## 1. Distribution of Individual Financial Inclusion Scores

## Source details

**Canonical URL:** [1. Distribution of Individual Financial Inclusion Scores](https://www.imf.org/-/media/files/publications/wp/2018/wp1838.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp1838.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp1838.pdf.json)

---

### Introduction: motivation and scope
- Financial inclusion increases individual savings, female empowerment, consumption and productive investment, and income for the poor.
- Evidence links financial inclusion with higher growth, lower income inequality, and lower aggregate consumption volatility.
- Tradeoffs: broader inclusion may impair asset quality for some newly included individuals.
- Gender gaps: worldwide, only 37 percent of women have a formal bank account compared with 46 percent of men.
- Study goal: understand drivers of financial inclusion, particularly for women, using microdata and macro variables, and examine determinants jointly and separately for men and women.
- Data source: World Bank’s Findex database (2011 and 2014 waves planned every three years).

### Data and index construction
- Microdata:
  - 2014 World Findex survey attached to the Gallup world poll.
  - Dataset contains 146,688 individual observations, nationally representative of 142 countries.
  - Survey comprises 44 questions on financial inclusion and individual demographics (gender, educational attainment, income quintile, age, formal and government employment).
- Financial inclusion index:
  - Uses the multi-dimensional index from Aslan and others (2017).
  - Constructed via correspondence analysis from 12 questions directly related to financial access (weights determined by the analysis).
  - Index captures access and intensity of use (e.g., payments, borrowing, saving) beyond a univariate account ownership measure.
  - The 12 questions used: Q1, Q3, Q4, Q6, Q9, Q11, Q14, Q16, Q18a, Q20, Q21a, Q24.
  - Q24 responses combined: “Very possible” and “Somewhat Possible” → “Very or somewhat possible”; “Not very possible” and “Not at all possible” → “Not very or at all possible”.

### Stylized facts from 146,688 observations
- Distributional patterns:
  - The index distribution is skewed towards low-level scores; intensity is concentrated in two low-inclusion bins, largely driven by low-income countries, particularly sub-Saharan Africa.
- Cross-country and income-level patterns:
  - Financial access (e.g., account ownership) increases with countries’ income level; account penetration is quasi universal in advanced economies (income levels beyond US$15,000).
  - Restricting analysis to the bottom 50 percent of economies by income leads to GDP per capita explaining only 22 percent of variation in account penetration.
- Individual-level associations:
  - Women are less likely to access financial services than men; differences are larger in emerging and developing markets.
  - Higher individual income is strongly associated with higher financial inclusion, but gender gaps remain significant even among the richest 40 or 20 percent.
  - In sub-Saharan Africa, access is low for both sexes at the lower end of income distribution.
  - Financial access increases with educational attainment, but gender gaps persist at higher education levels.
  - Individuals receiving wage payments or employed in the public sector have higher financial inclusion scores; women in public sector employment have higher scores than men in that sector.
  - Differences are statistically and economically significant individually and jointly.
- Macro-level associations (scatter R² values reported in source):
  - Real GDP per Capita: R² = 0.69 with financial inclusion.
  - Financial development (supply-side composite): R² = 0.67 with demand-side financial inclusion index.
  - Institutions (ICRG measure): R² = 0.572 with financial inclusion.
  - Mean years of schooling (aggregate human capital): R² = 0.5555 with financial inclusion.

### Key empirical insights
- Gender bias:
  - Regressions show strong evidence of gender bias in financial inclusion, robust to comprehensive individual and country-level controls.
  - Legal discrimination against women and gender norms explain part of cross-country variation in access for women.
  - Higher coverage of legislative protection against harassment and provision of basic economic rights is associated with higher financial inclusion for women.
- Importance of education:
  - Education is important for men and women alike in explaining financial inclusion.
- Complementarity of micro and macro analysis:
  - Combining individual-level demographics with country/regional macro characteristics identifies policy-relevant levers (legal rights, education, institutions, financial infrastructure).

### Methodology and estimation
- Estimation: Ordinary least squares (OLS) across countries (j) and individuals (i) with clustered standard errors at the country level.
- Baseline specification:
  - Financial Inclusion_ij = β′Demographics_ij + γ′WageEmployed_ij + ρ′Structural_j + τ′Policies_j + ε_ij
- Measures:
  - Financial Inclusion_ij: binary account ownership and an index of intensity of financial inclusion.
  - Demographics_ij: gender, education, age, income level (top 20 percent).
  - WageEmployed_ij: indicator for having received wage payments in the last 12 months.
  - Structural_j: GDP p.c., oil-exporter (Fuel Exporter) status, population density.
  - Policies_j: mean years of schooling, financial development, institutional quality measures, legal rights for women, Social Institutions and Gender Index (SIGI) components.

