## 1. Responses to Structural Deficit Shocks

## Source details

**Canonical URL:** [1. Responses to Structural Deficit Shocks](https://www.imf.org/-/media/files/publications/cr/2025/english/1colea2025002-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2025/english/1colea2025002-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2025/english/1colea2025002-source-pdf.pdf.json)

---

### Summary and objective
- Examines expansionary fiscal policy and its challenge to containing inflation in financially integrated Latin American economies (LA5: Brazil, Chile, Colombia, Mexico, Peru) during 2010-24.
- Frameworks and methods:
  - Open-economy New Keynesian Phillips curve and local projections.
  - Shocks measured as actual year-end structural deficit and public debt minus IMF WEO October forecasts (percent of previous year’s GDP).
  - Headline inflation shocks defined analogously (actual minus forecasted end-of-year headline inflation).

### Transmission channels highlighted
- Aggregate demand channel: fiscal expansion → higher domestic demand → upward pressure on prices (non-tradables) when short-run supply inelastic.
- Inflation expectations channel: expectations may de-anchor if central bank perceived reluctant to tighten due to higher public debt servicing costs.
- Exchange rate and confidence channel: higher demand and rates can appreciate currency, but erosion of investor confidence can dominate → higher risk premium → currency depreciation.
- Emerging market amplifiers: greater external shock exposure, weaker fiscal/central bank credibility, currency mismatches.

### Empirical specification and controls
- Dynamics: two lags of dependent variable; contemporaneous and one lag of real GDP growth and public debt.
- Monetary policy shocks: analysts’ forecast errors of policy rates.
- Baseline fiscal measure: end-of-year structural deficit and public debt (% of previous year’s GDP). Alternatives: structural primary deficit, cyclically adjusted deficit, cyclically adjusted primary deficit.
- Identification: forecast errors (actual − IMF October forecasts).
- Controls include country and year fixed effects; variables winsorized at 2nd and 98th percentiles.

### Key empirical findings — structural deficit shocks
- Headline inflation:
  - Headline inflation rises by 0.35 percentage points following a fiscal deficit shock equivalent to 1 percentage point of GDP.
  - Headline inflation remains elevated by 0.3 percentage points one year after the shock.
- Inflation expectations:
  - One-year-ahead inflation expectations rise by 1 percentage point in the current and the following year.
  - Five-year-ahead inflation expectations remain unchanged.
- Output gap and exchange rate:
  - Output gap increases by 0.38 percentage points in the current year.
  - Real exchange rate depreciates by 1 percent following the shock.
- Risk premium (CDS spreads):
  - One-year CDS spread increases by 6 basis points.
  - Five-year CDS spread increases by 11 basis points.
- Interpretation: aggregate demand channel is most influential; exchange rate depreciation and rising risk premiums contribute, particularly when confidence is eroded.

### Responses to alternative measures of fiscal deficit shocks
- Cyclically adjusted deficit shocks:
  - Inflation rises by 0.32 and 0.30 percentage points in the current and following year, respectively.
- Cyclically adjusted primary deficit shocks:
  - Inflation increases by 0.34 percentage points in both the current and following year.
- Channel nuances:
  - Risk premium response more muted following a primary structural deficit shock or cyclically adjusted primary deficit shock.
  - Exchange rate response more muted following a cyclically adjusted deficit shock.

### Responses to public debt shocks
- Inflation response:
  - A one-percentage-point-of-GDP shock to public debt → 0.25 percentage-point increase in inflation in the current year and 0.22 percentage-point increase in the following year.
- Output gap and exchange rate:
  - Output gap increases by 0.25 percentage points.
  - Real exchange rate depreciates by 1 percent.
- Debt-level-dependent amplification (debt ≥ 55 percent of GDP):
  - Real exchange rate depreciates by 12 percentage points.
  - One-year-ahead inflation expectations rise by 1.0 percentage point.
  - Five-year-ahead inflation expectations rise by 0.2 percentage points.
  - Five-year CDS spread increases by 54 basis points one year after the debt shock.
- When debt levels are low, responses are muted.

