## _wp16118 - Women’s Opportunities and Challenges in Sub-Saharan African Job Markets

## Source details

**Canonical URL:** [_wp16118 - Women’s Opportunities and Challenges in Sub-Saharan African Job Markets](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16118.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16118.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16118.pdf.json)

---

### Abstract and scope
- Uses household survey data to analyze determinants of the gender gap in the labor market and welfare implications in five SSA countries using multinomial logit models with propensity score matching.
- Confirms education opens opportunities for women to escape agricultural feminization and engage in formal wage employment, but these opportunities diminish when women marry.
- Opening a household enterprise offers women an alternative to escape low-paid agricultural jobs, but the increase in per capita income is lower than for male-owned household enterprises.
- Policy implication highlighted: improving women’s education needs to be supported by measures to allow married women to keep their jobs in the wage sector.

### Sample, focus, and methods
- Country sample: Burkina Faso, Rwanda, Zambia, Ghana, and Mauritius.
- Focus: determinants of employment sector structure (agriculture, household enterprise, wage) and welfare effect measured by household consumption per capita in low and lower-middle income SSA countries.
- Methods:
  - Multinomial logit model for employment sector choice (baseline: agriculture).
  - Propensity score matching to estimate average treatment effects on consumption per capita.
  - Consumption per capita converted to US dollar using the average exchange rate within the survey year.

### Employment-structure findings (by income level and country)
- Low/low-middle income patterns:
  - Labor force participation of women and men is very similar in low and lower-middle income countries.
  - Gender inequality mainly appears in unequal access to wage employment as wage-sector share rises with development.
- Agriculture shares (verbatim):
  - Burkina Faso: roughly 80% of workers in agriculture: 82% of women and 76% of men.
  - Rwanda: 82% of the female and 62% of the male work in the agriculture sector.
- Household enterprise percentages are similar between genders in Burkina Faso and Rwanda.
- Wage employment is dominated by men: men have roughly two to three times higher probability than females to be employed in the wage sector.
- Transition and urbanization:
  - As agriculture shrinks and urbanizes, males gain far more wage employment opportunities than females.
  - Females tend to move out of agriculture into household enterprises (female informal employment).
  - Zambia: women continue to work primarily in agriculture and do not gain from increasing employment opportunities.
  - Ghana: both genders have similar opportunities in non-farm jobs, but females are more likely to be informally employed in household enterprises.

### Determinants — Education
- Education is a key, non-linear determinant of employment sector outcomes.
- The percentage in agriculture decreases with more advanced education across four country cases.
- Marginal effects by education level:
  - The marginal impact of education, in terms of moving the population out of agriculture, reaches its peak at secondary education — the gateway to the formal wage sector for both genders.
  - Females benefit more from secondary education: the marginal increase in wage employment is largest when females gain secondary education.
  - Primary and secondary education levels have significant effects on women's chances of opening household enterprises.

### Determinants — Marriage and urbanization
- Marriage:
  - Marriage cancels out part of the education effect for women: higher education increases wage access for single men and women, but married women face reduced access to wage employment.
  - The gender gap widens among married populations at most education levels; the highest gender gap exists among the uneducated/low educated married population.
  - Country patterns post-marriage:
    - Burkina Faso and Zambia: females move out of formal wage and household enterprise sectors into agriculture after marriage.
    - Rwanda: females move out of wage sector into both household enterprise and agriculture.
    - Ghana: females drop out of wage sector and become informally employed in household enterprises.
- Urbanization and migration:
  - Urbanization generally increases non-agricultural job opportunities but tends to deepen gender gaps.
  - Burkina Faso: urban population share of 29% in 2014; urban men’s wage employment is 10 percentage points higher than women’s.
  - Rwanda: urban population share of 28%; in urban Rwanda, 55% of males versus 21% of females employed in the wage sector.
  - Ghana: 53% urban population; females more likely in household enterprise sector while males gain more than twice as many wage jobs as in rural areas.
  - Zambia: urbanization rate of 40% with a pattern similar to Ghana.
  - Migration facilitates moving out of agriculture but benefits males more; by migrating males gain significant wage jobs while females’ wage opportunities remain roughly the same or slightly drop in some cases.
  - Note: Zambia’s household survey has no migration data.

