The dynamics of financial inclusion in Ethiopia Diversifying delivery channels, expanding digital technology, and enhancing financial knowledge for greater financial inclusion © 2024 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. 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. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. 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. Any 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. Cover image: © Stephan Gladieu / World Bank The dynamics of financial inclusion in Ethiopia Diversifying delivery channels, expanding digital technology, and enhancing financial knowledge for greater financial inclusion Recommended citation Yonis, Manex B.; Ambel, Alemayehu A.; Di Salvo, Francesco and Rawlins, Marlon R. (2024). The dynamics of financial inclusion in Ethiopia: Diversifying delivery channels, expanding digital technology, and enhancing financial knowledge for greater financial inclusion (English). Washington, D.C. : World Bank Group Contents Abbreviations........................................................................................................vii Acknowledgements..............................................................................................vii Overview............................................................................................................... viii The dynamics of financial inclusion........................................................................................ xi Transitioning from being unbanked to banked.................................................................. xv Narrowing the gender gap.....................................................................................................xviii Expanding financial inclusion: broader policy options..................................................... xxi Chapter one: Introduction.................................................................................... 2 Chapter two: Understanding the dynamics: going beyond bank account ownership.......................................................................... 8 Measuring financial inclusion.................................................................................................... 9 Global Findex indicators: Financial access and use......................................................10 Index of financial inclusion................................................................................................11 Evolution of financial inclusion................................................................................................15 Growth in account ownership ..........................................................................................15 The rise of formal savings and stagnation in formal borrowing.................................18 Spatial-temporal evolution of Index of Financial Inclusion..........................................21 Chapter three: Pathways to financial inclusion............................................... 24 Correlates of financial inclusion..............................................................................................28 Entering the financial system: Gateway technology and financial knowledge as pillars..........................................................................................................33 Opportunities to boost financial inclusion...........................................................................41 Chapter four: Gender gap in financial Inclusion: trend, explanation, and bridging the divide................................................... 46 Explaining the gender gap........................................................................................................50 Bridging the gender gap and fostering financial inclusion..............................................52 Chapter five: Summary and priority policy actions......................................... 58 References ........................................................................................................... 62 Annexes................................................................................................................. 66 Annex 1. Additional Information on Measurements and Scores...................................67 Enumeration area level financial knowledge score........................................................67 Annex 2. Additional Descriptive Statistics............................................................................68 Annex 3. Additional Regression Results................................................................................71 Annex 4. Additional Figures.....................................................................................................80 iii List of Figures Figure O.1 Account ownership grew, but lags compared to regional and global comparators, 2011-2023...................................................................................................................................... xi Figure O.2 Education correlates with owning an account, over time................................................. xii Figure O.3 The poor in rural areas remains behind............................................................................... xii Figure O.4 The persistence of gaps of account ownership differs by gender and age group.... xiii Figure O.5 The share of adults saving formally increased, over time................................................ xiv Figure O.6 Formal borrowing is decreasing over time regardless of place of residence............. xiv Figure O.7 Movement into banked status from 2019 to 2022............................................................ xv Figure O.8 Movement from saving informally into saving formally between 2019 and 2022.....xv Figure O.9 Account ownership trajectories and individual characteristics...................................... xvi Figure O.10 Account ownership trajectories and financial knowledge............................................... xvi Figure O.11 Banks are the main financial access points for new entrants....................................... xvii Figure O.12 Payments are through cash, 2022........................................................................................ xvii Figure O.13 Gender gap on account ownership in Rural-Urban context, over time..................... xviii Figure O.14 What drives the gender gap? Socioeconomic status in 2022....................................... xviii Figure 1 Account ownership grew, but ranks lowest compared to neighboring countries, 2011-2023.................................................................................................................................... 15 Figure 2a Account ownership at individual and household level, over time................................. 16 Figure 2b Account ownership by place of residence, over time....................................................... 16 Figure 3a Rural-Urban gap in account ownership, by wealth, over time........................................ 17 Figure 3b Account ownership by gender, over time............................................................................ 17 Figure 4a Account ownership by age, over time................................................................................... 18 Figure 4b Account ownership by education, over time....................................................................... 18 Figure 5a Saving at formal institutions by place of residence, over time....................................... 19 Figure 5b Gender gap in saving at formal institutions, over time..................................................... 19 Figure 5c Saving frequency, over time..................................................................................................... 19 Figure 6a Household access to credit by place of residence and wealth, over time.................. 20 Figure 6b Source of loan by place of residence and wealth, over time.......................................... 20 Figure 7a Evolution of index of financial inclusion (IFI), over time.................................................... 22 Figure 7b Evolution of level of financial inclusion, over time............................................................. 22 Figure 7c Distribution of financial inclusion index, over time............................................................ 22 Figure 7d Distribution of financial inclusion index at woreda level, over time.............................. 22 Figure 7e Distribution of financial inclusion index in rural areas, over time................................. 23 Figure 7f Distribution of financial inclusion index in urban areas, over time............................... 23 Figure 7g Evolution of financial inclusion across regions, over time............................................... 23 Figure 8a Correlates of account ownership in 2019 and 2022......................................................... 28 iv Figure 8b Correlates of account ownership by place of residence, 2019 and 2022................... 29 Figure 9a Education effect of account ownership by gender in 2019............................................ 30 Figure 9b Education effect of account ownership by gender in 2022............................................ 30 Figure 10a Mobile phone ownership effect on account ownership by gender in 2019 .............. 31 Figure 10b Mobile phone ownership effect on account ownership by gender in 2022............... 31 Figure 11a Proximity to financial institution and account ownership by gender in 2019............ 32 Figure 11b Distance to nearest financial institution (Median).............................................................. 32 Figure 12a Account ownership trajectories and individual characteristics...................................... 36 Figure 12b Account ownership trajectories and financial knowledge............................................... 36 Figure 12c Account ownership trajectories by region .......................................................................... 37 Figure 13a Age and sex are associates with entering the financial system .................................... 38 Figure 13b Education matters to escape from being unbanked......................................................... 38 Figure 14 Mobile ownership plays a great role to accelerate financial inclusion ........................ 38 Figure 15a Financial knowledge facilitates the move towards financial inclusion.......................... 39 Figure 15b Financial knowledge gap between urban and rural areas in 2019............................... 39 Figure 16 Proximity to formal financial institutions matters to escape unbanked...................... 40 Figure 17 New entrants have a low-level financial inclusion.............................................................. 41 Figure 18a Banks are the main get way to the financial system......................................................... 42 Figure 18b Saving is a commonly used product by new entrants...................................................... 42 Figure 19a Households and individuals receive payments through cash, 2022............................. 43 Figure 19b Account holders hardly use digital mechanisms to make payments, 2022................ 43 Figure 20 Gender gap in financial inclusion level increased, over time ......................................... 48 Figure 21 Gender gap on account ownership in Rural-Urban context, over time ...................... 49 Figure 22 Gender gap in account ownership region context, over time........................................ 49 Figure 23a Lacks education, by gender, over time................................................................................. 50 Figure 23b Employed adult by gender, over time................................................................................... 50 Figure 23c Mobile use, by gender, over time........................................................................................... 50 Figure 24a Financial knowledge distribution, by gender, 2019........................................................... 51 Figure 24b Financial knowledge distribution, by gender, 2022........................................................... 51 Figure 25a Counterfactual density estimation, 2019............................................................................. 57 Figure 25b Counterfactual density estimation, 2022............................................................................. 57 Figure A.1 Account ownership, by wealth, over time............................................................................ 80 Figure A.2 Use of financial services by account holders, over time.................................................. 80 Figure A.3 Mobile phone penetration by gender, location, over time.............................................. 80 Figure A.4 Usage of financial institutions over time.............................................................................. 80 Figure A.5 Usage of financial inclusion by place of residence, over time........................................ 81 Figure A.6 Age and gender effect on account ownership in 2019.................................................... 81 Figure A.7 Age and gender effect on account ownership in 2022.................................................... 81 v List of Tables Table O.1 Decomposition of the gender gaps in account ownership..............................................xx Table 1 Profile of individual respondents, over time........................................................................ 27 Table 2 Account ownership transition matrix.................................................................................... 34 Table 3 Fairlie decomposition of the gender gaps in account ownership................................. 56 Table 4 Oaxaca-Blinder decomposition of the gender gaps in financial inclusion index....... 56 Table A.1 Account ownership by region, gender, place of residence, 2019 and 2022............... 68 Table A.2 Account owning individuals and financial institutions by region, gender and place of residence, 2022........................................................................................................... 69 Table A.3 Individual and household saving behavior by region, gender, place of residence, 2019 and 2022............................................................................................................................ 69 Table A.4 Access to loan and source of loan by region, place of residence and wealth, 2019 and 2022............................................................................................................................ 70 Table A.5 Estimates for probit model on account ownership........................................................... 71 Table A.6 Estimates for probit model on account ownership for urban-dominated regions and other regions....................................................................................................................... 72 Table A.7 Estimates for ordered logit model on level of financial inclusion ................................. 73 Table A.8 Linear regression estimated results on a z-score of financial inclusion index........... 75 Table A.9 Estimates for logit regression on moving out of financial exclusion............................. 76 Table A.10 Estimates for decomposition analysis on gender gaps in account ownership, 2019......................................................................................................................... 77 Table A.11 Estimates for decomposition analysis on gender gaps in account ownership, 2022......................................................................................................................... 78 Table A.12 Estimates for Oaxaca-Blinder decomposition analysis on gender gaps in financial inclusion, over time.................................................................................................................... 79 List of Boxes Box 1 The National Financial Inclusion Strategy................................................................................4 Box 2 Definitions and Measures of variables.................................................................................. 12 Box 3 Estimated models to examine the correlates of financial inclusion.............................. 26 Box 4 Robustness check....................................................................................................................... 32 Box 5 Transition matrix, financial knowledge index, and model estimation........................... 34 Box 6 Comparison between those who remained unbanked and those able to enter the financial system........................................................................................................................... 36 Box 7 Decomposition analysis combining with 2SLS regression method............................... 53 vi Abbreviations ESPS Ethiopian Socioeconomic Panel Surrey ESS Ethiopian Statistical Service HGERA Home-Grown Economic Reform Agenda IFI Index of Financial Inclusion LFI Level of Financial Inclusion LFMS Labor Force and Migration Survey LSMS-ISA Living Standards Measurement Study – Integrated Surveys on Agriculture MFIs Microfinance Intuitions NBE National Bank of Ethiopia NFIS National Financial Inclusion Strategy POS Point of Sale PSNP Productive Safety Net Program SACCOs Saving and Credit Cooperatives SDG Sustainable Development Goal • Acknowledgements The report was prepared by Manex Bule Yonis (Economist), Alemayehu A. Ambel (Senior Economist), Francesco Di Salvo (Senior Financial Sector Specialist) and Marlon Rolston Rawlins (Senior Financial Sector Specialist). The team also benefited from feedback from peer reviewers: Oya Pinar Ardic Alper (Senior Financial Sector Specialist), Heather G. Moylan (Senior Economist) and Tewodros Tassew Kebede (Financial Sector Specialist). This report leverages the data from the Ethiopia Socioeconomic Panel Survey (ESPS), which is being implemented by the Ethiopian Statistical Service since 2011/12, with technical support from the World Bank Living Standards Measurement Study (LSMS) program. Generous financial support to the survey was received from the Bill and Melinda Gates Foundation. vii viii Overview Image: © i_am_zews / Shutterstock.com Overview Ethiopia’s Home-Grown Economic Reform Agenda (HGERA) emphasizes financial inclusion as one of the key enabling conditions to foster the country’s inclusive growth and economic development. The National Financial Inclusion Strategy (NFIS-II) laid the groundwork for significant regulatory and financial infrastructure improvements, aiming for greater financial inclusion (NBE, 2021). The strategy emphasizes leveraging new opportunities and enablers emerging in the country’s social, economic, and political landscape, such as the rapid growth of mobile phone usage and the increasing availability of digital financial services, to scale account ownership. It also focuses on deepening financial inclusion by making credit, savings, and insurance accessible for those already in the financial system, as well as for new entrants. It emphasizes expanding gender and rural-poor tailored financial services and products as an integral element of the expansion strategy. This financial inclusion study seeks to inform how the country can expedite nationwide integration into the formal financial system to benefit from poverty reduction, equality, and other relevant reforms under the HGERA. Financial inclusion, an extensive concept, refers to access to and use of affordable financial products and services and participation in the financial market. Account ownership is not just a step, but a crucial gateway to broader financial inclusion. Gaining development outcomes requires simultaneous efforts towards promoting access to a financial account and realizing the benefit from its services, such as savings, credits, and insurance (Demirgüç-Kunt et al., 2017; Demirgüç-Kunt et al., 2022). Also, going beyond gender-neutral to gender-intentional financial services and products is a key pillar of women’s empowerment (Hendriks, 2019; Field et al., 2021; Arshad, 2023). Thus, the inclusion of diverse population groups in the financial system promotes gender equality, catalyzes inclusive and resilient growth, and enhances social well-being, thus underlining the broader impact of financial inclusion. ix The dynamics of financial inclusion in Ethiopia To identify policy options for greater financial inclusion of individuals, the report implements an analytical framework with three features. First, the status and dynamics of financial inclusion of individuals, in the context of gender and spatial disparities in the socioeconomic landscape, is examined in detail. The purpose is to better understand how fast the financial inclusion rate evolves across different groups and geographical spheres and who remains behind over time. Second, emphasis is placed on identifying and understanding the factors that explain the dynamics of financial inclusion, how these factors have evolved over time, how the new entrants have performed in the financial system, and the nature of the channels through which the new entrants access the financial system. Explaining the factors holding individuals back from entering and benefiting from the financial system from a dynamic perspective provides evidence-based pathways to greater financial inclusion. Third, a gender gap is estimated, and efforts are directed to explore what explains the gender gap by zooming in on how human capital endowment, resources, and economic opportunity define the gaps. Disaggregated status and dynamics updates and a deep understanding of the factors that foster financial inclusion and narrow the gender gap are essential for policymakers and practitioners to gauge their efforts towards disadvantageous groups, addressing structural impediments, and capitalizing on new opportunities. The report identifies three broad policy options to promote and deepen financial inclusion, anchoring on the preceding analytic farmwork: • Diversifying entry channels by overcoming constraints such as distance, time, cost, quality, and product type hurdles and exploiting opportunities; • Expanding digital technology that can offer more opportunities for previously unbanked people to access and use financial products and services; • Enhancing financial knowledge through financial education programs at higher schools and financial literacy training at the community-level. x Overview The dynamics of financial inclusion Over the last six years, the share of adults (age 18 and above) who own a financial account grew substantially. Account ownership nearly doubled, from 23% in 2016 to 41% in 2022. However, compared to neighboring countries and the average of Sub-Saharan African countries, Ethiopia ranks lowest in terms of the share of the adult population included in the financial system (Figure O.1). In 2022, Ethiopia was 14 percentage points below the 2021 average rate of the Sub-Saharan African countries. On average, account ownership increased by 35% every three years. The growth rate is higher than that of regional and global comparator