### Robustness and symmetry
- Results hold controlling for output gap and terms of trade.
- Patterns symmetric: unanticipated fiscal contractions → lower inflation; unexpected reductions in public debt → lower inflation.

### Quantitative model outputs (selected coefficients and diagnostics)
- Structural deficit shock (baseline, Year 0):
  - Headline inflation: 0.348*** (0.114)
  - Output gap: 0.108 (0.122)
  - One-year ahead inflation expectation: 0.099** (0.042)
  - Five-year ahead inflation expectation: 0.010 (0.013)
  - One-year CDS spread: 5.625** (2.642)
  - Five-year CDS spread: 10.607** (4.779)
  - Real exchange rate: -0.966** (0.464)
  - Observations: 45 for many variables; R-squared examples: 0.856 (headline inflation), 0.948 (1-yr expectation), 0.973 (5-yr expectation)
- Public debt shock (baseline, Year 0):
  - Headline inflation: 0.250** (0.094)
  - Output gap: 0.210** (0.091)
  - One-year ahead inflation expectation: 0.040 (0.032)
  - Five-year ahead inflation expectation: 0.012 (0.016)
  - One-year CDS spread: 1.861 (1.922)
  - Five-year CDS spread: 4.208 (3.207)
  - Real exchange rate: -0.976** (0.364)
  - Observations: 55 or 60 for many variables; R-squared examples: 0.806 (headline inflation), 0.921 (1-yr expectation), 0.959 (5-yr expectation)
- Responses to alternative fiscal measures (selected):
  - Cyclically adjusted deficit shock, Year 0 headline inflation: 0.317*** (0.093)
  - Primary structural deficit shock, Year 0 output gap: 0.423*** (0.088)
  - Cyclically adjusted primary deficit shock, Year 0 one-year inflation expectation: 0.100** (0.045)

### Methodology overview
- Local projection impulse response system with horizons h = 0, 1, 2 (years).
- Dependent variables: shocks to end-of-year headline inflation, output gap (percent of potential GDP), real effective exchange rate, one-year and five-year inflation expectations, one-year and five-year CDS spreads.
- Independent variable S_i,t: shock to fiscal deficit or public debt (actual − IMF WEO October forecast).
- Controls: one lag of shock, one and two lags of dependent variable, contemporaneous and one lag of real GDP growth and government debt-to-GDP, monetary policy shocks, country and year fixed effects.
- Debt-level interaction: HIGHDebt_i,t = 1 when public-debt-to-GDP ≥ 55 percent (baseline; robustness with 50 or 60 percent).

### Data sources and monetary shock identification
- Data: IMF WEO (fiscal deficit, public debt, output gap, real GDP growth); IMF IFS (real effective exchange rates); Consensus Economics (inflation expectations); monetary policy shocks estimated by Checo et al. (2024).
- Monetary shocks: analysts’ forecast errors of policy rate decisions; assumption that forecast errors reflect exogenous monetary policy variation.

### Policy implications and interpretation
- Fiscal stance materially affects inflation in LA5: deficit and debt shocks are inflationary with effects persisting up to one year.
- Loose fiscal policy and growing fiscal risks heighten inflationary pressures, complicating monetary policy normalization amid elevated inflation and external vulnerabilities.
- Monetary policy calibration should account for fiscal drivers and inflation risks; tighter fiscal policy can facilitate faster monetary easing and reduce fiscal/external imbalances.
- High public-debt regimes amplify financial market responses (notably CDS spreads) to debt shocks.

*Prepared by Sophia Chen (WHD); empirical details and data described in Annex I of the source.*

---

### A strong and credible policy framework is essential to effectively address the...
- Role of fiscal and monetary frameworks:
  - Robust fiscal framework: reassures markets of long-term fiscal discipline; prevents temporary fiscal expansions becoming unsustainable.
  - Credible monetary policy: anchors inflation expectations; supports cautious, data-driven policy; reinforces commitment to price stability.
- Empirical findings on inflation and expectations:
  - Actual inflation responds significantly to fiscal deficit and public debt shocks.
  - One-year-ahead inflation expectations react only moderately in countries with strong frameworks.
  - Five-year-ahead inflation expectations remain stable in countries with strong frameworks.
  - Interpretation: strong credibility and institutions can prevent an inflationary spiral by anchoring long-term expectations.