### Welfare implications (propensity score matching results)
- General results:
  - Household enterprise employment and wage employment are not significantly different in terms of consumption per capita.
  - Agricultural employment leads to a significantly lower consumption level compared to household enterprise employment.
  - Women gain less than men by moving from agriculture to household enterprise, reflecting lower productivity of female household enterprise workers due to constraints in access to land, capital, and other inputs.
- Summary statement (verbatim):
  - "Overall, the household enterprise sector increases the consumption per capita per year by 60-270 US dollar, depending on the specific country context, with the smallest impact in Zambia."
  - "Both female and male household enterprise workers increase their corresponding household members’ average consumption levels, but 'male household enterprise workers consistently increase the family consumption level more than the female counterparts.'"
  - "Moving from household enterprise to wage employment: 'the household consumption level is not significantly different between these two sectors.'"
- Key tabulated welfare estimates (verbatim entries preserved):
  - Burkina Faso:
    - "288.42 1***"
    - Sub-estimates: "272.416***", "294.697***", "270.727***", "256.297***", "323.316***"
    - Additional displayed numbers: "41.046", "21.526", "73.263", "9.675", "78.885", "-21.278"
  - Rwanda:
    - "196.35 2***"
    - Sub-estimates: "143.819***", "230.448***", "172.062***", "127.745***", "209.562***"
    - Additional displayed numbers: "91.108***", "248.705***", "-41.683", "-75.929", "5.850", "-203.915***"
  - Zambia:
    - "0.022* **", "0.020***", "0.022**"
    - Additional entries: "-0.014**", "-0.005", "-0.017", "0.017**", "0.018", "0.014", "-0.010", "-0.009", "-0.005"
    - Standard errors shown as "[0.00]", "[0.01]", "[0.01]", "[0.01]" in many cells.
  - Ghana:
    - "92.338***", "95.19***", "112.280***", "67.845***", "57.060***", "101.120***"
    - Additional displayed numbers: "-9.907", "-10.8723", "-60.683", "-27.422", "-48.848", "10.395"
    - Standard errors shown as "[13.63]", "[15.33]", "[23.58]", "[13.98]", "[15.73]", "[22.51]" and other bracketed values.
- Model specification note:
  - "Model 1 does not include the household size, household income and the household income dummy to avoid the concern about endogeneity, while model 2 does incorporate these three variables to control for the household income effect."

### Regression evidence — Multinomial logit key coefficients (verbatim excerpts)
- Estimation notes:
  - Estimations based on sub-sample of all employed individuals, using population weight; standard errors clustered at household level.
  - Significance notation: "* p<0.10, ** p<0.05, *** p<0.01".
- Selected country coefficients (relative to agriculture; entries preserved exactly):
  - Burkina Faso (N: "5277"):
    - Primary Education: "0.305**" (HH Enterprise), "1.227***" (Wage)
    - Secondary Education: "0.322*" (HH Enterprise), "2.565***" (Wage)
    - Higher Education: "15.262***" (HH Enterprise), "18.209***" (Wage)
    - Urban: "3.316***" (HH Enterprise), "2.784***" (Wage)
    - Female: "0.736***" (HH Enterprise), "1.320***" (Wage)
    - Female * Higher Education: "-14.076***" (HH Enterprise), "-13.556***" (Wage)
  - Rwanda (N: "28886"):
    - Primary Education: "0.742***" (HH Enterprise), "0.344***" (Wage)
    - Secondary Education: "1.719***" (HH Enterprise), "1.691***" (Wage)
    - Higher Education: "2.845***" (HH Enterprise), "3.367***" (Wage)
    - Urban: "1.389***" (HH Enterprise), "1.668***" (Wage)
    - Female: "-0.442***" (HH Enterprise), "-1.240***" (Wage)
    - Female * Higher Education: "-0.672" (HH Enterprise), "1.195**" (Wage)
  - Ghana (N: "13824"):
    - Primary Education: "0.465***" (HH Enterprise), "0.741***" (Wage)
    - Secondary Education: "0.835***" (HH Enterprise), "1.612***" (Wage)
    - Higher Education: "2.020***" (HH Enterprise), "3.609***" (Wage)
    - Urban: "2.230***" (HH Enterprise), "1.984***" (Wage)
    - Female: "1.229***" (HH Enterprise), "-0.189" (Wage)
    - Female * Primary Education: "0.236*" (HH Enterprise), "0.556***" (Wage)
  - Zambia (N: "30526"):
    - Primary Education: "0.301*" (HH Enterprise), "0.289" (Wage)
    - Secondary Education: "0.492***" (HH Enterprise), "1.144***" (Wage)
    - Higher Education: "2.144***" (HH Enterprise), "4.384***" (Wage)
    - Urban: "3.560***" (HH Enterprise), "3.566***" (Wage)
    - Female: "-0.15" (HH Enterprise), "-0.521*" (Wage)
    - Female * Higher Education: "0.277" (HH Enterprise), "1.467***" (Wage)
- Interpretive highlights:
  - Education and urban residence show strong positive associations with wage employment.
  - Female interactions with education and marital status show heterogeneous and sometimes large negative coefficients (e.g., Female * Higher Education in Burkina Faso).