trends; on average, financial inclusion grew only by 27% in Sub-Saharan African countries and by 14% in developing countries over the last seven years. FIGURE O.1 Account ownership grew, but lags compared to regional and global comparators, 2011-2023 100 PROPORTION OF INDIVIDUAL (%) 90 80 70 60 50 40 30 20 10 0 DEVELOPING COUNTRIES SUB-SAHARAN AFRICA ETHIOPIA 2011 2014 2016 2017 2019 2021 2022 Source: Authors’ estimates from ESPS and Findex. The comparator countries and regional and global rates are derived from the Findex dataset. The Ethiopian account ownership rate is estimated using the ESPS data, as the entire narration of the report is based on this data. Based on the Findex data, in 2022, account ownership rate is 46% in Ethiopia. xi The dynamics of financial inclusion in Ethiopia Account ownership remains low among women, less educated adults, and the rural poor. In 2022, almost more than half of the man adult population owned a financial account, while only 30% of women adults have an account. Education drives account ownership; in 2022, 94% of adults above secondary education level own an account, while only 25% of adults with no education have access to finance (Figure O.2). Similarly, the incidence of account ownership increases along the wealth distribution over the years. However, the evolution of financial inclusion among the rural poor (the bottom 40%) is weaker than that of other groups, as the share of these households in the financial system remains low between 2019 and 2022 (Figure O.3). This implies that the system leaves the most disadvantaged groups behind. FIGURE O.2 Education correlates with owning an account, over time 100 90 80 PROPORTION OF INDIVIDUALS (%) 70 60 50 40 30 20 10 0 2016 2019 2022 No education Primary Secondary Above secondary Source: Authors’ estimates from ESPS. FIGURE O.3 The poor in rural areas remains behind 100 90 80 PROPORTION OF INDIVIDUAL (%) 70 60 50 40 30 20 10 0 Poorest Poorer Middle Richer Richest Poorest Poorer Middle Richer Richest RURAL URBAN 2019 2022 Source: Authors’ estimates from ESPS. xii Overview Despite the progress in financial inclusion, the gender gap persisted and even doubled between 2016 and 2022. It grew from 10 percentage points to 21 percentage points, and the gap persisted across all age groups. It increased across every age category between 2016 and 2022; mainly almost doubled among young men, those between 18 and 24, and among older women, those 56 and older (Figure O.4). Regardless of the positive evolution across all age groups, however, the age gap between the younger (18-24) and the older (25-55) adults, remains high among men. For example, it was 22 percentage points in 2016 and 20 percentage points in 2022, while it remains 6 and 5 percentage points, respectively, among women over the same period. FIGURE O.4 The persistence of gaps of account ownership differs by gender and age group 100 PROPORTION OF INDIVIDUALS (%) 90 80 70 60 50 40 30 20 10 0 Ages [18,24] Ages [25,55] Ages (+55) Ages [18,24] Ages [25,55] Ages (+55) MALE FEMALE 2016 2019 2022 Source: Authors’ estimates from ESPS and Findex. xiii The dynamics of financial inclusion in Ethiopia A rise in savings at formal institutions and lack of progress in borrowing are trends that simultaneously occurred over time. Savings at formal institutions in Ethiopia grew faster than regional and global comparator averages. In Ethiopia, it increased from 18% in 2016 to 24% in 2022, while it remains, on average, 23% in developing countries and 16% in Sub-Saharan African regions (Figure O.5). In contrast, formal borrowing remains limited in the country, and it is getting worse over time. In 2022, only about 3% of adults received loans from formal financial institutions (Figure O.6). FIGURE O.5 FIGURE O.6 The share of adults saving formally Formal borrowing is decreasing over time increased, over time regardless of place of residence 25 5 4,5 PROPORTION OF PROPORTION OF INDIVIDUALS (%) INDIVIDUALS (%) 4 20 3,5 3 2,5 2 15 1,5 1 0,5 10 0 2014 2016 2017 2019 2021 2022 2016 2019 2022 Ethiopia Sub-Saharan Africa Developing Country Rural Urban Source: Authors’ estimates from ESPS and Global Findex. Source: Authors’ estimates from ESPS. We assume that household members managed the loans are the beneficiaries of the loans obtained by the household. Progress in financial inclusion is underway, and an improvement in the multidimensional nature of financial inclusion also manifests this. A comprehensive measure of financial inclusion reveals that the progress of including more adults into the financial system can be directly attributed to positive evolutions across all Woredas, particularly among people in rural areas who have successfully moved into the financial system between 2016 and 2022. Notably, the deepening of financial inclusion is up-and-coming, as the share of adults who access financial services and products beyond owning accounts increased from 20% in 2016 to 29% in 2022. xiv Overview Transitioning from being unbanked to banked Understanding the nature of moving out of financial exclusion and the possible channels that can led the previously unbanked population to enter the financial system is necessary to the monitoring of existing policies and strategies and redefining them based on present opportunities. Between 2019 and 2022, about 20% of the unbanked population entered the financial system, 24% in urban areas, and 19% in rural areas (Figure O.7). It is a significant achievement, though two-thirds of the rural population remained unbanked. Figure O.8 also shows positive dynamics in saving at formal financial institutions among adults, which also sheds light on the progress of financial inclusion over the years. About 24% of urban and 13% of rural adults started saving formally between 2019 and 2022. Despite the progress, eight out of ten rural adults maintained informal savings, while only about half of the urban adults did the same. FIGURE O.7 FIGURE O.8 Movement into banked status Movement from saving informally into from 2019 to 2022 saving formally between 2019 and 2022 7 15 13 28 19 PROPORTION OF PROPORTION OF INDIVIDUALS (%) INDIVIDUALS (%) 56 24 80 67 24 48 20 RURAL URBAN RURAL URBAN UU UB BB IS-IS IS-FS FS-FS Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. UU refers to those who remained unbanked, IS-IS refers to those who kept saving informally, UB refers to those who transitioned to banked, IS-FS refers to those who transitioned to formal savings, and BB refers to those who remained banked. and FS-FS refers to those who remained saving formally. xv The dynamics of financial inclusion in Ethiopia Access to digital technology, financial knowledge, and other socioeconomic and demographic factors are the primary drivers of the move up from being unbanked. Prior to the transition, those who entered the financial system were better off in their education level, use of mobile phones, and level of financial knowledge compared to those who remained unbanked (Figure O.9 and O.10). According to estimates in chapter three, the probability of entering the financial system increases by almost 50% for those who were unbanked and had higher financial awareness than for those with no prior awareness. Similarly, previously unbanked women who have a mobile phone have a 50% higher chance of entering the financial system; and this rate is 40% for men. Notably, the study finds that the rural population that remains unbanked also had low levels of financial awareness and limited access to mobile phones. FIGURE O.9 FIGURE O.10 Account ownership trajectories and Account ownership trajectories and individual characteristics financial knowledge 75 2 69 66 PROPORTION OF INDIVIDUALS (%) 1,5 DENSITY 48 45 38 38 1 24 24 19 0,5 14 14 0 0 0,5 1 UU UB BB FINANCIAL KNOWLEDGE INDEX No education Female Age [18-24] Mobile UU UB BB Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. UU refers to those who remained unbanked, UU refers to those who remained unbanked, UB refers to those who transitioned to banked, UB refers to those who transitioned to banked, and BB refers to those who remained banked. and BB refers to those who remained banked. The channels through which new entrants access the financial system are limited. For instance, most of those who enter the financial system use banks as an entry into the system, such as public banks (65%) and private banks (33%) (Figure O.11). Agent banking model is supposed to be an ideal channel for the most disadvantageous group residing further from traditional financial access points. However, only two percent of people who moved out of the unbanked group started using agents following their entry into the system. xvi Overview Promoting and deepening financial inclusion requires diversifying channels and offering affordable financial services and products. Currently, only savings remain the main product used by banked adults. For instance, among those who desired financial inclusivity, 48% of them used savings at the entry point. Access to credit is constrained over time, suggesting a structural bottleneck in expanding financial inclusion. The rate of ATM users had almost doubled, from 16% to 29% between 2019 and 2022. However, only 15% of those who moved into the financial system started using ATMs. Thus, diversifying innovative channels and affordable services and products in hard-to-reach areas plays a significant role in enhancing financial inclusion. Digitalization is ideal for promoting financial inclusion; however, as of 2022, digital payment penetration is limited among account holders. Less than 10% of account holders use digital payment mechanisms at least once in three months to realize their main economic transactions. Cash for payment and transfer dominate economic transactions; around two-thirds of wage payments, assistances, and transfers are made by cash (Figure O.12). Opportunities exist to drive up financial inclusion through transposing payment methods across economic transactions from cash into digital payment, which has a spillover effect, as it requires those who necessitate the transaction to be included in the system. FIGURE O.11 FIGURE O.12 Banks are the main financial access points Payments are through cash, 2022 for new entrants 65 84 HOUSEHOLDS (%) PROPORTION OF PROPORTION OF PROPORTION OF INDIVIDUALS (%) INDIVIDUALS (%) 60 61 33 19 12 5 Public Private SACCOS Microfinance Mobile Wage Assistance Transfer Bank Bank Money Individual Household Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. Assistance refers direct PSNSP and other non-PSNP aid payments. Transfer indicates receiving cash transfer from friends and relatives from local or abroad. xvii The dynamics of financial inclusion in Ethiopia Narrowing the gender gap The gender gap is a financial inclusion trap. Identifying the gender disparities in the financial system and a deeper understanding of socioeconomic and structural factors responsible for the divide is essential for informing the broader policy options for fostering financial inclusion. Despite the positive dynamics in the financial inclusion landscape, the growth has yet to promote different segments of society equally. Women lag in financial inclusion regardless of place of residence. The gender gap in account ownership grew from 10 percentage points to 23 percentage points in rural areas over the last six years, while it remained at about 15 percentage points in urban areas (Figure O.13). Correlates of financial inclusion estimates shows that a woman with secondary education has an equal probability of account ownership as a man with no education. Also, women are at a lower level on the financial inclusion ladder than men. This is across all regions and worsens in rural areas. FIGURE O.13 FIGURE O.14 Gender gap on account ownership What drives the gender gap? in Rural-Urban context, over time Socioeconomic status in 2022 100 67 90 80 PROPORTION OF PROPORTION OF INDIVIDUALS (%) INDIVIDUAL (%) 70 47 41 60 35 50 40 23 30 15 20 10 0 Employed Owns mobile With some 2016 2019 2022 2016 2019 2022 phone education URBAN RURAL Female Male Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. xviii Overview Why do substantial gaps remain? The gap manifests other underlying divides between adult men and women. Estimates in Chapter Four explore the extent to which socioeconomic and structural divides contribute to the gender gap. Inequalities between adult men and adult women are manifest in observed endowments; adult women have less access to resources and opportunities, less economic participation, and less education than adult men. The estimates show the same: the share of women without education is higher than that of men, though the former decreased from 61% in 2019 to 53% in 2022 (Figure O.14). At the same time, adult men have a better chance of being employed in the labor market than adult women; the gap in employment remained above ten percentage points in 2022. Looking beyond socioeconomic characteristics, women disproportionately lack mobile phones, a gateway to technology access. The divide in mobile phone use is significant; only 15% of adult women use mobile phones, while 41% of men use them. Women are also the most vulnerable group regarding financial knowledge, as most women’s financial knowledge levels remained below the national average in 2019 and 2022. Financial knowledge is the new dimension of inequality, and contributes significantly to the gender financial inclusion gap, particularly in rural areas. About 88% of the gender gap in account ownership, in 2019 and 2022, is explained by persistent observed underlying inequalities in human capital, access to resources, and economic opportunity (Table O.1). For instance, financial knowledge distribution among adult men improved over time, while women remained below the national average in financial knowledge. Consequently, the contribution of financial knowledge to the gender gap increased from 57% in 2019 to 65% in 2022. This is manifested in a wide gender gap in financial inclusion, even though women have experienced growth in account ownership over time. The gender gap is also better explained by access to technology, employment opportunities, and income. As Table O.1 shows, access to technology plays a role; particularly in urban areas, and inequality in mobile phone usage explains about 23% and 20% of the gender gap in 2019 and 2022, respectively. The evidence presented in this report suggests that improving financial knowledge and expanding access to digital technology, including mobile phones, closes the gender gap and, in turn, promotes greater financial inclusion. By closing this gap, the country can benefit from unlocking the economic potential of millions of women. xix The dynamics of financial inclusion in Ethiopia TABLE O.1 Decomposition of the gender gaps in account ownership 2019 2022   National Urban Rural National Urban Rural Men 40.2 69 27.7 51.6 80.2 41.8 Women 22.5 49.2 9.7 30.4 64.2 18.8 Gender Gap 17.7 19.9 18 21.2 16 23 Explained Gap 15.5 18 14.1 18.4 14.5 19.4   88% 90% 78% 87% 91% 84% Contributions from gender differences (%) Demographic, socioeconomic 12% 31% 4% 13% 19% 10% Digital technology 20% 23% 16% 8% 19% 2% Financial knowledge 57% 37% 60% 65% 51% 69% Region -1% 0% -2% 2% 1% 3% Source: Authors’ estimates from ESPS. Note: The drivers of the gender gap are estimated using a multivariable Fairlie decomposition technique combining with 2SLS regression method. The gaps are in percentage points. Demographic and socioeconomic variables include age, education, employment status, and income of the individual. Region captures the contribution of estimates of group of regions. xx Overview Expanding financial inclusion: broader policy options The dynamics update and pathways to greater financial inclusion identified in this report reflect the progress in financial inclusion across different groups of society and evidence-based factors for scaling and deepening financial inclusion. One clear observation is that universal integration in financial inclusion in the country can only be realized through reaching the rural poor and women. As it is rooted in the transformation agenda and financial inclusion strategy, the study confirms that the prominent pathway to greater financial inclusion is the diversification of delivery mechanisms; a crucial need that underscores the importance of innovation, adaptability, and harnessing the digital finance system by improving financial awareness and digital literacy. DIVERSIFYING ENTRY CHANNELS Directing efforts towards expanding alternative financial access points and overcoming constraints such as distance, time, cost, quality, and product type hurdles, are reaching millions of un-/underbanked people. Traditional delivery channels that focus on geographic access only bring partial success in scaling and deepening financial inclusion in a country like Ethiopia, where three-quarters of the population resides in rural areas with limited mobile and electricity coverage - the fundamental enablers of digital finance. Hence, innovative, affordable, targeted, and efficient financial access points are required to exploit opportunities and smooth the transition from unbanked to banked through various delivery mechanisms. Agent banking model is an ideal channel, as it provides a limited range of banking services with easy access to and at low cost for rural and urban households. It is instrumental in reaching the disconnected poor in rural areas and, in turn, promoting financial inclusion (Lyman et al., 2006; Cull et al., 2018). Ethiopia applies the model and uses directives on agent banking; agent networks increased from 3 per 100,000 adults in 2016 to 77 agents per 100,000 adults in 2020 (NBE, 2021). However, demand-side evidence showed that the increase in the number of agents resulted in a slowdown in customer growth through this channel, as only very few of the new entrants used agents. This calls for an investigation into reasons why the model is lacking, and agents are not able to generate enough customers. Possible ways of enhancing its xxi The dynamics of financial inclusion in Ethiopia effectiveness include going beyond using the channel simply as a cash receiving and sending mechanism but also introducing poor- and women-tailored products (like micro- credit); diversifying agent banking providers; creating efficiency in the delivery of the services; and holding awareness creation campaigns (Cull et al., 2018). Shifting the focus to digital finance opens up new possibilities for financial access. With the rapid expansion of Mobile Money and Mobile Banking, a strong commitment to designing services and products suitable to the unbanked population, particularly women and the poor, is a high-priority approach. It is crucial for regulators to evaluate and ensure the expansion of the use of Mobile Money reaching unbanked adults. The proposed delivery channels in the NFIS-II, for instance, the expansion of interest-free banking, agent banking, and micro-insurance (e.g., linking Edir), are important enablers of financial inclusion. However, the design of digital and innovative delivery channels only benefits from considering the spatial heterogeneities of the country, such as the different economic activities in different ecological zones, the varying livelihood opportunities in different regions, and the disparities in well-being across the country. EXPANDING DIGITAL TECHNOLOGY The evolution of the digital economy in the country is inevitable. Accordingly, the contemporary trend in financial inclusion is to capitalize on the evolution of digital finance. The digital finance approach leads to expansion in account ownership and deepening financial inclusion through the provision of improved payment mechanisms, savings, and credit services. The most prominent example of this in Ethiopia is the trend seen with Telebirr over the last two years. Digital finance requires digital tools, and mobile phones are the appropriate gateway to digital technology with the potential of reaching a large customer base across broader geographic areas. Connecting the unconnected promotes financial inclusion through digitization. The analysis in this report clearly shows that owning a mobile phone is a crucial pathway to transitioning adults from unbanked to banked. The inequality in mobile phone use among adult women and adult men explains the financial inclusion gap. This means digital technology as a gateway is essential for everyone to gain alternative access to modern financial services and products in the digital era. Thus, policies addressing mobile phone availability and ownership barriers, particularly among women and the rural population, and expanding mobile phone networks are essential in promoting equity in financial inclusion. xxii Overview ENHANCING FINANCIAL KNOWLEDGE Financial inclusion correlates with human capital. The analysis in this report suggests a focus on improving the human capital among adult women and rural poor. Particularly, the level of knowledge and understanding of financial matters is minimal among women and the rural population, highlighting the need to address the gender gap in financial literacy. Promoting financial inclusion among women and rural areas is centered on enhancing the knowledge and understanding of disadvantaged groups. Given the change in the landscape of the financial system due to rapid digital finance development, direct efforts to improve the digital literacy of the public and awareness of new technologies is a precondition to realizing the most from the digitization process. Possible policy considerations to enhance financial knowledge include carrying out financial education programs at higher schools, financial literacy training at the community level, and cross-promotion using the media and social networks. Introduction Chapter one: Image: © Binyam Teshome / World Bank Within the broader context of inclusive development, financial inclusion is a critical aspect of economic development, poverty reduction, and equality. Financial inclusion refers to the access and use of financial services by individuals and businesses, particularly those underserved or excluded from the formal financial system. An inclusive financial system empowers the poor and other disadvantaged groups to gain access to a formal financial system and improve their livelihood through the opportunity to obtain loans, investments in life-enhancing strategies, savings, building resilience, risk management products, and conducting efficient and safe transactions (Bittencourt, 2010; Peres-Moreno, 2011; Brune et al., 2016; Aker et al., 2016; Demirguc-Kunt et al., 2017; Majid et al., 2019; Omar & Inaba, 2020; Field et al., 2021). More importantly, it is a key enabler for eight crucial SDGs, i.e., No poverty (SDG 1), Zero Hunger (SDG 2), Gender Equality (SDG 5), Decent Work and Economic Growth (SDG 8), and Industry, Innovation, and Infrastructure (SDG 9). As a result, financial inclusion is widely considered an important force in promoting inclusive growth and reducing poverty and inequality. The financial market in Ethiopia is characterized by a diverse array of institutions that cater to various segments of the economy and are supervised by the National Bank of Ethiopia (NBE). The sector includes 32 banks that provide a range of banking services to individuals and businesses, 18 insurers offering risk management products, and 53 microfinance institutes that extend credit to smaller enterprises and individuals who may not have access to traditional banking services. Additionally, the market comprises six capital goods finance/lease companies that enable the acquisition of machinery and equipment through leasing arrangements, and ten payment instrument issuers/system operators that facilitate transactions and digital payments. A single re-insurance company indicates a nascent stage in developing a more comprehensive risk management infrastructure within the country’s financial system. These institutions collectively form the backbone of Ethiopia’s financial market, each playing a diverse role in the mobilization of capital, provision of credit, and management of financial risks, thereby contributing to the nation’s overall economic growth and stability. 