## Annex I. Data Details

## Source details

**Canonical URL:** [Annex I. Data Details](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023211-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023211-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023211-print-pdf.pdf.json)

---

### Figures and Content Inventory
- Figures listed in the source (titles preserved as presented):
  - 1. Impulse Responses for Macroeconomic Variables
  - 2. Impulse Responses for Women and Men Employment, and Gender Total Employment Gap
  - 3. Impulse Responses for Total Employment by Sectors
  - 4. Impulse Responses got the Gender Employment Gap by Sectors
  - 5. Impulse Responses for the Gender Gap in Labor Force Participation and Unemployment Rate
  - 6. Impulse Responses for the Gender Gap in Total Employment by Labor Market Regulation (EPL)
  - 7. Impulse Responses for the Gender Gap in Total Employment Depending on Gender Wage Gap
  - 8. Impulse Responses for the Gender Employment Gap Depending on Job Informality Gap
  - 9. Impulse Responses for the Gender Employment Gap Depending on the Type of Monetary Policy Shock
  - 10. Impulse Responses for the Gender Employment Gap in Expansions and Recessions
  - 11. Impulse Responses for the Gender Gap in the Unemployment Rate by Alternative Shock Method

### Key findings: scope and headline empirical results
- Sample and scope:
  - Panel of 22 advanced and emerging market economies from 1990Q1 to 2019Q4.
  - Gender gaps are defined as the female indicator minus the male indicator (a narrowing is represented as a positive impulse response for employment and labor force participation gender gaps and a negative impulse response for the unemployment gender gap).
- Main empirical findings:
  - An unexpected increase of 100-basis point in the interest rate:
    - reduces women’s employment less than men’s;
    - narrows the total gender employment gap starting ten quarters after the shock;
    - peak impact of about 0.3 percentage point (a quarter of its standard deviation in the sample).
  - Sectoral channels:
    - Industrial employment (male-dominated) contracts more and faster than services (female-disproportionate).
    - In services, monetary policy shocks have an immediate and persistent impact, slightly narrowing the gender employment gap.
    - In industry, the gender employment gap initially widens but over time reverses and experiences a larger narrowing than in services.
  - Labor market adjustment channels:
    - Following contractionary monetary policy shocks, the unemployment gap declines in the short term.
    - In the medium term, men more than women drop out of the labor market, contributing to a narrowing in the labor force participation gap up to about 0.2 percentage point (almost a quarter of the standard deviation) at the end of the projection period.
  - Heterogeneity by labor market characteristics:
    - Effects are larger in countries with more flexible overall employment protection regulation (lower firing costs and constraints).
    - Effects are larger in countries with larger wage gaps (women’s lower earnings than men).
    - Impact is relatively larger when gender gaps in informal employment are lower (informality mutes monetary policy transmission).
  - Asymmetry of shocks:
    - A 100-basis point contractionary monetary policy shock narrows total gender employment gaps after six quarters, with a peak impact of 0.6 percentage points after 12 quarters.
    - Impact of expansionary monetary policy shocks is muted.
    - The longer-term effect of monetary policy shocks is driven by their impact during expansions.
  - Robustness:
    - Results are reported as robust to various sensitivity tests.

### Methodology and identification
- Shock identification:
  - Monetary policy shocks identified country-by-country following Brandao-Marques and others (2020), extending Romer and Romer (2005).
  - Short-term (3-month) nominal interest rate denoted i_{k,t}; quarterly GDP growth g_{k,t}; quarterly inflation π_{k,t}.
  - One-year-ahead market forecasts indicated by superscript F (e.g., g_{k,t+4}^F, π_{k,t+4}^F).
  - Country-specific OLS regression (equation (1)) is estimated and the residual ε̂_{k,t} is the identified monetary policy shock.
- Local projections:
  - Responses estimated using local projections à la Jordà (2005).
  - Index n denotes outcome population (entire, women, men, or gender gap); horizon h ranges up to twenty quarters (h = 0, ... ,19) after the shock at t−1.
  - Fixed-effects panel regressions estimated for each horizon (equation (2)); country fixed effects λ_{k,h}^n and quarter-year fixed effects θ_{t,h}^n included; standard errors robust and clustered at the country level.
  - Estimated coefficients β quantify the percentage (point) change at horizon h in response to a monetary policy shock of 100 basis points or 1 percentage point.
- Tests for state dependence and heterogeneity:
  - Interaction with country-level labor market characteristics implemented via a smooth transition function G(d_i) = exp(−γ z_i)/(1+exp(−γ z_i)), with γ>0 and z_i normalized country indicator; equation (3) estimates β^− and β^+ for low versus high regimes.
  - Asymmetry between contractionary versus expansionary shocks implemented with a dummy D_{i,t} taking value one for positive monetary policy shocks; equation (4) estimates β^− and β^+ for negative versus positive shocks of 100 basis points.
  - Business cycle state dependence implemented with G(d_{i,t}) = exp(−γ z_{i,t})/(1+exp(−γ z_{i,t})) and γ = 1.5 (following Auerbach and Gorodnichenko, 2013). z_{i,t} normalized country growth indicator. Equation (5) estimates β^− and β^+ for low versus high growth periods and includes additional lags of ∆gdpg_{n,k,t−j} terms.
- Data timing and estimation:
  - Equations (2) to (5) estimated using quarterly data for the panel of 22 countries from 1990Q1 to 2019Q4.
  - Local projections estimate responses up to twenty quarters (h = 0, ... ,19).
  - Standard errors are robust and clustered at the country level.

