegovina 0.65 0.41 0.20 Botswana 0.16 0.61 0.37 Brazil 0.21 0.46 0.20 British Virgin Islands Brunei Darussalam 0.72 Bulgaria 0.18 0.54 0.61 Burkina Faso 0.45 0.45 0.09 Burundi 0.59 0.32 0.00 Cabo Verde 0.32 Cambodia 0.56 0.83 0.12 Cameroon 0.26 0.38 0.09 Canada 0.77 0.63 Cayman Islands Central African Republic Chad 0.27 0.37 0.03 Channel Islands Chile 0.03 0.29 0.59 0.33 China 0.94 0.86 Colombia 0.31 0.59 0.16 Comoros 0.65 0.51 0.10 Congo, Dem. Rep. 0.28 0.37 1.00 Congo, Rep. Costa Rica 0.31 0.74 0.27 Côte d’Ivoire 0.33 0.39 0.02 Croatia 0.14 0.82 0.48 Cuba Curacao Cyprus 0.66 0.43 0.66 Czech Republic 0.66 0.76 0.33 0.87 0.51 Denmark 0.79 0.89 0.37 0.72 Djibouti Dominica Dominican Republic 0.21 0.67 0.08 Ecuador 0.28 0.71 0.13 Egypt 0.84 0.68 0.03 El Salvador 0.48 0.76 0.13 Equatorial Guinea Eritrea Estonia 0.22 0.81 0.34 0.85 0.52 Eswatini 0.26 0.64 0.49 Ethiopia 0.41 0.34 0.07 Faroe Islands Fiji 0.18 Finland 0.53 0.87 0.23 0.63 France 1.00 0.73 0.34 0.67 0.52 41 French Polynesia Gabon 0.13 0.19 Gambia 0.29 0.36 0.12 Georgia 0.79 0.70 0.20 Germany 0.69 0.71 0.22 0.70 Ghana 0.55 0.49 0.09 Gibraltar Greece 0.06 0.52 0.69 0.48 0.35 Greenland Grenada Guam Guatemala 0.42 0.71 0.12 Guinea 0.32 0.24 Guinea-Bissau 1.00 Guyana 0.28 Haiti 0.36 0.15 0.00 Honduras 0.45 0.70 0.05 Hong Kong 0.75 0.86 Hungary 0.57 0.68 0.70 0.72 0.49 Iceland 0.85 0.61 India 0.57 0.75 0.03 Indonesia 0.79 0.79 Iran 0.66 0.63 0.03 Iraq 0.58 0.55 0.00 Ireland 0.17 0.70 0.14 0.58 Isle of Man Israel 0.75 0.63 0.36 Italy 0.47 0.61 0.53 0.38 Jamaica 0.60 0.69 0.30 Japan 1.00 0.77 0.59 0.46 Jordan 0.80 0.63 0.06 Kazakhstan 0.58 0.61 Kenya 0.41 0.49 0.38 Kiribati 0.53 Kosovo 0.75 0.62 0.17 Kuwait 0.91 0.78 Kyrgyzstan 0.57 0.70 0.07 Lao PDR 0.53 0.82 0.20 Latvia 0.19 0.67 0.35 0.77 Lebanon 0.39 0.31 0.09 Lesotho 0.14 0.53 0.09 Liberia 0.18 0.02 0.31 Libya 0.44 0.41 Liechtenstein Lithuania 0.12 0.69 0.37 0.69 Luxembourg 0.48 0.85 0.00 0.82 Macao SAR, China Madagascar 0.23 0.38 0.11 Malawi 0.29 0.34 42 Malaysia 0.46 0.63 Maldives 0.26 Mali 0.45 0.38 0.00 Malta 0.72 0.63 0.42 Marshall Islands 0.26 Mauritania 0.32 0.22 0.06 Mauritius 0.56 0.70 0.29 Mexico 0.00 0.27 0.60 0.18 Micronesia, Fed. Sts. Moldova 0.01 0.56 0.60 0.54 Monaco Mongolia 0.39 0.41 0.33 Montenegro 0.06 0.76 0.41 0.39 Morocco 0.55 0.50 Mozambique 0.34 0.51 0.10 Myanmar 0.52 0.75 0.07 Namibia 0.24 0.48 0.19 Nauru 0.00 Nepal 0.49 0.67 0.05 Netherlands 0.55 0.81 0.36 0.80 New Caledonia 0.60 New Zealand 0.55 0.56 Nicaragua 0.42 0.70 Niger 0.56 0.71 0.00 Nigeria 0.39 0.49 0.04 North Macedonia 0.01 0.66 0.48 0.24 Northern Mariana Islands Norway 0.30 0.95 0.28 0.65 Oman Pakistan 0.59 0.60 0.00 Palau 0.87 Palestine 0.62 0.75 Panama 0.40 0.68 0.22 Papua New Guinea Paraguay 0.32 0.41 Peru 0.28 0.45 0.13 Philippines 0.57 0.61 0.08 Poland 0.42 0.72 0.44 0.71 Portugal 0.28 0.85 0.45 0.60 Puerto Rico Qatar Romania 0.58 0.65 Russia 0.53 0.68 0.44 Rwanda 0.87 0.48 0.06 Samoa 0.20 San Marino São Tomé and Príncipe Saudi Arabia 0.84 0.78 Senegal 0.40 0.42 0.07 43 Serbia 0.03 0.77 0.53 0.40 Seychelles 0.57 Sierra Leone 0.37 0.16 0.00 Singapore 1.00 1.00 0.49 Sint Maarten (Dutch part) Slovak Republic Slovenia 0.26 0.92 0.12 0.65 0.32 Solomon Islands 0.13 Somalia 0.19 South Africa 0.13 0.65 South Korea 0.76 0.79 South Sudan Spain 0.34 0.80 0.86 0.65 Sri Lanka 0.58 0.67 0.07 St. Kitts and Nevis St. Lucia St. Martin (French part) St. Vincent and the Grenadines Sudan 0.04 Suriname 0.34 Sweden 0.78 0.77 0.24 0.74 Switzerland 1.00 0.88 0.39 0.91 0.52 Syrian Arab Republic Tajikistan 0.91 0.86 Tanzania 0.64 0.59 0.06 Thailand 0.56 0.71 0.38 Timor-Leste 1.00 Togo 0.40 0.21 0.06 Tonga 0.35 Trinidad and Tobago 0.24 Tunisia 0.49 0.32 0.20 Turkey 0.13 0.44 0.57 Turkmenistan 0.95 0.63 Turks and Caicos Islands Tuvalu 0.39 Uganda 0.34 0.42 0.02 Ukraine 0.28 0.47 0.59 United Arab Emirates 0.95 0.92 0.23 United Kingdom 0.74 0.18 0.72 United States 0.05 0.72 0.60 0.58 Uruguay 0.32 0.71 0.35 Uzbekistan 0.87 0.87 Vanuatu 0.17 Venezuela 0.05 0.00 Viet Nam 0.60 0.68 0.20 Virgin Islands (U.S.) Yemen 0.39 0.30 Zambia 0.27 0.43 0.08 44 Zimbabwe 0.28 0.35 0.17 45