3 The dynamics of financial inclusion in Ethiopia The country developed a national financial inclusion strategy in 2016, and this was renewed in 2021, creating a solid framework for progress in this space. The National Financial Inclusion Strategy (NFIS) 2016 laid the groundwork for significant regulatory and financial infrastructure improvements, doubling financial inclusion levels (Box 1). As the implementation period for the first strategy ended in 2020, a refreshed National Financial Inclusion Strategy has been developed to build on these achievements and address new opportunities and challenges. The 2021 Financial Inclusion Strategy for Ethiopia sets clear targets and indicators to track and communicate overall progress, understand the impact of specific actions, and identify unexpected challenges, opportunities, novel trends, and diverging developments. The strategy follows a three- tiered approach for target and indicator definition under the demand and supply side, encompassing overall targets, headline targets, and supporting targets; thereby instilling confidence in its effectiveness. In response to the multiple changes in the market and society and the increasing use of mobile phones and digital technology, there has been a significant shift in how financial services in Ethiopia are accessed and utilized. These changes, in addition to their impact on poverty reduction and equality, have prompted the need for a comprehensive study to understand the evolution of financial inclusion and the factors that enable greater inclusion. BOX 1 The National Financial Inclusion Strategy The strategy defines clear objectives on the demand and supply side, identifies strategic priorities and types of interventions, and emphasizes the importance of data. The demand side’s overall target is to increase the proportion of adults who own a formal financial account from 45% in 2020 to 70% by 2025. On the supply side, the overall target is to increase transaction accounts per 100 adults from the baseline of 159 in 2019 to 337 by the end of 2025. The headline targets for the supply side include increasing digital accounts per adult, reducing the transaction account gap between regions and genders, and increasing access to credit for micro, small, and medium-sized enterprises. Finally, the strategy emphasizes the importance of reliable, fine-grained, and updated data on financial inclusion in Ethiopia. It outlines key supply-side and demand-side data sources, including data from financial institutions, consumer surveys, and government entities. The financial inclusion objectives are supported by strategic priorities identified under three different approaches: scaling, deepening, and cross-cutting enablers for financial inclusion. The strategy recognizes three fundamental types of interventions: a) consumer protection, capacity building, and market interventions; b) financial infrastructure, including financial access points; and c) innovative products and services. In addition, the strategy emphasizes the need for strategic actions to address specific challenges, such as the lack of focus on severely underserved regions and areas, the widening gap between men and women in financial inclusion, and the expansion of digital financial services, including mobile money. 4 Chapter one The report seeks to inform how the country can accelerate nationwide integration into the formal financial system to enable progress in poverty reduction, equality, and other relevant reforms under the HGERA. The report supports the ongoing reforms in the financial sector, identified strategic priorities, and types of interventions. The goal is to fully understand the extent to which individuals in the country access and use formal financial services. Measuring financial inclusion; quantifying the effectiveness of initiatives such as mobile banking, banking agents, and financial education programs; and identifying and understanding the pathways to enabling people to enter the financial system are essential for policymakers and practitioners working to develop and design policies and programs meant to promote financial inclusion. In this regard, the report emphasizes looking at various indicators simultaneously to measure the progress of financial inclusion. The objective of the study is to examine the evolution of financial inclusion of individuals over time, identify the pathways that enable an individual’s decision to move out of being unbanked, define the channels people use to enter the financial system, and explore opportunities to drive up equitable financial inclusion. The study also focuses on measuring financial inclusion in different dimensions and reviewing their simultaneous development. It provides evidence to validate constraints to and enablers of financial inclusion in the context of gender and spatial disparities in the socioeconomic landscape. While the study explores the factors that contribute to or hinder financial inclusion, it emphasizes the role of human capital, which is financial knowledge and awareness, as well as technology, in promoting financial inclusion. The study uses the Ethiopian Socioeconomic Panel Survey (ESPS). ESPS is a multitopic, longitudinal survey representing regions, as well as urban and rural areas (ESS and World Bank, 2022). The survey collects information from all household members, 18 years and above, on various aspects of financial inclusion, such as account ownership, savings, access to credit, usage of digital financial services, and financial knowledge. This report utilizes the last three rounds of the ESPS1 - ESPS3, ESPS4, and ESPS5. ESPS3, in 2015/16, interviewed 4,954 households and 11,810 adult household members, 18 and above. In 2018/19, ESPS4 collected information from 6,770 households and 14,887 adult members. The last round, ESPS5, happened in 2021/22 and interviewed 4,959 households and 12,315 adult members. ESPS5 was not implemented in Tigray region due to security reasons, hence, the analysis and discussion of this report excludes information from Tigray region from all rounds. 1 ESPS4 and ESPS5 are representative at the regional level. The region disaggregation follows the country’s regional state structure developed in the 2018 pre-census cartographic database. Thus, the SNNP region refers to the geographic coverage that existed in 2018. The analysis and discussions of this report exclude the 2019 information from the Tigray region, as the 2022 survey was not conducted in the region due to the conflict. The ESPS in 2015/16 was not representative at the regional level; the financial inclusion information from this round only applied to national and rural-urban discussions. 5 The dynamics of financial inclusion in Ethiopia The report has five chapters, including the introduction. Chapter 2 presents the status and evolution of financial inclusion in the country since 2016. The chapter starts by explaining the evolution from different dimensions using descriptive presentations. Other subtopics addressed include evidence of account ownership, savings, and changes in access to credit over time, spatially and by gender. Chapter 3 identifies pathways to greater financial inclusion, focusing on technology and human capital endowments as enablers of inclusion into the financial system. The discussion of the pathways extends to identifying the channels through which new entrants access the financial system and highlights opportunities that could enhance financial inclusion. Chapter 4 shows how the gender gap in financial inclusion can be bridged. Also presented is evidence from decomposition analyses on how gender differences, in terms of human capital, resources, and economic status, contribute to the gender gap in financial inclusion. Chapter 5 presents policy agendas focused on findings from the previous chapters. 6 Image: © Dominic Chavez / World Bank Chapter one 7 Chapter two: Understanding the dynamics: going beyond bank account ownership Image: © Boijonell Prod / Shutterstock.com Measuring financial inclusion2 Measuring and monitoring financial inclusion is not straightforward. Chakravarty and Pal (2013) note that it is crucial to measure financial inclusion in a way that assesses access to and use of financial services and products, as well as the dynamics and variation across groups and geographic regions. A large body of research has addressed measurements and monitoring mechanisms from the perspectives of the demand-side and supply-side. The Global Findex Database constructed a set of indicators that are widely used to measure financial inclusion, and these are i) account ownership, ii) savings in formal institutions, and iii) borrowing (Demirgüç-Kunt & Klapper, 2013). Some other studies have constructed a multidimensional indicator that combines i) ease of access, ii) availability, and iii) usage (Sarma, 2008 & 2012). Besides, many scholars have proposed new multidimensional methods to measure the level of financial inclusion mathematically, attempting to improve the weight allocation given to each sub-index (Amidžić et al., 2014; Mukhopadhyay, 2016; Wang & Guan, 2017; Tram et al., 2023; Dao Ha et al., 2024). For example, Cámara and Tuesta (2014) have constructed a multidimensional financial inclusion index covering 82 countries using the Global Findex indicators simultaneously with supply-side indicators. Similarly, Nguyen (2021) has measured a composite index for 40 developing countries using the supply and demand side indicators simultaneously to define financial inclusion. Also, Selvarajan and Chandran (2023) have used this mechanism to capture the progress of financial inclusion in 149 counties. Financial inclusion is a broader concept than mere account ownership at a formal financial institution. Many argue that achieving development outcomes requires simultaneous efforts towards not only promoting access to a financial account but also enhancing the use of its services, like savings, credits, and insurance, which maximizes the benefit of being in the system (Demirgüç-Kunt et al., 2017; Demirgüç- 2 Note: the region disaggregation follows the country’s regional state structure existed before 2019. Also, the SNNP region refers to the geographic coverage exited before 2019. The analysis and discussions of this report exclude the 2016 and 2019 information from the Tigray region, as the 2022 survey was not conducted in the region due to the conflict. Also, regional level dynamics analysis is based on the 2019 and 2022 data, as the 2016 ESPS is not regionally representative. 9 The dynamics of financial inclusion in Ethiopia Kunt et al., 2022). Account ownership is a gateway for the use of financial services and products, and it is the headline indicator when it comes to financial inclusion. However, it does not, for example, provide information on using credit, insurance, savings, and investments nor does it give insights into individuals’ use of digital financial services and other innovative delivery channels, such as technologies and networks that enable efficient and safe transactions. The core of the strategic approach and directions of the country’s NFIS is to scale up and deepen financial inclusion. Hence, understanding the depth of financial inclusion is as crucial as the extent of access and use of a financial account for policymakers to conceptualize how far the disadvantaged groups are from financial inclusion. A nuanced understanding of the depth of financial inclusion in a country requires a comprehensive financial inclusion indicator. One way to measure the depth of financial inclusion is to harmonize various dimensions. Though various studies have provided different approaches to measuring financial inclusion, there is consensus on the importance of identifying appropriate measurements to carefully assess the extent to which individuals have access, use various financial services and products, and participate in the financial market. Using demand side information, the report measures the financial inclusion applying three methods. First, we use the Global Findex indicators, then we construct a comprehensive measure at the individual and Woreda level. This chapter reviews the evolution of financial inclusion using and analyzing different financial inclusion measures at individual and Woreda (District) levels. GLOBAL FINDEX INDICATORS: FINANCIAL ACCESS AND USE We use the three most commonly used indicators to measure financial inclusion, in line with Demirgüç-Kunt & Klapper (2013). The first one measures ownership of an account in formal institutions (formal account) at the time of the survey. In this work, we consider as formal account owners, those who (18 years and above old) own a bank account, a microfinance account, a savings and credit association account, and are mobile money users. The second indicator is savings in formal financial institutions. This indicator measures if a respondent saves cash in the last 12 months in their formal account. The third indicator is having a loan from formal financial institutions. These indicators provide 10 Chapter two valuable information to define the inclusiveness of the financial system. This report’s dynamics and core deep dive analysis mainly depend on these indicators. However, as the above discussions lay out, using them individually may provide a partial picture of the state of the financial inclusion level. In this regard, with an aim to produce a better picture of where the country’s level of financial inclusion is positioned, this study goes beyond a single-dimension discussion by incorporating findings from the composite financial inclusion index. INDEX OF FINANCIAL INCLUSION We apply a comprehensive measure of financial inclusion, i.e. a multidimensional indicator, which looks at information on several aspects of financial inclusion. The new index captures the multidimensional nature of financial inclusion by focusing on three distinct dimensions. These are access to finance, using financial services and products, and participation in the financial market. This measurement follows the method of Zhang and Posso (2019), Nguyen NT et al. (2021), Nguyen T. A.N et al. (2023), and Nguyen H. S et al. (2023). The process incorporates people’s savings, usage of credit, usage of digital financial services and products, and insurance coverage. We also construct a Woreda-level index to study intra-Woreda differences in financial inclusion over time, capturing the spatial-temporal dynamics. In this endeavor, we follow the approach of Nguyen NT et al. (2021) to measuring the geographical distribution of financial inclusion. Based on the index we created above at the individual level; we sum up all the scores across Woredas. Finally, the total is divided by the maximum possible point (the product of the total number of adult individuals in each Woreda and the total number of questions). 11 The dynamics of financial inclusion in Ethiopia BOX 2 Definitions and Measures of variables A. Global Findex financial inclusion indicators Account ownership: Account ownership is ownership of an individual or jointly owned account at a regulated institution, such as a bank, credit union, microfinance institution, post office, or mobile money service provider. An account can be used to take a loan, save, transfer and receive money, or receive wages. Also, this report considers those individuals who accessed credit from formal institutions, including for agricultural inputs, as formal account holders but reported as not having a formal account. Savings: Saving or setting aside any money at a bank or another type of financial institution in the last 12 months. Credit: Borrowing money from a bank or another type of financial institution in the last 12 months. B. Financial inclusion indexes and levels The index is based on the definition of financial inclusion as access to useful and affordable financial products and services that meet an individual’s needs, i.e., transactions, payments, savings, credit, and insurance. We break this definition into two crucial dimensions: access and usage. The access dimension measures adult financial access through account ownership, and the usage dimension measures the availability and use of affordable financial products and services, including savings, credit, ATM, and mobile banking. The ATM and Mobile Banking indicators represent the usage of digital financing. Additionally, we consider insurance to incorporate individuals’ market participation. To measure financial inclusion, capturing different dimensions, we derive six sets of questions from the financial inclusion module of the ESPS. The six questions are selected in a way that captures a) access, b) usage, and c) participation in the financial market. Also, the index construction method is conditioned on the availability of data in ESPS among a broad list of financial inclusion dimensions. The questions are in the form of 0 or 1. We develop the index using a non-parametric approach, which assigns equal significance to each question instead of each dimension, as each of them has an equal contribution to deepening financial inclusion in the era of digitalization. Hence, following the novel approach of Nguyen NT et al. (2021), we use an equal weighting method. 12 Chapter two Dimension/Variable Description Measurement Access Account ownership It measures adult ownership Dummy of a formal account at banks, microfinance institutions, SACCOs, or mobile money service providers. Usage Saving at formal It measures whether an adult Dummy institutions practiced saving money in the last 12 months at a formal financial institution. Obtaining credit from It measures whether an adult Dummy formal institutions obtained a loan from a formal financial institution in the past 12 months. This question was asked at the household level. We assume that household members who manage the loans are the beneficiaries of the loans obtained by the household. Using ATM card It measures whether an adult Dummy used an ATM card in the last 12 months. Using mobile banking It measures whether an adult Dummy used mobile banking in the last 12 months. Note: Mobile money is a recent introduction in the financial market and ESPS only collected information on the usage of mobile money in the 2021/22 round. Participation in financial Market Owns insurance It measures whether an adult Dummy obtained risk management products from insurance providers in the last 12 months. The sum of these six indicators makes up the financial inclusion index, which ranges from 0 (lowest financial inclusion) to 6 (highest financial inclusion). For ease of discussion, we categorize the index into four levels of financial inclusion as shown in the table below. We also calculate the z-score value of the index for use of econometric analysis. 13 The dynamics of financial inclusion in Ethiopia The level of financial inclusion for each person is then obtained by categorizing the financial inclusion index as follows. Financial inclusion index Financial depth levels 0 Unbanked – value 0 From 1 to 2 Low level – value 1 From 3 to 4 Medium level – value 2 From 5 to 6 High level – value 3 C. Woreda-level financial inclusion index The woreda level index is an aggregation of the individual level index, constructed above (Box 3). %" ! % ∑$&' ∑#&' %#$ !"!!" = " !" × %& %!&( ∑$&' ∑#&' %#$ !"!!" = &!" × &( where the IFIit represents the index of financial inclusion in Woreda i in year t. Djn refers the nth individual response to question jth. Nit refers to the total number of adult individuals in Woreda i in year t. Nq refers the total number of questions. We rescale the range of index of financial inclusion across Woredas at the national average by subtracting the logarithmic form of the national average financial inclusion from the i th Woreda index in year t. . )!" (!"!!" = )*(!"!)!" − )*(!"! . )!" (!"!!" = )*(!"!)!" − )*(!"! where the TIFIit represents the transformed index of financial inclusion at Woreda i in year t. IFI is the national average in year t. 14 Chapter two Evolution of financial inclusion GROWTH IN ACCOUNT OWNERSHIP Account ownership in Ethiopia has nearly doubled in the last six years, increasing from 23% in 2016 to 41% in 2022. However, compared to neighboring countries and regional and global comparators, Ethiopia ranks lowest in terms of the share of the adult population included in the financial system (Figure 1). In 2022, Ethiopia was 14 percentage points below the 2021 average rate of Sub-Saharan African countries. The region, particularly East African countries, has witnessed a steady growth in account ownership over time; Tanzania and Uganda are the two countries that enjoyed the fastest growth between 2011 and 2021. In Ethiopia, on average, account ownership increased by 35% every three years; this is higher than the average rate the global and regional comparator countries grew. FIGURE 1 Account ownership grew, but ranks lowest compared to neighboring countries, 2011-2023 100 PROPORTION OF INDIVIDUAL (%) 90 80 70 60 50 40 30 20 10 0 DEVELOPING SUB-SAHARAN KENYA UGANDA TANZANIA ETHIOPIA COUNTRIES AFRICA 2011 2014 2016 2017 2019 2021 2022 Source: Authors’ estimates from ESPS and Global Findex. The comparator countries and regional and global rates are derived from the Findex dataset. The Ethiopian account ownership rate is estimated using the ESPS data, as the entire narration of the report is based on this data. Based on the Findex data, in 2022, account ownership rate is 46% in Ethiopia. 