### Key quantitative responses (selected table outputs)
- Structural deficit shock (Table 1, baseline, Year 0):
  - Headline inflation: 0.348*** (0.114)
  - Output gap: 0.108 (0.122)
  - One-year ahead expectation: 0.099** (0.042)
  - Five-year ahead expectation: 0.010 (0.013)
  - One-year CDS spread: 5.625** (2.642)
  - Five-year CDS spread: 10.607** (4.779)
  - Real exchange rate: -0.966** (0.464)
  - Observations: 45 for many variables; R-squared examples: 0.856 (headline inflation)
- Public debt shock (Table 1, baseline, Year 0):
  - Headline inflation: 0.250** (0.094)
  - Output gap: 0.210** (0.091)
  - Real exchange rate: -0.976** (0.364)
  - Observations: 55 or 60; R-squared examples: 0.806 (headline inflation)

---

### Energy transition, fiscal impacts, and labor markets (selected findings)

### Global scenarios and methodology
- IEA scenarios used: Stated Policies Scenario (STEPS), Announced Pledges Scenario (APS), Net Zero Emissions by 2050 (NZE).
- Project-level modeling: FARI Methodology (IMF’s FARI project-level cash flow model); asset-level data (notably Rystad Energy for oil).
- Historical revenues: authorities’ reported royalty and income tax collections for 2015–2023.

### Oil production projections (IEA scenarios)
- Current estimated oil production: 736 thousand barrels per day (bpd).
- STEPS: 510 thousand bpd by 2030; 184 thousand bpd by 2040; 44 thousand bpd by 2050.
- APS: 460 thousand bpd by 2030; 150 thousand bpd by 2040; 35 thousand bpd by 250 (APS converges to 20 percent lower than STEPS in 2050).
- NZE: 350 thousand bpd by 2030; 18 thousand bpd by 2040; 3 thousand bpd by 2050.

### Fiscal implications of declining oil production
- Under NZE: royalty and income tax revenues from oil fall to 0.14 percent of GDP by 2030, near zero by 2050.
- Steep decline reduces Colombia’s royalties and income tax revenues by 0.9-1.2 ppts of GDP per year by 2030.
- Losses higher once lower dividends from Ecopetrol are factored in.

### Coal production projections and fiscal impact
- Mining production costs: US$40 to US$70 per ton; staff assumes operating cost US$30-45 per ton plus US$20 per ton transport.
- APS: Cerrejon mine closes 4 years earlier than STEPS; other mines closed by 2040.
- Under APS: government revenues from coal = 0.07 percent of GDP by 2030 and 0.01 percent of GDP in 2039 before cessation.
- Under NZE: all Colombian coal production assumed cease by 2032 when European coal price projected at US$54/ton.

### Subnational fiscal impacts
- Royalties distributed to subnational departments/municipalities by formula (population, poverty, unemployment).
- APS impacts: Antioquia, La Guajira, Huila see marked declines; Meta, Casanare, Cesar see direct royalty allocations decrease by 38 to 46 percent.

### Employment: direct and indirect impacts
- Jobs in oil and coal sectors in 2023: around 105,000 jobs (~1 percent of total labor force).
- Interlinkages:
  - Oil: support services, land transport, refining/fuel blending = 75 percent of total inputs; refining and gas distribution = 91 percent of outputs.
  - Coal: storage/transport, support services, refining, land transport = 70 percent of inputs; refining and electricity generation = >80 percent of outputs.
- APS preliminary estimates: annual job losses could average over 10,000 over next decade; oil sector most impacted.
- Indirect job losses ≈ half of total job losses in any year; main indirect losses in land transport, pipelines, support services, storage, refining, fuel blending, gas distribution.