### Data sources and variables
- Labor market variables (푦푦푛푛,푘푘,푡푡) from the International Labour Organization (ILO): total and sectoral employment, unemployment rates, labor force participation rates.
- Short-term interest rates (푖푖푘푘,푡푡) and three-month money market rates from the OECD.
- Actual real GDP growth rates (푔푔푘푘,푡푡) and inflation rates (휋휋푘푘,푡푡) from the IMF World Economic Outlook (WEO).
- Market forecasts for GDP growth (푔푔푘푘,푡푡+4퐹퐹) and inflation (휋휋푘푘,푡푡+4퐹퐹) from Consensus Economics, reported as 12-month-ahead forecasts and aggregated quarterly by averaging monthly forecasts within the quarter.
- Annual OECD data on employment protection legislation and the wage gap (relative difference between median men and women earnings).
- Annual (with gaps) ILO data on the share of informal women and men workers in total workers.
- Sample inclusion driven by availability of Consensus Economics data.
- Baseline results correspond to 훾훾 = 1.5; alternative values available upon request.

### Macroeconomic impulse responses to a 100 basis points monetary tightening
- GDP: falls gradually by about 0.4 percent at the peak of 8-10 quarters and remains at 0.3 percent below its initial value thereafter.
- CPI: starts falling after 6-10 quarters, with prices declining by 0.3 percent over the horizon.
- Notes: Responses are to a 100-basis point monetary policy shock; confidence intervals shown are 90 percent and one standard deviation.

### III. Overall results — Total employment
- A 100 basis points monetary policy shock decreases log employment for both men and women.
- Female employment:
  - Reduces by about 0.3 percent at the trough (20 percent of its standard deviation) 8 to 12 quarters after the shock.
  - Recovers 17 quarters after the shock.
- Male employment:
  - Cuts faster and recovers later, bottoming out at about 0.5 percent (40 percent of its standard deviation) 12 quarters after the shock.
- Gender gap in employment:
  - Starts to narrow around ten quarters after the shock, up to 0.3 percentage point (a quarter of its standard deviation) at the peak.

### III. Overall results — Sectoral employment
- Average workforce shares:
  - Services: 69.5 percent.
  - Industry: 25.2 percent.
  - Agriculture excluded (4.9 percent average; negligible).
- Industry sector:
  - Men account for 77.6 percent and women for 22.4 percent of industry employment.
  - Employment contracts by about 0.8 percent two quarters after the shock (a quarter of its standard deviation), converging to initial value 12 quarters later.
  - Short-run effect in industry initially lowers women’s employment more than men’s, but reverses over time — gender gap narrows by about 0.4 percentage point (15 percent of the standard deviation) at the end of the projection period.
- Services sector:
  - Women account for 53.1 percent and men for 46.9 percent of services employment on average.
  - Employment response begins around six to seven quarters, contracts by about 0.6 percent (a quarter of its standard deviation), and recovers later.
  - Impact on narrowing the gender gap in services employment is immediate and persistent, about 0.1 percentage point (10 percent of the standard deviation) over the projection period.
- Interpretation:
  - Larger and faster industry reaction aligns with investment and durable goods being more interest-sensitive than non-durable goods in services.
  - Within services, activities like education and healthcare (where women dominate) are less influenced by monetary policy.

### III. Overall results — Unemployment and labor force participation
- Decomposition indicates employment dynamics after monetary policy shocks are driven mainly by:
  - Short term: decline (narrowing) in the unemployment gap.
  - Medium term: increase (narrowing) in the labor force participation gap.
- Quantitative impacts:
  - A 100-basis point increase in the policy rate reduces the unemployment gap by 0.1 percentage point through 8 quarters following the shock (about 20 percent of its standard deviation).
  - By the end of the projection period, labor force participation gap narrows by about 0.2 percentage point (about 20 percent of its standard deviation).

