## _wp16140

## Source details

**Canonical URL:** [_wp16140](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16140.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16140.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16140.pdf.json)

---

### I. Motivation and stylized facts
- Recent decline in commodity prices increases pressure on resource-intensive countries to diversify output and export bases.
- Oil prices fell to "less than 30 USD/barrel in early 2016" and remain "significantly lower than their peak 2013 levels."
- Commodity prices are expected to remain at only a fraction of their high levels in the medium term.
- Diversification and structural transformation are associated with economic growth, particularly at early stages of development.
- Empirical patterns (1962–2010 and 1970–2010 illustrations):
  - "Export Product Diversification and Output Growth, 1962-2010" (higher diversification values = less diversification).
  - "Export Diversification and Output Volatility, 1962-2010" (higher diversification values = less diversification, volatility = standard deviation over 1962-2010).
- Structural characteristics associated with diversification: level of development, institutional quality, stronger infrastructure, higher degree of globalization, and education.
- Human capital accumulation:
  - Primary and secondary education enable imitation of frontier technology.
  - Tertiary education increases possibility of innovating.

### Main hypotheses linking gender equality and diversification
- Human capital channel:
  - "Gender gaps in opportunity, such as in education, harm diversification directly by constraining the potential pool of human capital."
  - Where girls’ education lags, female human capital cannot accumulate optimally, slowing technology adoption and innovation.
- Resource allocation channel:
  - "Gender gaps in labor force participation shrink the pool of talent from which employers can hire and limit the number of female entrepreneurs."
  - This reduces the ability to create and execute ideas, impeding diversification.

### Data and definitions (Box 1)
- Export product diversification:
  - Uses "Theil index of export diversification from IMF (2014b)" following Cadot and others (2011).
  - Index decomposable into "between" (extensive margin, number of products) and "within" (intensive margin, product shares) sub-indices.
  - "Lower values of the Theil index indicate higher levels of export product diversification."
  - Index available for "188 countries from 1962-2010."
- Output diversification:
  - Constructed for real subsectors from the UN’s sectoral database in IMF (2014b), following the export Theil methodology.
  - "The index covers 188 countries from 1970-2010."
- Gender Inequality Index (GII):
  - Extended version of the United Nations Gender Inequality Index.
  - Captures inequality across health (maternal mortality ratios and adolescent fertility rates), empowerment (share of parliamentary seats and education attainment at the secondary level for both males and females), and labor force participation (rates by sex).
  - "The index spans values between 0 and 1, with higher values indicating higher gender inequality."
  - "The index is available for 141 countries from 1990-2013."

### Empirical strategy and identification
- Baseline period: 1990-2010.
- Panel specification includes country fixed effects and time fixed effects.
- Main dependent variables: export diversification Theil index and output diversification Theil index.
- Main regressor: extended United Nations Gender Inequality Index (GII).
- Additional individual gender measures: female-to-male gross enrollment ratio in secondary school, female labor force participation rate, share of female seats in parliament, adolescent fertility rate, risk of maternal death.
- Controls include: real GDP per capita and its square, population size, human capital index, resource dependence (share of mining in GDP or share of fuel exports), institutional indices (Fraser Institute Summary Index, legal systems and property rights), macro/cyclical variables (terms-of-trade, real effective exchange rates, real GDP growth), trade openness, financial development, investment, infrastructure.
- Endogeneity addressed with IV-GMM using gender-based legal restrictions as instruments (Women, Business and the Law measures).

### Key results — Export diversification (selected findings and coefficients preserved)
- Main finding: gender inequality is strongly and negatively associated with export diversification in low-income and developing countries, controlling for structural characteristics, policies, and cyclical factors.
- Magnitude: moving from absolute gender inequality to perfect gender equality measured by the index could decrease the Theil index of export diversification by 0.6 to 2 units in low-income and developing countries (equivalent to up to about two standard deviations across LIDC).
- Confirmed structural relationships:
  - U-shaped relationship between export diversification and development.
  - Higher share of mining in output associated with less diversified export base.
  - Population size and human capital associated with higher export diversification in many specifications.
- Selected coefficient estimates (all specifications include country and time fixed effects; standard errors in parentheses):
  - Gender Inequality Index (examples): 0.703** (0.273); 0.752*** (0.278); 0.776*** (0.277); 1.156*** (0.319); 0.665** (0.264).
  - Interaction -- in LIDC: 1.014** (0.431); 0.983** (0.438); 1.113** (0.435).
  - Log(Population): -0.707*** (0.133); -0.560*** (0.135); -0.568*** (0.136).
  - Log(Real GDP per capita): -1.838*** (0.294); -2.371*** (0.289); -1.712*** (0.308).
  - Log(Real GDP per capita) -- squared: 0.114*** (0.0174); 0.140*** (0.0172); 0.103*** (0.0182).
  - Mining as share of GDP: 0.00937** (0.00396); 0.00694* (0.00398); 0.0119*** (0.00416).
  - Policy/institutional (selected): Fraser Institute Sum. Index -0.116*** (0.0137); Legal Systems & Property Rights -0.0358*** (0.0102).
  - Openness (selected): Freedom to trade -0.0646*** (0.00858); Globalization Index -0.0123*** (0.00268).
  - Infrastructure (selected): Log(landlines/1000 workers) -0.129*** (0.0177); Length of road network -0.0300** (0.0144).
  - Macro: Terms of Trade 0.00313*** (0.000347); Log(REER) 0.186*** (0.0519).
- Sample example: Observations 1,841; Countries 100; R-squared 0.181 (in column 1).