### Mauritius (upper-middle income case) — distinct patterns
- Contrasting features:
  - "The gender gap has quite different features in upper middle income countries like Mauritius than in low and lower-middle income countries analyzed in previous chapters."
  - Higher income enables families to reduce number employed in household, leading to lower female labor force participation.
- Key quantitative/descriptive findings:
  - "Low female labor force participation rate at below 50 percent."
  - For employed population only: "women are more employed in the wage sector than men."
  - For whole working age population 15-64: "More than half of the working age females are outside the labor force and female unemployment rate is very high."
  - Education:
    - Post-secondary education is critical for women to stay in the labor force.
    - Men and women have similar access to primary and secondary education, but "females are left behind in the post-secondary education and have a higher share among the uneducated."
    - Marginal effects: for men primary education has the highest marginal impact on labor force participation; for women the marginal effect peaks at post-secondary education.
  - Marital status and household responsibility:
    - "Females massively drop out of the labor force after they get married. Men, instead, substitute the female in the labor market."
    - "Zero percent of the male claim [household responsibility] is due to the household responsibility, while 63 percent of the female respond that this is the major reason for not working."
    - "More than 50 percent of respondents reply that lack of funding explains why they do not start their own business."
- Figures referenced (titles preserved): Figure 7 through Figure 12 (figure contents not reproduced here).

### Labor market participation, unemployment, and job-search (Mauritius and cross-country observations)
- Reasons not to work, by gender:
  - High female unemployment indicates willingness to work but inability to find a job.
  - Females try harder than males across job-search effort measures.
  - Time between current and last job is much longer for females than males.
  - When quitting last jobs, household responsibility is the major reason females report; this is negligible for males.
- Effort among unemployed (Figure 13 indicators):
  - Females exhibit higher effort on: time in job search (year); Register Employment Service use; length of registration; availability to work immediately.
- Main reason for leaving last job (Figure 14):
  - For females, household responsibility is a major reported reason; for males it is negligible.

### Policy implications (preserved recommendations and rationale)
- Education:
  - Improve education quality and secondary education coverage among females.
  - Education is the major contributor to move workers out of agriculture into household enterprise and wage sectors and is critical for productivity in agriculture and household enterprises.
- Reduce household-responsibility burdens and gender discrimination in social institutions and norms:
  - Changes in social and cultural factors needed to alter women's role.
  - Government actions: take into account needs of women in charge of family duties when allocating public investment; develop child care capacity.
  - Expected outcomes:
    - Allow females to stay in the wage sector after marriage, preserving the positive impact of education on wage employment.
    - Improve women’s productivity in household enterprises and agriculture by facilitating access to land, capital, and other inputs.
- Support for household enterprises:
  - Informal household enterprises are a major source of female employment in rural and urban areas; raising their productivity is essential to improve welfare for women and families.
  - Policy tools: simple and fair tax regimes for small enterprises; develop alternative funding sources like property taxes to reduce fees and levies on household enterprises.
- Migration and urbanization measures:
  - Spatial differences suggest immigration can help individuals find better employment opportunities.
  - To help women benefit from urbanization-related opportunities: protect independent female immigrants’ safety and provide job training tailored to wage employment.