### IV. Role of labor market characteristics
- Employment Protection Legislation (EPL):
  - Monetary policy has a limited impact on gender employment gaps in countries with more rigid EPL.
  - Rigid EPL (higher firing/hiring costs and stringent procedures) makes employment less responsive to demand fluctuations.
- Gender wage gap:
  - In countries with larger gender wage gaps, monetary policy shocks more significantly and persistently narrow gender employment gaps.
  - Mechanism: firms reduce relatively more expensive workers during low demand.
- Informality gap:
  - Higher informality gap (larger share of informal female workers than informal male workers) implies no statistically significant impact of monetary policy on the gender employment gap.
  - Lower informality gap: gender employment gap narrows by about 0.5 percentage point at the peak.
  - Interpretation: larger informality allows adjustment in informal employment to absorb shocks, muting effects on formal employment.

### V. Asymmetric effects
- Contractionary (positive) versus expansionary (negative) shocks:
  - Gender employment gap narrows more after contractionary shocks: over 0.4 percentage point eight quarters after the shock and peaks at 0.6 percentage point after 11 quarters.
  - Expansionary shocks yield no significant effects on the gender gap in employment.
- Business cycle state:
  - Monetary policy shocks have larger effects on gender employment gaps during expansions than recessions.
  - In expansions, monetary tightening narrows the total gender employment gap by up to 0.5 percentage point at the peak.

### VI. Robustness
- Alternative shock identification using residuals of short-term interest rate forecast errors (controlling for GDP and CPI forecast errors) yields comparable results to baseline.
- Results remain robust to:
  - Including additional controls (e.g., fiscal policy shocks).
  - Excluding extreme values of labor market outcomes and monetary policy shocks.
  - Changing the number of countries.

### VII. Policy implications and recommendations
- Incorporate gender-disaggregated sectoral data into central bank forecasts to improve growth and inflation projections and better calibrate monetary policy actions.
- Communicate documented gender-differentiated impacts of monetary policy to improve public understanding and effectiveness of central bank actions.
- Use gender-disaggregated monitoring to design targeted interventions to mitigate adverse effects (e.g., long-term unemployment, labor-force exit) for affected groups:
  - Active labor market programs: training, upskilling.
  - Increase part-time opportunities via targeted subsidies or tax incentives.
  - Extend coverage of unemployment benefits to vulnerable groups to mitigate short-term distributional effects.
- Tailor policies to country-specific circumstances and pursue further research on regional, country, and sector-specific dynamics considering economic and social development, policy design, macro conditions, and labor market structures.

### Country coverage and data availability (1990-2019)
- Table 1 lists countries and the first year & quarter to last year & quarter for three variables: Unemployment rate, Labor Force Participation rate, Employment.
- Country-level coverage (variable time spans preserved exactly):
  - Australia 1991q1-2019q4 1991q1-2019q4 1991q1-2019q4
  - Canada 1990q1-2019q4 1990q1-2019q4 1990q1-2019q4
  - Chile 2010q1-2019q3 2001q1-2019q3 2010q1-2019q3
  - Colombia 2016q2-2019q4 2016q2-2019q4 2016q2-2019q4
  - Czech Republic 1998q3-2019q4 1998q3-2019q4 1998q3-2019q4
  - France 1998q1-2019q4 2003q1-2019q4 1998q1-2019q4
  - Germany 1998q2-2019q4 2005q1-2019q4 1998q2-2019q4
  - Hungary 1999q1-2019q4 1999q1-2019q4 1999q1-2019q4
  - Italy 1990q1-2019q4 1998q1-2019q4 1998q1-2019q4
  - Japan 2002q2-2019q4 2002q2-2019q4 2002q2-2019q4
  - Mexico 2001q2-2019q4 2005q1-2019q4 2002q1-2019q4
  - Netherlands 1998q2-2019q4 2000q1-2019q4 1998q2-2019q4
  - New Zealand 1990q1-2019q4 1990q1-2019q4 2003q1-2019q4
  - Norway 1999q2-2019q4 2000q1-2019q4 1999q2-2019q4
  - Poland 1999q1-2019q4 2000q1-2019q4 2000q1-2019q4
  - Slovak Republic 1998q3-2019q4 1999q1-2019q4 1998q3-2019q4
  - South Korea 1995q1-2019q4 1999q3-2019q4 1995q1-2019q4
  - Spain 1992q2-2019q4 1999q1-2019q4 1998q1-2019q4
  - Sweden 1998q2-2019q4 2001q1-2019q4 1998q2-2019q4
  - Switzerland 2010q1-2019q4 2010q1-2019q4 2010q1-2019q4
  - United Kingdom 1992q2-2019q4 1990q1-2019q4 1998q2-2019q4
  - United States 1990q1-2019q4 1990q1-2019q4 2003q1-2019q4