### Electricity sector and renewable expansion
- Generation matrix: hydro ≈ two thirds of installed capacity; thermal ≈ 30 percent (mainly natural gas and coal); solar and wind ≈ 4 percent.
- If planned renewables completed by 2037, generation capacity could increase by almost 65 percent vs 2023; solar and wind main expansion sources.
- Staff uses average of four hypothetical generation-capacity scenarios over next 10 years.

### Employment opportunities from power sector expansion
- Current power sector employment: nearly 70,000 workers.
- Additional full-time jobs if planned renewables completed: 40,000 (subject to uncertainty).
  - 3.3 percent permanent (O&M).
  - ~65 percent of new positions from construction and installation.
  - ~one third of new employment from indirect employment.
- Average new job creation in subsequent decade: 22,000 per year (mainly construction/installation).
- Net impact in year T: ~24,000 new full-time jobs including construction/installation; excluding those implies net loss of 2.5 thousand jobs.
- Results driven by ambitious solar and wind expansion; absent such growth, positive impact lower.

### Regional heterogeneity in labor impacts
- Coal workers concentrated: La Guajira, Cesar, Boyacá, Cundinamarca.
- Oil workers concentrated: Santander, Meta, Casanare, Putumayo, Antioquia, Bogota.
- Wind potential concentrated: La Guajira; solar potential distributed across Santander, Tolima, Atlántico, Magdalena, Cesar, La Guajira.
- Subnational outcomes (conditional on projects proceeding):
  - La Guajira and Cesar: job losses could be offset by electricity sector jobs.
  - Antioquia and Tolima: likely benefit from hydroelectric and solar potential.
  - Boyacá, Huila, Norte de Santander, Bogotá: expected net job losses without compensation.
- Policy implication: assess skill gaps; regions facing net losses should diversify and strengthen social safety nets.

### Key social and policy considerations for transition
- Two key requirements for positive labor-market outcome:
  - Develop future skills.
  - Secure local community support.
- Monetary compensation alone insufficient; need targeted skill programs, community engagement, and effective communication.
- Short-run measures: leverage social safety nets; set up reskilling programs with conditional government support.
- Targeted policies: reskilling and unemployment insurance as part of active labor market policy for oil/coal workers and related sectors.

### Macroeconomic modeling of oil depletion and role of reforms (GIMF-GVC)
- Model: GIMF-GVC; baseline: oil production declines by 95 percent over 30 years; depletion fully anticipated.
- Key model results (relative to counterfactual of constant current production):
  - Real GDP 12 percent lower compared to counterfactual.
  - Potential growth declines by 0.4 percentage point annually over 30 years.
  - Declines in exports, consumption, and investment lower GDP by 6 percent, 5 percent, and 2 percent, respectively.
  - Capital in oil, coal, energy declines by 20 percent of GDP; with capital income share 23 percent, explains ~half of long-term GDP losses.
  - Employment in oil, coal, energy falls sharply; total labor supply reduced by 0.8 percent as workers shift sectors.
  - GDP losses: 9 percent attributable to lower labor productivity overall; 3 percent attributable to labor reallocation away from oil, coal, energy.

### Real exchange rate and fiscal dynamics
- Real exchange rate weakens by about 12 percent over same period; including “news” of future depletion, total depreciation is 21 percent.
- Export losses dominated by oil/coal/energy; depreciation helps tradables exports but small vs losses.
- Government revenue losses ≈ 3 percent of GDP; under fiscal rule government debt remains broadly stable but reduced transfers lower consumption among liquidity-constrained households.