### Key results — Output diversification (selected findings and coefficients preserved)
- Gender inequality is negatively associated with output diversification in low-income and developing countries in all specifications.
- Mixed results outside low-income/developing group: in some regressions gender inequality is significantly and positively associated with output diversification, possibly reflecting increased female participation in services where production re-concentrates as countries develop.
- Selected coefficient estimates (all specifications include country and time fixed effects; standard errors in parentheses):
  - Gender Inequality Index (examples): -0.0552* (0.0310); -0.0344 (0.0315); -0.0867** (0.0425); -0.1000*** (0.0369).
  - Interaction -- in LIDC: 0.188*** (0.0488); 0.203*** (0.0495); 0.212*** (0.0709).
  - Log(Population): -0.0376** (0.0150); -0.0318** (0.0154).
  - Log(Real GDP per capita): -0.215*** (0.0336); -0.238*** (0.0353); -0.225*** (0.0431).
  - Log(Real GDP per capita) -- squared: 0.0103*** (0.00199); 0.0112*** (0.00209).
  - Institutions: Fraser Institute Sum. Index -0.00961*** (0.00155).
  - Openness/Trade: Freedom to trade -0.00224** (0.000976); Average Tariff Rates 0.0290*** (0.0108).
  - Infrastructure/Investment: Length of road network -0.00464*** (0.00153); Log(Landlines/1000 workers) -0.00716** (0.00193); Investment per worker -3.79e-06*** (7.98e-07).
  - Financial development: Financial reform index -0.0760*** (0.0126).
- Sample example: Observations 1,880; Countries 102; R-squared 0.165 (in column 1).

### Identified channels (evidence)
- Human capital channel:
  - Higher female-to-male enrollment ratio is significantly and positively related to export diversification, particularly in low-income and developing countries.
- Resource allocation / labor participation channel:
  - Higher female labor force participation rates associated with higher export diversification levels in low-income and developing economies.
- Health channel:
  - Better female health outcomes (lower maternal mortality ratios and adolescent fertility rates) are positively associated with export diversification.
- Similar channel evidence for output diversification:
  - Higher female labor force participation and higher female-to-male educational enrollment ratios in low-income and developing countries are associated with higher output diversification when controlling for policies and institutions.

### Instrumental-variable strategy and diagnostics (causality evidence)
- Instruments: gender-based legal rights (e.g., right to be the head of a household, full community marital property rights).
- First-stage and diagnostics:
  - Instruments are individually significant in first-stage regressions.
  - Instrument F-statistics are "well above the rule-of-thumb threshold value of 10" in many specifications.
  - In specifications with two or more instruments, p-values of the Hansen J-statistic do not allow rejection of instrument exogeneity.
- Table 5 (IV-GMM selected results, significance notation preserved):
  - Export Diversification (columns (1)(2)):
    - GII Index: 5.785*** 3.534**
    - Log(Population): -0.976*** -0.252
    - Lag Human capital index: 0.0251 0.420***
    - Log(GDP per capita): -1.307*** -0.666*
    - - squared: 0.0931*** 0.0360*
    - Mining as share of GDP: 0.0318*** 0.0105
    - Observations: 1,552 1,204
    - P-value of Hansen J statistic: 0.296 0.248
    - Instrument F-test: 13.271 2.85
  - Output Diversification (columns (1)(2)):
    - GII Index: 1.778*** 0.153***
    - Log(Population): -0.0830** -0.134***
    - Lag Human capital index: 0.131*** -0.00844
    - Log(GDP per capita): -0.390*** -0.222***
    - - squared: 0.0230*** 0.0141***
    - Observations: 1,554 833
    - P-value of Hansen J statistic: 0.548 0.276
    - Instrument F-test: 16.283 3.44
  - Note: Lesotho and Mauritania are dropped from the output diversification estimation due to insufficient observations.
- Implication: IV diagnostics support instrument validity and suggest gender inequality may be a cause of lower economic diversification.

### Policy-relevant implications and conclusions
- Enhancing gender equality in opportunities (education, legal rights) and outcomes (labor force participation, health) can foster greater economic diversification.
- Legal reforms that remove restrictions on women’s economic rights are a valid policy lever to improve resource allocation and diversification, supported by IV evidence using legal-rights instruments.
- This paper presents the first empirical evidence, to the authors' knowledge, that gender inequality impacts both export and output diversification.
- Both gender equity in opportunities and in outcomes matter for economic diversification.
- Linking gender inequality to lower economic diversification highlights a new channel through which gender equality boosts growth.

### Annex I — Selected summary statistics (exact values preserved)
- Export Diversification Theil (IMF Diversification Toolkit)
  - Full Sample: Obs 6378; Mean 3.5; Std. Dev. 1.2; Min 1.0; Max 6.4
  - LIDC: Obs 2159; Mean 4.2; Std. Dev. 0.9; Min 1.8; Max 6.4
- Output Diversification Theil (IMF Diversification Toolkit)
  - Full Sample: Obs 7065; Mean 0.3; Std. Dev. 0.2; Min 0.0; Max 1.7
  - LIDC: Obs 2259; Mean 0.3; Std. Dev. 0.2; Min 0.0; Max 1.6
- Log(GDP per capita) (WEO)
  - Full Sample: Obs 6141; Mean 8.5; Std. Dev. 1.2; Min 5.2; Max 11.7
  - LIDC: Obs 1910; Mean 7.2; Std. Dev. 0.5; Min 5.2; Max 8.8
- GII Index (IMF GDI GII database)
  - Full Sample: Obs 2580; Mean 0.5; Std. Dev. 0.2; Min 0.0; Max 0.8
  - LIDC: Obs 774; Mean 0.6; Std. Dev. 0.1; Min 0.3; Max 0.8
- Female labor force participation rate (WDI)
  - Full Sample: Obs 3591; Mean 0.5; Std. Dev. 0.2; Min 0.1; Max 0.9
  - LIDC: Obs 1197; Mean 0.6; Std. Dev. 0.2; Min 0.1; Max 0.9
- Secondary enrollment ratio (WDI)
  - Full Sample: Obs 4371; Mean 0.9; Std. Dev. 0.3; Min 0.0; Max 3.1
  - LIDC: Obs 1230; Mean 0.7; Std. Dev. 0.3; Min 0.0; Max 2.1
- Maternal mortality ratio (WDI)
  - Full Sample: Obs 3591; Mean 272.0; Std. Dev. 374.7; Min 3.0; Max 2900
  - LIDC: Obs 1218; Mean 623.4; Std. Dev. 422.8; Min 29.0; Max 2900.0
- Adolescent fertility rate (WDI)
  - Full Sample: Obs 3696; Mean 65.0; Std. Dev. 49.3; Min 3.1; Max 229
  - LIDC: Obs 1218; Mean 106.6; Std. Dev. 48.3; Min 18.0; Max 222.4
- Fraser Institute Summary Index
  - Full Sample: Obs 3655; Mean 5.9; Std. Dev. 1.4; Min 2.0; Max 9.2
  - LIDC: Obs 1100; Mean 5.1; Std. Dev. 1.1; Min 2.0; Max 7.5
- Length of road network (Calderon-Serven database)
  - Full Sample: Obs 3755; Mean -1.2; Std. Dev. 1.4; Min -5.2; Max 1.6
  - LIDC: Obs 1043; Mean -2.0; Std. Dev. 1.4; Min -5.2; Max 0.0
- Terms of Trade (WEO)
  - Full Sample: Obs 4334; Mean 109.7; Std. Dev. 48.7; Min 5.5; Max 602.9
  - LIDC: Obs 1477; Mean 124.5; Std. Dev. 69.5; Min 5.5; Max 602.9
- Real GDP per capita growth rate (World Economic Outlook)
  - Full Sample: Obs 5981; Mean 0.0; Std. Dev. 0.1; Min -1.1; Max 1.0
  - LIDC: Obs 1861; Mean 0.0; Std. Dev. 0.1; Min -0.7; Max 0.7

