## 1. Introduction

## Source details

**Canonical URL:** [1. Introduction](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025145.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025145.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025145.pdf.json)

---

### Study purpose and contribution
- Provides a global, disaggregated analysis of labor market responses to oil supply shocks using a panel of quarterly labor market data across 89 countries from 1975 to 2022.
- Combines ILO labor market series (employment-to-population ratios, unemployment rates, labor force participation rates, population aged 25 and older, sex-disaggregated, and 21 one-digit ISIC sectors) with Känzig’s (2021) oil supply news shock series (extended through 2022).
- Estimates dynamic effects over a five-year horizon using local projections (Jordà, 2005).
- Decomposes labor market responses along four axes:
  - (i) asymmetries between contractionary and expansionary oil shocks;
  - (ii) differences between oil-importing and oil-exporting countries;
  - (iii) variation across oil-intensive and oil-moderate sectors;
  - (iv) sex-disaggregated labor market effects.

### Key mechanisms and contextual factors highlighted
- Elevated energy costs may pass through to higher wages and suppress hiring (Baba and Lee, 2022).
- Sectoral reallocation with search and matching inefficiency may be slow and frictional (Herrera and Karaki, 2015).
- Limited social protections may keep workers in the labor force despite deteriorating job prospects (Ugargol and Parvathy, 2023).
- Occupational distribution, sectoral exposure, and access to job security contribute to sex-disaggregated differences (Ologbenla, 2020; Bacon and Kojima, 2008).

### Novelty and scope
- Extends Känzig (2021) oil supply news shock series to 2022 and applies it in a country-quarter panel to measure labor market margins globally.
- Focuses on population aged 25 and older to maximize country coverage and data completeness.
- Uses sectoral ISIC one-digit data and sex-disaggregated ILO series to provide granular evidence on transmission channels and distributional consequences.

### Data and identification (methodological highlights)
- Oil supply news shock construction:
  - Based on Känzig (2021) approach, using changes in spot crude oil futures prices within tight windows around OPEC announcements.
  - Updated series extends through 2022 and covers 150 OPEC announcements.
  - Final estimated monthly shock series spans January 1975 to December 2022 and is calibrated to reflect an unexpected 10 percent increase in the spot crude oil price.
  - Känzig (2021) SVAR uses real oil price, world oil production, world oil inventories, world industrial production, industrial production in the US, and the US CPI, with lag order of 12. Identification sample ranges from July 1983 to December 2022; estimation sample extends from January 1974 to December 2022.
  - Shock series shows no evidence of autocorrelation and is not predictable by macroeconomic or financial variables.
- Panel data:
  - Quarterly ILO data for employment-to-population ratios, unemployment rates, labor force participation rates, population aged 25 and older, disaggregated by sex.
  - Covers 89 countries from 1975 to 2022; panel is unbalanced.
  - 21 disaggregated sectors at ISIC one-digit level used for sectoral heterogeneity tests.
  - Net oil trade (volume) from IMF WEO used to classify country-quarter observations as oil exporters if net oil trade is positive, oil importers otherwise; in matched sample, 84.4 percent are oil importers and 15.6 percent are oil exporters.
- Empirical strategy:
  - Local projections à la Jordà (2005) estimate horizon responses up to twenty quarters (h = 0, ..., 19).
  - Baseline fixed-effects panel regression (separate for each horizon h) includes country fixed effects and time fixed effects, controls up to four quarters of lags of the shock and outcome variables, and uses robust standard errors clustered at the country level.
  - Estimated coefficients β_h quantify accumulated change at horizon h in response to an oil supply news shock normalized to a 10 percent oil price increase.
  - Heterogeneity analyses use the same empirical approach on restricted sub-samples (asymmetric transmission, oil importers vs exporters, sectoral impacts, sex-disaggregation).