### Selected coefficients and summary statistics (exact values preserved)
- Table 2 — Determinants of the Financial Inclusion Index (selected coefficients with reported standard errors)
  - Female: -0.011**, -0.023***, -0.023***, -0.022***, -0.019***, -0.024***, -0.020*** (std. errors: (0.005), (0.003), (0.004), (0.004), (0.004), (0.004), (0.004))
  - Secondary education: 0.220***, 0.106***, 0.123***, 0.118***, 0.118***, 0.122***, 0.114*** (std. errors: (0.015), (0.009), (0.007), (0.007), (0.006), (0.007), (0.006))
  - Age: 0.009*** (std. errors: (0.001) etc.)
  - Age squared: -0.000*** (consistent)
  - Wage employment: 0.213***, 0.144***, 0.138***, 0.132***, 0.134***, 0.136***, 0.137*** (std. errors: (0.009), (0.008), (0.007), (0.007), (0.007), (0.007), (0.008))
  - Top 20 percent income: 0.052***, 0.083***, 0.083***, 0.084***, 0.082***, 0.084***, 0.084*** (std. errors: (0.006), (0.004), (0.005), (0.005), (0.005), (0.005), (0.004))
  - GDP p.c.: 0.127***, 0.078***, 0.070**, 0.097***, 0.060***, 0.056*** (std. errors vary)
  - Fuel Exporter: -0.132***, -0.070***, -0.058*, -0.054**, -0.030, -0.021 (std. errors vary)
  - Financial development: 0.213***, 0.199***, 0.134**, 0.229***, 0.208*** (std. errors vary)
  - Institutions (ICRG): 0.003**, 0.002, 0.001 (std. errors vary)
  - Political risk (ICRG): 0.005*** (0.001)
  - Corruption Perceptions Index 2014: 0.004*** (0.001)
  - Adjusted R squared (selected): 0.321, 0.479, 0.485, 0.504, 0.504, 0.491, 0.508
- Table 3 — Determinants restricted to Female Sample (selected)
  - Secondary education: 0.212***, 0.102***, 0.115***, 0.133***, 0.114***, 0.117***, 0.133***, 0.115***, 0.118*** (std. errors reported)
  - Age: 0.008***, 0.007***, 0.007***, 0.010***, 0.007***, 0.007***, 0.010***, 0.007***, 0.007*** (std. errors)
  - Wage employment: 0.241***, 0.169***, 0.160***, 0.139***, 0.148***, 0.151***, 0.139***, 0.148***, 0.151*** (std. errors)
  - GDP p.c.: 0.127***, 0.067***, 0.085***, 0.080***, 0.094***, 0.084***, 0.077***, 0.093*** (std. errors)
  - Fuel Exporter: -0.139***, -0.081***, -0.049, -0.053**, -0.043*, -0.051*, -0.054**, -0.044* (std. errors)
  - Financial development: 0.209***, 0.168***, 0.184***, 0.118**, 0.165***, 0.180***, 0.116** (std. errors)
  - Institutions (ICRG): 0.004**, 0.003*, 0.002, 0.000, 0.003*, 0.002, 0.000 (std. errors)
  - Legal rights (Civil remedies for sexual harassment): 0.032*, 0.054***, 0.055*** (std. errors shown)
  - Adjusted R squared (selected): 0.337, 0.485, 0.491, 0.480, 0.518, 0.518, 0.478, 0.518, 0.518
- Table 4 — Accounting for Social Norms and Legal Rights (selected)
  - Female: -0.025***, -0.023***, -0.026***, -0.024***, -0.021***, -0.022***, -0.025***, -0.022***, -0.023***, -0.024*** (std. errors)
  - GDP p.c.: 0.063**, 0.040*, 0.077***, 0.043*, 0.065**, 0.036, 0.078***, 0.045*, 0.065***, 0.039 (std. errors)
  - Financial development: 0.236***, 0.268***, 0.213***, 0.240***, 0.273***, 0.310***, 0.205***, 0.240***, 0.253***, 0.293*** (std. errors)
  - Social discrimination against women (SIGI): -0.188*, -0.192** (std. errors: (0.096), (0.089))
  - Discriminatory family code Value: -0.108*, -0.121*, -0.116*, -0.098 (std. errors vary)
  - Son bias Value: -0.131***, -0.076*, -0.114***, -0.055 (std. errors vary)
  - Restricted civil liberties Value: -0.076**, -0.062, -0.041, -0.065 (std. errors vary)
  - Adjusted R squared (selected): 0.382, 0.286, 0.489, 0.286, 0.490, 0.289, 0.486, 0.287, 0.497, 0.292

### Analysis and interpretation
- Individual-level predictors:
  - Gender, education, age, income, and wage employment are strong predictors of financial inclusion.
  - Example magnitude: being female—everything else equal—lowers an individual’s financial inclusion score by just 0.01 points in a quoted specification.
- Country-level factors:
  - GDP p.c. and oil-dependence (Fuel Exporter status) significantly influence average financial inclusion.
  - Financial development and institutional quality strongly associated with higher inclusion.
- Legal protections and social norms:
  - Legal remedies against sexual harassment (general and in employment) are positively and significantly associated with female financial inclusion in female-restricted samples.
  - Higher SIGI scores and components (discriminatory family code, son bias, restricted civil liberties) are associated with lower female financial inclusion.
- Robustness:
  - Results hold across specifications including income dummies, regional dummies, alternative institutional measures, interaction terms, and probit estimations for account ownership.

### Policy recommendations (as presented in source)
- Advance financial sector development and improve institutional quality to raise financial inclusion overall (strengthen financial infrastructure and business environment).
- Promote economic diversification in oil-exporting countries to raise average financial inclusion levels.
- Strengthen legal protections against harassment, including at the workplace (civil remedies), as these laws are associated with higher female financial inclusion.
- Address social norms and discriminatory practices (reduce son bias, reform discriminatory family codes, expand civil liberties) to improve women’s access to financial services.
- Equalize educational attainment between men and women to narrow gender gaps in financial access.

### Annex summaries (selected)
- Annex I — Correspondence Analysis:
  - CA gives nominal categorical variables a notion of distance; uses chi-square distances and SVD to find low-rank representations.
  - MCA and JCA described; JCA used for robustness though broadly similar results to MCA.
- Annex II — Country sample:
  - Whole sample: list of countries preserved (e.g., Afghanistan, Albania, Algeria, ... Zimbabwe).
  - Emerging and developing economies: exact list preserved (e.g., Argentina, Armenia, Azerbaijan, ... Vietnam, Republic of Yemen, Zambia, Zimbabwe).
  - Sub-Saharan Africa: exact list preserved (e.g., Angola, Benin, Botswana, ... Zambia, Zimbabwe).
- Annex III — Variables list (selected):
  - Financial Inclusion: Staff calculation from WB survey.
  - Female: WB Survey binary (female = 1).
  - Education, at least secondary: WB Survey binary (completed secondary or higher = 1).
  - Age; Top 20 percent income; Wage employed; Population density; lgdppcppp (Penn World Table 8.0); Financial Development (IMF, Sahay 2015); Fuel Exporters; Fertility rate; Women’s mean age at marriage; Mean years of schooling (Barro & Lee, 2013); ICRG risk rating; Legislation against domestic violence; Legal rights index; SIGI; Corruption Perception Index.
- Annex IV — Summary statistics (selected exact entries)
  - Proper2014a Financial Inclusion Index: N = 141188, Mean = 0.421, StDev = 0.318, Min = 0.0, Median = 0.449, Max = 1
  - female Female: N = 141188, Mean = 0.532, StDev = 0.499, Min = 0, Median = 1, Max = 1
  - educ_s Secondary Education: N = 141188, Mean = 0.668, StDev = 0.471, Min = 0, Median = 1, Max = 1
  - inc_top_20 Top 20% income: N = 141172, Mean = 0.247, StDev = 0.431, Min = 0, Median = 0, Max = 1
  - lgdppcppp Real GDP p.c.: N = 140168, Mean = 9.241, StDev = 1.145, Min = 6.569, Median = 9.351, Max = 11.419
  - findepth13 Financial development: N = 135180, Mean = 0.399, StDev = 0.237, Min = 0.064, Median = 0.322, Max = 0.945
  - ICRG_crr ICRG: N = 121612, Mean = 69.483, StDev = 8.282, Min = 47.756, Median = 68.759, Max = 90
  - sigi_2014 Social discrimination against women: N = 89411, Mean = 0.18, StDev = 0.139, Min = 0.002, Median = 0.147, Max = 0.563
  - account Account at financial institution: N = 141188, Mean = 0.58, StDev = 0.493, Min = 0, Median = 1, Max = 1
- Annex V — Correlation highlights (selected pairwise correlations)
  - account — Proper2014a1: 0.882***
  - account — female: -0.0598***
  - account — educ_s: 0.282***
  - account — lgdppcppp: 0.362***
  - account — findepth13: 0.358***
  - account — ICRG_crr: 0.229***
  - account — fertility_rate: -0.324***
  - meanyrsch — lgdppcppp: 0.802***
  - findepth13 — lgdppcppp: 0.671***
  - sigi1 — sigi_2014: 0.825***
  - Full matrix of pairwise correlations with significance levels reported in Table A3.
- Annex VI — Robustness checks (selected)
  - Table A4 (Female interactions): Secondary education coefficients include 0.225*** (0.016), 0.111*** (0.009), and female interaction terms such as -0.013** (0.006).
  - Age and age squared across columns show 0.010*** (0.001) and -0.000*** (0.000) patterns with female interactions often negative and significant.
  - Financial development: 0.240*** (0.042) with female interaction -0.043** (0.018) in one specification.
  - Civil remedies for sexual harassment: 0.054*** (0.018).
  - Adjusted R squared: (1) 0.322, (2) 0.480, (3) 0.487, (4) 0.495, (5) 0.495.
  - Table A5 (Probit for women’s account probability — marginal effects, std. errors):
    - Secondary education: (1) 0.838*** (0.076) ... (6) 0.521*** (0.033)
    - Age: (1) 0.031*** (0.005) ... (6) 0.033*** (0.003)
    - Wage employment: (1) 1.008*** (0.051) ... (6) 0.827*** (0.047)
    - GDP p.c.: (1) 0.566*** (0.047) ... (5) 0.514*** (0.141)
    - Fuel Exporter: (1) -0.666*** (0.119) ... (5) -0.199 (0.124)
    - Financial development: (1) 1.218*** (0.266) ... (4) 0.703** (0.347)
    - Civil remedies for sexual harassment: (3) 0.337*** (0.114), (4) 0.362*** (0.106), (5) 0.348*** (0.097)
    - Significance: * p<0.10;  ** p<0.05; *** p<0.010