*Source: IMF Working Paper excerpt (content unit: _wp16140).*

### REFERENCES.............................................................................................22

### _wp16140 - REFERENCES.............................................................................................22

### I. Motivation
- Recent decline in commodity prices increases pressure on resource-intensive countries to diversify output and export bases.
- Oil prices fell to "less than 30 USD/barrel in early 2016" and remain "significantly lower than their peak 2013 levels."
- Commodity prices are expected to remain at only a fraction of their high levels in the medium term.
- Diversification and structural transformation are associated with economic growth, particularly at early stages of development (references to IMF 2014a; Papageorgiou and Spatafora 2012).
- Empirical patterns (Figures 1 and 2):
  - "Export Product Diversification and Output Growth, 1962-2010" (higher diversification values = less diversification).
  - "Export Diversification and Output Volatility, 1962-2010" (higher diversification values = less diversification, volatility = standard deviation over 1962-2010).
- Key structural characteristics associated with diversification include: level of development, institutional quality, stronger infrastructure, higher degree of globalization, and education (IMF 2014b; IMF 2013).
- Human capital accumulation:
  - Primary and secondary education enable imitation of frontier technology.
  - Tertiary education increases possibility of innovating (Aghion and Howitt 2006).

### Main hypotheses linking gender equality and diversification
- Human capital channel:
  - "Gender gaps in opportunity, such as in education, harm diversification directly by constraining the potential pool of human capital."
  - Where girls’ education lags, female human capital cannot accumulate optimally, slowing technology adoption and innovation.
- Resource allocation channel:
  - "Gender gaps in labor force participation shrink the pool of talent from which employers can hire and limit the number of female entrepreneurs."
  - This reduces the ability to create and execute ideas, impeding diversification.

### Stylized data relationships
- Figures illustrate correlations for 1990-2010:
  - "Export Diversification and Gender Inequality, 1990-2010."
  - "Output Diversification and Gender Inequality, 1990-2010."
- Observed patterns:
  - Higher values of the re-estimated UN Gender Inequality Index are associated with higher export diversification index values (which indicate lower diversification).
  - Similar negative relationship for output diversification, particularly in low-income and developing countries.

### Empirical contributions (summary of paper's findings)
- Three-fold contribution:
  - Evidence that "gender inequality is negatively associated with both output and export diversification in low-income and developing economies," beyond standard drivers of diversification.
  - Both "inequality of opportunities" and "lower female labor force participation" are associated with lower economic diversification, supporting the human capital and resource allocation channels.
  - Provides evidence on causality and addresses endogeneity concerns in regressions (see Section II and Section III for methods).

### Box 1 — Definitions used in the analysis
- Export product diversification:
  - Uses "Theil index of export diversification from IMF (2014b)" following Cadot and others (2011).
  - Index decomposable into "between" (extensive margin, number of products) and "within" (intensive margin, product shares) sub-indices.
  - "Lower values of the Theil index indicate higher levels of export product diversification."
  - Index available for "188 countries from 1962-2010."
- Output diversification:
  - Constructed for real subsectors from the UN’s sectoral database in IMF (2014b), following the export Theil methodology.
  - "The index covers 188 countries from 1970-2010."
- Gender Inequality Index (GII):
  - Extended version of the United Nations Gender Inequality Index (Gonzales and others 2015b; Stotsky and others 2016).
  - Captures inequality across health (maternal mortality ratios and adolescent fertility rates), empowerment (share of parliamentary seats and education attainment at the secondary level for both males and females), and labor force participation (rates by sex).
  - "The index spans values between 0 and 1, with higher values indicating higher gender inequality."
  - "The index is available for 141 countries from 1990-2013."
- Mathematical notation:
  - The document includes the Theil index decomposition formulas exactly as presented in Box 1.