### Reform packages and outcomes
- Reform packages:
  - Domestic reforms (30 years): higher TFP (5 percent), higher labor productivity, lower markups.
  - Financial reforms (15 years): lower verification costs and borrower riskiness by 100 bps in tradables and non-tradables.
  - Export reforms (30 years): increase Colombia’s trade weights in partners’ import demand for tradables by 50 percent.
- Impacts:
  - Domestic reforms: GDP decline reduced from 12 percent to 4 percent over 30 years; government debt falls from ~53 percent of GDP to ~49 percent of GDP; non-tradable productivity reduces domestic inflation and further depreciates real exchange rate.
  - Financial reforms: GDP ~1.5 percent higher after 30 years vs no-policy scenario (about 10.5 percent below baseline); stronger investment effects.
  - Export reforms: GDP 9.5 percent smaller than baseline after 30 years; exchange rate declines by 4.5 percent instead of 12 percent.
- Combined reforms:
  - Real GDP recovers beyond depletion losses: instead of 12 percent loss, real GDP would be 0.8 percent larger than baseline.
  - Consumption and investment exceed baseline; about 40 percent of export losses compensated by higher tradables exports.
  - Peso depreciates by ~13 percent (similar to no-reform depreciation).

### Financing the energy transition (Annex I highlights)
- Additional investment needed by 2050: about US$92 billion.
- This equals 1.2 percent of Colombia’s discounted cumulative GDP over 2030-2050.
- Private sector would need to contribute at least 80 percent of additional investment needs.

### ESG finance landscape and market facts
- Institutional strengthening: National Climate Change System, Regional Climate Change Nodes, SFC’s biennial ESG survey (since 2019).
- Green portfolio: 2.4 percent of total loans (as of 2023).
- ESG bonds (2020-2024): Colombia issued US$15.3 billion—nearly 60 percent social bonds.
- Issuers: ~80 percent of ESG bonds issued by government; private sector 18 percent.
- Colombia’s ESG bond market = 6 percent of total ESG issuance in LAC (underdeveloped vs peers).

### Effects of ESG labeling and disclosure on bond yields
- Matched-sample (2,105 bonds, 2016-2024): ESG bonds have yields to maturity (YTM) on average 100 basis points lower than matched non-ESG bonds (ATT = -1.0 in Annex I Table 1, Full Sample).
- Temporal patterns:
  - Before Covid: no significant difference (ATT = 0.24).
  - After Covid: ESG bonds yield 190 basis points less (ATT = -1.9).
- By issuer type: strongest effect among corporate issuers (ATT = -0.97, P-Value 0.04); no significant difference for government or supranational bonds in full sample table.
- Selected table entries:
  - Full Sample: ESG bond 6.38; Non-ESG bond 7.38; ATT -1.0; T-Statistics -2.51; P-Value 0.01.
  - After Covid: ESG bond 6.05; Non-ESG bond 7.96; ATT -1.9; T-Statistics -3.65; P-Value 0.0.
  - Corporate: ESG bond 6.77; Non-ESG bond 7.73; ATT -0.97; T-Statistics -2.02; P-Value 0.04.

### Role of ESG disclosure frameworks
- Composite disclosure score (PCA of disclosure approach, applicability, assurance requirements, materiality) associated with lower ESG bond spreads.
- Regression results: stronger disclosure frameworks associated with average reduction of 33 basis points in ESG bond spreads (Panel A: Disclosure -0.33***).
- Subsample magnitudes:
  - Panel A (Full Sample): Disclosure -0.33*** (t-stat -6.19); Observations 1794; R-Squared 0.41.
  - Panel A (After Covid): Disclosure -0.6*** (t-stat -8.35); Observations 1017; R-Squared 0.47.
  - Panel B (Domestic Currency): Disclosure -0.65*** (t-stat -7.16); Observations 494; R-Squared 0.52.
  - Panel B (High Yield): Disclosure -0.56*** (t-stat -3.92); Observations 375; R-Squared 0.33.

### Policy recommendations to strengthen ESG financial market
- Ensure macro-financial stability to reduce investor risk premia.
- Improve regulatory framework: standardized reporting and third-party verification to enhance transparency.
- Expand issuer participation: support municipalities and SMEs via credit enhancement and technical assistance to lower borrowing costs.
- Foster regional and international cooperation: integrate with Latin American sustainable finance platforms and partnerships with multilaterals for technical and financial support.