### Sample summary statistics (1990-2019)
- Table 2 reports the sample mean and standard deviation over the estimation period. Gender labor market gaps are defined as the female indicator minus the male indicator.
- Variables (change over the period) Mean Standard deviation
  - Total Employment 0.29 1.28
  - Total Women Employment  0.39 1.46
  - Total Men Employment 0.17 1.16
  - Total Gender Employment Gap 0.18 1.16
  - Total Industry Employment 0.00 3.17
  - Women Industry Employment  -0.013.07
  - Men Industry Employment 0.07 2.38
  - Gender Industry Employment Gap -0.082.69
  - Total Service Employment  0.47 2.14
  - Women Service Employment 0.48 1.26
  - Men Service Employment 0.36 1.28
  - Gender Service Employment Gap 0.12 1.23
  - Labor Market Force Participation 0.26 1.00
  - Women Labor Force Participation 0.33 1.24
  - Men Labor Force Participation 0.19 0.99
  - Gender Labor Market Force Participation Gap 0.14 0.87
  - Total Unemployment Rate -0.030.59
  - Women Unemployment Rate -0.040.65
  - Men Unemployment Rate -0.280.67
  - Gender Unemployment Rate Gap -0.120.58
  - Employment Protection Legislation 0.00 0.05
  - Wage Gap -0.060.76
  - Informality Gap -0.054.75
  - Real GDP  0.61 1.18
  - CPI 0.63 0.81
- Additional definitions and measurement notes:
  - Employment protection legislation captures the cost and procedures of individual and group dismissal on regular contracts and hiring workers on temporary contracts.
  - The wage gap is the relative difference between median men and women earnings.
  - The informality gap is the share of informal women and men workers in total workers.

### Estimation details: effect of monetary policy on gender gap in total employment (1990-2019)
- Table 3 shows the impact of monetary policy shocks measured by lagged β-coefficients after 4, 8, 12, 16, and 19 quarters (q) using equation (2). T-statistics based on robust clustered standard errors in parentheses. ***/**/* denote significance at 1, 5, 10 percent, respectively.
- Coefficients and robust clustered standard errors (rows correspond to regressors; columns to horizons q=4, q=8, q=12, q=16, q=19):
  - Monetary policy shock (t−1)
    - 0.201 (0.145)
    - 0.130 (0.124)
    - 0.218* (0.126)
    - 0.294** (0.128)
    - 0.058 (0.072)
  - Monetary policy shock (t−2)
    - -0.116 (0.227)
    - 0.052 (0.116)
    - 0.084 (0.097)
    - 0.130 (0.119)
    - 0.029 (0.089)
  - Monetary policy shock (t−3)
    - 0.024 (0.093)
    - 0.010 (0.125)
    - 0.162 (0.121)
    - -0.014 (0.121)
    - 0.001 (0.103)
  - Monetary policy shock (t−4)
    - -0.167 (0.137)
    - 0.032 (0.136)
    - -0.035 (0.125)
    - -0.128 (0.092)
    - -0.157 (0.106)
  - Dependent variable (t−1)
    - 0.415*** (0.425)
    - 0.373*** (0.060)
    - 0.284*** (0.081)
    - 0.195** (0.085)
    - 0.273*** (0.072)
  - Dependent variable (t−2)
    - -0.174*** (0.510)
    - -0.027*** (0.046)
    - -0.312*** (0.043)
    - -0.361*** (0.038)
    - 0.112*** (0.072)
  - Dependent variable (t−3)
    - 0.121*** (0.028)
    - 0.086* (0.048)
    - 0.021 (0.064)
    - -0.092* (0.053)
    - -0.348*** (0.042)
  - Dependent variable (t−4)
    - 0.312*** (0.052)
    - 0.199*** (0.060)
    - 0.118** (0.051)
    - 0.0547 (0.048)
    - -0.195*** (0.061)
- Sample and fit statistics by horizon:
  - N 1379 1297 1219 1143 1079
  - R2 0.39 0.16 0.06 0.08 0.17

*Source: wpiea2023211-print-pdf — Annex I. Data Details (excerpts preserved from the source).*

### Annex I. Data Details ..................................................................................................