### II. Literature review — key points
- Diversification, development, and growth:
  - Greater diversification associated with higher growth, reduced volatility, and increased resilience (Koren and Tenreyro 2007; Cadot and others 2011).
  - Singer (1950) showed initial diversification positively correlated with economic growth.
  - IMF (2014a) finds extensive and intensive diversification and output diversification drive growth in low-income countries.
  - Diversification shifts resources away from high-volatility sectors (mining, agriculture) toward lower-volatility sectors (manufacturing).
  - Non-linear relationship between diversification and development (Imbs and Wacziarg 2003): countries diversify up to a critical point, then specialize.
- Gender inequality and growth:
  - Negative link between real GDP per capita growth and gender inequality (Elborgh-Woytek and others 2013).
  - Education channel:
    - Studies confirm negative effect of gender inequality in education on growth (Hill and King 1995; Engelbrecht 1997; Forbes 1998; Dollar and Gatti 1999; Klasen 1999; Knowles and others 2002; Klasen and Lamanna 2009; Seguino 2010).
    - Berge and Wood (1994) support that educated female labor force determines manufacturing exports growth.
    - Amin and others (2015) find strong negative impact on growth in poor countries using broader measures.
    - Hypothesis: negative effects on growth partly operate by obstructing diversification and structural transformation.
  - Occupation and entrepreneurship:
    - Occupational choice models assume same distribution of talent between men and women (Cuberes and Teignier 2012; Esteve-Volart 2004).
    - Gender gaps in entrepreneurship distort allocation of talent and reduce aggregate productivity.
    - Cuberes and Teignier (2014a) model includes self-employment restrictions and lower wages for female workers.
    - Cavalcanti and Tavares (2016): "an increase of 50 percent in the gender wage gap could lead to a decrease in income per capita by 35 percent."
  - Aggregate measures:
    - Extended GII evidence shows dimensions of gender inequality strongly associated with lower growth, particularly in low-income countries (Gonzales and others 2015b; Hakura and others 2016).
  - Gender wage inequality results are mixed by context (Seguino 2010; Schober and Winter-Ebmer 2011).
- Data constraints:
  - Lack of extensive, reliable data on wage inequality in low-income and developing countries leads this paper to focus on GII subcomponents: reproductive health, empowerment, and labor market participation.

* _wp16140 - REFERENCES.............................................................................................22

### 11.      Structural transformation has been shown to coincide with episodes of decreases

### _wp16140 - 11.      Structural transformation has been shown to coincide with episodes of decreases

### Structural transformation and women’s economic participation
- Several studies find structural transformation, particularly growth of the service sector, coincides with decreases in gender inequality (Akbulut 2011; Olivetti and Petrongolo 2014; Ngai and Petrongolo 2014; Rendall 2013).
- Rendall (2013): structural transformation reduces labor demand for physical (“brawn”) attributes, favoring women’s comparative advantage in less physical (“brain”) attributes.
- Empirical example: In Mauritius, development of the textile industry coincided with an increase in female labor force participation of nearly 60 percent between 1983 and 1999 (Svirydzenka and Petri 2014).
- Cavalcanti and Tavares (2007): increases in female labor force participation linked to increases in government expenditures and demand for government-provided services; public sector typically employs more women.
- Prior literature emphasizes causation from structural transformation → women’s participation. This paper investigates the reverse: whether greater gender equality can enhance and support structural transformation (diversification).

### Gender-based legal restrictions as instruments
- Gender-based legal restrictions significantly impact women’s economic participation: access to finance, employment, labor force participation, asset ownership, property rights, and technology adoption (Demirgüç-Kunt and others 2013; Amin and Islam 2014; Gonzales and others 2015b; Deere and others 2013; Razavi 2003; Quisumbing and Pandofelli 2010).
- World Bank Women, Business and the Law data show restrictions on inheritance, property, banking access, and professional freedom significantly exacerbate gender gaps in labor force participation (World Bank 2013; World Bank 2015; Gonzales and others 2015a).
- The paper uses gender-based legal restrictions as instruments to address endogeneity: legal restrictions exacerbate gender inequality, which in turn impedes output and export diversification.

### Empirical strategy (baseline and identification)
- Baseline period: 1990-2010.
- Panel specification includes country fixed effects and time fixed effects to control for unobservable variables across countries and common time effects.
- Main dependent variables: measures of export or output diversification (as defined in Box 1).
- Main regressor: extended United Nations Gender Inequality Index (combines gaps in labor force participation, education, reproductive health, and female seats in parliament).
- Additional individual gender measures used in robustness: female-to-male gross enrollment ratio in secondary school, female labor force participation rate, share of female seats in parliament, adolescent fertility rate, risk of maternal death.
- Controls included:
  - Development: real GDP per capita and its square (to capture U-shaped relationship and turning point), population size, human capital index.
  - Resource dependence: share of mining in GDP or share of fuel exports.
  - Institutions/regulatory environment: Fraser Institute Summary Index, legal systems and property rights.
  - Macroeconomic/cyclical: terms-of-trade, real effective exchange rates, real GDP growth.
  - Policy dimensions: openness to trade (globalization index, freedom to trade, average tariffs), financial development (financial reform index, interest rate controls, private sector credit-to-GDP), investment (investment percent of GDP and per worker), infrastructure (density of landlines, length of road network).
- Endogeneity addressed with IV-GMM:
  - Instrumental variables: legal rights for women (gender-based legal restrictions).
  - Validity arguments: (i) legal restrictions on the books do not exert a direct impact on export/output diversification (exogeneity; Hansen test confirms), (ii) legal restrictions are strongly correlated with gender inequality (relevance; confirmed in results).

### Key results — Export diversification
- Main finding: gender inequality is strongly and negatively associated with export diversification in low-income and developing countries, controlling for structural characteristics, policies, and cyclical factors.
- Magnitude statement: moving from absolute gender inequality to perfect gender equality measured by the index could decrease the Theil index of export diversification by 0.6 to 2 units in low-income and developing countries; this magnitude is equivalent to up to about two standard deviations of the index across low-income and developing countries.
- Gender inequality effect also significant across all levels of development in some specifications.
- Confirmed structural relationships:
  - U-shaped relationship between export diversification and development (countries diversify then re-concentrate).
  - Higher share of mining in output associated with less diversified export base.
  - Population size (in most specifications) and human capital (in some specifications) associated with higher export diversification.
- Policy and institutional controls:
  - Better institutions (Fraser Summary Index; legal systems and property rights) significantly and positively associated with higher export diversification.
  - Greater openness to international trade and better infrastructure significantly and positively associated with export diversification.
- Macroeconomic factors:
  - Real exchange rate appreciation and terms-of-trade improvement associated with lower degrees of export diversification.
- Selected coefficient estimates from Table 1 (all specifications include country and time fixed effects; standard errors in parentheses):
  - Gender Inequality Index coefficients (examples across columns): 0.703** (0.273); 0.752*** (0.278); 0.776*** (0.277); 1.156*** (0.319); 0.665** (0.264).
  - Interaction -- in LIDC: 1.014** (0.431); 0.983** (0.438); 1.113** (0.435) in some columns.
  - Log(Population): -0.707*** (0.133); -0.560*** (0.135); -0.568*** (0.136).
  - Log(Real GDP per capita): -1.838*** (0.294); -2.371*** (0.289); -1.712*** (0.308).
  - Log(Real GDP per capita) -- squared: 0.114*** (0.0174); 0.140*** (0.0172); 0.103*** (0.0182).
  - Mining as share of GDP: 0.00937** (0.00396); 0.00694* (0.00398); 0.0119*** (0.00416).
  - Policy/institutional coefficients (selected): Fraser Institute Sum. Index -0.116*** (0.0137) in one specification; Legal Systems & Property Rights -0.0358*** (0.0102) in another.
  - Openness measures (selected): Freedom to trade -0.0646*** (0.00858); Globalization Index -0.0123*** (0.00268).
  - Infrastructure (selected): Log(landlines/1000 workers) -0.129*** (0.0177); Length of road network -0.0300** (0.0144).
  - Macro: Terms of Trade 0.00313*** (0.000347); Log(REER) 0.186*** (0.0519).
- Sample sizes and model fit (examples): Observations 1,841; Countries 100; R-squared 0.181 (in column 1).