15 The dynamics of financial inclusion in Ethiopia The demographic and socioeconomic features of the population are dynamic, and it is crucial that the evolution of financial inclusion across different groups over time be investigated. The improvement in account ownership over time was seen in all regions, though the top three regions in 2022 were the urban-dominated regions: Addis Ababa was 86%, Harer was 65%, and Dire Dawa was 60% (Annex 2, Table A.1). In contrast, Afar and SNNP had the smallest financially included adult population. In terms of growth rate, however, Somali grew by 500%, Gambela by 50%, and Benishangul Gumuz by 48%, and these were the top three regions that showed rapid growth between 2019 and 2022. One way of reviewing the progress of account ownership is by measuring it at the household level3, and in 2022, about six out of ten households were financially included, up from three out of ten in 2016 (Figure 2a). The share of households with at least one adult owning an account increased in all regions (Annex 2, Table A.1). In Afar, it grew from 28% in 2016 to 53% in 2022. In 2022, financial inclusion at the household level was almost 100% in Addis Ababa. The gender and rural-urban gap persisted, and the gender gap doubled between 2016 and 2022. Account ownership increased from 28% to 51% between 2016 and 2022 among men, and from 18% to 30% among women, implying that over time, more men have benefitted from financial inclusion than women (Figure 3b). Despite the persistent rural-urban gap, the growth of account ownership in rural areas is significant. In rural areas, the share of the adult population that owned accounts grew from 13% in 2016 to 30% in 2022 -- a substantial increase. This growth rate was higher than in urban areas, where it grew from 51% to 72% (Figure 2b). Similarly, the rural areas enjoyed a 62% increase in the share of households with at least one adult owning an account, with 33% to 63% (Annex 2, Table A.1). FIGURE 2a FIGURE 2b Account ownership at individual and Account ownership by place household level, over time of residence, over time 70 80 PROPORTION OF INDIVIDUALS / 60 70 50 60 HOUSEHOLDS (%) PROPORTION OF INDIVIDUALS (%) 50 40 40 30 30 20 20 10 10 0 0 2016 2019 2022 2016 2019 2022 Adult At least one Adult Rural Urban Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. 3 We estimate household level account ownership rate based on measuring if at least one adult member owns an account in each household. 16 Chapter two Account ownership incidence increases with wealth; in 2019 and 2022, more than 50% of financially included adults were from the top 40% of households. Account ownership grew across all income levels, and it was relatively faster among the poorest as they started from a low level; it increased by 62%, from 10% in 2019 to 16% in 2022 (Annex 4, Figure A). However, the income gap in account ownership remained high. For instance, in 2022, wealthier adults (top 60%) were about 25 percentage points more likely to have access to finance than the poorer segment (bottom 40%). The spatial disparities in account ownership worsened among the bottom 40% between 2019 and 2022. For instance, account ownership among the bottom 40% in urban areas increased by 25 percentage points, while it only grew by six percentage points in rural areas (Figure 3a). As a result, the rural-urban gap among the bottom 40% of the adult population increased from 26 percentage points in 2019 to 45 percentage points in 2022, while it decreased from 46 to 41 percentage points among the top 60% during the same period. FIGURE 3a FIGURE 3b Rural-Urban gap in account Account ownership by gender, over time ownership, by wealth, over time U, 77 55 U, 67 U, 62 PROPORTION OF PROPORTION OF INDIVIDUALS (%) INDIVIDUALS (%) 45 U, 37 R, 36 35 R, 21 25 R, 17 R, 11 15 2019 2020 2019 2020 2016 2019 2022 POOREST 40% RICHEST 60% Male Female Source: Authors’ estimates from ESPS. U Source: Authors’ estimates from ESPS. refers Urban and R refers Rural. The account ownership rate grew across all age groups between 2016 and 2019, with the younger and the older cohorts still less likely to access finance. However, it is encouraging to see that these segments showed an improvement. Account ownership increased by 140% among younger people, those between 18 and 24, and almost doubled among older people, those 56 and above, between 2016 and 2022 (Figure 4a). The double-digit age gap in account ownership between the younger (18-24) and the older (25-55) adults persisted, showing a slight decrease from 13 percentage points in 2016 to 12 percentage points in 2022. 17 The dynamics of financial inclusion in Ethiopia Education drives account ownership; less-educated adults are more likely to be excluded from the financial system. There has been notable progress over time, with the share of account ownership growing across all education levels. It even doubled among adults without an education and those with at least a primary-level education (Figure 4b). It grew from 13% to 25% among adults without an education and from 20% to 41% among adults with at least a primary-level education between 2016 and 2022. Despite this progress, the education gap in access to finance remains high, at around 70 percentage points in all years between adults without an education and those with an above secondary education level. For instance, in 2022, 94% of adults with an above secondary education level own an account, while only 25% of adults with no education have access to finance. FIGURE 4a FIGURE 4b Account ownership by age, over time Account ownership by education, over time PROPORTION OF INDIVIDUALS (%) PROPORTION OF INDIVIDUAL (%) 50 100 45 40 80 35 30 60 25 20 40 15 10 20 5 0 0 18, 25, 31, 36, 46, 56, >64 2016 2019 2022 24 30 35 45 55 64 No education Primary 2016 2019 2022 Secondary Above secondary Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. THE RISE OF FORMAL SAVINGS AND STAGNATION IN FORMAL BORROWING Savings at formal institutions remained the primary way of saving, and this showed growth over time. In 2022, 24% of the adult population engaged in savings at formal institutions, up from 20% in 2019 (Figure 5a). The overall share of savings by adults increased from 26% in 2019 to 34% in 2022. The improvement is mainly a result of a growth in informal savings which grew by 48%, while savings at formal institutions grew 18 Chapter two by only 19%. The urban-rural gap in savings at formal institutions is persistent, though it decreased from 28 percentage points in 2019 to 17 percentage points in 2022, due to fast growth in rural areas (savings saw an increase of 57% in rural areas, while in urban areas it increased by 6%). Similarly, the gender gap in savings persists, and the downward trend was reversed between 2019 and 2022, as it grew from 14 percentage points to 18 percentage points (Figure 5b). This reversal underscores the need to emphasize gender- tailored interventions. Despite the change in savings at formal institutions, the trend in the frequency of savings at these institutions did not change, and the majority, six out of ten, expressed that they only saved whenever they got money (Figure 5c). FIGURE 5a Saving at formal institutions by place of residence, over time Country Any Urban Rural Country Formally Urban Rural Country Informally Urban Rural 0 10 20 30 40 50 2019 2022 PROPORTION OF INDIVIDUALS (%) Source: Authors’ estimates from ESPS. FIGURE 5b FIGURE 5c Gender gap in saving at formal Saving frequency, over time institutions, over time 40 2019 1 28 63 Public PROPORTION OF INDIVIDUALS (%) banks 30 2022 2 26 65 30 Private 2019 3 26 62 10 banks 2022 4 23 64 0 PROPORTION OF INDIVIDUALS (%) 2016 2019 2022 Weekly Monthly Once in 3 Months Male Female Once in 6 Months Once a year Whenever I get Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. 19 The dynamics of financial inclusion in Ethiopia The improvement in savings at formal financial institutions is a trend that is evident across all regions, with the exception of the urban-dominated regions (Annex 2, Table A.3). The top three regions in account ownership registered a decline in savings at formal institutions. Savings in any form (formal and/or informal) decreased in Addis Ababa and Dire Dawa, and more importantly, the share of adults saving at formal institutions decreased in Addis Ababa, Dire Dawa, and Harer. In disadvantaged regions, there is a promising trend of greater improvement in savings at formal institutions than in others. In Somali, it grew by 79%, in Gambela by 66%, and in Benishangul Gumuz by 42%. The promising fact is that these three regions are those that showed a rise in account ownership rate, indicating the realization of the expansion of financial inclusion in various dimensions across disadvantageous areas. Households’ overall access to credit4 increased from 16% in 2019 to 27% in 2022 (Figure 6a). This is evident in all groups; urban and rural areas, and among both the poor and the wealthy (Annex 2, Table A.4). The share of access to credit among urban households grew from 15% in 2019 to 28% in 2022, and from 16% to 28% in rural areas over the same period (Figure 6a). Moreover, access to credit increased from 17% in 2019 to 28% in 2022 among the bottom 40% of households in terms of their consumption expenditure, while it increased from 15% to 26% among the top 60% during the same period. Though all regions witnessed an improvement in access to credit, three regions showed greater improvement: Dire Dawa, where it grew by 173%, Oromia by 130%, and Somali by 80%. FIGURE 6a FIGURE 6b Household access to credit by place Source of loan by place of residence and of residence and wealth, over time wealth, over time 30 2019 64 24 12 Country 2022 75 16 9 HOUSEHOLDS (%) PROPORTION OF 25 2019 69 20 11 20 Urban 2022 79 15 6 15 2019 62 25 12 Rural 10 2022 74 16 10 5 2019 68 21 11 Top 60% 2022 75 15 10 0 Bottom 2019 59 27 14 y an l 40 m % ra tr 40% 60 2022 tto 75 18 7 rb Ru un % U p Bo Co To PROPORTION OF HOUSEHOLDS (%) 2019 2022 Relatives/Neighbor/ Formal Other Grocery institution Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. 4 Access to credit in this discussion refers access at household level, i.e., if any household member obtained loan from any source during the last 12 months. 20 Chapter two Accessing loans from formal institutions remains difficult for households. The share of households among those who obtained credit or who received loans from such institutions decreased from 24% to 16% between 2019 and 2022. This is similar across all types of households, regardless of place of residence and wealth status (Figure 6b). Relatives, neighbors, and groceries remain the primary sources of credit for households. The share of households accessing loans from such sources increased by 11 percentage points. Similarly, this share increased by 10 percentage points in urban areas and 12 percentage points in rural areas. The present chapter explores the dynamics of financial inclusion and its geographical dispersion over time, aiming to account for various dimensions of financial inclusion and its diverse geographical dispersion. Using multiple dimensions of the financial system, the findings show that the financially included adult population (age 18 and above) grew over the last six years. Account ownership nearly doubled, from 23% in 2016 to 41% in 2022. Also, the share of individuals who access financial services and products beyond account ownership increased from 20% in 2016 to 29% in 2022. In all measurements, however, the gender gap persisted over time, regardless of the place of residence. Savings at formal institutions showed only minimal growth, and in contrast, access to credit from formal financial institutions worsened over time. Overall, the findings reinforce the argument that looking beyond an increase in account ownership is crucial to understanding the extent of financial inclusion across different groups over time. SPATIAL-TEMPORAL EVOLUTION OF INDEX OF FINANCIAL INCLUSION From a multidimensional perspective, the distribution of financial inclusion (IFI) shifts to the right over time, indicating an overall improvement in the rate of financial inclusion. Though the majority lies closer to zero, the distribution higher than 0.3 appears wider in 2022 (Figure 2c). The improvement is further illustrated by the increase in the share of individuals who access financial services beyond account ownership from 20% in 2016 to 29% in 2022 (Figure 2a). As the IFI captures various dimensions of the financial system, the result demonstrates that the depth of financial inclusion showed progress, including using digital finance. However, most financially included people still remain at a low level, i.e., they only own formal accounts and use one additional financial service or product. In 2022, out of those included, 68% were at a low level of financial inclusion, down from 75% in 2016 (Figure 2b). Also, the proportion of people at a medium level of financial inclusion grew from 25% in 2016 to 31% in 2022. 21 The dynamics of financial inclusion in Ethiopia FIGURE 7a FIGURE 7b Evolution of index of financial inclusion Evolution of level of financial inclusion, (IFI), over time over time 80 PROPORTION OF 75 25 1 INDIVIDUAL (%) 2016 60 40 2019 71 29 1 20 2022 68 31 1 0 0 1 2 3 4 5 6 PROPORTION OF INDIVIDUAL (%) IFI 2016 2019 2022 Low Middle High Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. The distribution of the Woreda-level financial inclusion index also showed a forward shift between 2016 and 2022, implying an increasing trend among most Woredas whose people are financially included in the national average. More importantly, the number of Woredas with a low financial inclusion index decreased over time, as the 2022 curve is steeper than the previous years (Figure 2d). However, the distribution of the indexes proves that the rural population is at a disadvantage as far as access to the financial system is concerned. Most people living in urban areas of each Woreda were more likely to be included, as the peaks of the IFI distributions were always around one, compared to rural areas where the IFI distribution peak remained around zero (Figure 2e & 2f). However, rural areas registered a forward shift over time, while urban areas experienced a backward shift between 2016 and 2022. Such a trend indicates that the primary source of the increase in the national average emanated from the rural parts of each Woreda. FIGURE 7c FIGURE 7d Distribution of financial inclusion index, Distribution of financial inclusion index at over time woreda level, over time 2.5 0.4 2 0.3 1.5 DENSITY DENSITY 2016 2016 2019 0.2 2019 1 2022 2022 0.1 0.5 0 0 0 0.5 1 -8 -6 -4 -2 0 2 IFI IFI Source: Authors’ estimates from ESPS. We use a kernel smoothing function to define a nonparametric representation of the probability density function. 22 Chapter two FIGURE 7e FIGURE 7f Distribution of financial inclusion index in Distribution of financial inclusion index in rural areas, over time urban areas, over time 0.4 0.6 0.3 0.4 DENSITY DENSITY 2016 2016 0.2 2019 2019 2022 2022 0.2 0.1 0 0 -6 -4 -2 0 2 -6 -4 -2 0 2 IFI IFI Source: Authors’ estimates from ESPS. We use a kernel smoothing function to define a nonparametric representation of the probability density function. Addis Ababa, Dire Dawa, and Harer are predominantly urban areas and were the top three regions with a significant financially included population; their financial inclusion index was above 0.6 standard deviations from the average in 2019 and 2022. Over time, six regions showed an improvement in a z-score financial inclusion index: Afar, Somali, Gambela, Benishangul Gumuz, Oromia, and Harer, in this very order (Figure 2g). The SNNP and Somali regions were less financially included; as of 2022, they were below 0.12 and 0.25 standard deviations from the average. FIGURE 7g Evolution of financial inclusion across regions, over time 1.5 1 IFI (Z-SCORE) 0.5 0 ab is a ar la uz l ra um u aw Ab dd be G ang ar ha a -0.5 D H A am Am sh ire G ni D Be ia ar P i al -1 N m Af m SN ro So O 2019 2022 Source: Authors’ estimates from ESPS. Region level IFI is aggregated from the Woreda-level financial inclusion index. 23 Chapter three: Pathways to financial inclusion Image: © Stephan Bachenheimer / World Bank This chapter focuses on the trajectory of financial inclusion over time and explores factors that explain its dynamics. It also explores the process of moving out of financial exclusion and the possible channels that can led the previously unbanked population to enter the financial system. Because spatial heterogeneities in economic activity and livelihood opportunities are profound in Ethiopia (World Bank, 2020), we examine whether the same covariate has a different effect on financial inclusion in different areas. To explore the heterogeneity effect, we classify the sample into rural, urban, urban-dominated regions5, and other regions. The analysis uses account ownership and the constructed level of financial depth as the dependent variables. First, we study the direction and extent to which socioeconomic factors correlate with financial inclusion. Second, we investigate the dynamics of financial inclusion and pathways out of financial exclusion. The first and second analyses also answer the question about how heterogenous the correlates of financial inclusion are across places of residence. Third, we define the channels people use to enter the financial system and explore opportunities to drive up equitable financial inclusion. We use three different models to discuss the correlates of financial inclusion and explore how the relationship of the factors with financial inclusion varies over time. Each model defines the determining factors using three different financial inclusion measures, which we discussed in Chapter One. Box 3 presents the models, and a list of covariates used in each model. Tables E, F, and G in Annex 3 show the average marginal effects of the explanatory variables. Results are discussed using predicted marginal probabilities of owning a formal account and probabilities of being at a given level of financial inclusion based on several factors. 5 We classify Harer, Dire Dawa, and Addis Ababa regions as urban-dominated regions. In these regions, more than half of the population resides in urban areas, i.e., about 57% (Harer), 64% (Dire Dawa), and 100% (Addis Ababa) (ESPS, 2022) 25 The dynamics of financial inclusion in Ethiopia BOX 3 Estimated models to examine the correlates of financial inclusion First, we estimate a probit model to explain associations between individuals’ socioeconomic characteristics and financial inclusion (Eq. 1). In this regard, we run three separate models, one for the national adult sample and another two for the rural and urban adult population. Second, to examine the relationship between the factors and the level of financial inclusion, we estimate an ordered logit model (OLM) using the same explanatory variables (Eq. 2). The use of the ordered logit model is justifiable given that the level of financial inclusion does represent natural ordering. In the First, we estimate model,a the levelmodel probit to explain of financial associations inclusion between (LFI) ranges from individuals’ 1 to 3, where 1 is lowsocioeconomic level, 2 is mediumcharacteristics level, and financial inclusion (Eq. and 3 is high1). In this level regard, financial we run inclusion. three Third, separate to examine the stability one models, for the national and robustness of our adult sample and results, another two for the rural and urban adult population. Second, to examine the relationship between the factors and we run a model with a modified financial inclusion index. The financial inclusion index in this model is the z-score of the level of financial inclusion. We run an ordinary least square (OLS) model to estimate the socioeconomic characteristics effects of financial inclusion (Eq. 3). Box 4 highlights the outcome of the robustness check. "!!" = /0!" + 23!" + 4!" (1) 5"!!" = /0!" + 23!" + 4!" (2) "!)*+,-./,!" = /0!" + 23!" + 4!" (3) where subscripts denote i: individual and t: time. FIit refers i th individual account ownership status or IFI at time t - it is equal to 1 if an individual holds a formal account, otherwise it is 0; LFIit refers ith individual level of financial inclusion at time t; and FIz-score,it refers the z-score value of index of financial inclusion. Xit represents individual characteristics; and Hit presents household characteristics. The model for the full sample controls for the interaction of regions and urban/ rural areas fixed effects. The vector of the covariates is selected based on evidence (Alle et al., 2012; Fungáčová & Weill, 2015). 26 Chapter three TABLE 1 Profile of individual respondents, over time Variables Definition Measurement Descriptive Statistics-Mean Individual characteristics 2019 2022 Age Age: 18 and above Continuous (Years) 36 years 36 years Sex Sex of an individual Dummy: Men=1 48% 50% Education Highest education level of Categorical: an individual No education=1 51.1% 43% Primary =2 29.7% 34.5% Secondary =3 12.9% 15.8% Post secondary=4 6.4% 6.7% Employment status Employment status of an Dummy: 25.3% 28.7% individual Employed=1 Income Income level/ Quantile Continuous (Birr) 15574 Birr 26811 Birr Mobile phone use6 Working mobile phone use Dummy: Owns 35.9% 43.7% mobile=1 Household characteristics Proximity to a Measures how far the Continuous (Km) 16 km 16 km financial institution household member is from the nearest formal financial institution Household education Household average years Continuous (Years) 3.5 Years 3.9 Years level of education: total years of education divide by the HH size 6 In 2019, the ESPS collected detailed information on mobile phone ownership and usage using a dedicated module. In 2022, this information was aggregated from the 2019 data set and from different modules that collected information on mobile phone ownership and use in 2022. 27 The dynamics of financial inclusion in Ethiopia Correlates of financial inclusion The socioeconomic and demographic status of an individual are associated with the likelihood of being financially included. Significant demographic variables include age and gender, and men and older adults are more likely to have an account (Annex 3, Table A.5 and Table A.6). Other individual-level characteristics associated with account ownership include education, employment status and mobile ownership. Other significant factors include proximity to a financial institution and household level of education (Figure 8a). The marginal effect of gender increases over time, highlighting the widening gender gap. Women and young people are less likely to be included, and this is an issue in rural and urban areas. However, women in rural areas and in non-urban-dominated regions have a lower chance of being financially included than those in other groups (Figure 8b & Annex 3, Table A.6). As shown in the analyses, men and older adults have more access to account ownership as an economic opportunity. FIGURE 8a Correlates of account ownership in 2019 and 2022 Age Age (Sqr) Female Primary education 2019 Secondary education 2022 Post secondary education Employed Owns mobile phone Distance to Fl (Km) Log of income HH years of education (average) -0.1 0 0.1 0.2 0.3 0.4 Source: Authors’ estimates from ESPS. 28 Chapter three FIGURE 8b Correlates of account ownership by place of residence, 2019 and 2022 Urban Rural Age Age (Sqr) Female Primary education 2019 Secondary education 2022 Post secondary education Employed Owns mobile phone Distance to Fl (Km) Log of income HH years of education (average) -0.2 0 0.2 0.4 0.6 -0.2 0 0.2 0.4 0.6 Source: Authors’ estimates from ESPS. Even when women are financially included, they are more likely to be at a low financial inclusion level compared to men. In 2019, the probability of being at a low level of financial depth was 79% for women and 72% for men, and in 2022, the probability of being at a low level decreased to 67% for men while it remained at 79% for women (Annex 3, Table A.7). The result shows that, over time, the chance of being above a low level of financial depth increased for men, e.g., they are more likely to be at a medium level with 11 percentage points in 2022, up from 7 percentage points in 2019. Age does not influence the probability of being at various levels of financial inclusion. Better economic opportunities are associated with financial inclusion. For instance, being employed increased the chance of owning an account in 2019 and 2022, both in rural and urban areas (Figure 8b). The marginal effect of being employed is higher in urban areas, and the magnitude of this increased from 7% in 2019, to 13% in 2022. 