*Source: WP/16/118, “Women’s Opportunities and Challenges in Sub-Saharan African Job Markets,” Prepared by Christine Dieterich, Anni Huang, Alun Thomas (Sections 1–4).*

### Section 1

### _wp16118 - Section 1

### Abstract and scope
- Uses household survey data to analyze determinants of the gender gap in the labor market and welfare implications in five SSA countries using multinomial logit models with propensity score matching.
- Confirms education opens opportunities for women to escape agricultural feminization and engage in formal wage employment, but these opportunities diminish when women marry.
- Opening a household enterprise offers women an alternative to escape low-paid agricultural jobs, but the increase in per capita income is lower than for male-owned household enterprises.
- Policy implication highlighted: improving women’s education needs to be supported by measures to allow married women to keep their jobs in the wage sector.

### Sample and research focus
- Country sample: Burkina Faso, Rwanda, Zambia, Ghana, and Mauritius.
- Focus for low and lower-middle income countries: determinants of employment sector structure (agriculture, household enterprise, wage) and welfare effect measured by household consumption per capita.
- Methods: multinomial logit model for employment sector choice; propensity score matching for welfare (average treatment effect) analysis.

### Key empirical findings — employment structure by income level
- Low/low-middle income pattern:
  - Labor force participation of women and men is very similar in low and lower-middle income countries.
  - Gender inequality mainly appears in unequal access to wage employment as wage-sector share rises with development.
- Agriculture feminization and sector shares:
  - In Burkina Faso, roughly 80% of workers are employed in the agriculture sector: 82% of women and 76% of men.
  - In Rwanda, 82% of the female and 62% of the male work in the agriculture sector.
  - Household enterprise percentages are similar between genders in Burkina Faso and Rwanda.
  - Wage employment is dominated by men, with men having roughly two to three times higher probability than females to be employed in the wage sector.
- Lower-middle income transition:
  - With shrinking agriculture and urbanization, males gain far more opportunities in wage employment than females.
  - Females tend to move out of agriculture into household enterprises (female informal employment).
  - In Zambia, women continue to work primarily in agriculture and do not gain from increasing employment opportunities.
  - In Ghana, both genders have similar opportunities in non-farm jobs, but females are more likely to be informally employed in household enterprises.

### Determinants of employment sector outcomes — education
- Education is a key determinant of employment sector outcomes; impact is non-linear.
- The percentage in agriculture decreases with more advanced education in all four country cases.
- Marginal effects by education level:
  - The marginal impact of education, in terms of moving the population out of agriculture, reaches its peak at secondary education which is the gateway to the formal wage sector for both genders.
  - Females benefit more from secondary education as the marginal increase in wage employment is largest when females gain secondary education.
  - Primary and secondary education levels have significant effects on women's chances of opening household enterprises.

### Determinants — marriage and urbanization (summary from Section I and III)
- Marriage cancels out part of the education effect for women: while higher education increases access to wage employment for single men and women, married women face reduced access to wage employment.
- As countries urbanize and the wage sector expands, this marriage-related barrier causes a deepening gender gap in labor markets.
- Urbanization is generally Pareto-improving for both genders, but females do not gain their fair share of new employment opportunities in the transition process.

### Welfare implications (propensity score matching results summary)
- Welfare effect measured as consumption per capita:
  - Household enterprise employment and wage employment are not significantly different in terms of consumption per capita.
  - Agricultural employment leads to a significantly lower consumption level compared to household enterprise employment.
  - Women gain less than men by moving from the agriculture sector to the household enterprise sector, likely reflecting lower productivity of female household enterprise workers due to constraints in access to land, capital, and other inputs.

### Contributions relative to existing literature
- Focuses on employment sector structure rather than labor force participation and unemployment rates — argued to be more relevant for low and lower-middle income SSA countries.
- First paper to model gender-split employment sector structure (per authors’ claim) and to add marital status and urbanization effects to employment sector outcomes.
- Uses propensity score matching to quantify monetary welfare differences between employment sectors and to check robustness of Fox and Sohnesen (2012) findings.