### Key results — Output diversification
- Gender inequality is negatively associated with output diversification in low-income and developing countries in all specifications.
- Mixed results for non-low-income/developing countries: in some regressions gender inequality is significantly and positively associated with output diversification, possibly because lower gender inequality increases female participation in the service sector where production re-concentrates as countries develop.
- Selected coefficient estimates from Table 2 (all specifications include country and time fixed effects; standard errors in parentheses):
  - Gender Inequality Index coefficients (examples): -0.0552* (0.0310); -0.0344 (0.0315); -0.0867** (0.0425); -0.1000*** (0.0369).
  - Interaction -- in LIDC: 0.188*** (0.0488); 0.203*** (0.0495); 0.212*** (0.0709).
  - Log(Population): -0.0376** (0.0150); -0.0318** (0.0154).
  - Log(Real GDP per capita): -0.215*** (0.0336); -0.238*** (0.0353); -0.225*** (0.0431).
  - Log(Real GDP per capita) -- squared: 0.0103*** (0.00199); 0.0112*** (0.00209).
  - Mining as share of GDP: mixed coefficients (examples): 0.000214 (0.000449); -0.00216** (0.000583).
  - Institutions: Fraser Institute Sum. Index -0.00961*** (0.00155).
  - Openness/Trade: Freedom to trade -0.00224** (0.000976); Average Tariff Rates 0.0290*** (0.0108).
  - Infrastructure/Investment: Length of road network -0.00464*** (0.00153); Log(Landlines/1000 workers) -0.00716** (0.00193); Investment per worker -3.79e-06*** (7.98e-07).
  - Financial development: Financial reform index -0.0760*** (0.0126) in one specification.
- Sample sizes and model fit (examples): Observations 1,880; Countries 102; R-squared 0.165 (in column 1).

### Identified channels through which gender inequality inhibits diversification
- Human capital channel:
  - Higher female-to-male enrollment ratio is significantly and positively related to export diversification, particularly in low-income and developing countries (Table 3 results summarized).
- Resource allocation / labor participation channel:
  - Higher female labor force participation rates associated with higher export diversification levels in low-income and developing economies.
- Health channel:
  - Better female health outcomes (lower maternal mortality ratios and adolescent fertility rates) are positively associated with export diversification.
- Similar channel evidence for output diversification:
  - Higher female labor force participation and higher female-to-male educational enrollment ratios in low-income and developing countries are associated with higher output diversification when controlling for policies and institutions (Table 4 results summarized).

*Italic: IMF Working Paper excerpt (content unit: _wp16140 - 11.      Structural transformation has been shown to coincide with episodes of decreases).*

### 20.      Finally, we also find evidence for causality in the specifications by instrumenting

### 20.      Finally, we also find evidence for causality in the specifications by instrumenting

### Instrumental-variable strategy and diagnostics
- Gender inequality is instrumented with legal rights for women, such as the right to be the head of a household or full community marital property rights, used as instruments for gender inequality in GMM regressions.
- The instruments:
  - Are individually significant in the first-stage regressions.
  - Produce F-statistics of the IV regressions that are "well above the rule-of-thumb threshold value of 10."
  - In specifications with two or more instruments, the p-values of the Hansen J-statistic do not allow rejection of the joint null hypothesis that the instruments are uncorrelated with the error term.
- Implication: These diagnostics support the validity and exclusion of the legal-rights instruments and suggest that gender inequality may be a cause of lower economic diversification.

### Key empirical findings from Tables (selected results preserved exactly)
- Table 3 (Explaining Export Diversification – Dimensions of Gender Inequality)
  - Female labor force participation rate coefficients (columns as presented): 0.473 0.970** 0.758 1.762*** 0.995** 0.859* 1.562*** 1.478*** -0.0324
  - Maternal mortality ratio coefficients: 0.00142** 0.00156** 0.00151** 0.00154* 0.00130* 0.00104 0.00145** 0.00152** 0.00169***
  - Mining as share of GDP coefficients: 0.0114** 0.008740 0.0151** 0.0122** 0.0142** 0.0151*** 0.0143*** 0.0191*** 0.0390***
  - Log(Real GDP per capita) coefficients: -2.059*** -2.261*** -2.051*** 1.137* -1.698*** -1.626*** 0.248 -0.766 -0.848
  - Observations (columns): 1,033 1,034 1,032 954 989 989 1,083 1,084 927
  - Countries reported (across specs): 96 97 96 101 86 86 96 96 88
  - R-squared examples: 0.203 0.162 0.194 0.133 0.174 0.175 0.149 0.168 0.354