### Annex I. Data Details

### Figures and Content Inventory
- Figures listed in the source (titles preserved as presented):
  - 1. Impulse Responses for Macroeconomic Variables
  - 2. Impulse Responses for Women and Men Employment, and Gender Total Employment Gap
  - 3. Impulse Responses for Total Employment by Sectors
  - 4. Impulse Responses got the Gender Employment Gap by Sectors
  - 5. Impulse Responses for the Gender Gap in Labor Force Participation and Unemployment Rate
  - 6. Impulse Responses for the Gender Gap in Total Employment by Labor Market Regulation (EPL)
  - 7. Impulse Responses for the Gender Gap in Total Employment Depending on Gender Wage Gap
  - 8. Impulse Responses for the Gender Employment Gap Depending on Job Informality Gap
  - 9. Impulse Responses for the Gender Employment Gap Depending on the Type of Monetary Policy Shock
  - 10. Impulse Responses for the Gender Employment Gap in Expansions and Recessions
  - 11. Impulse Responses for the Gender Gap in the Unemployment Rate by Alternative Shock Method

### Key Findings (summary of substantive results)
- Sample and scope:
  - Panel of 22 advanced and emerging market economies from 1990Q1 to 2019Q4.
  - Gender gaps are defined as the female indicator minus the male indicator (a narrowing is represented as a positive impulse response for employment and labor force participation gender gaps and a negative impulse response for the unemployment gender gap).

- Main empirical findings:
  - An unexpected increase of 100-basis point in the interest rate:
    - reduces women’s employment less than men’s;
    - narrows the total gender employment gap starting ten quarters after the shock;
    - peak impact of about 0.3 percentage point (a quarter of its standard deviation in the sample).
  - Sectoral channels:
    - Employment in the industrial sector (where men’s employment dominates) contracts more and faster than in services (where women are disproportionally represented).
    - In services, monetary policy shocks have an immediate and persistent impact, slightly narrowing the gender employment gap.
    - In industry, the gender employment gap initially widens but over time reverses and experiences a larger narrowing than in services.
  - Labor market adjustment channels:
    - Following contractionary monetary policy shocks, the unemployment gap declines in the short term.
    - In the medium term, men more than women drop out of the labor market, contributing to a narrowing in the labor force participation gap up to about 0.2 percentage point (almost a quarter of the standard deviation) at the end of the projection period.
  - Heterogeneity by labor market characteristics:
    - Effects are larger in countries with more flexible overall employment protection regulation (lower firing costs and constraints).
    - Effects are larger in countries with larger wage gaps (women’s lower earnings than men).
    - Impact is relatively larger when gender gaps in informal employment are lower (informality mutes monetary policy transmission).
  - Asymmetry of shocks:
    - A 100-basis point contractionary monetary policy shock narrows total gender employment gaps after six quarters, with a peak impact of 0.6 percentage points after 12 quarters.
    - Impact of expansionary monetary policy shocks is muted.
    - The longer-term effect of monetary policy shocks is driven by their impact during expansions.
  - Robustness:
    - Results are reported as robust to various sensitivity tests.

### Methodology and Data Details (preserved technical elements)
- Shock identification:
  - Monetary policy shocks identified country-by-country following Brandao-Marques and others (2020), extending Romer and Romer (2005).
  - Short-term (3-month) nominal interest rate denoted i_{k,t}; quarterly GDP growth g_{k,t}; quarterly inflation π_{k,t}.
  - One-year-ahead market forecasts indicated by superscript F (e.g., g_{k,t+4}^F, π_{k,t+4}^F).
  - Country-specific OLS regression (equation (1)) is estimated and the residual ε̂_{k,t} is the identified monetary policy shock.
- Local projections:
  - Responses estimated using local projections à la Jordà (2005).
  - Index n denotes outcome population (entire, women, men, or gender gap); horizon h ranges up to twenty quarters (h = 0, ... ,19) after the shock at t−1.
  - Fixed-effects panel regressions estimated for each horizon (equation (2)); country fixed effects λ_{k,h}^n and quarter-year fixed effects θ_{t,h}^n included; standard errors robust and clustered at the country level.
  - Estimated coefficients β quantify the percentage (point) change at horizon h in response to a monetary policy shock of 100 basis points or 1 percentage point.
- Tests for state dependence and heterogeneity:
  - Interaction with country-level labor market characteristics implemented via a smooth transition function G(d_i) = exp(−γ z_i)/(1+exp(−γ z_i)), with γ>0 and z_i normalized country indicator; equation (3) estimates β^− and β^+ for low versus high regimes.
  - Asymmetry between contractionary versus expansionary shocks implemented with a dummy D_{i,t} taking value one for positive monetary policy shocks; equation (4) estimates β^− and β^+ for negative versus positive shocks of 100 basis points.
  - Business cycle state dependence implemented with G(d_{i,t}) = exp(−γ z_{i,t})/(1+exp(−γ z_{i,t})) and γ = 1.5 (following Auerbach and Gorodnichenko, 2013). z_{i,t} normalized country growth indicator. Equation (5) estimates β^− and β^+ for low versus high growth periods and includes additional lags of ∆gdpg_{n,k,t−j} terms.
- Data timing and estimation details:
  - Equations (2) to (5) estimated using quarterly data for the panel of 22 countries from 1990Q1 to 2019Q4.
  - Local projections estimate responses up to twenty quarters (h = 0, ... ,19).
  - Standard errors are robust and clustered at the country level.