---

### 3. Results

### 3.1. Baseline
- A positive oil supply news shock normalized to reflect a 10 percent increase in oil prices leads to a persistent weakening of aggregate labor market conditions.
- Employment-to-population ratio:
  - Falls sharply in the initial quarters.
  - Trough of approximately 0.25 percentage points below baseline in the seventh quarter.
  - By the 20th quarter, it is roughly 0.14 percentage points below baseline.
- Unemployment rate:
  - Rises rapidly, peaking at approximately 0.23 percentage points above baseline in the eighth quarter.
  - Remains elevated, hovering around 0.1 percentage points above baseline at quarter 20.
- Labor force participation rate:
  - More muted and less stable response; fluctuates in the short run.
  - Begins to decline visibly around the fourth quarter.
  - Trough of approximately 0.11 percentage points below baseline by the seventh quarter.
  - Statistically significant in the medium term; longer-run dynamics uncertain with wide confidence intervals and signs of partial recovery.
- Overall interpretation:
  - Oil supply shocks diminish employment opportunities and prompt some labor force exits, producing broad-based deterioration and lasting labor market scarring.

### 3.2. Asymmetric Transmission
- Definitions:
  - Contractionary oil supply news shocks: OPEC announcement of reduced oil supply and accompanying increase in oil price.
  - Expansionary oil supply news shocks: OPEC announcement of increased oil supply and accompanying decrease in oil price.
- Contractionary shocks (OPEC tightening):
  - Employment-to-population ratio:
    - Modest initial dip in first two quarters, then statistically indistinguishable from zero through quarter ten.
    - Begins to decline around quarter 11; by end of 20-quarter horizon stands nearly 0.45 percentage points below baseline.
  - Unemployment rate:
    - Increases sharply and persistently, peaking at around 0.6 percentage points above baseline in quarter 18.
  - Labor force participation rate:
    - Rises in response, peaking at about 0.25 percentage points above baseline around quarter 11.
    - Effect is transitory and trends downward later, returning close to baseline by the end of the horizon.
  - Interpretation: adverse employment effects emerge with delay but persist; temporary increase in participation possibly due to households seeking additional work.
- Expansionary shocks (OPEC loosening):
  - Employment-to-population ratio:
    - Increases modestly short term, peaking at approximately 0.22 percentage points above baseline around quarter four.
    - Declines to near -0.22 percentage points by quarter ten, then rebounds toward pre-shock levels later.
  - Unemployment rate:
    - Falls sharply short term, reaching near -0.27 percentage points by quarter four.
    - Rises back toward baseline by quarter 11, then declines again to roughly -0.2 percentage points by quarter 20.
  - Labor force participation rate:
    - Significant decline in medium term, trough of about -0.3 percentage points around quarter ten.
    - Gradual recovery but remains below baseline until end of projection window.
  - Interpretation: expansionary shocks bring smaller, transitory improvements with more volatile and drawn-out adjustments.
- Overall asymmetry:
  - Contractionary shocks cause more persistent and pronounced adverse effects, especially for employment and unemployment.
  - Expansionary shocks yield short-term improvements but no sustained participation changes in either case.

### 3.3. Oil Importers versus Oil Exporters
- Classification:
  - Oil-importers: countries with negative net trade of oil.
  - Oil-exporters: countries with positive net trade of oil.
- Oil-importing countries:
  - Employment-to-population ratio:
    - Declines notably in first two years; trough of approximately 0.28 percentage points below baseline by the seventh quarter.
    - Partial recovery; about 0.16 percentage points below baseline by quarter 20.
  - Unemployment rate:
    - Rises sharply, peaking at about 0.24 percentage points above baseline around the seventh quarter.
    - Gradual decline but remains approximately 0.12 percentage points above baseline by quarter 20.
  - Labor force participation rate:
    - Contracts by about 0.1 percentage points that persists through the medium-term horizon.
  - Interpretation: higher oil prices tighten production budgets, raise input costs, reduce labor demand, and impede job search and participation for households.
- Oil-exporting countries:
  - Employment-to-population ratio:
    - Rises gradually, peaking at about 0.2 percentage points above baseline toward the end of the forecast horizon.
  - Unemployment rate:
    - Ticks up modestly, peaking at around 0.13 percentage points above baseline seven quarters after the shock.
    - Returns to baseline within three years and remains stable long term with no sustained deviation.
  - Labor force participation rate:
    - Trends upward, with an increase of nearly 0.2 percentage points by the third year after the shock.
  - Interpretation: oil trade balance is a key determinant of vulnerability; oil-importers are adversely affected while oil-exporters tend to benefit gradually.