- Table 4 (Explaining Output Diversification – Dimensions of Gender Inequality)
  - Secondary enrollment ratio coefficients: 0.124*** 0.133*** 0.0947*** 0.119*** 0.110*** 0.117*** 0.0922*** 0.0751** 0.0313
  - Maternal mortality ratio coefficients: 0.000162** 0.000171*** 6.44e-05 6.27e-05 0.000203*** 0.000174** 0.000117 5.40e-05 1.97e-07
  - Adolescent fertility rate coefficients: 0.000925*** 0.00101*** 0.000931** 0.000769*** 0.000327 0.000428 0.000535* 0.000401 0.000758
  - Log(GDP per capita) coefficients: -0.0755 -0.0619 -0.196** -0.217*** -0.124** -0.115* -0.240*** 0.0433 -0.243
  - Mining as share of GDP examples: 0.000103 -0.000482 -0.00373*** -9.53e-05 0.00241*** 0.00252*** 0.000476 -0.00552*** -0.00855***
  - Observations (columns): 1,063 1,062 681 987 1,014 1,014 942 552 485
  - Countries reported (across specs): 98 98 95 103 87 87 104 73 65
  - R-squared examples: 0.245 0.229 0.231 0.330 0.259 0.258 0.294 0.276 0.341

- Table 5 (Explaining Diversification – Instrumental Variable GMM)
  - Export Diversification (columns (1)(2)):
    - GII Index: 5.785*** 3.534**
    - Log(Population): -0.976*** -0.252
    - Lag Human capital index: 0.0251 0.420***
    - Log(GDP per capita): -1.307*** -0.666*
    - - squared: 0.0931*** 0.0360*
    - Mining as share of GDP: 0.0318*** 0.0105
    - Observations: 1,552 1,204
    - P-value of Hansen J statistic: 0.296 0.248
    - Instrument F-test: 13.271 2.85
  - Output Diversification (columns (1)(2)):
    - GII Index: 1.778*** 0.153***
    - Log(Population): -0.0830** -0.134***
    - Lag Human capital index: 0.131*** -0.00844
    - Log(GDP per capita): -0.390*** -0.222***
    - - squared: 0.0230*** 0.0141***
    - Observations: 1,554 833
    - P-value of Hansen J statistic: 0.548 0.276
    - Instrument F-test: 16.283 3.44
  - Note: Lesotho and Mauritania are dropped from the output diversification estimation due to insufficient observations.
  - Significance notation preserved: *** p<0.01, ** p<0.05, * p<0.1

### Interpretation and mechanisms
- Gender inequality negatively impacts both export and output diversification in low-income and developing countries, based on:
  - A multi-dimensional GII Index.
  - Individual indicators: female labor force participation, education (secondary enrollment), maternal mortality, adolescent fertility, representation (women in parliament).
- Mechanisms highlighted:
  - Inequality constrains the level of human capital, limiting diversification (evidence: negative association between gender inequalities in education and diversification).
  - Lower female labor force participation implies inefficient allocation of resources, leading to suboptimal creation of ideas and sector development.
  - De jure legal restrictions (e.g., restrictions to the right to be the head of a household) skew resource allocation by impeding women’s economic participation and preventing equal opportunities for daughters and sons.

### Policy-relevant implications
- Enhancing gender equality in opportunities (education, legal rights) and outcomes (labor force participation, health) can foster greater economic diversification.
- Legal reforms that remove restrictions on women’s economic rights are a valid policy lever to improve resource allocation and diversification, supported by IV evidence using legal-rights instruments.

### Conclusion (section V. CONCLUSIONS, preserved text points)
- This paper presents the first empirical evidence, to the authors' knowledge, that gender inequality impacts both export and output diversification.
- Both gender equity in opportunities and in outcomes matter for economic diversification.
- The empirical estimation strategy using country-specific de jure laws and regulations as instruments provides support for causality from gender inequality to lower diversification.
- Linking gender inequality to lower economic diversification highlights a new channel through which gender equality boosts growth.

*Source: _wp16140 - 20. Finally, we also find evidence for causality in the specifications by instrumenting*