*Source: IMF working paper content unit “1. Distribution of Individual Financial Inclusion Scores” (based on Findex 2014 and Aslan and others (2017) index construction).*

### 1. Distribution of Individual Financial Inclusion Scores _____________________________7

### 1. Distribution of Individual Financial Inclusion Scores _____________________________7

### Introduction: motivation and scope
- Financial inclusion has important macroeconomic consequences: increases in individual savings, female empowerment, consumption and productive investment, and income for the poor.
- Evidence links financial inclusion with higher growth, lower income inequality, and lower aggregate consumption volatility.
- Tradeoffs exist: broader inclusion may impair asset quality for some newly included individuals.
- Gender gaps are systematic and persistent: worldwide, only 37 percent of women have a formal bank account compared with 46 percent of men.
- Study goal: understand drivers of financial inclusion, particularly for women, using microdata and macro variables, and examine determinants jointly and separately for men and women.
- Data source: World Bank’s Findex database (2011 and 2014 waves planned every three years).

### Data and index construction
- Microdata:
  - 2014 World Findex survey attached to the Gallup world poll.
  - Dataset contains 146,688 individual observations, nationally representative of 142 countries.
  - Survey comprises 44 questions on financial inclusion and individual demographics (gender, educational attainment, income quintile, age, formal and government employment).
- Financial inclusion index:
  - Uses the multi-dimensional index from Aslan and others (2017).
  - Constructed via correspondence analysis from 12 questions directly related to financial access (weights determined by the analysis).
  - Index captures access and intensity of use (e.g., payments, borrowing, saving) beyond a univariate account ownership measure.
  - Table 1 lists the 12 questions used (Q1, Q3, Q4, Q6, Q9, Q11, Q14, Q16, Q18a, Q20, Q21a, Q24) and their response categories as presented in the source.
  - Q24 responses were combined: “Very possible” and “Somewhat Possible” → “Very or somewhat possible”; “Not very possible” and “Not at all possible” → “Not very or at all possible”.

### Stylized facts from 146,688 observations
- Distributional patterns:
  - The index distribution is skewed towards low-level scores; intensity is concentrated in two low-inclusion bins, largely driven by low-income countries, particularly sub-Saharan Africa.
- Cross-country and income-level patterns:
  - Financial access (e.g., account ownership) increases with countries’ income level; account penetration is quasi universal in advanced economies (income levels beyond US$15,000).
  - Restricting analysis to the bottom 50 percent of economies by income leads to GDP per capita explaining only 22 percent of variation in account penetration (as noted in the literature cited).
- Individual-level associations (global, emerging/developing, and sub-Saharan Africa):
  - Women are less likely to access financial services than men; differences are larger in emerging and developing markets.
  - Higher individual income is strongly associated with higher financial inclusion, but gender gaps remain significant even among the richest 40 or 20 percent.
  - In sub-Saharan Africa, access is low for both sexes at the lower end of income distribution.
  - Financial access increases with educational attainment, but gender gaps persist at higher education levels.
  - Individuals receiving wage payments or employed in the public sector have higher financial inclusion scores; women in public sector employment have higher scores than men in that sector.
  - These differences are statistically and economically significant individually and jointly (as shown in later regressions).
- Macro-level associations:
  - Development (Real GDP per Capita) is strongly associated with higher financial inclusion (Figure 4; scatter shows R² = 0.69).
  - Financial development (supply-side composite of depth, efficiency, access) is closely related to the demand-side financial inclusion index (R² = 0.67).
  - Quality of institutions (ICRG measure) correlates with higher financial inclusion (R² = 0.572).
  - Aggregate human capital (mean years of schooling) is strongly related to higher financial inclusion (R² = 0.5555).

### Key empirical insights highlighted
- Gender bias:
  - Regressions show strong evidence of gender bias in financial inclusion, robust to comprehensive individual and country-level controls.
  - Legal discrimination against women and gender norms explain part of cross-country variation in access for women.
  - Higher coverage of legislative protection against harassment and provision of basic economic rights is associated with higher financial inclusion for women.
- Importance of education:
  - Education is important for men and women alike in explaining financial inclusion.
- Complementarity of micro and macro analysis:
  - Combining individual-level demographics with country/regional macro characteristics identifies policy-relevant levers (e.g., legal rights, education, institutions, financial infrastructure).

### Methodological notes relevant to interpretation
- Index construction:
  - Correspondence analysis (categorical PCA equivalent) used to derive weights for the 12 access-related questions; index excludes individual personal characteristics or environment.
- Robustness:
  - Study uses the index as primary measure and single measures (e.g., having an account) in robustness checks.
- Sample coverage:
  - Global sample expands beyond prior studies focused on developing countries; controls include resource-rich status and policy-relevant indicators such as education level, institutional quality, and financial depth.