*IMF Working Paper — Annex I. Data Details (excerpts preserved from the source).*

### Annex I lists the countries included in our analysis and the periods available for each of the three labor market

### wpiea2023211-print-pdf - Annex I lists the countries included in our analysis and the periods available for each of the three labor market

### Data and identification
- Labor market variables (푦푦푛푛,푘푘,푡푡) from the International Labour Organization (ILO): total and sectoral employment, unemployment rates, labor force participation rates.
- Short-term interest rates (푖푖푘푘,푡푡) and three-month money market rates from the OECD.
- Actual real GDP growth rates (푔푔푘푘,푡푡) and inflation rates (휋휋푘푘,푡푡) from the IMF World Economic Outlook (WEO).
- Market forecasts for GDP growth (푔푔푘푘,푡푡+4퐹퐹) and inflation (휋휋푘푘,푡푡+4퐹퐹) from Consensus Economics, reported as 12-month-ahead forecasts and aggregated quarterly by averaging monthly forecasts within the quarter.
- Annual OECD data on employment protection legislation and the wage gap (relative difference between median men and women earnings).
- Annual (with gaps) ILO data on the share of informal women and men workers in total workers.
- Sample inclusion driven by availability of Consensus Economics data.
- Baseline results correspond to 훾훾 = 1.5; alternative values available upon request.

### Macroeconomic impulse responses to a 100 basis points monetary tightening
- GDP: falls gradually by about 0.4 percent at the peak of 8-10 quarters and remains at 0.3 percent below its initial value thereafter.
- CPI: starts falling after 6-10 quarters, with prices declining by 0.3 percent over the horizon.
- Notes: Responses are to a 100-basis point monetary policy shock; confidence intervals shown are 90 percent and one standard deviation.

### III. Overall results — Total employment
- A 100 basis points monetary policy shock decreases log employment for both men and women.
- Female employment:
  - Reduces by about 0.3 percent at the trough (20 percent of its standard deviation) 8 to 12 quarters after the shock.
  - Recovers 17 quarters after the shock.
- Male employment:
  - Cuts faster and recovers later, bottoming out at about 0.5 percent (40 percent of its standard deviation) 12 quarters after the shock.
- Gender gap in employment:
  - Starts to narrow around ten quarters after the shock, up to 0.3 percentage point (a quarter of its standard deviation) at the peak.

### III. Overall results — Sectoral employment
- Average workforce shares:
  - Services: 69.5 percent.
  - Industry: 25.2 percent.
  - Agriculture excluded (4.9 percent average; negligible).
- Industry sector:
  - Men account for 77.6 percent and women for 22.4 percent of industry employment.
  - Employment contracts by about 0.8 percent two quarters after the shock (a quarter of its standard deviation), converging to initial value 12 quarters later.
  - Short-run effect in industry initially lowers women’s employment more than men’s, but reverses over time — gender gap narrows by about 0.4 percentage point (15 percent of the standard deviation) at the end of the projection period.
- Services sector:
  - Women account for 53.1 percent and men for 46.9 percent of services employment on average.
  - Employment response begins around six to seven quarters, contracts by about 0.6 percent (a quarter of its standard deviation), and recovers later.
  - Impact on narrowing the gender gap in services employment is immediate and persistent, about 0.1 percentage point (10 percent of the standard deviation) over the projection period.
- Interpretation:
  - Larger and faster industry reaction aligns with investment and durable goods being more interest-sensitive than non-durable goods in services.
  - Within services, activities like education and healthcare (where women dominate) are less influenced by monetary policy.

### III. Overall results — Unemployment and labor force participation
- Decomposition indicates employment dynamics after monetary policy shocks are driven mainly by:
  - Short term: decline (narrowing) in the unemployment gap.
  - Medium term: increase (narrowing) in the labor force participation gap.
- Quantitative impacts:
  - A 100-basis point increase in the policy rate reduces the unemployment gap by 0.1 percentage point through 8 quarters following the shock (about 20 percent of its standard deviation).
  - By the end of the projection period, labor force participation gap narrows by about 0.2 percentage point (about 20 percent of its standard deviation).