### 3.4. Sectoral Analysis
- Sectoral approach:
  - Use ISIC one-digit codes; take logarithm of employment.
  - Classification:
    - Oil-intensive sectors: mining and quarrying, construction, and transportation and storage.
    - Oil-moderate sectors: all other sectors.
- Employment responses:
  - Oil-intensive sectors:
    - Employment drops sharply almost immediately.
    - Reaches a trough of about one percent below baseline by the 11th quarter.
    - Immediate response reflects direct pass-through of energy cost increases in fuel-, transportation-, or oil-based materials–intensive sectors.
  - Oil-moderate sectors:
    - More delayed response; employment stable first few quarters.
    - Declines become pronounced from around quarter six.
    - By quarter 11, employment losses approach one percent, similar in magnitude to oil-intensive sectors.
- Medium- to long-term finding:
  - Employment impacts converge across both groups; both remain below baseline toward end of forecast horizon.
  - Interpretation: even sectors with lower direct oil dependence eventually experience comparable employment erosion via second-round effects and demand channels.

### 3.5. Sex-Disaggregated Results
- Definitions:
  - Gaps defined as female indicator minus male indicator for each variable.
- Employment-to-population ratio:
  - Men:
    - Significant decline after shock; negative effect more pronounced from quarter two.
    - Lowest point around quarter eight with shock effect of nearly -0.25 percentage points.
    - Gradual recovery toward zero by quarter 17, then stabilizes.
  - Women:
    - Deep decline intensifying between quarters two and eight, hitting a minimum between -0.2 and -0.25 percentage points.
    - Less recovery in long term; remains negative throughout the 20-quarter horizon.
  - Gap (female minus male):
    - Mostly positive, indicating narrowing of the gap post-shock.
    - Peaks at quarter eight; effect fades and fluctuates closer to zero thereafter as men recover while women’s response remains more persistent.
  - Interpretation: initial impact larger for men, leading to temporary narrowing of the employment gap.
- Unemployment rate:
  - Men:
    - Sharp increase, peaking around quarter eight at approximately 0.23 percentage points.
    - Gradual decline but remains elevated throughout 20 quarters.
  - Women:
    - Rapid increase peaking near quarter eight at around 0.2 percentage points.
    - Declines afterwards but remains higher than baseline over 20 quarters.
  - Gap (female minus male):
    - Initially negative in first few quarters (male unemployment rises more).
    - Reverses after quarter eight and fluctuates around zero subsequently.
  - Interpretation: shock initially hits men harder on unemployment, with effects becoming more balanced over time.
- Labor force participation rates:
  - Men:
    - Fluctuates around zero; modest and variable impact.
  - Women:
    - More notable downward trend with less significant rebound; more persistent negative impact.
  - Gap (female minus male):
    - Slight positive response on average over almost the entire period.
  - Interpretation: participation effects are more uncertain; women experience more persistent declines while male participation is more variable.

---

### 4. Robustness Checks

### Overview
- Baseline conclusions survive a broad set of robustness checks, reported in Annex II, confirming stability across various specifications and assumptions.

### Robustness exercises and implementations
- Lag length sensitivity:
  - Extended the lag structure in the local projections from four quarters (1 year) to six quarters (1.5 years).
  - Adjustment applied to both the oil shock variable and the outcome variables.
- Additional controls in local projections:
  - Introduced quarterly inflation (calculated from consumer price indexes) and the quarterly percent change in GDP (constant prices, national currency).
  - Both controls collected from the WEO database.
  - Maintained the original four-quarter lag structure as in the baseline analysis.
- Alternative age range:
  - Re-estimated results including individuals aged 15 and older.
- Alternative oil supply shock estimates (global-only):
  - Re-estimated the oil supply shock by focusing exclusively on global variables and removing U.S.-specific factors.
  - Modified Känzig (2021) six-variable model (four global variables and two U.S.-specific variables) to generate a new “global” oil shock estimate using only the four standard global variables (real oil price, world oil production, world oil inventories, and world industrial production).