29 The dynamics of financial inclusion in Ethiopia The least financially included groups are the young, women, and the uneducated. Education is correlated with account ownership, both in rural and urban areas (Figure 8b). The prevalence of account ownership is significantly higher for people with more education, and the probability is higher for men and increases with age. Returns of human capital increased between 2019 and 2022 among the young and adult cohort of the population. For example, in 2019, at age 25, the probability of account ownership among those above secondary completion was 31% for women and 40% for men, while in 2022, it was 53% and 64%, respectively (Figure 9a & 9b). As evidence of the significance of the gender gap, a woman with secondary education had an equal probability of account ownership as a man with no education. Overall, the finding highlights that the more women are educated, the higher the chances of narrowing the gender financial inclusion gap. Similarly, relatively educated banked people had a better chance of having a higher financial inclusion level than uneducated banked people (Annex 3, Table A.7). FIGURE 9a FIGURE 9b Education effect of account ownership by Education effect of account ownership by gender in 2019 gender in 2022 1 1 PROBABILITY OF ACCOUNT OWNERSHIP PROBABILITY OF ACCOUNT OWNERSHIP .8 .8 Female, No education Female, No education Female, Primary Female, Primary .6 Female, Secondary .6 Female, Secondary Female, Above secondary Female, Above secondary Male, No education Male, No education .4 Male, Primary .4 Male, Primary Male, Secondary Male, Secondary Male, Above secondary Male, Above secondary .2 .2 0 0 15 20 25 30 35 40 45 50 55 60 15 20 25 30 35 40 45 50 55 60 AGE AGE Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. Note: Predicted as marginal probability of account ownership Note: Predicted as marginal probability of account ownership based on education, tabulated by gender and age. based on education, tabulated by gender and age. Mobile phone use is the second main correlate of account ownership. This facilitates access to financial services and products, and increases the probability of account ownership, especially among women. For example, in 2019, women without mobile phones had a 45% lower probability of having an account compared to mobile owners. For men, this rate was 39% (Figure 10a). The mobile phone effect of financial inclusion remained positive in 2022 (Figure 10b), and this effect is more potent in rural 30 Chapter three areas than in urban areas (Annex 3, Table A.6). For example, in 2022, rural women mobile owners had a 128% higher probability than those without mobile phones, while urban women had only a 58% higher probability than non-owners. FIGURE 10a FIGURE 10b Mobile phone ownership effect on account Mobile phone ownership effect on account ownership by gender in 2019 ownership by gender in 2022 DO NOT OWN DO NOT OWN Female Female MOBILE MOBILE Male Male Female Female MOBILE MOBILE OWNS OWNS Male Male .2 .3 .4 .5 .3 .35 .4 .45 .5 .55 PROBABILITY OF ACCOUNT OWNERSHIP PROBABILITY OF ACCOUNT OWNERSHIP Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. Note: Predicted as marginal probability of account Note: Predicted as marginal probability of account ownership based on mobile phone, tabulated by gender. ownership based on mobile phone, tabulated by gender. Mobile phone ownership also determines the level of financial depth, though this has less of an impact over time. In 2019, the probability of being at a medium level of financial inclusion was six percentage points higher for mobile owners than for those without mobile phones. However, in 2022, mobile phone ownership appeared to have no association with the level of financial inclusion (Annex 3, Table A.7). The result implies that at some point, the mobile phone effect of financial inclusion will be negligible unless the rapid expansion of mobile phones is accompanied by a growth in the use of digital payment mechanisms. Proximity to formal institutions was an advantage, as it increased the probability of account ownership. For example, in 2019, the probability of owning an account was 17% for women and 25% for men who were 50 km away from the nearest financial institution, and this probability increased to 28% for women and 38% for men who were five km away from nearest financial institution (Figure 11a). In 2022, there was no association between financial inclusion and proximity to the nearest financial institution. This could be explained by the fact that there was an increase in financial inclusion even though, on average, people lived nine km away from the nearest formal financial institution in both 2019 and 2022 (Figure 11b). 31 The dynamics of financial inclusion in Ethiopia FIGURE 11a FIGURE 11b Proximity to financial institution and Distance to nearest financial institution account ownership by gender in 2019 (Median) .4 16 FINANCIAL INSTITUTION (KM) PROBABILITY OF ACCOUNT OWNERSHIP DISTANCE TO THE NEAREST 14 12 .3 10 8 .2 Male 4 Female 2 0 Country Urban Rural Country Urban Rural .1 0 5 20 35 50 65 80 95 110 125 140 2019 2022 Distance to the nearest financial institution (Km) Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. BOX 4 Robustness check Robustness addresses whether the results for the correlates estimates are stable to alternative financial inclusion indexes and estimation techniques. The financial inclusion index is the z-score of the level of financial depth. We run an ordinary least square (OLS) model. The results confirm the probit and ordered logit model findings for various groups. Individuals’ socioeconomic and demographic features correlate with financial inclusion (Annex 3, Table A.8). Education and the use of mobile phones are the main correlates of financial inclusion in 2019 and 2022, as people with higher levels of education and using a mobile phone are more likely to be included in the financial system. It also confirms the effects of proximity and economic opportunity, being far away from formal financial institutions and being unemployed is associated with being unbanked. 32 Chapter three Entering the financial system: Gateway technology and financial knowledge as pillars Exploring the dynamics of financial inclusion is as essential as those of poverty, price, and other economic indices. The previous discussions highlighted the evolution and correlates of financial inclusion. However, from a policy perspective, what matters is: understanding the trajectory of financial inclusion over time; understanding and supporting pathways that enable an individual’s decision to move out of financial exclusion; defining the channels that people use to enter the financial system; and exploring opportunities to drive up equitable financial inclusion. Every year, on average, about three percent of the total population turns 18 years and those in this group are expected to join the financial system.7 Hence, understanding the trajectories of financial inclusion and exploring factors that enable financial inclusion from a dynamic perspective is essential to promoting equitable inclusion for new entrants as well as for those who remained unbanked for many years. Such investigation requires multitopic, individual- level longitudinal data, and we use the 2019 and 2022 ESPS. The present study applies the Spell approach to exploring the financial inclusion trajectory. Box 5 presents the transition matrix and how we estimate factors affecting the transition. A 2x2 matrix that shows people's entry into and exit out of the financial system was drawn. The diagonal cells show the probability of remaining unbanked and being stable banked, and the off-diagonal cells present the likelihood of moving in and out of financial inclusion (Table 2). Based on the matrix, there are four possible financial inclusion trajectories, and these are: a) remain unbanked, b) enter banked, c) exit banked, and d) stable banked (Box 5). The present work assumes that account deactivations are unlikely within the context of the country, and therefore, people are not likely to go from being banked to being unbanked. Based on this, the analysis excluded 500 adults from the analysis who reported going from being banked to being unbanked. Box 6 compares the 2019 baseline characteristics of those who escaped unbanked and remained unbanked for years. 7 Authors’ estimates from the National Labor Force and Migration Survey, 2021. 33 The dynamics of financial inclusion in Ethiopia About 20% of adults entered the formal financial system between 2019 and 2022. The rate of escape unbanked in rural areas was 19% and 23% in urban areas (Table 2). Over half of the adult population (52%) remained unbanked between 2019 and 2022. In rural areas, about seven out of ten adults remained unbanked between 2019 and 2022. In contrast, more than half of the urban adult population remained banked. TABLE 2 Account ownership transition matrix 2022 National Rural Urban Status Unbanked Banked Unbanked Banked Unbanked Banked Unbanked 52.2 20.2 66.5 18.7 20.7 23.4 2019 Banked 27.7 14.9 55.9 Source: Authors’ estimates from ESPS. Note: The dynamics matrix uses the account ownership indicator. BOX 5 Transition matrix, financial knowledge index, and model estimation Transition matrix: The panel nature of ESPS allows us to develop possible individual-level account ownership trajectories. The trajectory has four groups: 1. Remain unbanked refers to the experience of being unbanked for an extended period. The present study classifies unbanked individuals in 2019 and 2022 as the remain unbanked population. 2. Enter banked refers to individuals who succeeded in moving out of financial exclusion. In this work, those who were unbanked in 2019 and owned a formal account in 2022 are classified as escape unbanked. We define this as a financial inclusion ‘entry’ scenario. 3. Exit banked refers to individuals who decided to leave financial inclusion in 2022. We define this as a financial inclusion ‘exit’ scenario. 4. Persistent banked refers to individuals who remained banked over many years. In this work, these are who were banked in 2019 and 2022. 34 Chapter three Financial knowledge index: Financial knowledge is a key predictor of financial inclusion (Cole et al., 2011; Hasan and Hoque, 2021, Ansar et al. 2023). Hence, we estimate the effect of financial knowledge on the probability of entering the financial system. We have constructed a comprehensive financial knowledge index using 11 indicators, each assessing whether the respondent had heard about key financial concepts prior to the interview day. If an individual had heard about an indicator, 1 is selected, otherwise it is 0. These indicators are grouped into four dimensions, each carrying an equal weight (25%). The dimensions are: 1. Knowledge of financial institutions: Three indicators ask about knowledge of public banks, private banks, and insurance companies. Each indicator has a weight of (8.33)%; 2. Knowledge of financial services and products: Four indicators that ask about knowledge of money transfer, mobile money agents, bank agents, and ATM. Each indicator has a weight of (6.25)%; 3. Knowledge of financial terms: Three indicators ask about knowledge of collateral, interest, and inflation. Each indicator has a weight of (8.33)%; 4. How to open an account: Measures whether the respondent knows how to open an account in formal institutions, and the dimension is worth 25% of the total weight. Thereafter the weighted indicators are summed up to obtain an individual’s financial knowledge score. Estimating transition probabilities: We estimate logit models to examine the correlates of moving out of exclusion for different subsamples. We model the logit models to estimate the correlates of moving out of exclusion for those who started unbanked in 2019, i.e., financial inclusion ‘entry’ factors. All models estimate factors in 2019, aiming to limit reverse causality. 89!" = /0!"*' + 23!"*' + 4!" where subscripts denote i: individual and t: time. UBit refers ith individual trajectory status at time t - it is equal to 1 if individual i escape unbanked and 0 otherwise; Xit represents individual characteristics; and Hit presents household characteristics (Table 1). In addition to variables in Table 1, we include the financial knowledge index in the model to examine the effect of the level of financial knowledge on financial inclusion ‘entry’. The estimation controls for regions and urban/rural interaction fixed effects. Annex 3, Table A.9 presents the average marginal effect of the explanatory variables. Results are discussed using predicted marginal probabilities of owning a formal account based on several factors. 35 The dynamics of financial inclusion in Ethiopia BOX 6 Comparison between those who remained unbanked and those able to enter the financial system Prior to the transition, those who entered financial inclusion were better off in their education level, use of mobile phones, and level of financial knowledge compared to those who remained unbanked. More than half of those who moved into inclusion are men, a quarter are youth between the ages of 18 and 24 (Figure 12a) and about four out of ten use mobile phones. Meanwhile, those who remain unbanked are primarily women, uneducated people, people with poor financial knowledge, and mainly without mobile phones (Figure 12a & 12b). The characteristics of the persistent banked people show that women and the youth are the more disadvantaged segments of the population in the financial system, as 15% are youth and 40% are women. FIGURE 12a FIGURE 12b Account ownership trajectories and Account ownership trajectories and individual characteristics financial knowledge 2 75 PROPORTION OF 69 INDIVIDUAL (%) 66 1.5 DENSITY 48 45 1 38 38 19 24 24 0.5 14 14 0 UU UB BB 0 0.5 1 No education Female FINANCIAL KNOWLEDGE INDEX Age (18-24) Mobile UU UB BB Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. A higher share of those who remained unbanked live in regions other than urban-dominated regions (Addis Ababa, Dire Dawa, and Harer). In urban-dominated regions, only one out of ten adults remained unbanked for years; in contrast, more than half of the population of the other regions remained excluded from the financial system (Figure 12c). However, the rate of moving out of exclusion appears to be equal between these two groups. Such features of the dynamics of financial inclusion reaffirm the notion that the gender and spatial gap are not narrowing, as those escaping exclusion are more likely to be men and from mostly urban areas. Overall, the result highlights the gap in education, financial knowledge, and the use of mobile phones between those remaining unbanked and those able to escape exclusion. 36 Chapter three FIGURE 12c Account ownership trajectories by region URBAN-DOMINATED REGION 10 18 71 UU UB 55 20 24 BB OTHER REGIONS PROPORTION OF INDIVIDUALS (%) Source: Authors’ estimates from ESPS. The youth, women, and less educated are the least likely to enter the financial system. Those aged 25 to 44 have a better chance of entering the formal financial system than the youth. For example, men between the ages of 18 and 24 have a probability of roughly eight percentage points lower than those aged 25 to 44 (Figure 12a). However, at any age level, the probability of remaining unbanked for years is higher for women. About 3 million8 people turn 18 each year, and that combined with the fact that the youth are less likely to enter the financial system at a younger age, underscores the need for youth and women-focused strategies that improve financial accessibility and usability for them. It is worth noting that, as the heterogeneity analysis highlights, age in urban areas and both gender and age in urban-dominated regions did not have any effect on the financial inclusion trajectory (Annex 3, Table A.9). The education gap adequately explains why some remain unbanked while some enter the financial system. People above the secondary education level are more likely to escape years of being unbanked. For example, the probability of moving out of exclusion for those with no education is only 20% (women) and 34% (men), and this increases two-fold for women with above secondary education level and doubles for men with the same (Figure 12b). The result also reveals that education is the most important determinant of how quickly women can enter the financial system, as the rate of the probability of moving out of exclusion grows with their education level. 8 In Ethiopia, according to the recent national labor force survey, every year about 3 million young people (3% of the total population) reach age 18 (NLFS, 2021). 37 The dynamics of financial inclusion in Ethiopia FIGURE 13a FIGURE 13b Age and sex are associates with entering Education matters to escape from being the financial system unbanked Female Female AGE, NO 18-24 Male EDUCATION Male Female Female AGE, 25-34 PRIMARY Male Male Female Female AGE, 35-44 SECONDARY Male Male Female Female AGE, ABOVE >44 OLD Male SECONDARY Male .2 .25 .3 .35 .4 .2 .4 .6 .8 PROBABILITY OF MOVING OUT OF EXCLUSION PROBABILITY OF MOVING OUT OF EXCLUSION Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. Note: Predicted as marginal probability of moving out of Note: Predicted as marginal probability of moving out of exclusion based on age group, tabulated by gender. exclusion based on education level, tabulated by gender. Mobile phone utilization increases the probability of escaping from being unbanked for years and, therefore, drives up financial inclusion. An expansion of digital technology offers more opportunities for previously unbanked people to access and use financial products and services anytime without visiting a physical institution (Aker et al. 2016; Lenka and Barik, 2018; Senou et al. 2019; Shen et al., 2020; Demirgüç- Kunt 2022). This study proves that unbanked people who use mobile phones have a better chance of entering the financial system. For example, previously unbanked women, who have a mobile phone, have a 50% higher chance of entering the financial system; and this rate is 40% for men (Figure 14). The correlation between the prevalence of mobile phones and entering the financial system is evident in rural and urban areas (Annex 3, Table A.10). However, the use of mobile phones, particularly in rural areas, is still limited, despite an expansion between 2019 and 2022. Mobile phone usage increased from 35% to 52% among men, and from 9% to 19% among women (Annex 3, Figure C). Hence, the result reveals that expanding mobile phone use in rural areas could offer more access to the rural unbanked population, particularly the women. FIGURE 14 Mobile ownership plays a great role to accelerate financial inclusion Female DO NOT OWN MOBILE Male OWNS Female MOBILE Male .2 .3 .4 .5 Source: Authors’ estimates from ESPS. Note: Predicted as marginal probability of moving out of exclusion based on mobile ownership, tabulated by gender 38 Chapter three A low level of financial knowledge contributes to keeping people unbanked for lengthy periods, as a lack of financial awareness prevents people from taking advantage of services and products offered by formal institutions. Financially literate customers demand financial inclusion for better financial decisions (Grohmann et al. 2018). For example, the probability of entering the financial system increases by almost 50% for those who were unbanked and had higher financial awareness than for those with no prior awareness. The effect of financial knowledge on easing the pathway to financial inclusion is demonstrated among men and women (Figure 15a). Unbanked women with no financial knowledge have a 19% probability of moving out of unbanked, while it increases to 30% for those who are financially conscious. FIGURE 15a FIGURE 15b Financial knowledge facilitates the move Financial knowledge gap between urban towards financial inclusion and rural areas in 2019 .4 .5 PROBABILITY OF MOVING OUT OF EXCLUSION .4 .3 .3 Female DENSITY .2 Rural Male Urban .2 .1 .1 0 0 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 -2 -1 0 1 2 3 FINANCIAL KNOWLEDGE INDEX FINANCIAL KNOWLEDGE (Z-SCORE) Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. We use a Note: Predicted as marginal probability of moving out of kernel smoothing function to define a nonparametric exclusion based on financial knowledge, tabulated by gender. representation of the probability density function. Financial knowledge as a pathway to entering the financial system was more prevalent in urban areas. Rural areas have low levels of financial knowledge, as most of the adult population live with a low level of financial knowledge, i.e., <-1 standard deviation from the average, in contrast to urban areas where the majority have high financial awareness (Figure 15b). Chapter One documented that the rural adult population is one those who are excluded from financial inclusion. Hence, expanding financial education programs across rural areas means enabling the rural unbanked people to move out of exclusion. 39 The dynamics of financial inclusion in Ethiopia The distance from the nearest financial institution is also a determining factor in entering the financial system. People proximate to formal institutions, e.g. within a 5 km radius, had a higher probability (35% for women and 31% for men) of entering the financial system than those farther away from formal institutions e.g. 100 km (Figure 16). FIGURE 16 Proximity to formal financial institutions matters to escape unbanked .4 PROBABILITY OF MOVING OUT OF .35 .3 EXCLUSION Female Male .25 .2 .15 5.00 20.00 35.00 50.00 65.00 80.00 95.00 110.00 125.00 140.00 DISTANCE TO THE NEAREST FORMAL FINANCIAL INSTITUTION (KM) Source: Authors’ estimates from ESPS. Note: Predicted as marginal probability of moving out of exclusion based on proximity, tabulated by gender. 40 Chapter three Opportunities to boost financial inclusion Enablers are required to boost the use of more financial services. New entrants to the financial system tend to use only minimal services. Eight out of ten adults who became financially included in 2021, UB in Figure 17, were still at a low level of financial inclusion9. Also, less than half of those who stayed in the system over the years (BB) are above the low level of financial inclusion. FIGURE 17 New entrants have a low-level financial inclusion BB 57 41 2 LOW MIDDLE 82 18 0 HIGH UB PROPORTION OF INDIVIDUALS (%) Source: Authors’ estimates from ESPS. There is a limited number of channels used by people to access the financial system and the range of affordable services and products. For example, in 2022, the highest share of account owners used public banks (74%), followed by private banks (42%) (Annex 2, Table A.2). However, over time, the share of adults using private banks doubled between 2016 and 2022 (Annex 4, Figure D). The expansion of private bank use is dominant in rural areas, e.g., it increased from 17% in 2019 to 34% in 2022 (Annex 4, Figure E). Similarly, most of those who enter the financial system use banks as an entry into the system, such as public banks (65%) and private banks (33%) (Figure 18a). In some regions, however, the use of MFIs and SACCOs is more common. For example, 34% of account owners in Somali own an account at MFIs, and 34% of account owners in Amhara use SACCOs (Annex 2, Table A.2). The diversity in available institutions in some areas could play a significant role in enhancing financial inclusion. 