*Source: IMF working paper content unit “1. Distribution of Individual Financial Inclusion Scores” (based on Findex 2014 and Aslan and others (2017) index construction).*

### 17.      Based on the above stylized facts, we combine the detailed information on

### 17.      Based on the above stylized facts, we combine the detailed information on

### Methodology
- Estimation: Ordinary least squares (OLS) across countries (j) and individuals (i) with clustered standard errors at the country level.
- Baseline specification:
  - Financial Inclusion_ij = β′Demographics_ij + γ′WageEmployed_ij + ρ′Structural_j + τ′Policies_j + ε_ij
- Data sources and variables: Individual-level data from the Findex database combined with macroeconomic country-level data (see Annex II and III in source).
- Measures:
  - Financial Inclusion_ij: binary account ownership and an index of intensity of financial inclusion.
  - Demographics_ij: gender, education, age, income level (top 20 percent).
  - WageEmployed_ij: indicator for having received wage payments in the last 12 months.
  - Structural_j: GDP p.c., oil-exporter (Fuel Exporter) status, population density.
  - Policies_j: mean years of schooling, financial development, institutional quality measures, legal rights for women (e.g., civil remedies against harassment), Social Institutions and Gender Index (SIGI) components.

### Key Empirical Results
- Individual characteristics:
  - Men, higher-income individuals, and wage-employed individuals have higher intensity of using financial services.
  - Being female is associated with lower financial inclusion; equalizing education levels could narrow the gender gap. Text example: "being female—everything else equal—lowering an individual’s financial inclusion score by just 0.01 points."
- Structural country characteristics:
  - Higher GDP p.c. is strongly related to higher financial inclusion.
  - Fuel Exporter (oil-exporter) status is associated with lower average levels of financial inclusion.
- Policies:
  - Greater financial development and stronger institutional quality are strongly associated with higher financial inclusion.
  - Results robust to inclusion of income-group dummies, regional dummies, and alternative institutional quality measures.

### Determinants — Selected Coefficients (from Tables)
- Table 2. Determinants of the Financial Inclusion Index (multiple specifications):
  - Female: -0.011**, -0.023***, -0.023***, -0.022***, -0.019***, -0.024***, -0.020*** (standard errors in parentheses: (0.005), (0.003), (0.004), (0.004), (0.004), (0.004), (0.004))
  - Secondary education: 0.220***, 0.106***, 0.123***, 0.118***, 0.118***, 0.122***, 0.114*** (std. errors: (0.015), (0.009), (0.007), (0.007), (0.006), (0.007), (0.006))
  - Age: 0.009*** (consistent across columns) (std. errors: (0.001) etc.)
  - Age squared: -0.000*** (consistent)
  - Wage employment: 0.213***, 0.144***, 0.138***, 0.132***, 0.134***, 0.136***, 0.137*** (std. errors: (0.009), (0.008), (0.007), (0.007), (0.007), (0.007), (0.008))
  - Top 20 percent income: 0.052***, 0.083***, 0.083***, 0.084***, 0.082***, 0.084***, 0.084*** (std. errors: (0.006), (0.004), (0.005), (0.005), (0.005), (0.005), (0.004))
  - GDP p.c.: 0.127***, 0.078***, 0.070**, 0.097***, 0.060***, 0.056*** (std. errors vary)
  - Fuel Exporter: -0.132***, -0.070***, -0.058*, -0.054**, -0.030, -0.021 (std. errors vary)
  - Financial development: 0.213***, 0.199***, 0.134**, 0.229***, 0.208*** (std. errors vary)
  - Institutions (ICRG): 0.003**, 0.002, 0.001 (std. errors vary)
  - Political risk (ICRG): 0.005*** (0.001)
  - Corruption Perceptions Index 2014: 0.004*** (0.001)
  - Adjusted R squared (selected): 0.321, 0.479, 0.485, 0.504, 0.504, 0.491, 0.508
- Table 3. Determinants — Restricted to Female Sample (selected):
  - Secondary education: 0.212***, 0.102***, 0.115***, 0.133***, 0.114***, 0.117***, 0.133***, 0.115***, 0.118*** (standard errors reported)
  - Age: 0.008***, 0.007***, 0.007***, 0.010***, 0.007***, 0.007***, 0.010***, 0.007***, 0.007*** (std. errors)
  - Wage employment: 0.241***, 0.169***, 0.160***, 0.139***, 0.148***, 0.151***, 0.139***, 0.148***, 0.151*** (std. errors)
  - GDP p.c.: 0.127***, 0.067***, 0.085***, 0.080***, 0.094***, 0.084***, 0.077***, 0.093*** (std. errors)
  - Fuel Exporter: -0.139***, -0.081***, -0.049, -0.053**, -0.043*, -0.051*, -0.054**, -0.044* (std. errors)
  - Financial development: 0.209***, 0.168***, 0.184***, 0.118**, 0.165***, 0.180***, 0.116** (std. errors)
  - Institutions (ICRG): 0.004**, 0.003*, 0.002, 0.000, 0.003*, 0.002, 0.000 (std. errors)
  - Legal rights (Civil remedies for sexual harassment): 0.032*, 0.054***, 0.055*** (std. errors: -0.019, -0.017, -0.016 shown as parentheses in table)
  - Adjusted R squared (selected): 0.337, 0.485, 0.491, 0.480, 0.518, 0.518, 0.478, 0.518, 0.518
- Table 4. Determinants — Accounting for Social Norms and Legal Rights (selected Global and EMDE columns):
  - Female: -0.025***, -0.023***, -0.026***, -0.024***, -0.021***, -0.022***, -0.025***, -0.022***, -0.023***, -0.024*** (std. errors: (0.005) etc.)
  - GDP p.c.: 0.063**, 0.040*, 0.077***, 0.043*, 0.065**, 0.036, 0.078***, 0.045*, 0.065***, 0.039 (std. errors vary)
  - Financial development: 0.236***, 0.268***, 0.213***, 0.240***, 0.273***, 0.310***, 0.205***, 0.240***, 0.253***, 0.293*** (std. errors vary)
  - Social discrimination against women (SIGI): -0.188*, -0.192** (std. errors: (0.096), (0.089))
  - Discriminatory family code Value: -0.108*, -0.121*, -0.116*, -0.098 (std. errors vary)
  - Son bias Value: -0.131***, -0.076*, -0.114***, -0.055 (std. errors vary)
  - Restricted civil liberties Value: -0.076**, -0.062, -0.041, -0.065 (std. errors vary)
  - Adjusted R squared (selected): 0.382, 0.286, 0.489, 0.286, 0.490, 0.289, 0.486, 0.287, 0.497, 0.292

### Analysis and Interpretation
- Individual-level factors (gender, education, age, income, wage employment) are strong predictors of financial inclusion.
- Country-level structural characteristics (GDP p.c., oil-dependence) and policy variables (financial development, institutions) significantly explain cross-country variation in financial inclusion beyond individual characteristics.
- Legal protections and social norms specifically matter for female financial inclusion:
  - Legal remedies against sexual harassment (general and in employment) are positively and significantly associated with female financial inclusion in female-restricted samples.
  - Higher SIGI scores (stronger social discrimination against women) and specific SIGI components (discriminatory family code, son bias, restricted civil liberties) are associated with lower female financial inclusion.
- Robustness: results hold across specifications including income dummies, regional dummies, alternative institutional measures, interaction terms, and probit estimations for account ownership.