### Reporting and supporting material
- Robustness results are presented in Annex II (Figures 9–12), which include:
  - Figure 9. Impulse responses using alternative lag length
  - Figure 10. Impulse responses using additional controls
  - Figure 11. Impulse responses using alternative age range
  - Figure 12. Impulse responses using alternative shock estimates

*Source: IMF Working Paper — Section 1 and Sections 3–4 (content unit wpiea2025145).*

### 1. Introduction ........................................................................................................

### 1. Introduction

### Study purpose and contribution
- Provides a global, disaggregated analysis of labor market responses to oil supply shocks using a panel of quarterly labor market data across 89 countries from 1975 to 2022.
- Combines ILO labor market series (employment-to-population ratios, unemployment rates, labor force participation rates, population aged 25 and older, sex-disaggregated, and 21 one-digit ISIC sectors) with Känzig’s (2021) oil supply news shock series (extended through 2022).
- Estimates dynamic effects over a five-year horizon using local projections (Jordà, 2005).
- Decomposes labor market responses along four axes:
  - (i) asymmetries between contractionary and expansionary oil shocks;
  - (ii) differences between oil-importing and oil-exporting countries;
  - (iii) variation across oil-intensive and oil-moderate sectors;
  - (iv) sex-disaggregated labor market effects.

### Key mechanisms and contextual factors highlighted
- Elevated energy costs may pass through to higher wages and suppress hiring (Baba and Lee, 2022).
- Sectoral reallocation with search and matching inefficiency may be slow and frictional (Herrera and Karaki, 2015).
- Limited social protections may keep workers in the labor force despite deteriorating job prospects (Ugargol and Parvathy, 2023).
- Occupational distribution, sectoral exposure, and access to job security contribute to sex-disaggregated differences (Ologbenla, 2020; Bacon and Kojima, 2008).

### Main empirical findings (high-level)
- Employment-to-population ratio: sharp decline after contractionary shocks with partial recovery; employment falls sharply and recovers only partially in baseline.
- Unemployment rate: rapid rise after contractionary shocks and remains elevated; unemployment rises quickly and remains elevated in baseline.
- Labor force participation rate: declines moderately and unevenly in baseline; responses differ by shock sign and country oil-trade status.
- Asymmetry: Contractionary shocks (reductions in oil supply / rising oil prices) cause delayed but persistent employment declines, immediate and lasting unemployment rises, and a temporary rise in participation. Expansionary shocks produce short-term employment gains and falling unemployment but medium-run declines in participation.
- Net oil trade heterogeneity:
  - Oil importers (84.4 percent of country-quarter observations): sharp employment losses with partial recovery; unemployment rises quickly and stays elevated; persistent decline in participation.
  - Oil exporters (15.6 percent of country-quarter observations): more gradual and modest employment gains with lag; unemployment increases slightly then returns to baseline; participation grows steadily.
- Sectoral heterogeneity:
  - Oil-intensive sectors: immediate, sharp drop in employment that remains below baseline long-term.
  - Oil-moderate sectors: delayed decline with medium- and long-term drop, indicating propagation beyond directly exposed sectors.
- Sex-disaggregated heterogeneity:
  - Men: sharp short-term employment drops with faster rebound; unemployment mirrors employment with steep rise and faster recovery; participation falls initially and stabilizes.
  - Women: less severe but more persistent employment declines with limited recovery; unemployment mirrors employment with milder increase and slow recovery; participation shows gradual, sustained decline.

### Novelty and importance
- Extends Känzig (2021) oil supply news shock series to 2022 and applies it in a country-quarter panel to measure labor market margins globally.
- Focuses on population aged 25 and older to maximize country coverage and data completeness.
- Uses sectoral ISIC one-digit data and sex-disaggregated ILO series to provide granular evidence on transmission channels and distributional consequences.

*Table 1 (Summary of Findings) presents the response patterns normalized to a 10 percent oil price increase: employment-to-population ratio, unemployment rate, and labor force participation rate across Baseline, Asymmetric Transmission (Contractionary vs Expansionary), Net Oil Trade (Oil Importers vs Oil Exporters), Sectoral Analysis (Oil-Intensive vs Oil-Moderate), and Sex-Disaggregated (Men vs Women).*

### Organization of the paper
- Section 2: Data and empirical strategy.
- Section 3: Main results on labor market outcomes.
- Section 4: Robustness checks.
- Section 5: Conclusion.