*Source: WP/16/118, “Women’s Opportunities and Challenges in Sub-Saharan African Job Markets,” Prepared by Christine Dieterich, Anni Huang, Alun Thomas (Section 1).*

### Section 2

### _wp16118 - Section 2

### The literature
- Better education increases productivity in the agricultural sector and reduces food insecurity and extreme poverty among females who have more constraints in access to other production inputs.
- Quisumbing, A. R., & Pandolfelli, L. (2010) provide an extensive literature review summarizing promising approaches to address the needs of poor female farmers.

### B. Women’s Marriage Cancels out Education Effect
- Marital status (single or married) is used as a proxy for female household responsibility because:
  - Change in marital status implies a significant jump in female domestic responsibility.
  - Household size and number of children can have non-linear or endogenous implications for household responsibility and are harder to compare cross-country.
- Single vs. married wage employment patterns (four-country comparison) are diverse:
  - Burkina Faso: single women have better access to wage employment than single men at the same education level, except at higher education where percentages are almost equal.
  - Rwanda and Zambia: single men show a higher percentage of wage employment than single women at the same education, except at higher education where women have a higher percentage.
  - Ghana: split pattern — single men have higher wage access at lower education levels; single women at higher education levels.
- For married populations, the gender gap widens at most education levels:
  - The highest gender gap exists among the uneducated/low educated married population.
  - Married females with no education or primary education have much less opportunity than married males in the wage sector.
  - The negative bias against married women is strong enough to produce a large gender gap even among the married with secondary education (except Burkina Faso).
  - Conclusion: the return to female secondary education in the labor market is often cancelled out by marriage due to unequal division of family responsibility.
- Post-marriage sector shifts (country-specific):
  - Burkina Faso and Zambia: females move out of formal wage sector and household enterprise sector into agriculture after marriage.
  - Rwanda: females move out of wage sector into both household enterprise and agriculture.
  - Ghana: females drop out of wage sector and become informally employed in household enterprises.
- Welfare implication from Section D (summary):
  - Household enterprise employment is associated with higher consumption per capita within the household compared to agriculture.
  - Therefore, in Burkina Faso and Zambia, married females moving into agriculture may suffer larger welfare losses than in Rwanda and Ghana where some move into household enterprises.

### C. Urbanization and Job Opportunity
- Urbanization generally provides more non-agricultural job opportunities but tends to deepen gender gaps in employment structure:
  - Rural areas: gender disparity in employment sector is narrower than in urban areas because few wage jobs exist in rural areas.
  - Burkina Faso: share of urban population of 29% in 2014 (World Bank Database); rural employment sector distribution similar between males and females; in urban areas, men’s wage employment is 10 percentage points higher than women’s.
  - Rwanda: share of urban population of 28% (very close to Burkina Faso); in urban Rwanda, 55% of males versus 21% of females are employed in the wage sector.
  - Ghana: urbanization more advanced with 53% of the population living in urban areas; females are much more likely to work in the household enterprise sector while males have more than twice as many wage jobs as in rural areas.
  - Zambia: similar pattern to Ghana despite a lower urbanization rate of 40%.
- Migration and dynamic effects:
  - Migration facilitates moving out of agriculture, but males benefit more.
  - By migrating, males gain significant amounts of wage jobs while females’ opportunities in the wage sector remain roughly the same or slightly drop in some cases.
  - This pattern suggests migration decisions often prioritize better job prospects for males, sometimes at the cost of female career plans.
- Note: Zambia’s household survey has no migration data.