### Policy Recommendations
- Advance financial sector development and improve institutional quality to raise financial inclusion overall (e.g., strengthen financial infrastructure and business environment).
- Promote economic diversification in oil-exporting countries to raise average financial inclusion levels.
- Strengthen legal protections against harassment, including at the workplace (civil remedies), as these laws are associated with higher female financial inclusion and will likely increase women’s economic participation and bargaining power.
- Address social norms and discriminatory practices (reduce son bias, reform discriminatory family codes, expand civil liberties) to improve women’s access to financial services.
- Policies that equalize educational attainment between men and women can narrow gender gaps in financial access.

*Italic: Source: IMF working paper content (excerpt provided).*

### References

### wp1838 - References

### Major referenced topics
- Financial inclusion and its measurement (e.g., “Measuring Financial Inclusion: The Global Findex Database”; “The Foundations of Financial Inclusion: Understanding Ownership and Use of Formal Accounts”; “Assessing Countries’ Financial Inclusion Standing—A new Composite Index”).
- Gender, women’s empowerment, and gender gaps in access to finance and labor markets (e.g., “Empowering the Third Billion. Women and the World of Work in 2012.”; “Financial Inclusion and Legal Discrimination Against Women: Evidence from Developing Countries”; “Catalyst for Change: Empowering Women and Tackling Income Inequality”).
- Links between finance, inequality, and growth (e.g., “Finance and Inequality: Channels and Evidence”; “Causes and Consequences of Income Inequality: A Global Perspective”; “Financial Deepening, Inequality, and Growth: A Model-Based Quantitative Evaluation”).
- Financial stability and policy trade-offs related to inclusion (e.g., “The Nexus of Financial Inclusion and Financial Stability. A Study of Trade-Offs and Synergies”; “Financial Inclusion-Issues for Central Banks”).
- Empirical and methodological contributions (e.g., correspondence and multiple correspondence analysis methods; country-level indices such as the Financial Development Index and broad-based index of financial development).

### Selected cited works (authors and year highlights)
- Aguirre, DeAnne; Hoteit, Leila; Rupp, Christine; Sabbagh, Karim. 2012.
- Aportela, F. 1999.
- Ashraf, N.; Karlan, D.; Yin, W. 2010.
- Allen, Franklin; Demirguc-Kunt, Asli; Klapper, Leora; Martinez Peria, Maria Soledad. 2012.
- Amidžić, Goran; Massara, Alexander; Mialou, André. 2014.
- Aslan, Goksu; Deléchat, Corinne; Newiak, Monique; Yang, Fan. 2017.
- Aterido, Reyes; Beck, Thorsten; Iacovone, Leonardo. 2013.
- Beck, Thorsten; Levine, Ross; Levkov, Alexey. 2007.
- Beck, Thorsten; Demirguc-Kunt, Asli. 2008.
- Cihak, Martin; Mare, Davide S.; Melecky, Martin. 2016.
- Bellucci, Andrea; Borisov, Alexander; Zazzaro, Alberto. 2010.
- Claessens, Stijn; Perotti, Enrico. 2007.
- Dabla-Norris, Era et al. 2015a; Dabla-Norris, Era et al. 2015b.
- De Koker, Louis; Jentzsch, Nicola. 2013.
- Demirguc-Kunt, Asli; Klapper, Leora. 2012a; 2012b; Demirguc-Kunt, Asli; Klapper, Leora; Singer, Dorothe. 2013.
- Dieterich, Christine; Huang, Anni; Thomas, Alun. 2015.
- Gine, Xavier; Townsend, Robert M. 2004.
- Gonzales, Christian; Jain-Chandra, Sonali; Kochhar, Kalpana; Newiak, Monique; Zeinullayev, Tlek. 2015b.
- Hakura, Dalia et al. 2016.
- Han, Rui; Melecky, Martin. 2013.
- Honohan, Patrick. 2007.
- Klapper, Leora; Parker, Simon. 2010.
- Klasen, Stephan. 1999; Klasen, Stephan; Lamanna, Francesca. 2009.
- Konte, Maty. 2015.
- Leroux, Brigitte; Rouanet, Henry. 2010.
- Mehrotra, Aaron N.; Yetman, James. 2015.
- Miller, G. 2008.
- Mookerjee, Rajen; Kalipioni, Paul. 2010.
- Muravyev, Alexander; Talavera, Oleksandr; Schäfer, Dorothea. 2009.
- Sahay, Ratna et al. 2015.
- Svirydzenka, Katsiaryna. 2016.
- Swamy, Vighneswara. 2014.
- Townsend, Robert M.; Ueda, Kenichi. 2006.
- WorldBank. 2012.
- Zins, Alexandra; Weill, Laurent. 2016.

### Methodological reference highlights
- Multiple Correspondence Analysis: Leroux, Brigitte; Rouanet, Henry. 2010. “Multiple Correspondence Analysis: Quantitative Applications in the Social Sciences, Volume 163 of Sage University papers / Quantitative applications in the social sciences.”
- Joint Correspondence Analysis (JCA) and Multiple Correspondence Analysis (MCA): explanation and choice of JCA for robustness in this work (Annex I).

---

### Annex I. Correspondence Analysis (summary)
- Motivation
  - Categorical variables often lack comparable scale and distance properties; Correspondence Analysis (CA) gives nominal variables a notion of distance to analyze associations across many categorical questions.
  - CA is established in biometrics, psychometrics, marketing, ecology, and computational sciences.

- Simple Correspondence Analysis (step summary)
  - Example: two questions, each with three responses (“yes”, “no”, “missing”).
  - Construct a contingency table (Table A2) with notation: n_rc, n_r+, n_+c, n_++.
  - Weighted chi-square distance between columns c and c′:
    - d_chi^2(c,c′) = sqrt( sum_{r=1}^{R} ( n_++ / n_r+ * ( n_rc / n_+c − n_rc′ / n_+c′ )^2 ) )
    - Each response has chi-square distance to the “average” response c*:
    - d_chi^2^2(c,c*) = sum_{r=1}^{R} ( n_++ / n_r+ * ( n_rc / n_+c − n_r+ / n_++ )^2 )
  - Column mass weights distances to obtain inertia:
    - Inertia = sum_{c=1}^{J} ( n_+c / n_++ * d_chi^2^2(c,c*) )
  - Use Singular Value Decomposition (SVD) to find a rank K ≤ min(M−1, N−1) subspace maximizing total inertia (analogous to PCA).