### Annex I: Summary Statistics

### Annex I: Summary Statistics

### Key variables and summary statistics
- Export Diversification Theil (IMF Diversification Toolkit)
  - Full Sample: Obs 6378; Mean 3.5; Std. Dev. 1.2; Min 1.0; Max 6.4
  - LIDC: Obs 2159; Mean 4.2; Std. Dev. 0.9; Min 1.8; Max 6.4
- Output Diversification Theil (IMF Diversification Toolkit)
  - Full Sample: Obs 7065; Mean 0.3; Std. Dev. 0.2; Min 0.0; Max 1.7
  - LIDC: Obs 2259; Mean 0.3; Std. Dev. 0.2; Min 0.0; Max 1.6
- Log(GDP per capita) (WEO)
  - Full Sample: Obs 6141; Mean 8.5; Std. Dev. 1.2; Min 5.2; Max 11.7
  - LIDC: Obs 1910; Mean 7.2; Std. Dev. 0.5; Min 5.2; Max 8.8
- Log(Population) (PWT 8.1)
  - Full Sample: Obs 6141; Mean 1.7; Std. Dev. 1.9; Min -3.2; Max 7.2
  - LIDC: Obs 1910; Mean 1.8; Std. Dev. 1.4; Min -2.6; Max 5.1
- Human capital index (5-year lag) (PWT 8.1/ Barro Lee)
  - Full Sample: Obs 4385; Mean 2.1; Std. Dev. 0.6; Min 1.0; Max 3.6
  - LIDC: Obs 1289; Mean 1.6; Std. Dev. 0.4; Min 1.0; Max 2.9
- Mining as share of GDP (IMF Jobs and Income Surveillance toolkit)
  - Full Sample: Obs 4831; Mean 21.0; Std. Dev. 11.6; Min 0.8; Max 85.6
  - LIDC: Obs 1865; Mean 17.7; Std. Dev. 11.6; Min 0.8; Max 75.9
- GII Index (IMF GDI GII database)
  - Full Sample: Obs 2580; Mean 0.5; Std. Dev. 0.2; Min 0.0; Max 0.8
  - LIDC: Obs 774; Mean 0.6; Std. Dev. 0.1; Min 0.3; Max 0.8
- Ratio of female tertiary teachers (WDI)
  - Full Sample: Obs 2105; Mean 0.3; Std. Dev. 0.1; Min 0.0; Max 0.8
  - LIDC: Obs 521; Mean 0.2; Std. Dev. 0.1; Min 0.0; Max 0.8
- Unmarried women; equal property rights (Women, Business, and the Law)
  - Full Sample: Obs 3707; Mean 0.9; Std. Dev. 0.3; Min 0.0; Max 1.0
  - LIDC: Obs 1470; Mean 0.9; Std. Dev. 0.3; Min 0.0; Max 1.0
- Married women; equal property rights (Women, Business, and the Law)
  - Full Sample: Obs 3688; Mean 0.8; Std. Dev. 0.4; Min 0.0; Max 1.0
  - LIDC: Obs 1431; Mean 0.7; Std. Dev. 0.5; Min 0.0; Max 1.0
- Married women; head household (Women, Business, and the Law)
  - Full Sample: Obs 3723; Mean 0.6; Std. Dev. 0.5; Min 0.0; Max 1.0
  - LIDC: Obs 1466; Mean 0.5; Std. Dev. 0.5; Min 0.0; Max 1.0
- Married women; legal proceedings (Women, Business, and the Law)
  - Full Sample: Obs 3763; Mean 0.9; Std. Dev. 0.3; Min 0.0; Max 1.0
  - LIDC: Obs 1506; Mean 0.8; Std. Dev. 0.4; Min 0.0; Max 1.0
- Married women; bank account (Women, Business, and the Law)
  - Full Sample: Obs 3742; Mean 0.9; Std. Dev. 0.3; Min 0.0; Max 1.0
  - LIDC: Obs 1490; Mean 0.8; Std. Dev. 0.4; Min 0.0; Max 1.0
- Equal inheritance, sons and daughters (Women, Business, and the Law)
  - Full Sample: Obs 3688; Mean 0.7; Std. Dev. 0.5; Min 0.0; Max 1.0
  - LIDC: Obs 1431; Mean 0.6; Std. Dev. 0.5; Min 0.0; Max 1.0
- Joint titling of property (Women, Business, and the Law)
  - Full Sample: Obs 3582; Mean 0.4; Std. Dev. 0.5; Min 0.0; Max 1.0
  - LIDC: Obs 1354; Mean 0.4; Std. Dev. 0.5; Min 0.0; Max 1.0
- Full community marital property regime (Women, Business, and the Law)
  - Full Sample: Obs 3589; Mean 0.1; Std. Dev. 0.2; Min 0.0; Max 1.0
  - LIDC: Obs 1351; Mean 0.0; Std. Dev. 0.2; Min 0.0; Max 1.0
- Partial community marital property regime (Women, Business, and the Law)
  - Full Sample: Obs 3589; Mean 0.4; Std. Dev. 0.5; Min 0.0; Max 1.0
  - LIDC: Obs 1351; Mean 0.3; Std. Dev. 0.5; Min 0.0; Max 1.0
- Separate property marital property regime (Women, Business, and the Law)
  - Full Sample: Obs 3589; Mean 0.4; Std. Dev. 0.5; Min 0.0; Max 1.0
  - LIDC: Obs 1351; Mean 0.5; Std. Dev. 0.5; Min 0.0; Max 1.0
- Guaranteed equity (Women, Business, and the Law)
  - Full Sample: Obs 3734; Mean 0.9; Std. Dev. 0.3; Min 0.0; Max 1.0
  - LIDC: Obs 1501; Mean 0.9; Std. Dev. 0.3; Min 0.0; Max 1.0
- Nondiscrimination clause (Women, Business, and the Law)
  - Full Sample: Obs 3734; Mean 0.4; Std. Dev. 0.5; Min 0.0; Max 1.0
  - LIDC: Obs 1501; Mean 0.4; Std. Dev. 0.5; Min 0.0; Max 1.0
- Valid customary law (Women, Business, and the Law)
  - Full Sample: Obs 3734; Mean 0.3; Std. Dev. 0.5; Min 0.0; Max 1.0
  - LIDC: Obs 1501; Mean 0.5; Std. Dev. 0.5; Min 0.0; Max 1.0