---

### The decline in labor force participation (LFP) and macro implications

### Summary of trends and components (2000-19)
- Colombia’s labor grew on average 2.3 percent per year—0.3 ppts faster than LA8 peers.
- Components (2000-19):
  - Population growth: 1.3 percent in Colombia; 1.1 percent in LA8.
  - Working-age population: 0.7 percent in Colombia; 0.6 percent in LA8.
  - Employment growth: 0.5 percent in Colombia; 0.3 percent in LA8.
  - Labor force participation growth: -0.3 percent in Colombia (lowest among LA8).
- Over two decades to pandemic, labor dynamics contributed 0.7 percentage points to annual real growth; projected negative population changes reduce long-term labor contribution by 0.1-0.2 ppts.

### Data and methodology
- Data: Colombian household survey (Gran Encuesta Integrada de Hogares) with migration modules; >576,000 observations for 2023; 54 percent women.
- Method: Probit regressions for LFP (women and men) with household and individual controls; robustness checks for 2019, 2024, and 2013–2019.

### Key quantitative findings (2023 marginal contributions to individual participation probability)
- Female drivers:
  - head: 19.0%
  - education: 9.7%
  - ven_migrant: 8.8%
  - children: 3.8%
  - age: 0.2%
  - married: -1.9%
  - elders: -2.4%
  - rural: -5.6%
  - young_children: -6.4%
- Male drivers:
  - married: 16.1%
  - ven_migrant: 12.8%
  - head: 8.3%
  - young_children: 6.7%
  - rural: 5.8%
  - education: 3.5%
  - age: 0.4%
  - elders: -1.4%
  - children: -1.7%
- Childcare and eldercare:
  - Young children (<5) reduce women’s participation by ~6 percentage points.
  - Elders living at home reduce women’s participation by ~2 percentage points.

### Venezuelan migrants and LFP
- Venezuelan migrants: average age 28 vs Colombians 32; ~74 percent actively sought a job vs 63 percent of locals.
- After controls, being a Venezuelan migrant increases probability of participating by ~10 percentage points on average (9 ppts for women; 12 ppts for men).
- Migration may have heterogeneous effects across skill and gender groups.

### Counterfactuals and macro impacts
- If LFP returned to its peak: long-term growth could increase by 0.3 ppts.
- If gender gap narrowed to 10 percent: labor contribution could boost long-term growth by 0.6 ppts.
- Better migrant integration adds further gains.
- Increasing LFP would reduce pension funding pressures.

### Policy recommendations to raise participation
- Enhance work flexibility, especially for women with caregiving responsibilities.
- Expand childcare and eldercare services to mitigate negative effects on female LFP.
- Integrate migrants into formal labor markets: expand access, active labor market policies, job search assistance, labor protections, formalization to reduce misallocation.
- Accelerate and strengthen labor market integration measures to reap migration gains.

*Source: IMF chapter text and associated figures and tables as provided in the content unit.*

### 1.   Responses to Structural Deficit Shocks _______________________________________________ 8

### 1.   Responses to Structural Deficit Shocks

### Summary and Objective
- Examines the extent to which expansionary fiscal policy poses a challenge to containing inflation in financially integrated Latin American economies (LA5: Brazil, Chile, Colombia, Mexico, Peru) during 2010-24.
- Uses an open-economy New Keynesian Phillips curve framework and local projections to assess dynamic responses of inflation and its drivers to shocks to the end-of-year structural deficit and public debt (both expressed as a percentage of the previous year’s GDP).
- Fiscal deficit and public debt shocks are measured as the difference between actual year-end values and IMF World Economic Outlook October forecasted values. Headline inflation shocks are analogously defined as differences between actual and forecasted end-of-year headline inflation.

### Transmission Channels Highlighted
- Aggregate demand channel: fiscal expansion boosts domestic demand, exerting upward pressure on prices—particularly non-tradables—when short-run supply is relatively inelastic.
- Inflation expectations channel: expectations may de-anchor if the central bank is perceived as reluctant to tighten policy due to concerns that higher interest rates raise public debt servicing costs.
- Exchange rate and confidence channel: while higher domestic demand and interest rates can cause appreciation, erosion of investor confidence can dominate, raising the risk premium and triggering currency depreciation.
- Additional amplification in emerging markets due to greater exposure to external shocks, weaker fiscal and central bank credibility, and persistent currency mismatches.