### IV. Role of labor market characteristics
- Employment Protection Legislation (EPL):
  - Monetary policy has a limited impact on gender employment gaps in countries with more rigid EPL.
  - Rigid EPL (higher firing/hiring costs and stringent procedures) makes employment less responsive to demand fluctuations.
- Gender wage gap:
  - In countries with larger gender wage gaps, monetary policy shocks more significantly and persistently narrow gender employment gaps.
  - Mechanism: firms reduce relatively more expensive workers during low demand.
- Informality gap:
  - Higher informality gap (larger share of informal female workers than informal male workers) implies no statistically significant impact of monetary policy on the gender employment gap.
  - Lower informality gap: gender employment gap narrows by about 0.5 percentage point at the peak.
  - Interpretation: larger informality allows adjustment in informal employment to absorb shocks, muting effects on formal employment.

### V. Asymmetric effects
- Contractionary (positive) versus expansionary (negative) shocks:
  - Gender employment gap narrows more after contractionary shocks: over 0.4 percentage point eight quarters after the shock and peaks at 0.6 percentage point after 11 quarters.
  - Expansionary shocks yield no significant effects on the gender gap in employment.
- Business cycle state:
  - Monetary policy shocks have larger effects on gender employment gaps during expansions than recessions.
  - In expansions, monetary tightening narrows the total gender employment gap by up to 0.5 percentage point at the peak.

### VI. Robustness
- Alternative shock identification using residuals of short-term interest rate forecast errors (controlling for GDP and CPI forecast errors) yields comparable results to baseline.
- Results remain robust to:
  - Including additional controls (e.g., fiscal policy shocks).
  - Excluding extreme values of labor market outcomes and monetary policy shocks.
  - Changing the number of countries.

### VII. Policy implications and recommendations
- Incorporate gender-disaggregated sectoral data into central bank forecasts to improve growth and inflation projections and better calibrate monetary policy actions.
- Communicate documented gender-differentiated impacts of monetary policy to improve public understanding and effectiveness of central bank actions.
- Use gender-disaggregated monitoring to design targeted interventions to mitigate adverse effects (e.g., long-term unemployment, labor-force exit) for affected groups:
  - Active labor market programs: training, upskilling.
  - Increase part-time opportunities via targeted subsidies or tax incentives.
  - Extend coverage of unemployment benefits to vulnerable groups to mitigate short-term distributional effects.
- Tailor policies to country-specific circumstances and pursue further research on regional, country, and sector-specific dynamics considering economic and social development, policy design, macro conditions, and labor market structures.

*Source: wpiea2023211-print-pdf*

### Annex I.  Data Details

### Annex I.  Data Details

### Country coverage and data availability (1990-2019)
- Table 1 lists countries and the first year & quarter to last year & quarter for three variables: Unemployment rate, Labor Force Participation rate, Employment.
- Country-level coverage (variable time spans preserved exactly):
  - Australia 1991q1-2019q4 1991q1-2019q4 1991q1-2019q4
  - Canada 1990q1-2019q4 1990q1-2019q4 1990q1-2019q4
  - Chile 2010q1-2019q3 2001q1-2019q3 2010q1-2019q3
  - Colombia 2016q2-2019q4 2016q2-2019q4 2016q2-2019q4
  - Czech Republic 1998q3-2019q4 1998q3-2019q4 1998q3-2019q4
  - France 1998q1-2019q4 2003q1-2019q4 1998q1-2019q4
  - Germany 1998q2-2019q4 2005q1-2019q4 1998q2-2019q4
  - Hungary 1999q1-2019q4 1999q1-2019q4 1999q1-2019q4
  - Italy 1990q1-2019q4 1998q1-2019q4 1998q1-2019q4
  - Japan 2002q2-2019q4 2002q2-2019q4 2002q2-2019q4
  - Mexico 2001q2-2019q4 2005q1-2019q4 2002q1-2019q4
  - Netherlands 1998q2-2019q4 2000q1-2019q4 1998q2-2019q4
  - New Zealand 1990q1-2019q4 1990q1-2019q4 2003q1-2019q4
  - Norway 1999q2-2019q4 2000q1-2019q4 1999q2-2019q4
  - Poland 1999q1-2019q4 2000q1-2019q4 2000q1-2019q4
  - Slovak Republic 1998q3-2019q4 1999q1-2019q4 1998q3-2019q4
  - South Korea 1995q1-2019q4 1999q3-2019q4 1995q1-2019q4
  - Spain 1992q2-2019q4 1999q1-2019q4 1998q1-2019q4
  - Sweden 1998q2-2019q4 2001q1-2019q4 1998q2-2019q4
  - Switzerland 2010q1-2019q4 2010q1-2019q4 2010q1-2019q4
  - United Kingdom 1992q2-2019q4 1990q1-2019q4 1998q2-2019q4
  - United States 1990q1-2019q4 1990q1-2019q4 2003q1-2019q4