- Multiple Correspondence Analysis (MCA) and Joint Correspondence Analysis (JCA)
  - Construct N x Q complete disjunctive indicator matrix Z (1 if selected response, 0 otherwise).
  - Form Q x Q matrix B = Z^T Z; center and standardize using marginal totals to find S; apply SVD-based algorithm.
  - JCA addresses influence of diagonals of B by iteratively fitting using only off-diagonal values.
  - The paper tried both MCA and JCA; ultimately used JCA for robustness (broadly similar results).

---

### Annex II. Country Sample
- Whole Sample (alphabetical, exact list preserved): Afghanistan, Albania, Algeria, Angola, Argentina, Armenia, Australia, Azerbaijan, Bahrain, Bangladesh, Belarus, Belgium, Belize, Benin, Bhutan, Bolivia, Bosnia and Herzegovina, Botswana, Brazil, Bulgaria, Burkina Faso, Burundi, Cambodia, Cameroon, Canada, Chad, Chile, China, Columbia, Democratic Republic of Congo, Republic of Congo, Costa Rica, Croatia, Cyprus, Czech Republic, Cote d’Ivoire, Denmark, Dominican Republic, Ecuador, Egypt, Arab Rep., El Salvador, Estonia, Ethiopia, Finland, France, Gabon, Georgia, Germany, Ghana, Greece, Guatemala, Guinea, Haiti, Honduras, Hong Kong SAR, China, Hungary, India, Indonesia, Iraq, Ireland, Israel, Italy, Jamaica, Japan, Jordan, Kazakhstan Kenya, Korea, Rep., Kosovo, Kuwait, Kyrgyz Republic, Latvia, Lebanon, Lithuania, Luxembourg, Macedonia, FYR, Madagascar, Malawi, Malaysia, Mali, Malta, Mauritania, Mauritius, Mexico, Moldova, Mongolia, Montenegro, Myanmar, Namibia, Nepal, Netherlands, New Zealand, Nicaragua, Niger, Nigeria, Norway, Pakistan, Panama, Peru, Philippines, Poland, Portugal, Puerto Rico, Romania, Russian Federation, Rwanda, Saudi Arabia, Senegal, Serbia, Sierra Leone, Singapore, Slovak Republic, Slovenia, Somalia, South Africa, Spain, Sri Lanka, Sudan, Sweden, Switzerland, Taiwan, China, Tajikistan, Tanzania, Thailand, Togo, Tunisia, Turkey, Turkmenistan, Uganda, Ukraine, United Arab Emirates, United Kingdom, United States, Uruguay, Uzbekistan, Venezuela, RB, Vietnam, West Bank and Gaza, Yemen, Rep., Zambia, Zimbabwe.

- Emerging and developing economies (exact list preserved): Argentina, Armenia, Azerbaijan, Bahrain, Bangladesh, Belarus, Belize, Benin, Bhutan, Bolivia, Bosnia and Herzegovina, Botswana, Brazil, Bulgaria, Burkina Faso, Burundi, Cambodia, Cameroon, Chad, Chile, China, Columbia, Democratic Republic of Congo, Republic of Congo, Costa Rica, Croatia, Dominican Republic, Ecuador, Arab Republic of Egypt, El Salvador, Ethiopia, Gabon, Georgia, Ghana, Guatemala, Guinea, Haiti, Honduras, Hungary, India, Indonesia, Iraq, Jamaica, Jordan, Kazakhstan, Kenya, Kosovo, Kuwait, Kyrgyz Republic, Lebanon, Macedonia, Madagascar, Malawi, Malaysia, Mali, Mauritania, Mauritius, Mexico, Moldova, Mongolia, Montenegro, Myanmar, Namibia, Nepal, Nicaragua, Niger, Nigeria, Pakistan, Panama, Peru, Philippines, Poland, Romania, Russian Federation, Rwanda, Saudi Arabia, Senegal, Serbia, Sierra Leone, South Africa, Sri Lanka, Sudan, Tajikistan, Tanzania, Thailand, Togo, Tunisia, Turkey, Turkmenistan, Uganda, Ukraine, United Arab Emirates, Uruguay, Uzbekistan, Venezuela, Vietnam, Republic of Yemen, Zambia, Zimbabwe.

- Sub-Saharan Africa (exact list preserved): Angola, Benin, Botswana, Burkina Faso, Burundi, Cameroon, Chad, Democratic Republic of Congo, Republic of Congo, Cote d’Ivoire, Ethiopia, Gabon, Ghana, Guinea, Kenya, Madagascar, Malawi, Mali, Mauritania, Mauritius, Namibia, Niger, Nigeria, Rwanda, Senegal, Sierra Leone, Somalia, South Africa, Sudan, Tanzania, Togo, Uganda, Zambia, Zimbabwe.

---

### Annex III. Variables List (Table A1 descriptions — selected fields preserved exactly)
- Financial Inclusion
  - Staff calculation from WB survey: Individual financial inclusion access scores calculated from WB Survey.
- Female
  - WB Survey: Binary, being female equals to 1.
- Education, at least secondary
  - WB Survey: Binary as having completed secondary or higher education equals to 1.
- Age
  - WB Survey.
- Top 20 percent income
  - WB Survey and staff calculations: Binary, being in the top 20% income quintile equals to 1.
- Wage employed
  - WB Survey: Binary, having received wage payment in the last 12 months equals to 1.
- Population density
  - World Bank: Land area divided by adult population defined as WB Survey.
- lgdppcppp
  - Penn World Table 8.0: Log of output-side real GDP at chained PPPs (in mil. 2011US$) divided by population.
- Financial Development
  - Financial Development Index: IMF, Sahay 2015.
- Fuel Exporters
  - (listed variable name).
- Fertility rate
  - World Bank: Fertility rate, total (births per woman).
- Women’s mean age at marriage
  - United Nations UNPD World Fertility and Marriage Database.
- Mean years of schooling
  - Barro & Lee, 2013: Mean years of schooling.
- ICRG risk rating (composite index and political risk index)
  - International Country Risk Guide: Composite of Political, Financial, Economic Risk Rating for a country (CPFER) = 0.5 ((Political Risk + Financial Risk + Economic Risk). Ranging from Very High Risk (00.0 - 49.5) to Very Low Risk (80.0 - 100). The higher the points, the lower the risk.
- Legislation against domestic violence (average; and civil remedies for sexual harassment)
  - Women Business and the Law: Binary as 1 if there is a legislation against domestic violence.
- Legal rights index
  - Women Business and the law: 0 to 1, with 1 denoting full range of rights.
- Social discrimination against women (SIGI)
  - Social Institutions and Gender Index: 0 to 1, with 1 indicating high inequality.
- Corruption Perception Index
  - Transparency International: 0 (highly corrupt) to 100 (very clean).