### D. Regression Analysis and Welfare Implication
- Methodology:
  - Multinomial logit model used to study determinants of employment sector structure (agriculture, household enterprise, wage employment) because outcomes are discrete without clear ordering.
  - Baseline case in the multinomial logit model is the agriculture sector; coefficients are interpreted relative to agriculture.
  - Controls: education levels, marital status, urbanization factors, and interactions with gender.
  - Baseline for “urban” variable is “rural”; baseline for “immigrant” variable is “stayer”.
  - Interaction terms with gender: coefficients for factors alone reflect impacts on males; female impacts require combining main effect and interaction term.
  - Coefficient of gender captures gender-specific characteristics not explained by controls.
- Key regression findings:
  - Education strongly defines employment:
    - Burkina Faso (relative to no education baseline): obtaining primary education increases probability to get a wage job by 7.7%; obtaining secondary education increases probability by 28.1%; obtaining higher education increases probability by 62.5%.
    - Rwanda, Ghana, and Zambia: steady and comparable increases in chance of wage employment when moving up the education ladder.
    - Impact of education is larger among females, indicating females rely on education to narrow a wider gender gap among uneducated/less educated populations.
  - Education also improves chances to be employed in the household enterprise sector, but parameter estimates are smaller on average.
  - No observed gender difference in the impact of education on household enterprise employment.
  - Marriage effects:
    - Marriage does not have significant implication for males’ employment sector structure.
    - Marriage massively decreases chances for females to move out of agriculture across all four countries.
    - The marginal effect of marriage on females’ chances to work in the wage sector appears larger in countries with lower urbanization levels.
    - Impact of marriage on females’ household enterprise employment is indecisive; household enterprise offers time flexibility enabling women to stay in or re-enter after marriage, but limited flexibility may also reduce productivity of female-owned household enterprises.
  - Migration effects:
    - Mixed results: in Rwanda, female migrants increase chances to obtain wage employment more than male counterparts; in Ghana the reverse holds.
    - Differences may reflect social standards or motives behind migration decisions (female-driven vs. husband-driven migration).
- Welfare analysis and propensity score matching:
  - Building on Fox and Sohnesen (2012) who find household enterprise employment generates roughly the same level of consumption per capita as wage employment (OLS), this study uses propensity score matching with average treatment effect estimation to provide additional insights.
  - Advantages of propensity score matching:
    - More accurately pairs individuals with similar observed characteristics who choose different employment sectors.
    - Puts more weight on comparable individuals who choose different sectors.
    - Allows estimation of dollar value impact of employment sector outcome on consumption per capita.
  - Conversion: consumption per capita converted from domestic currency to US dollar using the average exchange rate within the survey year.
  - Results (Table 3 summary):
    - Moving from agriculture to household enterprise employment significantly increases consumption per capita in the worker’s household (reported as average treatment effect of moving from one sector to another on consumption per year per capita).
    - Moving from household enterprise to wage employment has no significant impact on consumption per capita.

*Source: _wp16118 - Section 2*

### Section 3

### _wp16118 - Section 3

### Welfare effects of moving between employment sectors
- Summary statement from analysis:
  - "Overall, the household enterprise sector increases the consumption per capita per year by 60-270 US dollar, depending on the specific country context, with the smallest impact in Zambia."
  - Both female and male household enterprise workers increase their corresponding household members’ average consumption levels, but "male household enterprise workers consistently increase the family consumption level more than the female counterparts." The text notes this pattern suggests lower female productivity in the household enterprise sector, potentially due to lack of production inputs like land, credit and other factors.
  - Moving from household enterprise to wage employment: "the household consumption level is not significantly different between these two sectors."

- Key tabulated welfare estimates (as shown in Table 3; verbatim entries preserved):
  - Burkina Faso:
    - "288.42 1***"
    - Sub-estimates: "272.416***", "294.697***", "270.727***", "256.297***", "323.316***"
    - Additional displayed numbers: "41.046", "21.526", "73.263", "9.675", "78.885", "-21.278"
  - Rwanda:
    - "196.35 2***"
    - Sub-estimates: "143.819***", "230.448***", "172.062***", "127.745***", "209.562***"
    - Additional displayed numbers: "91.108***", "248.705***", "-41.683", "-75.929", "5.850", "-203.915***"
  - Zambia:
    - "0.022* **", "0.020***", "0.022**"
    - Additional entries: "-0.014**", "-0.005", "-0.017", "0.017**", "0.018", "0.014", "-0.010", "-0.009", "-0.005"
    - Standard errors columns show "[0.00]", "[0.01]", "[0.01]", "[0.01]" consistently for many cells.
  - Ghana:
    - "92.338***", "95.19***", "112.280***", "67.845***", "57.060***", "101.120***"
    - Additional displayed numbers: "-9.907", "-10.8723", "-60.683", "-27.422", "-48.848", "10.395"
    - Standard errors as shown: "[13.63]", "[15.33]", "[23.58]", "[13.98]", "[15.73]", "[22.51]" and other bracketed values.
- Note on models:
  - "Model 1 does not include the household size, household income and the household income dummy to avoid the concern about endogeneity, while model 2 does incorporate these three variables to control for the household income effect."