### Data and identification details (methodological highlights)
- Oil supply news shock construction:
  - Based on Känzig (2021) approach, using changes in spot crude oil futures prices within tight windows around OPEC announcements.
  - Updated series extends through 2022 and covers 150 OPEC announcements.
  - Final estimated monthly shock series spans January 1975 to December 2022 and is calibrated to reflect an unexpected 10 percent increase in the spot crude oil price.
  - Känzig (2021) SVAR uses real oil price, world oil production, world oil inventories, world industrial production, industrial production in the US, and the US CPI, with lag order of 12. Identification sample ranges from July 1983 to December 2022; estimation sample extends from January 1974 to December 2022.
  - Shock series shows no evidence of autocorrelation and is not predictable by macroeconomic or financial variables.
- Panel data:
  - Quarterly ILO data for employment-to-population ratios, unemployment rates, labor force participation rates, population aged 25 and older, disaggregated by sex.
  - Covers 89 countries from 1975 to 2022; panel is unbalanced (number of observations varies by country and variable).
  - 21 disaggregated sectors at ISIC one-digit level used for sectoral heterogeneity tests.
  - Net oil trade (volume) from IMF WEO used to classify country-quarter observations as oil exporters if net oil trade is positive, oil importers otherwise; in matched sample, 84.4 percent are oil importers and 15.6 percent are oil exporters.
- Empirical strategy:
  - Local projections à la Jordà (2005) estimate horizon responses up to twenty quarters (h = 0, ..., 19).
  - Baseline fixed-effects panel regression (separate for each horizon h) includes country fixed effects and time fixed effects, controls up to four quarters of lags of the shock and outcome variables, and uses robust standard errors clustered at the country level.
  - Estimated coefficients β_h quantify accumulated change at horizon h in response to an oil supply news shock normalized to a 10 percent oil price increase.
  - Heterogeneity analyses use the same empirical approach on restricted sub-samples (asymmetric transmission, oil importers vs exporters, sectoral impacts, sex-disaggregation).

*Source: IMF Working Paper — 1. Introduction (content unit wpiea2025145).*

### 3.  Results

### 3. Results

### 3.1. Baseline
- A positive oil supply news shock normalized to reflect a 10 percent increase in oil prices leads to a persistent weakening of aggregate labor market conditions.
- Employment-to-population ratio:
  - Falls sharply in the initial quarters.
  - Trough of approximately 0.25 percentage points below baseline in the seventh quarter.
  - By the 20th quarter, it is roughly 0.14 percentage points below baseline.
- Unemployment rate:
  - Rises rapidly, peaking at approximately 0.23 percentage points above baseline in the eighth quarter.
  - Remains elevated, hovering around 0.1 percentage points above baseline at quarter 20.
- Labor force participation rate:
  - More muted and less stable response; fluctuates in the short run.
  - Begins to decline visibly around the fourth quarter.
  - Trough of approximately 0.11 percentage points below baseline by the seventh quarter.
  - Statistically significant in the medium term; longer-run dynamics uncertain with wide confidence intervals and signs of partial recovery.
- Overall interpretation:
  - Oil supply shocks diminish employment opportunities and prompt some labor force exits, producing broad-based deterioration and lasting labor market scarring.

*Notes: Impulse responses are to an oil supply news shock normalized to reflect a 10 percent increase in the oil price. Confidence bands described in source figures.*