(continued)
- Female labor force participation rate (WDI)
  - Full Sample: Obs 3591; Mean 0.5; Std. Dev. 0.2; Min 0.1; Max 0.9
  - LIDC: Obs 1197; Mean 0.6; Std. Dev. 0.2; Min 0.1; Max 0.9
- Secondary enrollment ratio (WDI)
  - Full Sample: Obs 4371; Mean 0.9; Std. Dev. 0.3; Min 0.0; Max 3.1
  - LIDC: Obs 1230; Mean 0.7; Std. Dev. 0.3; Min 0.0; Max 2.1
- Women in parliament (WDI)
  - Full Sample: Obs 2425; Mean 14.1; Std. Dev. 9.9; Min 0.0; Max 56.3
  - LIDC: Obs 753; Mean 12.0; Std. Dev. 9.1; Min 0.0; Max 56.3
- Maternal mortality ratio (WDI)
  - Full Sample: Obs 3591; Mean 272.0; Std. Dev. 374.7; Min 3.0; Max 2900
  - LIDC: Obs 1218; Mean 623.4; Std. Dev. 422.8; Min 29.0; Max 2900.0
- Adolescent fertility rate (WDI)
  - Full Sample: Obs 3696; Mean 65.0; Std. Dev. 49.3; Min 3.1; Max 229
  - LIDC: Obs 1218; Mean 106.6; Std. Dev. 48.3; Min 18.0; Max 222.4
- Fraser Institute Summary Index (Fraser Institute)
  - Full Sample: Obs 3655; Mean 5.9; Std. Dev. 1.4; Min 2.0; Max 9.2
  - LIDC: Obs 1100; Mean 5.1; Std. Dev. 1.1; Min 2.0; Max 7.5
- Legal system and property rights (Fraser Institute)
  - Full Sample: Obs 3509; Mean 5.3; Std. Dev. 1.9; Min 1.1; Max 9.6
  - LIDC: Obs 989; Mean 4.0; Std. Dev. 1.1; Min 1.6; Max 6.8
- Freedom to trade (Fraser Institute)
  - Full Sample: Obs 3820; Mean 5.8; Std. Dev. 2.4; Min 0.0; Max 10.0
  - LIDC: Obs 1215; Mean 4.3; Std. Dev. 2.1; Min 0.0; Max 8.8
- Globalization index (KOF Index of Globalization)
  - Full Sample: Obs 4451; Mean 46.3; Std. Dev. 19.2; Min 9.6; Max 92.9
  - LIDC: Obs 1728; Mean 31.4; Std. Dev. 10.2; Min 9.6; Max 63.1
- Length of road network (Calderon-Serven database)
  - Full Sample: Obs 3755; Mean -1.2; Std. Dev. 1.4; Min -5.2; Max 1.6
  - LIDC: Obs 1043; Mean -2.0; Std. Dev. 1.4; Min -5.2; Max 0.0
- Log(Landlines per 1000 workers) (Calderon-Serven database)
  - Full Sample: Obs 3765; Mean 3.7; Std. Dev. 2.0; Min -0.6; Max 7.2
  - LIDC: Obs 1043; Mean 1.8; Std. Dev. 1.1; Min -0.6; Max 5.2
- Terms of Trade (WEO)
  - Full Sample: Obs 4334; Mean 109.7; Std. Dev. 48.7; Min 5.5; Max 602.9
  - LIDC: Obs 1477; Mean 124.5; Std. Dev. 69.5; Min 5.5; Max 602.9
- Log(REER) (IFS)
  - Full Sample: Obs 3350; Mean 4.7; Std. Dev. 0.7; Min 0.7; Max 15.3
  - LIDC: Obs 1171; Mean 4.9; Std. Dev. 1.0; Min 0.7; Max 15.3
- Average Tariff Rates (Trade Index)
  - Full Sample: Obs 3194; Mean 0.7; Std. Dev. 0.2; Min 0.0; Max 1.0
  - LIDC: Obs 999; Mean 0.7; Std. Dev. 0.2; Min 0.0; Max 1.0
- Investment per worker (PWT)
  - Full Sample: Source PWT; Obs 401245895362-832460861500 538555-8325208
- Financial reform index (IMF Index of Financial reform)
  - Full Sample: Obs 2527; Mean 0.5; Std. Dev. 0.3; Min 0.0; Max 1.0
  - LIDC: Obs 558; Mean 0.3; Std. Dev. 0.2; Min 0.0; Max 0.9
- Gini index (WDI)
  - Full Sample: Obs 1035; Mean 40.7; Std. Dev. 10.3; Min 16.2; Max 99.9
  - LIDC: Obs 233; Mean 43.0; Std. Dev. 9.1; Min 25.9; Max 69.5
- Income ratio (top 20%/bottom 20%) (WDI)
  - Full Sample: Obs 1034; Mean 10.5; Std. Dev. 11.6; Min 2.2; Max 278.2
  - LIDC: Obs 233; Mean 12.6; Std. Dev. 20.0; Min 3.7; Max 278.2
- Agriculture, value added (% of GDP) (WDI)
  - Full Sample: Obs 5099; Mean 19.0; Std. Dev. 15.5; Min 0.0; Max 74.3
  - LIDC: Obs 1806; Mean 33.5; Std. Dev. 13.6; Min 3.1; Max 74.3
- Rural population (WDI)
  - Full Sample: Obs 7044; Mean 50.9; Std. Dev. 24.8; Min 0.0; Max 97.2
  - LIDC: Obs 2259; Mean 71.3; Std. Dev. 14.8; Min 23.0; Max 97.2
- Fuel exports (WDI)
  - Full Sample: Obs 4516; Mean 16.2; Std. Dev. 29.1; Min 0.0; Max 359.3
  - LIDC: Obs 1025; Mean 13.2; Std. Dev. 30.1; Min 0.0; Max 359.3
- Domestic credit to private sector (WDI)
  - Full Sample: Obs 5731; Mean 37.8; Std. Dev. 35.8; Min 0.1; Max 312.2
  - LIDC: Obs 1781; Mean 15.3; Std. Dev. 11.5; Min 0.2; Max 114.7
- Real GDP per capita growth rate (World Economic Outlook)
  - Full Sample: Obs 5981; Mean 0.0; Std. Dev. 0.1; Min -1.1; Max 1.0
  - LIDC: Obs 1861; Mean 0.0; Std. Dev. 0.1; Min -0.7; Max 0.7

### Country sample (Annex II)
- Non-LIDC Countries:
  - Albania, Argentina, Armenia, Australia, Austria, Belgium, Brazil, Bulgaria, Canada, Chile, China, Colombia, Costa Rica, Croatia, Denmark, Dominican Republic, Ecuador, Arab, Republic of Egypt, El Salvador, Estonia, Finland, France, Germany, Greece, Guatemala, Hungary, India, Indonesia, Islamic Republic of Iran, Iraq, Ireland, Israel, Italy, Jamaica, Japan, Jordan, Kazakhstan, Latvia, Lithuania, Malaysia, Mexico, Morocco, Namibia*, Netherlands, New Zealand, Norway, Pakistan, Panama, Paraguay, Peru, Philippines, Poland, Portugal, Romania, Saudi Arabia, Singapore, Slovak Republic, Slovenia, South Africa, Spain, Sri Lanka, Sweden, Switzerland, Syrian Arab Republic, Thailand, Tunisia, Turkey, Ukraine, United Kingdom, United States, Uruguay, Venezuela
- LIDC Countries:
  - Bangladesh, Benin, Bolivia, Burundi, Cambodia, Cameroon, Central African Republic, Democratic Republic of Congo, Republic of Congo, Côte d'Ivoire, Ghana, Honduras, Kenya, Kyrgyz Republic, People's Democratic Republic of Lao, Lesotho*, Liberia, Malawi, Mali, Mauritania, Moldova, Mongolia, Mozambique, Nepal, Niger, Rwanda, Senegal, Sierra Leone, Sudan, Tajikistan, Tanzania, Togo, Uganda, Republic of Yemen, Zambia, Zimbabwe

* Available for output diversification only.

*Annex I: Summary Statistics — source PDF: _wp16140 - Annex I: Summary Statistics*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16140.pdf_