### Determinants of employment sectors (Multinomial Logit results from Tables 1 and 2)
- Estimation and inference:
  - "The estimations are based on the sub-sample of all the employed individuals, using population weight in the survey design. Standard errors are clustered at the household level to control for the correlation within the household."
  - Significance notation used: "* p<0.10, ** p<0.05, *** p<0.01"
- Selected coefficient patterns preserved exactly (country columns retained as presented):
  - Burkina Faso (Table 1):
    - Primary Education: "0.305**" (HH Enterprise) and "1.227***" (Wage Employment)
    - Secondary Education: "0.322*" (HH Enterprise) and "2.565***" (Wage Employment)
    - Higher Education: "15.262***" (HH Enterprise) and "18.209***" (Wage Employment)
    - Urban: "3.316***" (HH Enterprise) and "2.784***" (Wage Employment)
    - Female: "0.736***" (HH Enterprise) and "1.320***" (Wage Employment)
    - Female * Higher Education: "-14.076***" (HH Enterprise) and "-13.556***" (Wage Employment)
    - N: "5277"
  - Rwanda (Table 1):
    - Primary Education: "0.742***" (HH Enterprise) and "0.344***" (Wage Employment)
    - Secondary Education: "1.719***" (HH Enterprise) and "1.691***" (Wage Employment)
    - Higher Education: "2.845***" (HH Enterprise) and "3.367***" (Wage Employment)
    - Urban: "1.389***" (HH Enterprise) and "1.668***" (Wage Employment)
    - Female: "-0.442***" (HH Enterprise) and "-1.240***" (Wage Employment)
    - Female * Higher Education: "-0.672" (HH Enterprise) and "1.195**" (Wage Employment)
    - N: "28886"
  - Ghana (Table 2):
    - Primary Education: "0.465***" (HH Enterprise) and "0.741***" (Wage Employment)
    - Secondary Education: "0.835***" (HH Enterprise) and "1.612***" (Wage Employment)
    - Higher Education: "2.020***" (HH Enterprise) and "3.609***" (Wage Employment)
    - Urban: "2.230***" (HH Enterprise) and "1.984***" (Wage Employment)
    - Female: "1.229***" (HH Enterprise) and "-0.189" (Wage Employment)
    - Female * Primary Education: "0.236*" (HH Enterprise) and "0.556***" (Wage Employment)
    - N: "13824"
  - Zambia (Table 2):
    - Primary Education: "0.301*" (HH Enterprise) and "0.289" (Wage Employment)
    - Secondary Education: "0.492***" (HH Enterprise) and "1.144***" (Wage Employment)
    - Higher Education: "2.144***" (HH Enterprise) and "4.384***" (Wage Employment)
    - Urban: "3.560***" (HH Enterprise) and "3.566***" (Wage Employment)
    - Female: "-0.15" (HH Enterprise) and "-0.521*" (Wage Employment)
    - Female * Higher Education: "0.277" (HH Enterprise) and "1.467***" (Wage Employment)
    - N: "30526"
- Interpretive highlights preserved from text:
  - Education and urban residence show strong positive associations with wage employment in many country samples.
  - Female interactions with education and marital status show heterogeneous effects across countries and sectors, including large negative Female * Higher Education coefficients in some low-income country estimations (e.g., Burkina Faso).