9 The level of financial inclusion is defined and discussed based on a multidimensional indicator of financial inclusion (see Box 2). 41 The dynamics of financial inclusion in Ethiopia FIGURE 18a FIGURE 18b Banks are the main get way Saving is a commonly used to the financial system product by new entrants 64 PROPORTION OF INDIVIDUALS (%) 80 65 PROPORTION OF INDIVIDUALS (%) 48 48 33 8 12 5 37 20 19 4 5 15 k k S ce ey 8 n n O an on Ba Ba CC fin M SA ic e Formal Formal ATM at bl ro ile iv Pu ic ob saving loan Pr M M BB UB BB UB Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. Banking agents are cheaper channels used by suppliers and users to increase access and use of financial products. With lower costs and closer proximity to society, agents are in a good position to serve the unbanked, particularly in rural areas where the transaction cost of access and supply is high (Afabde & Mbugua 2015; Cull et al., 2018). However, the use of banking agents to access the institution’s products and services has remained low over time -- less than three percent (Annex 4, Figure B). Likewise, only two percent of people who moved out of the unbanked group started using banking agents following their entry into the system. The channel is supposed to be efficient, and an expansion could boost financial inclusion, particularly in rural areas. The channels are not performing as well in linking the unbanked population and the institutions. A thorough review of the existing model is recommended to identify reasons the system is not working to bridge the gap between the unbanked and the institutions. Accessing loans from formal financial institutions remains difficult for households, and savings is a widely used product by account owners (Figure 2.4 & 2.5). Savings also appeared to be the main product for those who desired financial inclusivity; 48% of them used savings at the entry point, as it has been a widely used product by persistent banked people (64%) (Figure 18b). The probability of accessing credit from formal institutions for any group remained low, less than 10% (Figure 18b). This implies that access to credit is not only a problem for new entries but also a structural problem in the country. From a service perspective, the rate of ATM users among the persistently banked increased from 29% to 37% between 2019 and 2022. However, only 15% of those who moved into the financial system started using ATMs (Figure 18b). 42 Chapter three FIGURE 19a FIGURE 19b Households and individuals receive Account holders hardly use digital payments through cash, 2022 mechanisms to make payments, 2022 Bill payment 48 Agriculture transfer 37 CASH 84 HOUSEHOLDS (%) PROPORTION OF PROPORTION OF INDIVIDUALS (%) Business transfer 32 61 Family/Friend transfer 27 60 Bill payment DIGITAL PAYMENT 9 Agriculture transfer 8 Business transfer 5 Wage Assistance Transfer Family/Friend transfer 4 Individual Household PROPORTION OF ACCOUNT HOLDERS (%) Source: Authors’ estimates from ESPS. Assistance refers Source: Authors’ estimates from ESPS. Digital payment direct PSNP and other non-PSNP aid payments. Transfer method includes ATM, Online\Mobile Banking, and Mobile indicates receiving cash transfer from friends and relatives Money. We define as digital payment when an account from local or abroad. holder made the transactions using these methods at least once in 3 months in the last 12 months. Most individuals and households rely on cash for payment and transfer in their economic transactions. For example, in 2022, 60% of wage employees received cash for their payments (Figure 19a). Again, 61% of households who received transfers obtained cash payments. Likewise, 84% of direct PSNP 10 and other humanitarian beneficiary households received payments in cash. In Ethiopia, there are 9.7 million PSNP beneficiaries, of which about 82% are from rural areas. The facts reveal that opportunities exist to drive up financial inclusion through transposing payment methods across economic transactions from cash into digital payment. It is also worth noting that almost half of the account holders made bill payments only in cash. At least four out of ten rural people made fund transfers related to their agriculture transactions only in cash (Figure 19b). The result seems to indicate that individuals’ use of cash in making payments and transfers is lower in general. However, more than 90% of non-cash payments and transfers are conducted through bank tellers. Only less than 10% of account holders use digital payment mechanisms at least once in three months to realize their main economic transactions (Figure 18b). The results show that the use of digital payments, even among individuals who have formal accounts, is weak. 10 The Ethiopian government flagship Productive Safety Net Program (PSNP), as of 2021, benefits about 8 million people in rural areas and 1.7 million people in urban areas. Of which, 2.2 million are direct support beneficiaries. 43 The dynamics of financial inclusion in Ethiopia In the era of digital finance, people are in a better position to use their mobile phones to access financial services and products anytime and from anywhere. However, mobile money is a new introduction to the country’s financial system, and as of 2022, it has yet to appear as a channel for new entrants. In 2022, only six percent of men and four percent of women had mobile money accounts, and among those who moved out of unbanked, only five percent owned a mobile money account (Figure 17a). Since 2015, banks and microfinance institutions have been offering the service with limited success. However, the introduction of TeleBirr in 2021 and M-PESA in 2022 has played a vital role in boosting financial inclusion through scaling digital payments via mobile money service. For example, as of the end of 2023, TeleBirr had 41.1 million11 and M-PESA had 3.1 million12 subscribers, evidence that the platforms are emerging as the backbone of digital financial service. A move towards shifting payments and transfers from cash into a digitalized merchant system can serve as an entry channel into the financial system for the unbanked population, mainly in urban areas where digital technology is prevalent. The study reveals that saving accounts at banks are widely used channels for entry into the formal financial system, and most economic transactions are made in cash or through a teller window. The implication is that digital payment penetration in the country is light, even among the financially included population, which explains the wide opportunity for boosting financial inclusion beyond account ownership through digitalized merchant payments. There is a growing consensus over the fact that the evolution of the digital economy in the country is inevitable, and the digital payment system is a core enabling system to achieving the goal13. Hence, to smooth the transition, the study shows that there is a need to improve financial awareness and digital literacy, particularly in rural areas that have been historically excluded and where there is a lack of sufficient knowledge. Moreover, supply-side interventions to expand digital finance and increase the usability of digital payments will spill over onto individuals who have been unbanked for years, as it requires those who necessitate the transaction to be included in the system. 11 Ethio Telecom 2023/24 Semi-Annual Business Performance Report. https://www.ethiotelecom.et/ethio- telecom-2023-24-semi-annual-business-performance/ 12 Safaricom Telecommunications Ethiopia PLC (STE): Q3 update as of 31 Dec 2023. https://www.safaricom.co.ke/ images/Downloads/Safaricom-Ethiopia-Quarterly-Update-Q3-FY24.pdf 13 Digital Ethiopia 2025: Strategy Summary. https://mint.gov.et/wp-content/uploads/2022/01/Summary_of_Digital_ Strategy_Final_English1.pdf 44 Image: © Natalia Cieslik / World Bank Chapter three 45 Chapter four: Gender gap in financial inclusion: trend, expla- nation, and bridging the divide Image: © Dana Smillie / World Bank The effective participation of women in the socioeconomic and political dynamics is crucial to promoting inclusive and resilient economic growth. Despite their potential to positively contribute to the economy, women are unequal economically and socially; notably, they are currently the most disadvantaged group socioeconomically (Wodon & Brière, 2018; Santos & Klasen, 2021). This economic inequality is a manifestation of the fact that women have lower economic participation, and compared to men, are limited in terms of education, access to resources and opportunities, and decision-making power. These disparities hinder their ability to contribute to the economy (Wodon & Brière, 2018; Santos & Klasen, 2021). Similarly, women are less likely to be included in the financial system, a key factor in empowering them. Financial inclusion could provide them with access to opportunities to obtain loans, invest in life-enhancing strategies and savings, build resilience and risk management products, and conduct efficient and safe transactions (Demirgüç-Kunt et al., 2022). The financial inclusion inequality exacerbates their exclusion from economic opportunities. Thus, ending gender inequality in access to finance and the use of services and products empowers them and, in turn, enhances economic and social well-being, thereby underlining the broader impact of financial inclusion (Hendriks, 2019; Field et al., 2021; Arshad, 2023). The predictors of a gender gap in financial inclusion explain their contribution to fostering financial inclusion. Gender disparities in the financial system can be explained by socioeconomic status, structural context, norms, access to digital technology, and financial consciousness (Ndoya & Ysala, 2021; Özşuca, 2019; Mndolwa & Alhassan, 2020; Hundie & Tulu, 2023). Building on the discussion in the previous chapter on the correlates of account ownership and the financial inclusion score, this chapter centers on the contributors to the gender gap. The chapter discusses the factors using non-linear and linear decomposition frameworks to determine how human capital, resources, and economic opportunity define the gaps. 47 The dynamics of financial inclusion in Ethiopia Women lag in financial inclusion and are more likely to remain unbanked. The gap persists and grows over time, and grew, on average, from 10 percentage points in 2016 to 21 percentage points in 2022 (see Figure 3b). As the predicted marginal probability estimate from the previous chapter shows, the gap is seen across all age groups (Annex 4, Figure F & G). The issue is not only about women being excluded from the financial system, most financially included women are also at a lower level on the financial inclusion ladder than men. In 2016, only 27% of included men and 22% of included women were at an above-low level, highlighting a five-percentage point gap. However, this gap grew to 15 percentage points in 2022 (Figure 20). The gender gap is also evident in savings; it narrowed between 2016 and 2019 but increased by four percentage points between 2019 and 2022 (Figure 5b). FIGURE 20 Gender gap in financial inclusion level increased, over time 2016 73 27 1 MALE 2019 71 29 1 2022 66 33 2 2016 78 22 0 FEMALE 2019 81 19 1 2022 80 20 0 PROPORTION OF INDIVIDUALS (%) Low level Medium level High level Source: Authors’ estimates from ESPS. 48 Chapter four There is always a gender gap in account ownership, regardless of the location or area, and this doubled in rural areas between 2016 and 2022. However, the gap is narrowing in urban areas and urban-dominated regions (Figures 21 & 22). In other regions, except Afar, the gap increased between 2019 and 2022. In 2022, all regions except Addis Ababa and Dire Dawa showed a significant gap – higher than ten percentage points, and among these, SNNP and Gambela were the top two regions with a wider gender gap. In 2019, the top two regions where the gender gap was higher were SNNP and Harer, revealing that women in SNNP are at a bigger disadvantage in terms of financial inclusion than women in other regions. FIGURE 21 FIGURE 22 Gender gap on account ownership Gender gap in account ownership region in Rural-Urban context, over time context, over time 80 70 PROPORTION OF INDIVIDUALS (%) 60 40 PROPORTION OF INDIVIDUALS (%) 50 35 40 30 25 30 20 20 15 10 10 5 0 0 2016 2019 2022 2016 2019 2022 a a i ar an m r l G ra uz ia G NP la al A a ab aw m be Af ar gu ha m um SN Ab ro D H am So URBAN RURAL O ire s di D Ad Male Female sh 2019 2022 ni Be Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. 49 The dynamics of financial inclusion in Ethiopia Explaining the gender gap Do gender differences in income, education, employment status, financial knowledge, and access to mobile phones and the internet explain the existing gender gap in financial inclusion? Men are more likely than women to be educated and employed. For example, men are relatively better off in education and labor force status than women, and the divide in these characteristics has remained the same over time (Figures 23a & 23b). The share of women without an education decreased from 61% in 2019 to 53% in 2022, and for men, from 41% to 33% over the same period. Even though human capital among women increased, it was still below that of men, as the share of uneducated women was 20 percentage points higher. Similarly, men perform better in the labor market, as men are more likely to be employed than women. The employment gap stayed at about 10 percentage points between 2019 and 2022. Women are less likely to own a phone and hence less likely to take the opportunities digital technologies offer. The gender gap in mobile phone use was wider in 2022 than in 2019. It increased from 19 percentage points in 2019 to 26 percentage points in 2022 (Figure 23c). The gap in usage worsened in rural areas; it increased from 26 to 33 percentage points during the same period, while it narrowed in urban areas from 19 percentage points to 13 percentage points during the same period (Annex 4, Figure C). FIGURE 23a FIGURE 23b FIGURE 23c Lacks education, by Employed adult by Mobile use, by gender, over time gender, over time gender, over time 35 41 PROPORTION OF PROPORTION OF PROPORTION OF INDIVIDUALS (%) INDIVIDUALS (%) INDIVIDUALS (%) 61 31 53 23 30 41 21 33 15 11 2019 2022 2019 2022 2019 2022 Female Male Female Male Female Male Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. 50 Chapter four The previous chapter also shows that financial knowledge is a major determinant for entry into the financial system, however, women are less financially conscious. Figures 24a and 24b highlight the gender gap in financial knowledge, and this gap grew between 2019 and 2022. The distribution of a z-score financial knowledge index for men showed a substantial forward shift between 2019 and 2022. Accordingly, the distribution peak moved from 0.5 standard deviations below the average to one standard deviation above the average. Though the distribution for women also displayed a forward shift, most women’s level of financial knowledge remained below the average in 2019 and 2022. Overall, the distribution trends reveal that men experienced an improvement in financial knowledge over time, while women remained the most vulnerable group in financial knowledge. Is the financial knowledge divide responsible for the gender gap in financial inclusion? FIGURE 24a FIGURE 24b Financial knowledge distribution, by Financial knowledge distribution, by gender, 2019 gender, 2022 .4 .4 .3 .3 DENSITY DENSITY Female Female Male .2 Male .2 .1 .1 0 0 -2 -1 0 1 2 3 -2 -1 0 1 2 FINANCIAL KNOWLEDGE (Z-SCORE) FINANCIAL KNOWLEDGE (Z-SCORE) Source: Authors’ estimates from ESPS. We use a Source: Authors’ estimates from ESPS. We use a kernel smoothing function to define a nonparametric kernel smoothing function to define a nonparametric representation of the probability density function. representation of the probability density function. 51 The dynamics of financial inclusion in Ethiopia Bridging the gender gap and fostering financial inclusion The decomposition analysis is relevant for policy dialogue and implementation, as differences in characteristics largely explain the existing gaps. There is a divide between men and women in terms of human capital, resources, and economic status. To investigate how this divide contributes to the gender gap in financial inclusion, we use non-linear and linear decomposition methods. Using the Fairlie non- linear decomposition method, we explore the gap in account ownership. The Fairlie method is an extension of the standard Oaxaca-Blinder (1973) decomposition for logistic regression (Box 7). Aiming to explore the gap beyond account ownership and examine the stability and robustness of our results, we apply the Oaxaca-Blinder linear decomposition method by changing the dependent variable to the financial inclusion index (Box 7). We perform the gender decomposition for 2019 and 2022 to examine the change in the contribution of each covariate and the coefficients in explaining the gaps. We also examine whether the men and women characteristics contribute differently in different geographic locations to the gender gap. We classify the sample into rural and urban to explore the heterogeneity effect. A regression model with financial knowledge as a covariate has an endogeneity issue due to reverse causality. On the one hand, financially included individuals have more opportunities to participate in a wide range of financial activities, which in turn improves their financial knowledge. On the other hand, a low level of financial knowledge may constrain the inclusion of individuals in the financial system. To alleviate the endogeneity issue, we use a 2SLS model to estimate the average predicted probability of financial inclusion among men and women prior to the decomposition analysis (Box 7). 52 Chapter four BOX 7 Decomposition analysis combining with 2SLS regression method Modeling financial inclusion: 2SLS regression method The decomposition method starts with estimating the average predicted probability of account ownership among men and women using the covariates listed in Table 1. Also, the decomposition analysis includes the financial knowledge indicator as a covariate (see Box 5). Due to endogeneity concerns from reverse causality between financial inclusion and financial knowledge, we apply the 2SLS model. First stage: We use the financial knowledge score of the household head and the community at the national average as instrumental variables. We believe individual financial knowledge can improve through intra-household and intra-community interactions. However, the household head and community financial knowledge score does not directly affect the individual probability of being included in the financial system. ":!" = /0!" + 2":3!" + 2":;!" + 4!" where subscripts denote i: individual and t: time. FKit refers i th individual financial knowledge status <(=!" = 1|@!" ) = /0!" + 4!" at time t (see Box 5); Xit represents individual characteristics; FKHit presents household head financial knowledge score; and FKCit refers community level knowledge score at the national average (Annex 1). ":!" = /0!" + 2":3!" + 2":;!" + 4!" Second stage: We estimate a logistic regression model: <(=!" = 1|@!" ) = /0!" + 4!" where P(Yit = 1|xit) refers ith average predicted probability of account ownership at time t; Xit represents individual characteristics, including financial knowledge score (FKit). Both the first and second stage regression control for regions and urban/rural interaction fixed effects. Fairlie decomposition framework We use a multivariable Fairlie decomposition technique using the second stage estimate of the multivariate logistic regression from the 2SLS method. The decomposition for a non-linear equation is expressed as: %# %$ %$ %$ E 1F "D0!1 / E 1F "D0!2 / E1F "D0!2 / E2F "D0!2 / A 2 = BC A1 − = = −C G + BC −C G &1 &2 &2 &2 !&' !&' !&' !&' ¯ F is the average predicted probability of men’s and women’s financial ¯ M and Y where F(x) = P(Y = 1|x), Y inclusion, βM and βF is a vector of coefficient estimates for men and women, N are group sample size, and F is the logistic cumulative distribution function. The overall gap in the mean value between men and women is decomposed into the portion of the financial inclusion gap that is due to differences in the distribution of X, i.e., the first term, and the part due to differences in the group process determining levels of Y (i.e., differences in the coefficients), i.e., the second term. The second term also captures the portion of the gender gap due to differences in unobserved endowments, it represents the effect of discrimination. "!)*+,-./,!" = H + /0!" + 4!" 53 A1 − = = A1 − 0 A 2 )/ A 2 = (0 E1 + 0 E 2 F + (I E1 − / A 2 D/ J2) J1 − I %# %$ %$ %$ E 1F "D0!1 / E 1 F dynamics The "D0!2 / "D0of financial 2 E 1 2 E 2 in Ethiopia inclusion A1 A2 ! / F "D0 ! / F = − = = BC −C G + BC −C G &1 &2 &2 &2 !&' !&' !&' !&' %# %$ %$ %$ E 1F "D0!1 / E 1F "D0!2 / "D0!2 /E2F E1F "D0!2 / A1 − = = A2 = %#BC 1 1 −% C$ G + BC % − C %$ $ G 2 "D0 & E /1 F 2 "D0& !2 E / F1 "D0 2 E 1 &2 ! / F &"D0 2 2 E ! / F A 1 A 2 !&' ! !&' !&' !&' = − = = BC −C G + BC −C G &1 &2 &2 &2 !&' !&' !&' !&' The total contribution of gender differences in the covariates to the gender gap in account ownership lets us estimate the individual contribution of each explanatory variable to the overall gap. The contribution of each factor responsible for the gap is thus equal to the change in the average predicted probability from replacing the women distribution with the men distribution of a specific variable, while holding the distributions of the other variables constant. The separate contributions of explanatory variables are responsive to the ordering of variables because the contribution of each covariate depends on the distributions of all covariates. Following Fairlie (2017), we randomize the order of variables across the simulations; we draw 1000 sequences of each decomposition with the ordering of covariates being randomly determined and then average results over the draws. Oaxaca-Blinder decomposition framework To explore the source of the gender gap, we also follow the linear Oaxaca-Blinder decomposition method (OB) (Oaxaca, 1973; Blinder, 1973). Again, the decomposition uses the 2SLS estimation method. "!)*+,-./,!" = H + /0!" + 4!" where Xit represents individual characteristics at time t, including financial knowledge score, (FKit). "!)