### Empirical Specification and Controls
- Estimation includes two lags of the dependent variable and contemporaneous and one lag of real GDP growth and public debt to capture business cycle dynamics and debt stock influences.
- Incorporates monetary policy shocks measured as analysts’ forecast errors of policy rates.
- Baseline fiscal measures: end-of-year structural deficit and public debt (% of prior year GDP). Alternative measures examined: structural primary deficit, cyclically adjusted deficit, cyclically adjusted primary deficit.
- Identification via forecast errors (actual minus IMF October forecasts) to mitigate reverse causality and spurious correlation concerns.

### Key Empirical Findings — Structural Deficit Shocks
- Headline inflation response:
  - Headline inflation rises by 0.35 percentage points following a fiscal deficit shock equivalent to 1 percentage point of GDP.
  - Headline inflation remains elevated by 0.3 percentage points one year after the shock.
- Inflation expectations:
  - One-year-ahead inflation expectations rise by 1 percentage point both in the current and the following year.
  - Five-year-ahead inflation expectations remain unchanged.
- Output gap and exchange rate:
  - Output gap increases by 0.38 percentage points in the current year.
  - Real exchange rate depreciates by 1 percent following the shock.
- Risk premium (CDS spreads):
  - One-year CDS spread increases by 6 basis points.
  - Five-year CDS spread increases by 11 basis points.
- Interpretation: aggregate demand channel is most influential; exchange rate depreciation and rising risk premiums contribute, particularly when confidence is eroded.

### Responses to Alternative Measures of Fiscal Deficit Shocks
- Cyclically adjusted deficit shocks:
  - Inflation rises by 0.32 and 0.30 percentage points in the current and following year, respectively.
- Cyclically adjusted primary deficit shocks:
  - Inflation increases by 0.34 percentage points in both the current and following year.
- Channel responses:
  - Comparable output gap, exchange rate, and expectations responses as baseline, with some nuances:
    - Risk premium response is more muted following a primary structural deficit shock or a cyclically adjusted primary deficit shock.
    - Exchange rate response is more muted following a cyclically adjusted deficit shock.

### Responses to Public Debt Shocks
- Inflation response:
  - A one-percentage-point-of-GDP shock to public debt is linked to a 0.25 percentage-point increase in inflation in the current year and a 0.22 percentage-point increase in the following year.
- Output gap and exchange rate:
  - Output gap increases by 0.25 percentage points.
  - Real exchange rate depreciates by 1 percent.
- Debt-level-dependent amplification:
  - When debt level is at or above 55 percent of GDP, responses are larger:
    - Real exchange rate depreciates by 12 percentage points.
    - One-year-ahead inflation expectations rise by 1.0 percentage point.
    - Five-year-ahead inflation expectations rise by 0.2 percentage points.
    - Five-year CDS spread increases by 54 basis points one year after the debt shock.
  - When debt levels are low, responses are muted.

### Robustness and Additional Results
- Results hold across different cyclical positions of the economy (controlling for output gap) and when controlling for external trade positions (terms of trade).
- Patterns are symmetric for positive and negative shocks: unanticipated fiscal contractions are associated with lower inflation; unexpected reductions in public debt are associated with lower inflation.

### Policy Implications and Interpretation
- Fiscal policy stance has significant implications for inflation in the LA5 sample: shocks to the fiscal deficit and public debt are inflationary with effects persisting up to one year.
- Loose fiscal policy and increasing fiscal risks can heighten inflationary pressures, complicating monetary policy normalization especially amid elevated inflation and external vulnerabilities.
- Calibration of monetary policy should account for fiscal drivers and risks of inflation; tighter fiscal policy can facilitate faster monetary easing and reduce fiscal and external imbalances, while expansionary fiscal policy can limit scope for monetary policy normalization and exacerbate confidence and exchange rate pressures.