### Upper-middle income country case: Mauritius — low female labor participation and high female unemployment
- Overview statements preserved:
  - "The gender gap has quite different features in upper middle income countries like Mauritius than in low and lower-middle income countries analyzed in previous chapters."
  - "Higher income provides families the luxury to reduce the number of employed in the household, leading to lower female labor force participation."
- Key quantitative and descriptive findings:
  - "Low female labor force participation rate at below 50 percent requires widening the analysis to understand the underlying reasons."
  - For employed population only, "women are more employed in the wage sector than men." (Figure 7 Left Panel)
  - For the whole working age population 15-64, "More than half of the working age females are outside the labor force and female unemployment rate is very high." (Figure 7 Right Panel)
  - Education:
    - "Equal access to primary and secondary education in Mauritius has little impact on labor force participation, but post-secondary education is the critical condition for women to stay in the labor force."
    - Men and women have similar access to primary and secondary education, but "females are left behind in the post-secondary education and have a higher share among the uneducated."
    - Marginal effects: for men primary education has the highest marginal impact on labor force participation; for women the marginal effect peaks at post-secondary education.
  - Marital status and household responsibility:
    - "Females massively drop out of the labor force after they get married. Men, instead, substitute the female in the labor market."
    - "The unemployment rates of married males are much lower than any other group."
    - Household responsibility as reason for non-participation: "zero percent of the male claim it is due to the household responsibility, while 63 percent of the female respond that this is the major reason for not working."
    - Lack of funding explanation for not starting a business: "More than 50 percent of respondents reply that lack of funding explains why they do not start their own business."
- Figures referenced in text (verbatim figure titles preserved):
  - "Figure 7. Mauritius: Labor Market Structure, by Gender"
  - "Figure 8. Mauritius: Access to Education among the Working Age Population, by Gender"
  - "Figure 9. Mauritius: Employment Sectors and Education, Male"
  - "Figure 10. Mauritius: Employment Sectors and Education, Female"
  - "Figure 11. Mauritius: Economic Activity and Marital Status, by Gender"
  - "Figure 12." (figure referenced but content not included in supplied text)

*Source: _wp16118 - Section 3*

### Section 4

### _wp16118 - Section 4

### Reasons Not to Work, by Gender
- High female unemployment indicates that females are willing to work but could not find a job.
- Females try harder than males to get a job in almost all dimensions of the effort measures among the unemployed.
- The time length between current and last job is much longer for the female than the male.
- When asked why they quit their last jobs, other than job nature (temporary job), household responsibility is the major reason why females quit the last job; this is irrelevant in the answers from the male respondents.

### Effort in Job Searching Among the Unemployed (Figure 13) — key observations
- Females exhibit higher effort than males across multiple measures:
  - time in job search (year)
  - Register Employment Service (percentage answered "yes")
  - Length of Registeration in Employment Service (year)
  - Available to work immediately (percentage answered "yes")

### Mauritius: The Main Reason Why Leave Last Job, by Gender (Figure 14) — key observation
- For females, household responsibility is a major reported reason for leaving the last job; for males this reason is negligible.

### Policy Implications — summary of key messages for policy makers
- Improve education quality and secondary education coverage among the female.
  - Education is the major contributor to move workers out of the agricultural sector into household enterprise sector and wage sector.
  - Education is critical in explaining productivity in agriculture and household enterprise sectors.
- Reduce the burden from household responsibilities and remove gender discrimination in social institutions and social norms.
  - This requires changes in social and cultural factors that determine the role of women.
  - Governments can ease the burden through practical steps like taking into account the needs of women in charge of family duties when allocating resources for public investment, and by developing child care capacity.
  - Progress would:
    - allow females to stay in the wage sector after marriage, instead of massively dropping out of the wage sector, eliminating the positive impact of education on wage employment.
    - improve women’s productivity in household enterprises as well as agriculture because a lower burden from household duties and less discrimination through social norms would also facilitate access to land, capital, and other inputs.
- With informal employment in household enterprises being a major source of female employment in rural and, even more, in urban areas for years to come, improving their productivity will be essential to further improve welfare for women and their families.
  - Female owned household enterprises continue to be somewhat less productive compared to men owned.
  - The welfare analysis did not indicate a significant difference between the welfare implications of wage and household enterprise employment, measured as consumption per capita.
  - Women are often employed in lower-paid wage jobs than men.
  - Fiscal policy should be considerate of the major role these household enterprises play in diversifying employment and income beyond the agricultural sector:
    - include simple and fair tax regimes for these small enterprises.
    - developing alternative sources of funding like property taxes would reduce the appetite for fees and levies on household enterprises.
- Spatial differences in employment sector structure suggest that immigration might help individuals to find better employment opportunities.
  - As migration improves prospects to move out of agriculture, measures to protect independent female immigrants’ safety and provide job training tailored to wage employment would help women benefit from the job opportunities associated with urbanization.

*Source: _wp16118 - Section 4*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16118.pdf_