*+,-./,!" = H + /0!" + 4!" = A1 − 0 A 2 = (0 A1 − = E1 + 0 A 2 )/ A 2 D/ E 2 F + (I E1 − / 1 J2) −I The difference between the average financial inclusion score J of men and women could be expressed as: A1 − = = A1 − 0 A 2 )/ A 2 = (0 E1 + 0 E 2 F + (I E1 − / A 2 D/ J2) J1 − I "!)*+,-./,!" = H + /0!" + 4!" ¯ F refers the financial inclusion index (z-score) of men’s and women’s, βM, βF, âM, and ¯ M and Y where Y âF is a vector of coefficient estimates for men and women, X represents the vector of means of covariates predicting financial inclusion. A1 − = = A1 − 0 A 2 )/ A 2 = (0 E1 + 0 E 2 F + (I E1 − / A 2 D/ J2) J1 − I The decomposition analysis provides two sources of the gender gap: composition effect and coefficient effect. The first term on the right side of the equation shows the portion of the financial 4 "3 differences inclusion gap that is due to gender = O "3|6,7&8 inM" the P(0) distribution 6| 7&8 of X. The second and the third term measure the part due to differences in the process determining levels of Y (i.e., differences in "4mean the coefficients). Going beyond the = O" M" we analyze difference, P(0) differences of financial inclusion 3 3|6,7&8 6| 7&8 distributions between men and women and identify a counterfactual distribution for women. We use recentered influence function (RIF) decomposition in combination with a reweighted strategy to estimate the counterfactual distribution (Firpo et al., 2018). The counterfactual distribution of financial inclusion can be approximated as follows: 4 "3 = O "3|6,7&8 M"6| 7&8 P(0) where ω(X) is the reweighting factor and dFX| T = 0 refers the observed distribution of X for women. 54 Chapter four About 88% of the gender disparities in account ownership are explained by differences in observed endowment characteristics between men and women. The change in the gap over time is mainly a result of the inequality-increasing effect of financial awareness (Table 3). For about 57% in 2019 and 65% in 2022, the overall gap was explained by differences in financial knowledge between men and women. As the descriptive estimate explains, such a difference resulted from the fact that financial awareness among women remained low, while this increased over time for men (Figure 3.5a & 3.5b). Mobile phone use also contributes to the overall gender gap, though it decreased from 20% in 2019 to 8% in 2022. The result implies that mobile phone penetration is not likely to be a huge factor in the divide between women and men, suggesting that to foster financial inclusion, the use of digital finance needs to evolve in parallel with growing mobile phone penetration. Moreover, the fact that men earn higher incomes and have better chances of employment than women, explains the observed gap in gender account ownership (Annex 3, Table A.10 & A.11). The gender differences in characteristics have contributed in different ways to the overall gender gap in rural and urban areas. In 2019, the difference in financial knowledge explained 37% (urban) and 60% (rural) of the overall gap and increased to 51% (urban) and 69% (rural) in 2022. The result reveals that financial knowledge is the primary contributor to the gender gap in rural and urban areas. The difference in the labor market status for men and women was more relevant in urban areas than in rural areas. In urban areas, it accounted for 10% of the overall gap in 2019 and 13% in 2022, while it only accounted for about 1.4% of the gap in rural areas (Annex 3, Table A.10 & A.11). In contrast, the difference in income level accounts for more of the gender gap in rural areas than in urban areas. The Oaxaca-Blinder decomposition estimates show that the gap in a z-score index increased from 0.39 in 2019 to 0.46 in 2022. About 52% of the gap in a z-score index emanates from being above the average for men and 48% from being below the average for women (Table 4). In 2022, about 89% of the disparities are explained by differences in characteristics, down from 102% in 2019. The fact that men are better off in their socioeconomic status and financial knowledge largely explains the financial inclusion score gap in 2019 and 2022. In 2019, about 70% of the overall gap was driven by differences in financial knowledge and this was the main contributor in 2022, i.e., 60% (Table 4). In 2022, socioeconomic differences were the second largest contributor to the gap, accounting for 21% of the overall gap, up from 17% in 2019. Whereas the contribution of mobile phone usage to the overall gap decreased from 16% in 2019 to 8% in 2022 (Annex 3, Table A.12). 55 The dynamics of financial inclusion in Ethiopia TABLE 3 Fairlie decomposition of the gender gaps in account ownership 2019 2022   National Urban Rural National Urban Rural Men 40.2 69 27.7 51.6 80.2 41.8 Women 22.5 49.2 9.7 30.4 64.2 18.8 Gender Gap 17.7 19.9 18 21.2 16 23 Explained Gap 15.5 18 14.1 18.4 14.5 19.4   88% 90% 78% 87% 91% 84% Contributions from gender differences (%) Demographic, socioeconomic 12% 31% 4% 13% 19% 10% Digital technology 20% 23% 16% 8% 19% 2% Financial knowledge 57% 37% 60% 65% 51% 69% Region -1% 0% -2% 2% 1% 3% Source: Authors’ estimates from ESPS. Note: The gaps are in percentage points. Demographic and socioeconomic variables include age, education, employment status, and income of the individual. Region captures the contribution of estimates of group of regions. TABLE 4 Oaxaca-Blinder decomposition of the gender gaps in financial inclusion index 2019 2022 Men 0.2027 0.2330 Women -0.1873 -0.2315 Gender Gap 0.3901 0.4644 Explained Gap 0.3973 0.4148   101.8% 89.3% Contributions from gender differences (%) Demographic, socioeconomic 16.8% 21.4% Digital technology 16.0% 7.5% Financial knowledge 69.6% 60.2% Region -0.6% -0.2% Source: Authors’ estimates from ESPS. Note: The gaps are in a z-score index. Demographic and socioeconomic variables include age, education, employment status, and income of the individual. Region captures the contribution of estimates of group of regions. 56 Chapter four The empirical evidence shows that financial knowledge is the new dimension of inequality; socioeconomic and resource status also contribute significantly to gender disparities in financial inclusion. It is evident in the counterfactual women financial inclusion distribution, which nearly matched the distribution for men across the z-scores in both years (Figure 24a & 24b). The counterfactual curve explains how the distribution for women would prevail if they had the same characteristics as men. The result suggests that empowering women in financial knowledge, technology, labor market performance, and human capital will enhance financial inclusion. FIGURE 25a FIGURE 25b Counterfactual density estimation, 2019 Counterfactual density estimation, 2022 .5 .4 .4 .3 .3 DENSITY DENSITY .2 .2 .1 .1 0 0 -2 0 2 4 6 -2 0 2 4 6 FINANCIAL INCLUSION (Z-SCORE) FINANCIAL INCLUSION (Z-SCORE) Female Male Counterfactual Female Male Counterfactual Source: Authors’ estimates from ESPS. We use a Source: Authors’ estimates from ESPS. We use a kernel smoothing function to define a nonparametric kernel smoothing function to define a nonparametric representation of the probability density function. representation of the probability density function. 57 Chapter five: Summary and priority policy actions Image: © Natalia Cieslik / World Bank This study examines financial inclusion in Ethiopia over the six years from 2016 to 2022. It highlights stylized facts and identifies correlates and pathways to greater financial inclusion. The study finds that financial inclusion in the country has nearly doubled in the last six years. It increased across gender, age, and place of residence. There has also been progress in dimensions that are beyond merely measuring account ownership, such as usage of financial services including mobile and online banking services. However, over half of the adult population remains unbanked. The national average of account ownership (41% in 2022) is way below the average for Sub-Saharan Africa countries and developing economies. Additionally, there are substantial disparities by region, age, gender, education, and employment status, among other things. Thus, expediting progress in financial inclusion and closing the existing gaps requires intervention. Diversifying entry channels: Non-traditional approaches can be explored to provide financial services to people that may not be reached through the traditional commercial banking institutions. The majority of individuals who reported owning a financial account use the traditional commercial banking institutions. The business model of these banks is not always feasible for people in remote places or that have little connection with the formal sector. Thus, diversifying innovative, affordable, tailored, and efficient entry channels to the financial system will deliver a wide range of financial services and products to more people and deepen financial inclusion. A strong commitment to designing services and products suitable for the unbanked is as important as diversifying delivery channels. 59 The dynamics of financial inclusion in Ethiopia The proposed delivery channels in the NFIS-II, for instance, the expansion of interest-free banking, mobile money, agent banking, and micro-insurance (e.g., linking Edir), are important enablers of financial inclusion. Given the rapid expansion of mobile money and mobile banking, it is worth investing in digital services and products tailored to the unbanked, particularly women and the poor. Enhancing agent banking effectiveness should also be a priority in reaching the disconnected poor in rural areas, including • going beyond using the channel as a cash-receiving and cash-sending mechanism and introducing products tailored to women and the poor (like micro-credit), • diversifying agent banking providers, • creating efficiencies in the delivery of the services and • holding awareness campaigns. The design of digital and innovative delivery channels only benefits from considering the spatial heterogeneities of the country, such as the different economic activities in different ecological zones, the varying livelihood opportunities in different regions, and the disparities in well-being across the country. Expanding mobile phone coverage: Mobile phones are a perfect example of a gateway technology necessary for expanding digitalized payment systems. The expansion of mobile phones offers more opportunities for previously unbanked people to access and use financial products and services. The study showed that those who have used mobile phones have a higher probability of escaping from years of being unbanked. The report also finds that access to technology is a fundamental resource for narrowing the gender gap in the financial system. However, the use of mobile phones in rural areas is still limited, and women are more likely to lack access to mobile phones. The expansion of digital finance, in parallel with the growth of digital technology across the country, will foster financial inclusion. Thus, policies addressing mobile phone availability and ownership barriers, particularly among women and the rural population, and expanding mobile phone networks are essential in promoting financial inclusion among disadvantaged groups. 60 Chapter five Enhancing financial knowledge and awareness: Financial knowledge plays a pivotal role in transitioning individuals into the financial system and is also a gateway to women empowerment. However, poor financial consciousness traps many in an unbanked situation, particularly women and those in rural areas known for exclusion and insufficient knowledge. The study also shows that a low level of knowledge is a trap for most disadvantaged groups to remain unbanked and contributes to inequalities in financial inclusion. There is a need to enhance financial knowledge and awareness in order for most to take advantage of financial services and products. Given that the landscape of the financial system has been changing due to rapid digital finance development, direct effort to improve the digital literacy of the public and awareness of new technologies is a precondition to benefitting from the digitalization process. Policy considerations that could enhance financial knowledge include: • financial education programs at places of higher education; • financial literacy training at the community level by integrating gender-intentional practices in the design and delivery of training; and • well designed and coordinated cross-promotion using hybrid mechanisms to gain massive boosts in service and product brand awareness. 61 The dynamics of financial inclusion in Ethiopia References Afande, F. O., & Mbugua, S. W. (2015). Role of Agent Banking Services in Promotion of Financial Inclusion in Nyeri Town, Kenya. Research Journal of Finance and Accounting , 6(3). https://api. semanticscholar.org/CorpusID:212478714 Aker, J., Boumnijel, R., Mcclelland, A., Tierney, N. (2016). Payment Mechanisms and Antipoverty Programs. Economic Development and Cultural Change, 65(1) , 1-37. https://www.jstor.org/ stable/26545224 Allen, F., Demirguc-Kunt, A., Klapper, L., Martinez, M. S. (2012). The Foundations of Financial Inclusion Understanding Ownership and Use of Formal Accounts. Journal of financial Intermediation, 27, 1-30. https://doi.org/10.1016/j.jfi.2015.12.003 Amidžić, G., Massara, A., & Mialou, A. (2014). Assessing Countries’ Financial Inclusion Standing-A New Composite Index IMF Working Paper Statistics Department Assessing Countries’ Financial Inclusion Standing-A new Composite Index. Ansar, S., Klapper, L., & Singer, D. (2023). The Importance of Financial Education for the Effective Use of Formal Financial Services. Policy Research Working Paper N0. 10345. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/39520 Arshad, A. (2023). Nexus between financail inclusion and women empowerment: evidence from developing countries. Gender in Management, 38(4), 561-580. https://doi.org/10.1108/GM-04-2022- 0125 Bittencourt, M. (2010). Financial development and inequality: Brazil 1985-1994. Economic Change and Restructuring, 43(2), 113–130. https://doi.org/10.1007/s10644-009-9080-x Brune, L., Giné, X., Goldberg, J., & Yang, D. (2016). Facilitating Savings for Agriculture. Economic Development and Cultural Change, 64(2), 187–220. https://doi.org/10.2307/26544743 Cámara, N., & Tuesta, D. (2014). Measuring Financial Inclusion: A Multidimensional Index Measuring Financial Inclusion. Working Paper NO. 14/26. BBVA, Madrid. www.bbvaresearch.com Chakravarty, S. R., & Pal, R. (2013). Financial inclusion in India: An axiomatic approach. Journal of Policy Modeling, 35(5), 813–837. https://doi.org/10.1016/j.jpolmod.2012.12.007 Cole, S., Sampson, T., & Zia, B. (2011). Prices or Knowledge? What Drives Demand for Financial Services in Emerging Markets? The Journal of Finance, 66(6). https://www.jstor.org/stable/41305181 62 Cull, R., Gine, X., Harten, S., Heitmann, S., & Rusu, A. B. (2018). Agent Banking in a Highly Under- Developed Financial Sector: Evidence from Democratic Republic of Congo. World Development, 107, 54-74. https://doi.org/10.1016/j.worlddev.2018.02.001Get rights and content Demirgüç-Kunt, A., & Klapper, L. (2013). Measuring Financial Inclusion: Explaining Variation in Use of Financial Services across and within Countries. Brookings Papers on Economic Activity, 279–321. http://www.jstor.org/stable/23594869 Demirgüç-Kunt, A., Klapper, L., & Singer, D. (2017). Financial Inclusion and Inclusive Growth A Review of Recent Empirical Evidence. Policy Research Working Paper No. 8040. World Bank, Washington, DC. https://hdl.handle.net/10986/26479 Demirgüç-Kunt, A., Klapper, L., Singer, D., & Ansar, S. (2022). Financial Inclusion, Digital Payments, and Resilience in the Age of COVID-19. Washington, DC: World Bank. https://doi:10.1596/978-1-4648- 1897-4 Ethiopia Statistical Service and World Bank. (2023). Ethiopia Socioeconomic Panel Survey (ESPS) Report- Wave 5, 2021/22. https://microdata.worldbank.org/index.php/catalog/6161/download/78643 Field, E., Pande, R., Rigol, N., Schaner, S., & Moore, C. T. (2021). On Her Own Account: How Strengthening Women’s Financial Control Impacts Labor Supply and Gender Norms. American Economic Review, 111(7), 2342-2375. https://doi.org/10.2307/27086075 Firpo, S., Fortin, N. M., Lemieux, T. (2018). Decomposing wage distributions using recentered influence functions regressions. Econometrics, 6(2), 28. https://doi.org/10.3390/econometrics6020028 Fungáčová, Z., & Weill, L. (2015). Understanding Financial Inclusion in China. China Economic Review, 34, 196-206. https://doi.org/10.1016/j.chieco.2014.12.004 Grohmann, A., Klühs, T., & Menkhoff, L. (2018). Does financial literacy improve financial inclusion? Cross country evidence. World Development, 111, 84–96. https://doi.org/10.1016/j. worlddev.2018.06.020 Ha, D., Nguyen, M., Nguyen, K., & Şensoy, A. (2024). Unveiling financial inclusion dynamics: Fintech’s resonance in Association of Southeast Asian Nations (ASEAN). International Journal of Finance and Economics. https://doi.org/10.1002/ijfe.2963 Hasan, M., Le, T., & Hoque, A. (2021). How does financial literacy impact on inclusive finance? Financial Innovation, 7(40). https://doi.org/10.1186/s40854-021-00259-9 Hendriks, S. (2019). The role of financial inclusion in driving women’s economic empowerment. Development in Practice, 8. https://doi.org/10.1080/09614524.2019.1660308 63 The dynamics of financial inclusion in Ethiopia Hundie, S. K., & Tulu, D. T. (2023). Determinants of financial inclusion gender gap in Ethiopia: Evidence from decomposition analysis. Cogent Business and Management, 10. https://doi.org/10.10 80/23311975.2023.2238124 Khan, F., Siddiqui, M. A., & Imtiaz, S. (2022). Role of financial literacy in achieving financial inclusion: A review, synthesis and research agenda. Cogent Business and Management, 9(1). https://doi.org/10. 1080/23311975.2022.2034236 Kim, M., Zoo, H., Lee, H., & Kang, J. (2018). Mobile financial services, financial inclusion, and development: A systematic review of academic literature. Electronic Journal of Information Systems in Developing Countries, 84(5). https://doi.org/10.1002/isd2.12044 Lenka, S. K., & Barik, R. (2018a). Has expansion of mobile phone and internet use spurred financial inclusion in the SAARC countries? Financial Innovation, 4(5). https://doi.org/10.1186/s40854-018- 0089-x Majid, M. S. A., Dewi, S., Aliasuddin, & Kassim, S. H. (2019). Does Financial Development Reduce Poverty? Empirical Evidence from Indonesia. Journal of the Knowledge Economy, 10(3), 1019–1036. https://doi.org/10.1007/s13132-017-0509-6 Mndolwa, F. D., Alhassan, A. L. (2020). Gender disparities in financial inclusion: Insights from Tanzania. Africa Development Review, 32, 578-590. https://doi.org/10.1111/1467-8268.12462 Mukong, A., Shiwayu, N., & Kaulihowa, T. (2020). A decomposition of the gender gap in financial inclusion: Evidence from Namibia. African Journal of Business and Economic Research, 15(4), 149–169. https://doi.org/10.31920/1750-4562/2020/V15N4A7 National Bank of Ethiopia. (2021). National Financial Inclusion Strategy-II, 2021-2025. Addis Ababa, National Bank of Ethiopia. Ndoya, H. H., & Tsala, C. O. (2021). What drives gender gap in financial inclusion? Evidence from Cameroon. African Development Review, 33(4), 674–687. https://doi.org/10.1111/1467-8268.12608 Nguyen, H. S., Ho, C. M., Vo, A. T., & Vo, D. H. (2023). The determinants of financial inclusion from the households’ perspectives in Vietnam. PLoS ONE, 18(9). https://doi.org/10.1371/journal.pone.0291020 Nguyen, T. A. N., & Luong, T. T. H. (2023). The Determinants of Financial Inclusion in Vietnam: A Demand-Side Approach. SAGE Open, 13(4). https://doi.org/10.1177/21582440231210639 Omar, M. A., & Inaba, K. (2020). Does financial inclusion reduce poverty and income inequality in developing countries? A panel data analysis. Journal of Economic Structures, 9(1). https://doi. org/10.1186/s40008-020-00214-4 64 Özşuca, E. A. (2019). Gender gap in financial inclusion: Evidence from MENA. Economics and Business Letters, 8(4), 199–208. https://doi.org/10.17811/ebl.8.4.2019.199-208 Perez-Moreno, S. (2011). Financial development and poverty in developing countries: A causal analysis. Empirical Economics, 41(1), 57–80. https://doi.org/10.1007/s00181-010-0392-5 Moreno, S. (2011). Financial development and poverty in developing countries: A causal analysis. Empirical Economics, 41(1), 57–80. https://doi.org/10.1007/s00181-010-0392-5 Sarma, M. (2008). Index of Financial Inclusion. ICRIER Working Papers, Indian Council for Research on International Economic Relations No. 215. https://hdl.handle.net/10419/176233 Sarma, M. (2012). Index of Financial Inclusion-A measure of financial sector inclusiveness. Berlin Working Paper on Money, Finance, Trade and Development No. 07. https://ideas.repec.org/p/mtf/wpaper/1207.html Selvarajan, S. K., & Chandran, V. G. R. (2024). Financial Inclusion Trajectories: Geographical Dispersion, Convergence, and Development Implications. European Journal of Development Research, 36(4), 897–924. https://doi.org/10.1057/s41287-023-00616-7 Senou, M. M., Ouattara, W., & Houensou, D. A. (2019). Financial inclusion dynamics in WAEMU: Was digital technology the missing piece? Cogent Economics and Finance, 7(1). https://doi.org/10.1080/2 3322039.2019.1665432 Shen, Y., Hueng, C. J., & Hu, W. (2020). Using digital technology to improve financial inclusion in China. Applied Economics Letters, 27(1), 30–34. https://doi.org/10.1080/13504851.2019.1606401 Tram, T. X. H., Lai, T. D., & Nguyen, T. T. H. (2023). Constructing a composite financial inclusion index for developing economies. Quarterly Review of Economics and Finance, 87, 257–265. https:// doi.org/10.1016/j.qref.2021.01.003 Wang, X., & Guan, J. (2017). Financial inclusion: measurement, spatial effects and influencing factors. Applied Economics, 49(18), 1751–1762. https://doi.org/10.1080/000 36846.2016.1226488 Wodon, Q., & Brière, B. (2018). Unrealized Potential: The High Cost of Gender Inequality in Earnings. The Cost of Gender Inequality Notes Series. World Bank, Washington, DC. Zhang, Q., & Posso, A. (2019). Thinking Inside the Box: A Closer Look at Financial Inclusion and Household Income. Journal of Development Studies, 55(7), 1616–1631. https://doi.org /10.1080/00220388.2017.1380798 65 Annexes Image: © Binyam Teshome / World Bank Annex 1. Additional Information on Measurements and Scores ENUMERATION AREA LEVEL FINANCIAL KNOWLEDGE SCORE An EA level financial knowledge score is developed by summing up all individual level14 scores across EA. Finally, the total is divided by the maximum possible point (the product of the total number of adult individuals in each EA and a possible highest score, which is equal to one). $ : ∑#%& &# !"#!" = '!" × )'()*+ *!"# !" = +,(!"#) !" − index +,(!"# 0 where the FKIit represents financial knowledge at)EA !" i in year t. D refers the nth n individual score. We rescale the range of index of financial knowledge across EAs at the national average by subtracting the logarithmic $: form of the national average financial ∑#%& &# !"# knowledge from the ith ES index !" = t. in year '!" × )'()*+ 0 )!" *!"#!" = +,(!"#)!" − +,(!"