---

### Annex IV. Summary Statistics (Table A.2 — selected exact entries)
- Proper2014a Financial Inclusion Index: N = 141188, Mean = 0.421, StDev = 0.318, Min = 0.0, Median = 0.449, Max = 1
- female Female: N = 141188, Mean = 0.532, StDev = 0.499, Min = 0, Median = 1, Max = 1
- educ_s Secondary Education: N = 141188, Mean = 0.668, StDev = 0.471, Min = 0, Median = 1, Max = 1
- age Age: N = 1408784, Mean = 1.774, StDev = 17.713, Min = 0, Median = 154, Max = 099
- age2 Age squared: N = 1408782, Mean = 2058.816, StDev = 1655.252, Min = 2516, Median = 009801 (note: values preserved exactly as in source table)
- q34 Public employment: N = 140206, Mean = 0.339, StDev = 0.474, Min = 0, Median = 1
- inc_top_20 Top 20% income: N = 141172, Mean = 0.247, StDev = 0.431, Min = 0, Median = 0, Max = 1
- pop_dens Population density: N = 141188, Mean = 1.9757, StDev = 7.458, Min = 0.013, Median = 0.612, Max = 64.102
- lgdppcppp Real GDP p.c.: N = 140168, Mean = 9.241, StDev = 1.145, Min = 6.569, Median = 9.351, Max = 11.419
- Fuel_Exporters Fuel Exporters: N = 141188, Mean = 0.149, StDev = 0.356, Min = 0, Median = 0, Max = 1
- meanyrsch Mean years of schooling: N = 131575, Mean = 8.286, StDev = 3.019, Min = 1.38, Median = 8.6, Max = 12.9
- findepth13 Financial development: N = 135180, Mean = 0.399, StDev = 0.237, Min = 0.064, Median = 0.322, Max = 0.945
- ICRG_crr ICRG: N = 121612, Mean = 69.483, StDev = 8.282, Min = 47.756, Median = 68.759, Max = 90
- dm_avg Domestic violence legislation: N = 139188, Mean = 0.578, StDev = 0.237, Min = 0.034, Median = 0.667, Max = 0.893
- C_avg Other laws on female empowernment: N = 92068, Mean = 0.83, StDev = 0.164, Min = 0.333, Median = 0.867, Max = 1.133
- borrow Borrowed from financial institution: N = 141188, Mean = 0.427, StDev = 0.495, Min = 0, Median = 0, Max = 1
- account Account at financial institution: N = 141188, Mean = 0.58, StDev = 0.493, Min = 0, Median = 1, Max = 1
- fertility_rate Fertility rate: N = 131575, Mean = 2.699, StDev = 1.481, Min = 1.22, Median = 2.11, Max = 57.656
- smam_f Female age of marriage, average: N = 131575, Mean = 23.632, StDev = 6.293, Min = 0, Median = 24.433, Max = 3.2
- elc_accs_rural_2012 Access to electricity, rural: N = 130577, Mean = 75.991, StDev = 35.774, Min = 1.299, Median = 9.755, Max = 100
- edu_elec Access to electricity, rural*education: N = 130577, Mean = 56.855, StDev = 46.392, Min = 0, Median = 90.51, Max = 100
- sigi_2014 Social discrimination against women: N = 89411, Mean = 0.18, StDev = 0.139, Min = 0.002, Median = 0.147, Max = 0.563
- sigi1 Discriminatory family code: N = 128571, Mean = 0.309, StDev = 0.257, Min = 0, Median = 0.262, Max = 1
- sigi3 Son bias: N = 106445, Mean = 0.229, StDev = 0.224, Min = 0, Median = 0.143, Max = 1
- sigi5 Restricted civil liberties: N = 129574, Mean = 0.384, StDev = 0.234, Min = 0, Median = 0.354, Max = 1
- C1822000 Non-discrimination clause covering gender: N = 92068, Mean = 0.556, StDev = 0.497, Min = 0, Median = 1, Max = 1

*Source: wp1838 - References (selected content and annexes as presented in the source PDF).*

### Annex V. Correlation Matrix

### Annex V. Correlation Matrix

### Table A3. Correlation Matrix — key correlations and significance
- Note on significance: "* p<0.05 ** p<0.01 *** p<0.001"
- Selected correlations involving "account":
  - account — Proper2014a1: 0.882***
  - account — female: -0.0598***
  - account — educ_s: 0.282***
  - account — age: 0.112***
  - account — age2: 0.0800***
  - account — q3: 0.305***
  - account — inc_top_20: 0.162***
  - account — pop_dens: 0.0825***
  - account — lgdppcppp: 0.362***
  - account — Fuel_Exporters: -0.0681***
  - account — meanyrsch: 0.345***
  - account — findepth13: 0.358***
  - account — ICRG_crr: 0.229***
  - account — ICRG_prr: 0.259***
  - account — dm_avg: 0.130***
  - account — C_avg: 0.253***
  - account — fertility_rate: -0.324***
  - account — smam_f: 0.283***
  - account — sigi_2014: -0.307***
  - account — sigi1: -0.342***
  - account — sigi3: -0.00746
  - account — sigi5: -0.280***
  - account — C1822000: -0.0283***
- Selected other notable correlations:
  - fertility_rate — lgdppcppp: -0.773***
  - fertility_rate — findepth13: -0.707***
  - fertility_rate — ICRG_crr: -0.695***
  - fertility_rate — smam_f: -0.575***
  - fertility_rate — sigi_2014: 0.663***
  - smam_f — lgdppcppp: 0.595***
  - lgdppcppp — Proper2014a1: 0.439***
  - meanyrsch — lgdppcppp: 0.802***
  - findepth13 — lgdppcppp: 0.671***
  - ICRG_prr — ICRG_crr: 0.747***
  - C_avg — ICRG_prr: 0.556***
  - sigi1 — sigi_2014: 0.825***
  - sigi5 — sigi1: 0.769***
- Full matrix of pairwise correlations with significance levels is reported in Table A3.

*Source: wp1838 - Annex V. Correlation Matrix*

---

### Annex VI. Robustness Checks

### Table A4. Determinants of the Financial Inclusion Index - Full Sample with Female Interactions
- Columns: (1) through (5) — regressions with female interaction terms
- Coefficients (standard errors in parentheses). Significance indicators preserved.

- Secondary education:
  - (1) 0.225*** (0.016)
  - (2) -0.011* (0.006)
  - (3) 0.111*** (0.009)
  - (4) -0.011* (0.006)
  - (5) 0.128*** (0.008)
  - (5, female) -0.013** (0.006)
  - (4, female) -0.012* (0.006)
  - (3, female) -0.013** (0.006)
  - (2, female) -0.011* (0.006)
  - (1, female) -0.012* (0.006)

- Age:
  - (1) 0.010*** (0.001)
  - (2) -0.001*** (0.000)
  - (3) 0.010*** (0.001)
  - (4) -0.003*** (0.001)
  - (5) 0.011*** (0.001)
  - Female interactions:
    - (1) -0.001*** (0.000)
    - (3) -0.004*** (0.001)
    - (4) -0.003*** (0.001)
    - (5) -0.003*** (0.001)

- Age squared:
  - (1) -0.000*** (0.000)
  - (2) 0.000** (0.000)
  - (3) -0.000*** (0.000)
  - (4) 0.000*** (0.000)
  - (5) -0.000*** (0.000)
  - Female interactions generally small but significant in some columns (see table).