### Sample summary statistics (1990-2019)
- Table 2 reports the sample mean and standard deviation over the estimation period. Gender labor market gaps are defined as the female indicator minus the male indicator.
- Variables (change over the period) Mean Standard deviation
  - Total Employment 0.29 1.28
  - Total Women Employment  0.39 1.46
  - Total Men Employment 0.17 1.16
  - Total Gender Employment Gap 0.18 1.16
  - Total Industry Employment 0.00 3.17
  - Women Industry Employment  -0.013.07
  - Men Industry Employment 0.07 2.38
  - Gender Industry Employment Gap -0.082.69
  - Total Service Employment  0.47 2.14
  - Women Service Employment 0.48 1.26
  - Men Service Employment 0.36 1.28
  - Gender Service Employment Gap 0.12 1.23
  - Labor Market Force Participation 0.26 1.00
  - Women Labor Force Participation 0.33 1.24
  - Men Labor Force Participation 0.19 0.99
  - Gender Labor Market Force Participation Gap 0.14 0.87
  - Total Unemployment Rate -0.030.59
  - Women Unemployment Rate -0.040.65
  - Men Unemployment Rate -0.280.67
  - Gender Unemployment Rate Gap -0.120.58
  - Employment Protection Legislation 0.00 0.05
  - Wage Gap -0.060.76
  - Informality Gap -0.054.75
  - Real GDP  0.61 1.18
  - CPI 0.63 0.81

- Additional definitions and measurement notes:
  - Employment protection legislation captures the cost and procedures of individual and group dismissal on regular contracts and hiring workers on temporary contracts.
  - The wage gap is the relative difference between median men and women earnings.
  - The informality gap is the share of informal women and men workers in total workers.

### Estimation details: effect of monetary policy on gender gap in total employment (1990-2019)
- Table 3 shows the impact of monetary policy shocks measured by lagged β-coefficients after 4, 8, 12, 16, and 19 quarters (q) using equation (2). T-statistics based on robust clustered standard errors in parentheses. ***/**/* denote significance at 1, 5, 10 percent, respectively.
- Coefficients and robust clustered standard errors (rows correspond to regressors; columns to horizons q=4, q=8, q=12, q=16, q=19):
  - Monetary policy shock (t−1)
    - 0.201 (0.145)
    - 0.130 (0.124)
    - 0.218* (0.126)
    - 0.294** (0.128)
    - 0.058 (0.072)
  - Monetary policy shock (t−2)
    - -0.116 (0.227)
    - 0.052 (0.116)
    - 0.084 (0.097)
    - 0.130 (0.119)
    - 0.029 (0.089)
  - Monetary policy shock (t−3)
    - 0.024 (0.093)
    - 0.010 (0.125)
    - 0.162 (0.121)
    - -0.014 (0.121)
    - 0.001 (0.103)
  - Monetary policy shock (t−4)
    - -0.167 (0.137)
    - 0.032 (0.136)
    - -0.035 (0.125)
    - -0.128 (0.092)
    - -0.157 (0.106)
  - Dependent variable (t−1)
    - 0.415*** (0.425)
    - 0.373*** (0.060)
    - 0.284*** (0.081)
    - 0.195** (0.085)
    - 0.273*** (0.072)
  - Dependent variable (t−2)
    - -0.174*** (0.510)
    - -0.027*** (0.046)
    - -0.312*** (0.043)
    - -0.361*** (0.038)
    - 0.112*** (0.072)
  - Dependent variable (t−3)
    - 0.121*** (0.028)
    - 0.086* (0.048)
    - 0.021 (0.064)
    - -0.092* (0.053)
    - -0.348*** (0.042)
  - Dependent variable (t−4)
    - 0.312*** (0.052)
    - 0.199*** (0.060)
    - 0.118** (0.051)
    - 0.0547 (0.048)
    - -0.195*** (0.061)
- Sample and fit statistics by horizon:
  - N 1379 1297 1219 1143 1079
  - R2 0.39 0.16 0.06 0.08 0.17

*Source: wpiea2023211-print-pdf - Annex I.  Data Details*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023211-print-pdf.pdf_