# where the TIFIit represents the transformed financial knowledge score at EA i in year t. FKI is the national average in year t. 14 See Box 2 67 The dynamics of financial inclusion in Ethiopia Annex 2. Additional Descriptive Statistics TABLE A.1 Account ownership by region, gender and place of residence, 2019 and 2022 2019 2022 (Findex method) (Findex method) Any Household Any Household Select an Adult Select an Adult randomly randomly Member Member   Individual Individual All Men Women All Men Women National 31.0 40.2 22.5 32.6 47.5 40.8 51.3 30.3 42.9 63.3 Afar 20.5 28.4 12.9 23.8 27.9 28.2 35.0 19.7 36.6 52.9 Amhara 37.1 45.6 29.3 40.2 56.8 45.5 54.9 36.2 48.1 70.7 Oromia 25.9 36.1 16.0 27.9 42.5 37.6 49.7 25.6 41.3 59.9 Somali 6.3 8.7 4.0 6.5 9.6 37.6 43.0 31.0 42.6 58.6 Benishangul Gumuz 28.3 36.3 20.6 28.6 45.3 42.0 51.9 31.8 41.7 64.4 SNNP 24.8 37.1 13.7 25.5 41.4 31.7 45.5 18.5 28.9 53.7 Gambela 35.0 45.7 24.9 39.6 56.1 52.5 67.7 35.6 51.6 81.6 Harer 49.3 60.4 39.2 47.1 67.9 65.3 74.2 56.0 63.1 80.4 Addis Ababa 74.8 81.2 69.9 76.8 92.5 85.7 88.4 83.4 86.8 98.7 Dire Dawa 52.2 58.8 46.6 52.6 67.2 60.4 63.6 57.0 61.3 73.3 Urban 58.3 69.1 49.1 62.0 77.4 72.2 80.2 64.2 73.1 92.0 Rural 18.5 27.7 9.7 18.5 33.0 30.1 41.5 18.7 32.3 53.3 68 Annexes TABLE A.2 Account owning individuals and financial institutions by region, gender and place of residence, 2022   Private Bank Public Bank Microfinance SACCO Women Women Women Women Men Men Men Men All All All All Country 41.5 45.7 34.6 74.1 75.6 71.6 9 8.9 9.2 16.5 15.8 17.7 Afar 27.8 31.3 19.9 92.4 90.7 96 2.5 3.6 0 3.4 3.9 2.3 Amhara 25.5 30 18.9 65.2 68.2 61 9.4 9.9 8.6 34.1 32.6 36.2 Oromia 58.9 64.1 49 74.3 75.8 71.5 5.7 5.8 5.4 13.2 12.5 14.4 Somali 20.5 20.1 21.2 45.7 44 48.5 33.5 27.3 43.7 4.8 5.3 3.9 Benishangul Gumuz 33.7 42.1 19.4 85.8 89.6 79.3 6.1 9 1.1 15.8 16.4 14.7 SNNP 26.8 28.7 22.1 82.2 84.2 77.4 12.7 11.8 14.9 9.8 9.9 9.5 Gambela 60.3 61.9 56.8 82.5 85 77.1 3.5 4.7 1.1 4.7 4.3 5.7 Harer 59.4 63.1 54.3 92.1 92.4 91.7 4.6 6.2 2.3 5.6 6.2 4.9 Addis Ababa 58.9 68.3 50.3 92.6 92.7 92.4 2.1 2.1 2.2 3.8 3.5 4.1 Dire Dawa 55.8 61.1 49.8 84.3 86 82.4 2.8 1.7 3.9 3.6 3.9 3.2 Urban 51.1 56.8 44 83 84.6 81.1 5.9 7.2 4.3 10.3 10.8 9.8 Rural 33.5 38.1 23.3 66.5 69.4 60.2 11.6 10.1 14.9 21.6 19.1 27 TABLE A.3 Individual and household saving behavior by region, gender, place of residence, 2019 and 2022 2019 2022 Any Household   Individual member Individual Any Household member institutions institutions institutions institutions Informal Informal Formal Formal Formal Formal way way Any Any Any Any Country 25.6 20.4 40.3 33.6 33.6 23.8 20.4 56.9 44.3 37.9 Afar 18.2 14.2 28.6 21.9 25.3 19.5 13.4 51.7 38.8 28.0 Amhara 26.1 22.7 42.6 38.6 30.6 26.5 15.1 52.1 47.4 29.5 Oromia 22.7 17.3 36.3 29.3 40.9 23.9 29.7 65.4 45.4 49.9 Somali 3.9 2.9 6.9 5.2 9.8 5.2 6.2 22.8 12.1 15.3 Benishangul Gumuz 25.2 18.5 42.0 32.2 37.7 26.3 22.3 59.0 46.1 35.8 SNNP 26.5 16.1 41.4 27.7 28.3 21.1 17.1 52.7 41.2 36.2 Gambela 27.9 23.4 48.3 41.6 45.6 38.9 24.0 74.5 68.0 43.8 Harer 37.2 31.7 54.0 47.4 38.1 31.0 21.7 60.7 51.1 36.8 Addis Ababa 54.0 51.1 77.0 75.1 44.4 42.3 9.1 67.7 65.2 22.2 Dire Dawa 38.4 36.1 53.5 50.9 37.6 34.6 10.9 53.9 51.3 22.1 Urban 44.0 39.8 63.0 58.8 46.4 41.8 18.0 70.9 66.3 35.3 Rural 17.3 11.6 29.4 21.6 29.2 17.6 21.2 51.9 36.6 38.8 69 The dynamics of financial inclusion in Ethiopia TABLE A.4 Access to loan and source of loan by region, place of residence and wealth, 2019 and 2022   Source of Loan 2019, Access to Loan / 2022, Access to Loan / Religious Institutions Bank (Commercial) Household Level Household Level Money Lender Grocery/Local Microfinance Institutions Merchant Employer Neighbor Relatives SACCOS Other NGO Country 15.7 26.8 48.1 24.1 3.0 4.3 0.8 0.9 4.4 0.1 1.3 11.3 1.7 Afar 9.5 14.8 58.7 39.0 1.8 0.6 0.0 Amhara 22.6 29.5 47.2 13.6 1.3 7.7 1.1 2.1 8.8 0.3 17.6 0.3 Oromia 13.8 31.7 46.1 28.0 3.0 3.0 0.4 0.5 2.5 1.4 11.7 3.3 Somali 10.6 19.1 41.2 36.8 18.8 1.4 1.9 0.0 Benishangul Gumuz 12.2 16.5 63.1 22.5 2.3 0.3 0.2 11.5 0.0 SNNP 14.6 21.7 55.1 28.8 2.4 3.3 0.2 0.4 3.3 3.6 1.8 1.0 Gambela 12.3 11.3 45.3 48.9 5.8 0.0 Harer 10.1 16.0 42.1 38.0 12.7 1.7 5.4 0.0 Addis Ababa 8.6 15.4 51.5 23.8 3.9 8.0 3.3 1.9 7.5 0.0 Dire Dawa 8.2 22.4 43.4 36.2 12.0 2.1 0.7 0.8 3.2 1.6 Urban 14.6 28.1 59.4 17.2 2.7 0.2 2.3 2.2 2.7 0.2 0.9 12.3 0.1 Rural 15.9 26.3 44.4 26.4 3.1 5.7 0.3 0.5 5.0 1.4 11.0 2.3 Bottom 40% 16.9 27.5 47.4 25.7 2.2 4.2 0.5 1.1 2.9 0.7 14.6 0.7 Tom 60% 14.5 26.3 48.6 22.9 3.5 4.3 0.9 0.8 5.5 0.1 1.7 9.0 2.5 70 Annexes Annex 3. Additional Regression Results TABLE A.5 Estimates for probit model on account ownership   2019 2022 National Urban Rural National Urban Rural Age 0.0203*** 0.0196*** 0.0209*** 0.0147*** 0.0150*** 0.0112*** (0.0018) (0.0024) (0.0028) (0.0021) (0.0027) (0.0030) Age (Square) -0.0002*** -0.0002*** -0.0002*** -0.0002*** -0.0002*** -0.0001*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) Women -0.0975*** -0.1082*** -0.0784*** -0.1056*** -0.1217*** -0.0693*** (0.0100) (0.0130) (0.0158) (0.0138) (0.0173) (0.0201) Primary education 0.0505*** 0.0444*** 0.0575** 0.0623*** 0.0524** 0.0784*** (0.0140) (0.0171) (0.0239) (0.0187) (0.0224) (0.0298) Secondary education 0.0846*** 0.0530* 0.1194*** 0.1051*** 0.1020*** 0.1103*** (0.0191) (0.0279) (0.0290) (0.0262) (0.0343) (0.0372) Post secondary 0.2464*** 0.3226*** 0.2525*** 0.3651*** 0.4227*** 0.2852*** (0.0275) (0.0568) (0.0362) (0.0417) (0.0658) (0.0487) Employed 0.0847*** 0.0715*** 0.1171*** 0.0882*** 0.0719*** 0.1250*** (0.0106) (0.0149) (0.0155) (0.0144) (0.0188) (0.0199) Owns mobile 0.1481*** 0.1229*** 0.1906*** 0.1351*** 0.1227*** 0.1627*** (0.0107) (0.0141) (0.0161) (0.0139) (0.0177) (0.0198) Distance to financial -0.0027*** -0.0030*** -0.0011** 0.0003 0.0004* -0.0015** institution (Km) (0.0004) (0.0005) (0.0005) (0.0002) (0.0002) (0.0006) Log of income 0.0427*** 0.0309*** 0.0683*** 0.0813*** 0.0936*** 0.0396*** (0.0075) (0.0089) (0.0134) (0.0097) (0.0118) (0.0148) HH years of education 0.0131*** 0.0132*** 0.0135*** 0.0194*** 0.0229*** 0.0149*** (average) (0.0022) (0.0035) (0.0032) (0.0031) (0.0045) (0.0038) N 13402 5910 7492 12308 5762 6546 Pseudo R2 0.3549 0.2445 0.3027 0.2942 0.2068 0.2481 Chi2 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Robust standard error in parenthesis. Results are average marginal effect. The base level for education is No education. *p < .1; **p < .05; ***p=.01 71 The dynamics of financial inclusion in Ethiopia TABLE A.6 Estimates for probit model on account ownership for urban- dominated regions and other regions   2019 2022 Urban- Urban- dominated dominated regions Other regions regions Other regions Age 0.0171*** 0.0206*** 0.0068* 0.0152*** (0.0024) (0.0020) (0.0027) (0.0023) Age (Square) -0.0001*** -0.0002*** -0.0001* -0.0002*** (0.0000) (0.0000) (0.0000) (0.0000) Women -0.0109 -0.0973*** -0.0016 -0.1040*** (0.0159) (0.0109) (0.0177) (0.0148) Primary education 0.1056*** 0.0343** 0.0814*** 0.0516*** (0.0245) (0.0150) (0.0265) (0.0198) Secondary education 0.1634*** 0.07407*** 0.1106*** 0.1080*** (0.0277) (0.0214) (0.0329) (0.0281) Post secondary 0.3413*** 0.2283*** 0.1745*** 0.3655*** (0.0341) (0.0324) (0.0414) (0.0501) Employed 0.1260*** 0.09235*** 0.1213*** 0.1061*** (0.0153) (0.0115) (0.0182) (0.0153) Owns mobile 0.1833*** 0.1528*** 0.1313*** 0.1429*** (0.0167) (0.0111) (0.0177) (0.0148) Distance to financial -0.0080*** -0.0037*** -0.0024*** 0.0002 institution (Km) (0.0009) (0.0004) (0.0009) (0.0002) Log of income 0.0748*** 0.0451*** 0.0599*** 0.0991*** (0.0121) (0.0079) (0.0139) (0.0101) HH years of education 0.0020 0.0188*** 0.0181*** 0.0251*** (average) (0.0029) (0.0025) (0.0033) (0.0033) N 4438 8964 8285 12308 Pseudo R2 0.2454 0.2977 0.3851 0.2942 Chi2 0.0000 0.0000 0.0000 0.0000 Robust standard error in parenthesis. Results are average marginal effect. The base level for education is No education. *p < .1; **p < .05; ***p=.01. 72 Annexes TABLE A.7 Estimates for ordered logit model on level of financial inclusion   2019 2022 National Urban Rural National Urban Rural Age Low (LFI=1) -0.0224*** -0.0231*** -0.0218*** -0.0035 0.0055 -0.0175*** (0.0038) (0.0061) (0.0051) (0.0037) (0.0045) (0.0061) Medium 0.0216*** 0.0229*** 0.0206*** 0.0032 -0.0054 0.0155*** (LFI=2) (0.0037) (0.0060) (0.0048) (0.0034) (0.0044) (0.0054) High 0.0008*** 0.0002 0.0012*** 0.0002 -0.0001 0.0020*** (LFI=3) (0.0002) (0.0002) (0.0004) (0.0002) (0.0001) (0.0008) Age square Low (LFI=1) 0.0003*** 0.0003*** 0.0003*** 0.0001 -0.0001 0.0002*** (0.0000) (0.0001) (0.0001) (0.0000) (0.0001) (0.0001) Medium -0.0003*** -0.0003*** -0.0003*** -0.0001 0.0001 -0.0002*** (LFI=2) (0.0000) (0.0001) (0.0001) (0.0000) (0.0000) (0.0001) High -0.0000*** -0.0000 -0.0000*** -0.0000 -0.0000 -0.0000** (LFI=3) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) Women Low (LFI=1) 0.0656*** 0.0387 0.0892*** 0.1206*** 0.1093*** 0.1320*** (0.0170) (0.0316) (0.0200) (0.0191) (0.0321) (0.0221) Medium -0.0632*** -0.0383 -0.0843*** -0.1129*** -0.1074*** -0.1166*** (LFI=2) (0.0164) (0.0312) (0.0190) (0.0178) (0.0316) (0.0195) High -0.0025*** -0.0004 -0.0049*** -0.0077*** -0.0019* -0.0154*** (LFI=3) (0.0008) (0.0005) (0.0014) (0.0017) (0.0011) (0.0036) Primary Low (LFI=1) -0.0197 -0.02657 -0.0054 -0.0968*** -0.0835** -0.1200** education (0.0296) (0.0366) (0.0451) (0.0313) (0.0375) (0.0489) Medium 0.0190 0.02632 0.0051 0.0907*** 0.0821** 0.1060** (LFI=2) (0.0285) (0.0362) (0.0426) (0.0292) (0.0368) (0.0431) High 0.0008 0.0003 0.0003 0.0062*** 0.0014 0.014** (LFI=3) (0.0008) (0.0005) (0.0014) (0.0017) (0.0011) (0.0036) Secondary Low (LFI=1) -0.1314*** -0.1206** -0.1317*** -0.1691*** -0.0927* -0.2731*** education (0.0326) (0.0506) (0.0457) (0.0371) (0.0536) (0.0504) Medium 0.1265*** 0.1194** 0.1244*** 0.1583*** 0.0912* 0.2413*** (LFI=2) (0.0314) (0.0499) (0.0432) (0.0347) (0.0527) (0.0443) High 0.0049*** 0.0011 0.0072** 0.0108*** 0.0016 0.0318*** (LFI=3) (0.0016) (0.0012) (0.0029) (0.0029) (0.0012) (0.0081) Post Low (LFI=1) -0.2669*** -0.2498*** -0.2858*** -0.3167*** -0.2475*** -0.4087*** secondary (0.0356) (0.0628) (0.0474) (0.0410) (0.0582) (0.0550) Medium 0.2569*** 0.2475*** 0.2700*** 0.2965*** 0.2433*** 0.3612*** (LFI=2) (0.0342) (0.0622) (0.0447) (0.0384) (0.0574) (0.0483) High 0.0100*** 0.0024 0.0157*** 0.0203*** 0.0042* 0.0476*** (LFI=3) (0.0024) (0.0023) (0.0041) (0.0041) (0.0022) (0.0103) 73 The dynamics of financial inclusion in Ethiopia   2019 2022 National Urban Rural National Urban Rural Employed Low (LFI=1) -0.0646*** -0.0491* -0.0774*** -0.0784*** -0.0478* -0.1059*** (0.0163) (0.0272) (0.0205) (0.0177) (0.0263) (0.0235) Medium 0.0622*** 0.0486* 0.0731*** 0.0734*** 0.0470* 0.0936*** (LFI=2) (0.0157) (0.0270) (0.0194) (0.0166) (0.0258) (0.0207) High 0.0024*** 0.0005 0.0042*** 0.0050*** 0.0008 0.0123*** (LFI=3) (0.0008) (0.0005) (0.0014) (0.0014) (0.0006) (0.0034) Owns mobile Low (LFI=1) -0.0596** -0.0343 -0.1003*** -0.0311 -0.0264 -0.0465 (0.0238) (0.0314) (0.0369) (0.0227) (0.0296) (0.0330) Medium 0.0573** 0.0339 0.0948*** 0.0291 0.0260 0.04108 (LFI=2) (0.0229) (0.0311) (0.0348) (0.0212) (0.0291) (0.0292) High 0.0022** 0.0003 0.0055** 0.0020 0.0005 0.0054 (LFI=3) (0.0010) (0.0004) (0.0023) (0.0015) (0.0006) (0.0039) Distance Low (LFI=1) 0.0003 0.0002 0.0008 0.0021* 0.0026 0.0005 to financial (0.0009) (0.0010) (0.0013) (0.0013) (0.0019) (0.0009) institution (Km) Medium -0.0003 -0.0001 -0.0007 -0.0020* -0.0025 -0.0004 (LFI=2) (0.0008) (0.0010) (0.0012) (0.0012) (0.0018) (0.0008) High -0.0000 -0.0000 -0.0000 -0.0001 -0.0000 -0.0001 (LFI=3) (0.0000) (0.0000) (0.0001) (0.0001) (0.0000) (0.0001) Log of income Low (LFI=1) -0.0599*** -0.0559** -0.0658*** -0.0377*** -0.0260 -0.0560*** (0.0139) (0.0218) (0.0170) (0.0144) (0.0208) (0.0197) Medium 0.0576*** 0.0554** 0.0622*** 0.0352*** 0.0255 0.0495*** (LFI=2) (0.0134) (0.0216) (0.0160) (0.0135) (0.0205) (0.0173) High 0.0022*** 0.0005 0.0036*** 0.0024** 0.0004 0.0065** (LFI=3) (0.0007) (0.0005) (0.0012) (0.0010) (0.0004) (0.0026) HH years of Low (LFI=1) -0.0062** -0.0069 -0.0064* -0.0142*** -0.0148*** -0.0145*** education (0.0030) (0.0060) (0.0036) (0.0033) (0.0057) (0.0039) (average) Medium 0.0060** 0.0068 0.0060* 0.0133*** 0.0145*** 0.0128*** (LFI=2) (0.0029) (0.0060) (0.0034) (0.0031) (0.0056) (0.0035) High 0.0002* 0.0001 0.0004* 0.0009*** 0.0003 0.0017*** (LFI=3) (0.0001) (0.0001) (0.0002) (0.0002) (0.0002) (0.0005) N 5722 1283 4439 7161 2077 5084 Pseudo R 2 0.1905 0.1671 0.1744 0.1995 0.1673 0.2048 Chi2 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Robust standard error in parenthesis. Results are average marginal effect. The base level for education is No education. *p < .1; **p < .05; ***p=.01. 74 Annexes TABLE A.8 Linear regression estimated results on a z-score of financial inclusion index   2019 2022 National Rural Urban National Rural Urban Age 0.0437*** 0.0380*** 0.05172*** 0.02840*** 0.02565*** 0.03043*** (0.0032) (0.0040) (0.0055) (0.0036) (0.0043) (0.0065) Age (Square) -0.0004*** -0.0004*** -0.0005*** -0.0003*** -0.0003*** -0.0003*** (0.0000) (0.0000) (0.0001) (0.0000) (0.0000) (0.0001) Women -0.1924*** -0.1682*** -0.2511*** -0.2444*** -0.2294*** -0.2869*** (0.0218) (0.0271) (0.0370) (0.0271) (0.0333) (0.0455) Primary education 0.1071*** 0.1069*** 0.1106** 0.1305*** 0.1208*** 0.1694*** (0.0305) (0.0389) (0.0512) (0.0356) (0.0423) (0.0633) Secondary education 0.3300*** 0.2341*** 0.3858*** 0.4070*** 0.3460*** 0.4920*** (0.0504) (0.0793) (0.0665) (0.0558) (0.0766) (0.0816) Post secondary 1.0697*** 1.3605*** 0.9728*** 1.1416*** 1.2609*** 1.0821*** (0.0722) (0.1666) (0.0849) (0.0747) (0.1218) (0.1010) Employed 0.2939*** 0.2142*** 0.3672*** 0.2651*** 0.1997*** 0.3862*** (0.0280) (0.0407) (0.0379) (0.0324) (0.0428) (0.0459) Owns mobile 0.3732*** 0.3318*** 0.4305*** 0.2767*** 0.2550*** 0.3669*** (0.0306) (0.0418) (0.0418) (0.0321) (0.0396) (0.0536) Distance to financial -0.0020*** -0.0019** -0.0005 -0.0010*** -0.0010*** -0.0025* institution (Km) (0.0003) (0.0003) (0.0007) (0.0003) (0.0003) (0.0013) Log of income 0.1080*** 0.0730*** 0.1976*** 0.1555*** 0.1500*** 0.1748*** (0.0153) (0.0181) (0.0289) (0.0187) (0.0220) (0.0348) HH years of education 0.0349*** 0.0346*** 0.0303*** 0.0393*** 0.0420*** 0.0354*** (average) (0.0055) (0.0090) (0.0071) (0.0058) (0.0083) (0.0077) N 13402 5910 7492 12308 5762 6546 R2 0.4192 0.2494 0.4276 0.4189 0.2992 0.4017 F 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Robust standard error in parenthesis. The base level for education is No education. *p < .1; **p < .05; ***p=.01 75 The dynamics of financial inclusion in Ethiopia TABLE A.9 Estimates for logit regression on moving out of financial exclusion Urban- dominated Other National Rural Urban regions regions Age [25, 34] 0.0600** 0.0738*** 0.0014 0.0486 0.0477* (0.0239) (0.0282) (0.0459) (0.0555) (0.0251) Age [35, 44] 0.0661** 0.0801*** 0.0056 -0.0206 0.0555** (0.0260) (0.0303) (0.0513) (0.0644) (0.0271) Age >44 0.0394 0.0492 -0.0009 -0.0021 0.0241 (0.0264) (0.0309) (0.0523) (0.0707) (0.0276) Women -0.1388*** -0.1398*** -0.1384*** -0.0567 -0.1425*** (0.0170) (0.0193) (0.0353) (0.0431) (0.0176) Primary education 0.0339 0.0464* -0.0358 0.0533 0.0143 (0.0222) (0.0248) (0.0465) (0.0634) (0.0230) Secondary education 0.0819** 0.0887* 0.0366 0.0399 0.0764** (0.0368) (0.0454) (0.0684) (0.0817) (0.0389) Post secondary 0.3459*** 0.4192*** 0.2751*** 0.0770 0.4040*** (0.0847) (0.1352) (0.1034) (0.1263) (0.0895) Owns mobile 0.1019*** 0.0872*** 0.1538*** 0.1764*** 0.1225*** (0.0184) (0.0222) (0.0343) (0.0386) (0.0192) Financial knowledge 0.1171*** 0.0900* 0.2166*** 0.2956*** 0.1137*** (0.0401) (0.0496) (0.0730) (0.0757) (0.0419) Distance to financial -0.000*** -0.0006** -0.0018*** -0.0039*** 0.0000 institution (Km) (0.0002) (0.0003) (0.0006) (0.0014) (0.0002) Q2 -0.0540* -0.0531 -0.0366 -0.0848 -0.0550* (0.0314) (0.0337) (0.0682) (0.0867) (0.0319) Q3 0.0317 0.0258 0.0803 0.0321 0.0316 (0.0295) (0.0319) (0.0639) (0.0795) (0.0298) Q4 -0.0033 -0.0134 0.0547 0.0082 -0.0125 (0.0294) (0.0325) (0.0624) (0.0749) (0.0294) Q5 0.0044 -0.0180 0.1018 -0.0477 -0.0249 (0.0319) (0.0368) (0.0669) (0.0736) (0.0324) HH years of education 0.0126*** 0.0138** 0.0102 0.0184** 0.0146*** (average) (0.0045) (0.0058) (0.0076) (0.0085) (0.0048) Kebele relative income 0.0025*** 0.0032* 0.0023** 0.0013* 0.0061*** (0.0009) (0.0018) (0.0011) (0.0007) (0.0010) N 5377 3361 2016 1153 4224 Pseudo R2 0.1656 0.1011 0.1730 0.2109 0.1155 chi2 0.0000 0.0000 0.0000 0.0000 0.0000 Kebele relative income measure refers the difference between the average income of the Kebele that the household resides in and the average income level of the country. Income is proxied by household consumption expenditure. Financial knowledge refers to the z-score of the index. Robust standard error in parenthesis. Results are average marginal effect. The base level for education is No education, age is between 8 and 24, and for income it is Q1. *p < .1; **p < .05; ***p=.01. 76 Annexes TABLE A.10 Estimates for decomposition analysis on gender gaps in account ownership, 2019 National Urban Rural Men 40.2 69 27.7 Women 22.5 49.2 9.7 Gender Gap 17.7 19.9 18 Explained gap 15.5 87.6% 18 90.5% 14.1 78.3% Contributions from gender Contributions from gender Contributions from gender differences differences differences Contribution Contribution Contribution   AME AME AME (%) (%) (%) SE SE SE   Demographic, socioeconomic 4.6% 16.6% -1.3% Age 0.0045*** (0.0012) 2.5% 0.003** (0.0013) 1.5% 0.0040** (0.0017) 2.2% Primary education -0.0017 (0.0025) -1.0% 0.0003 (0.0012) 0.1% -0.0045 (0.0044) -2.5% Secondary education -0.0039** (0.0016) -2.2% 0.0035 (0.0049) 1.7% -0.0061*** (0.0020) -3.4% Post secondary 0.0020** (0.0009) 1.1% 0.0067** (0.0027) 3.4% 0.0019 (0.0013) 1.0% Employed 0.0072*** (0.0020) 4.1% 0.0196*** (0.0049) 9.9% 0.0023 (0.0019) 1.3% Income 7.5% 14.4% 4.9% Q2 -0.0024 (0.0024) -1.4% -0.0049* (0.0030) -2.4% -0.0024 (0.0029) -1.3% Q3 -0.0021 (0.0013) -1.2% -0.0020 (0.0019) -1.0% -0.0005 (0.0017) -0.3% Q4 0.0027* (0.0014) 1.5% 0.0089** (0.0037) 4.5% -0.0019 (0.0023) -1.1% Q5 0.0150*** (0.0041) 8.5% 0.0266*** (0.0064) 13.4% 0.0138* (0.0071) 7.6% Digital technology 19.7% 22.8% 16.3% Mobile use 0.0348*** (0.0049) 19.7% 0.0454*** (0.0072) 22.8% 0.0293*** (0.0067) 16.3% Financial knowledge 57.4% 37.1% 60.2% Financial knowledge index 0.1014*** (0.0074) 57.4% 0.0738*** (0.0151) 37.1% 0.1084*** (0.0095) 60.2% Region -0.0022*** (0.0008) -1.2% -0.0009 (0.0014) -0.5% -0.0033** (0.0014) -1.8% SE refers to robust standard errors. AME refers to the average marginal effect. The base level for education is No education and for income it is Q1. *p < .1; **p < .05; ***p=.01. 77 The dynamics of financial inclusion in Ethiopia TABLE A.11 Estimates for decomposition analysis on gender gaps in account ownership, 2022 National Urban Rural Men 51.6 80.2 41.8 Women 30.4 64.2 18.8 Gender Gap 21.2 16 23 Explained gap 18.4 86.8% 14.5 90.6% 19.4 84.3% Contributions from gender Contributions from gender Contributions from gender differences differences differences Contribution Contribution Contribution   AME AME AME (%) (%) (%) SE SE SE   Demographic, socioeconomic 0.8% 13.0% 0.3% Age -0.0010 (0.0008) -0.5% -0.0029 (0.0027) -1.8% 0.0005 (0.0014) 0.2% Primary education -0.0022 (0.0029) -1.0% 0.0000 (0.0012) 0.0% -0.0032 (0.0043) -1.4% Secondary education -0.0046 (0.0029) -2.2% -0.0012 (0.0119) -0.8% -0.0031 (0.0022) -1.4% Post secondary 0.0035*** (0.0012) 1.6% 0.0050** (0.0022) 3.1% 0.0033* (0.0018) 1.4% Employed 0.0059** (0.0029) 2.8% 0.0200*** (0.0065) 12.5% 0.0032 (0.0029) 1.4% Income 11.8% 6.1% 10.0% Q2 -0.0052 (0.0032) -2.5% 0.0035 (0.0043) 2.1% -0.0070 (0.0044) -3.0% Q3 -0.0038** (0.0017) -1.8% -0.0007 (0.0012) -0.4% -0.0069** (0.0027) -3.0% Q4 0.0023 (0.0016) 1.1% 0.0022 (0.0026) 1.4% 0.0006 (0.0023) 0.3% Q5 0.0317*** (0.0058) 14.9% 0.0048 (0.0062) 3.0% 0.0364*** (0.0078) 15.8% Digital technology 8.1% 19.1% 2.5% Mobile use 0.0172** (0.0070) 8.1% 0.0306*** (0.0067) 19.1% 0.0057 (0.0094) 2.5% Financial knowledge 64.5% 51.0% 68.8% Financial knowledge index 0.1370*** (0.0102) 64.5% 0.0817*** (0.0234) 51.0% 0.1585*** (0.0133) 68.8% Region 0.0038*** (0.0012) 1.8% 0.0017 (0.0017) 1.1% 0.0059*** (0.0022) 2.6% SE refers to robust standard errors. AME refers to the average marginal effect. Financial knowledge index refers to the z-score of the index (see Box 2). The base level for education is No education and for income is Q1. *p < .1; **p < .05; ***p=.01. 78 Annexes TABLE A.12 Estimates for Oaxaca-Blinder decomposition analysis on gender gaps in financial inclusion, over time 2019 2022 Men 0.2027 0.233 Women -0.1873 -0.215 Gender Gap 0.3901 0.4644 Explained gap 0.3973 101.8% 0.4148 89.3% Contributions from gender differences Contributions from gender differences Contribution Contribution Coefficients Coefficients   (%) (%) SE SE   Demographic, socioeconomic 9.8% 11.1% Age 0.0111*** (0.0036) 2.8% -0.0010 (0.0022) -0.2% Primary education -0.0066 (0.0053) -1.7% -0.0010 (0.0054) -0.2% Secondary education -0.0037 (0.0037) -0.9% 0.0081 (0.0064) 1.8% Post secondary 0.0162*** (0.0041) 4.1% 0.0215*** (0.0060) 4.6% Employed 0.0213*** (0.0049) 5.5% 0.0238*** (0.0067) 5.1% Income 7.0% 10.3% Q2 0.0018 (0.0037) 0.5% -0.0030 (0.0051) -0.7% Q3 -0.0000 (0.0013) 0.0% -0.0074* (0.0040) -1.6% Q4 0.0018 (0.0018) 0.5% 0.0029 (0.0025) 0.6% Q5 0.0237*** (0.0069) 6.1% 0.0553*** (0.0110) 11.9% Digital technology 16.0% 7.5% Mobile use 0.0623*** (0.0102) 16.0% 0.03476** (0.0137) 7.5% Financial knowledge 69.6% 60.2% Financial knowledge index 0.2716*** (0.0221) 69.6% 0.2798*** (0.0278) 60.2% Region -0.6% -0.2% SE refers to robust standard errors. Financial knowledge index refers to the z-score of the index (see Box 2). The base level for education is No education and for income is Q1. *p < .1; **p < .05; ***p=.01. 79 The dynamics of financial inclusion in Ethiopia Annex 4. Additional Figures FIGURE A.1 FIGURE A.2 Account ownership, by wealth, over time Use of financial services by account holders, over time 80 PROPORTION OF INDIVIDUALS (%) ATM/Debit Card 60 Mobile Banking 40 Interest Free Banking 20 Online Banking Agent Banking 0 0 10 20 30 40 t er e er t es es dl or ch or ch id Po Ri M Po PROPORTION OF INDIVIDUALS (%) Ri 2019 2022 2016 2019 2022 Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. FIGURE A.3 FIGURE A.4 Mobile phone penetration by Usage of financial institutions, over time gender, location, over time 100 PROPORTION OF INDIVIDUALS (%) PROPORTION OF INDIVIDUALS (%) 80 80 60 60 40 40 20 20 0 0 nk nk O ce Female Male Female Male CC Ba Ba an SA fin ic e at bl ro RURAL URBAN iv Pu ic Pr M 2019 2022 2016 2019 2022 Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. 80 Annexes FIGURE A.5 Usage of financial institutions by place of residence, over time PROPORTION OF INDIVIDUALS (%) 100 80 60 40 20 0 2019 2022 2019 2022 URBAN RURAL Public Bank Private Bank SACCO Microfinance Source: Authors’ estimates from ESPS. FIGURE A.6 FIGURE A.7 Age and gender effect on account Age and gender effect on account ownership in 2019 ownership in 2022 .8 .8 PROBABILITY OF ACCOUNT OWNERSHIP PROBABILITY OF ACCOUNT OWNERSHIP .6 .6 Female Female .4 Male Male .4 .2 .2 0 15 20 25 30 35 40 45 50 55 60 15 20 25 30 35 40 45 50 55 60 Age Age Source: Authors’ estimates from ESPS. Source: Authors’ estimates from ESPS. 81 World Bank Group | Living Standards Measurement Study The dynamics of financial inclusion in Ethiopia Diversifying delivery channels, expanding digital technology, and enhancing financial knowledge for greater financial inclusion