- Wage employment:
  - (1) 0.187*** (0.010)
  - (2) 0.053*** (0.007)
  - (3) 0.122*** (0.007)
  - (4) 0.045*** (0.006)
  - (5) 0.117*** (0.007)
  - Female interactions around 0.042–0.053 with significance as reported.

- Top 20 percent income:
  - (1) 0.058*** (0.006)
  - (2) -0.014*** (0.005)
  - (3) 0.089*** (0.004)
  - (4) -0.011*** (0.004)
  - (5) 0.087*** (0.005)
  - Female interactions negative in several specifications (e.g., -0.009**, -0.008*, -0.009**).

- Population density:
  - Coefficients small and not consistently significant across columns; e.g., (1) 0.000 (0.001), (2) -0.000 (0.000).

- GDP p.c.:
  - (1) 0.125*** (0.007)
  - (2) 0.004*** (0.001)
  - (3) 0.082*** (0.019)
  - (4) -0.010 (0.008)
  - (5) 0.085*** (0.019)
  - Female interaction coefficients small/negative in some specifications (e.g., -0.007, -0.006).

- Fuel Exporter:
  - (1) -0.124*** (0.027)
  - (2) -0.016 (0.013)
  - (3) -0.053* (0.029)
  - (4) -0.030** (0.015)
  - (5) -0.034 (0.030)
  - Female interactions: around -0.034** to -0.036 (see table).

- Mean years of schooling:
  - Coefficients near zero; example: -0.001 (0.005) and interactions 0.005** (0.002) in some columns.

- Financial development:
  - (1) 0.240*** (0.042)
  - Female interactions negative in several columns: e.g., -0.043** (0.018), -0.038** (0.017) repeated.

- Institutions (ICRG):
  - Coefficients small and sometimes significant: e.g., 0.002 (0.001), 0.002** (0.001), 0.001* (0.001).

- Civil remedies for sexual harassment (when included):
  - Civil remedies for sexual harassment: 0.054*** (0.018) with female interaction -0.000 (0.007)
  - Civil remedies for sexual harassment in employment: 0.052*** (0.019) with female interaction -0.001 (0.007)

- Constant terms:
  - (1) -0.081*** (0.017)
  - (2) -1.065*** (0.062)
  - (3) -0.929*** (0.116)
  - (4) -0.942*** (0.107)
  - (5) -0.935*** (0.106)

- Adjusted R squared:
  - (1) 0.322
  - (2) 0.480
  - (3) 0.487
  - (4) 0.495
  - (5) 0.495

- Standard errors are reported in parentheses. Significance: ="* p<0.10   *** p<0.010"

*Source: wp1838 - Annex VI. Robustness Checks (Table A4)*

---

### Table A5. Probit Estimations: Determinants of Women’s Probability of Having an Account at a Financial Institution
- Columns: (1) through (6). Coefficients are marginal effects (standard errors in parentheses). Significance preserved.

- Individual characteristics
  - Secondary education:
    - (1) 0.838*** (0.076)
    - (2) 0.451*** (0.055)
    - (3) 0.483*** (0.038)
    - (4) 0.518*** (0.035)
    - (5) 0.499*** (0.038)
    - (6) 0.521*** (0.033)

  - Age:
    - (1) 0.031*** (0.005)
    - (2) 0.030*** (0.004)
    - (3) 0.029*** (0.004)
    - (4) 0.029*** (0.004)
    - (5) 0.033*** (0.004)
    - (6) 0.033*** (0.003)

  - Age squared:
    - (1) -0.000** (0.000)
    - (2) -0.000*** (0.000)
    - (3) -0.000*** (0.000)
    - (4) -0.000*** (0.000)
    - (5) -0.000*** (0.000)
    - (6) -0.000*** (0.000)

  - Wage employment:
    - (1) 1.008*** (0.051)
    - (2) 0.837*** (0.047)
    - (3) 0.818*** (0.054)
    - (4) 0.821*** (0.052)
    - (5) 0.813*** (0.051)
    - (6) 0.827*** (0.047)

  - Top 20 percent income:
    - (1) 0.163*** (0.028)
    - (2) 0.342*** (0.026)
    - (3) 0.345*** (0.027)
    - (4) 0.342*** (0.028)
    - (5) 0.351*** (0.026)
    - (6) 0.341*** (0.027)

- Country characteristics
  - Population density:
    - (1) 0.015* (0.008)
    - (2) 0.008 (0.005)
    - (3) 0.006 (0.005)
    - (4) -0.002 (0.004)
    - (5) 0.005 (0.006)

  - GDP p.c.:
    - (1) 0.566*** (0.047)
    - (2) 0.270** (0.130)
    - (3) 0.317** (0.128)
    - (4) 0.525*** (0.184)
    - (5) 0.514*** (0.141)

  - Fuel Exporter:
    - (1) -0.666*** (0.119)
    - (2) -0.342*** (0.118)
    - (3) -0.265** (0.120)
    - (4) -0.231* (0.131)
    - (5) -0.199 (0.124)

- Policies
  - Mean years of schooling:
    - (1) 0.042 (0.036)
    - (2) 0.024 (0.035)
    - (3) 0.019 (0.034)
    - (4) 0.05 (0.038)

  - Financial development:
    - (1) 1.218*** (0.266)
    - (2) 1.144*** (0.244)
    - (3) 1.155*** (0.282)
    - (4) 0.703** (0.347)

  - Institutions (ICRG):
    - (1) 0.012 (0.009)
    - (2) 0.008 (0.009)
    - (3) 0.004 (0.009)
    - (4) -0.005 (0.009)

- Laws
  - Civil remedies for sexual harassment:
    - (3) 0.337*** (0.114)
    - (4) 0.362*** (0.106)
    - (5) 0.348*** (0.097)

- Constants:
  - (1) -1.715*** (0.10)
  - (2) -6.286*** (0.43)
  - (3) -5.183*** (0.86)
  - (4) -5.377*** (0.81)
  - (5) -6.311*** (1.20)
  - (6) -6.246*** (0.90)

- Model controls:
  - Income dummies: included in column (5) only (Yes shown)
  - Regional dummies: included in column (6) only (Yes shown)
  - Standard errors in parentheses
  - Significance: * p<0.10;  ** p<0.05; *** p<0.010

*Source: wp1838 - Annex VI. Robustness Checks (Table A5)*

*wp1838 - Annex V & VI. Correlation Matrix and Robustness Checks (tables extracted from the supplied PDF content)*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp1838.pdf_