### 3.2. Asymmetric Transmission
- Definitions:
  - Contractionary oil supply news shocks: OPEC announcement of reduced oil supply and accompanying increase in oil price.
  - Expansionary oil supply news shocks: OPEC announcement of increased oil supply and accompanying decrease in oil price.
- Contractionary shocks (OPEC tightening):
  - Employment-to-population ratio:
    - Modest initial dip in first two quarters, then statistically indistinguishable from zero through quarter ten.
    - Begins to decline around quarter 11; by end of 20-quarter horizon stands nearly 0.45 percentage points below baseline.
  - Unemployment rate:
    - Increases sharply and persistently, peaking at around 0.6 percentage points above baseline in quarter 18.
  - Labor force participation rate:
    - Rises in response, peaking at about 0.25 percentage points above baseline around quarter 11.
    - Effect is transitory and trends downward later, returning close to baseline by the end of the horizon.
  - Interpretation: adverse employment effects emerge with delay but persist; temporary increase in participation possibly due to households seeking additional work.
- Expansionary shocks (OPEC loosening):
  - Employment-to-population ratio:
    - Increases modestly short term, peaking at approximately 0.22 percentage points above baseline around quarter four.
    - Declines to near -0.22 percentage points by quarter ten, then rebounds toward pre-shock levels later.
  - Unemployment rate:
    - Falls sharply short term, reaching near -0.27 percentage points by quarter four.
    - Rises back toward baseline by quarter 11, then declines again to roughly -0.2 percentage points by quarter 20.
  - Labor force participation rate:
    - Significant decline in medium term, trough of about -0.3 percentage points around quarter ten.
    - Gradual recovery but remains below baseline until end of projection window.
  - Interpretation: expansionary shocks bring smaller, transitory improvements with more volatile and drawn-out adjustments.
- Overall asymmetry:
  - Contractionary shocks cause more persistent and pronounced adverse effects, especially for employment and unemployment.
  - Expansionary shocks yield short-term improvements but no sustained participation changes in either case.

*Notes: Impulse responses normalized to a 10 percent increase in the oil price; contractionary = OPEC tightening, expansionary = OPEC loosening.*

### 3.3. Oil Importers versus Oil Exporters
- Classification:
  - Oil-importers: countries with negative net trade of oil.
  - Oil-exporters: countries with positive net trade of oil.
- Oil-importing countries (left-hand panels, Figure 4):
  - Employment-to-population ratio:
    - Declines notably in first two years; trough of approximately 0.28 percentage points below baseline by the seventh quarter.
    - Partial recovery; about 0.16 percentage points below baseline by quarter 20.
  - Unemployment rate:
    - Rises sharply, peaking at about 0.24 percentage points above baseline around the seventh quarter.
    - Gradual decline but remains approximately 0.12 percentage points above baseline by quarter 20.
  - Labor force participation rate:
    - Contracts by about 0.1 percentage points that persists through the medium-term horizon.
  - Interpretation: higher oil prices tighten production budgets, raise input costs, reduce labor demand, and impede job search and participation for households.
- Oil-exporting countries (right-hand panels, Figure 4):
  - Employment-to-population ratio:
    - Rises gradually, peaking at about 0.2 percentage points above baseline toward the end of the forecast horizon.
  - Unemployment rate:
    - Ticks up modestly, peaking at around 0.13 percentage points above baseline seven quarters after the shock.
    - Returns to baseline within three years and remains stable long term with no sustained deviation.
  - Labor force participation rate:
    - Trends upward, with an increase of nearly 0.2 percentage points by the third year after the shock.
  - Interpretation: oil trade balance is a key determinant of vulnerability; oil-importers are adversely affected while oil-exporters tend to benefit gradually.

*Notes: Impulse responses normalized to a 10 percent increase in the oil price.*

### 3.4. Sectoral Analysis
- Sectoral approach:
  - Use ISIC one-digit codes; take logarithm of employment.
  - Classification:
    - Oil-intensive sectors: mining and quarrying, construction, and transportation and storage.
    - Oil-moderate sectors: all other sectors.
- Employment responses:
  - Oil-intensive sectors:
    - Employment drops sharply almost immediately.
    - Reaches a trough of about one percent below baseline by the 11th quarter.
    - Immediate response reflects direct pass-through of energy cost increases in fuel-, transportation-, or oil-based materials–intensive sectors.
  - Oil-moderate sectors:
    - More delayed response; employment stable first few quarters.
    - Declines become pronounced from around quarter six.
    - By quarter 11, employment losses approach one percent, similar in magnitude to oil-intensive sectors.
- Medium- to long-term finding:
  - Employment impacts converge across both groups; both remain below baseline toward end of forecast horizon.
  - Interpretation: even sectors with lower direct oil dependence eventually experience comparable employment erosion via second-round effects and demand channels.

*Notes: Classification supported by prior literature; impulse responses normalized to a 10 percent increase in the oil price.*

### 3.5. Sex-Disaggregated Results
- Definitions:
  - Gaps defined as female indicator minus male indicator for each variable.
- Employment-to-population ratio (Figure 6):
  - Men:
    - Significant decline after shock; negative effect more pronounced from quarter two.
    - Lowest point around quarter eight with shock effect of nearly -0.25 percentage points.
    - Gradual recovery toward zero by quarter 17, then stabilizes.
  - Women:
    - Deep decline intensifying between quarters two and eight, hitting a minimum between -0.2 and -0.25 percentage points.
    - Less recovery in long term; remains negative throughout the 20-quarter horizon.
  - Gap (female minus male):
    - Mostly positive, indicating narrowing of the gap post-shock.
    - Peaks at quarter eight; effect fades and fluctuates closer to zero thereafter as men recover while women’s response remains more persistent.
  - Interpretation: initial impact larger for men, leading to temporary narrowing of the employment gap.
- Unemployment rate (Figure 7):
  - Men:
    - Sharp increase, peaking around quarter eight at approximately 0.23 percentage points.
    - Gradual decline but remains elevated throughout 20 quarters.
  - Women:
    - Rapid increase peaking near quarter eight at around 0.2 percentage points.
    - Declines afterwards but remains higher than baseline over 20 quarters.
  - Gap (female minus male):
    - Initially negative in first few quarters (male unemployment rises more).
    - Reverses after quarter eight and fluctuates around zero subsequently.
  - Interpretation: shock initially hits men harder on unemployment, with effects becoming more balanced over time.
- Labor force participation rates (Figure 8):
  - Men:
    - Fluctuates around zero; modest and variable impact.
  - Women:
    - More notable downward trend with less significant rebound; more persistent negative impact.
  - Gap (female minus male):
    - Slight positive response on average over almost the entire period.
  - Interpretation: participation effects are more uncertain; women experience more persistent declines while male participation is more variable.

*Notes: Impulse responses normalized to a 10 percent increase in the oil price. Gaps are female minus male.*

*Italic: Source: IMF Working Paper — Section 3, “Results.”*

### 4.  Robustness Checks

### 4.  Robustness Checks

### Overview
- Baseline conclusions survive a broad set of robustness checks, reported in Annex II, confirming stability across various specifications and assumptions.
- Robustness exercises assess sensitivity to lag length, additional macro controls, alternative sample age range, and alternative oil supply shock estimates.

### Robustness exercises and implementations
- Lag length sensitivity
  - Extended the lag structure in the local projections from four quarters (1 year) to six quarters (1.5 years).
  - Adjustment applied to both the oil shock variable and the outcome variables to assess whether a longer lag alters shock dynamics.

- Additional controls in local projections
  - Introduced quarterly inflation (calculated from consumer price indexes) and the quarterly percent change in GDP (constant prices, national currency).
  - Both controls collected from the WEO database.
  - Maintained the original four-quarter lag structure as in the baseline analysis.

- Alternative age range
  - Re-estimated results including individuals aged 15 and older, broadening demographic scope relative to the baseline.

- Alternative oil supply shock estimates (global-only)
  - Re-estimated the oil supply shock by focusing exclusively on global variables and removing U.S.-specific factors.
  - Modified the approach used by Känzig (2021), who estimated shocks using a six-variable model that included:
    - Four global variables: real oil price, world oil production, world oil inventories, and world industrial production.
    - Two U.S.-specific variables: U.S. CPI and U.S. industrial production.
  - Generated a new “global” oil shock estimate using only the four standard global variables to assess whether removing the two U.S. variables materially changes results and to provide a more comprehensive view of global implications.

### Reporting and supporting material
- Robustness results are presented in Annex II (Figures 9–12), which include:
  - Figure 9. Impulse responses using alternative lag length
  - Figure 10. Impulse responses using additional controls
  - Figure 11. Impulse responses using alternative age range
  - Figure 12. Impulse responses using alternative shock estimates

*Source: Oil Shocks and Labor Market Developments, Working Paper No. WP/2025/145*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025145.pdf_