*Prepared by Sophia Chen (WHD); empirical details and data described in Annex I of the source.*

### 14. A strong and credible policy framework is essential to effectively address the

### 14. A strong and credible policy framework is essential to effectively address the

### Role of fiscal and monetary frameworks
- Robust fiscal framework:
  - Reassures markets of long-term fiscal discipline.
  - Helps ensure temporary fiscal expansions do not become unsustainable.
- Credible monetary policy:
  - Anchors inflation expectations.
  - Supports a cautious and data driven monetary policy stance.
  - Reinforces commitment to price stability.

### Empirical findings on inflation and expectations
- Actual inflation responds significantly to fiscal deficit and public debt shocks.
- One-year-ahead inflation expectations react only moderately in countries with strong frameworks.
- Five-year-ahead inflation expectations remain stable in countries with strong frameworks.
- Interpretation: Strong credibility and institutions can help prevent an inflationary spiral by anchoring long-term inflation expectations, even amid near-term fiscal pressures.

### Key quantitative responses (summary of model outputs)
- Structural deficit shock (Table 1, baseline estimates):
  - Year 0 headline inflation: 0.348*** (standard error 0.114)
  - Year 0 output gap: 0.108 (0.122)
  - Year 0 one-year ahead inflation expectation: 0.099** (0.042)
  - Year 0 five-year ahead inflation expectation: 0.010 (0.013)
  - Year 0 one-year CDS spread: 5.625** (2.642)
  - Year 0 five-year CDS spread: 10.607** (4.779)
  - Year 0 real exchange rate: -0.966** (0.464)
  - Observations: 45 for many variables; R-squared examples: 0.856 (headline inflation), 0.948 (1-yr expectation), 0.973 (5-yr expectation)
- Public debt shock (Table 1, baseline estimates):
  - Year 0 headline inflation: 0.250** (0.094)
  - Year 0 output gap: 0.210** (0.091)
  - Year 0 one-year ahead inflation expectation: 0.040 (0.032)
  - Year 0 five-year ahead inflation expectation: 0.012 (0.016)
  - Year 0 one-year CDS spread: 1.861 (1.922)
  - Year 0 five-year CDS spread: 4.208 (3.207)
  - Year 0 real exchange rate: -0.976** (0.364)
  - Observations: 55 or 60 for many variables; R-squared examples: 0.806 (headline inflation), 0.921 (1-yr expectation), 0.959 (5-yr expectation)
- Responses to alternative fiscal deficit measures (Table 2, selected estimates):
  - Cyclically adjusted deficit shock, Year 0 headline inflation: 0.317*** (0.093)
  - Primary structural deficit shock, Year 0 output gap: 0.423*** (0.088)
  - Cyclically adjusted primary deficit shock, Year 0 one-year inflation expectation: 0.100** (0.045)
  - R-squared examples: 0.868 (headline inflation for cyclically adjusted deficit), 0.960 (1-yr expectation for cyclically adjusted primary deficit)
- Public debt shocks by debt level (Table 3, interaction with high debt dummy):
  - Baseline high-debt dummy threshold: 55 percent of GDP (robustness tests use 50 or 60 percent).
  - Year 0 public debt shock coefficient: 0.200 (0.185)
  - Year 0 public debt shock * high debt: -0.443 (1.165)
  - Year 1 public debt shock * high debt: 1.299 (1.073)
  - Year 2 public debt shock * high debt: 1.519 (1.321)
  - Notable large point estimates for CDS spreads when interacting with high debt dummy (e.g., Year 0 five-year CDS interaction: 47.759* (24.316); Year 2 one-year CDS interaction: 27.557** (10.950); Year 2 five-year CDS interaction: 57.204*** (19.079))

### Methodology overview
- Estimation approach:
  - Local projection impulse response system (equations (1) and (2)) with horizons ℎ = 0, 1, 2 (years after t).
  - Dependent variables: shocks to end-of-year headline inflation, output gap (percent of potential GDP), real effective exchange rate, one-year and five-year inflation expectations, and one-year and five-year CDS spreads.
  - Independent variable of interest, S_i,t: shock to fiscal deficit or public debt defined as difference between actual year-end value and IMF WEO October forecast.
  - Baseline fiscal deficit measure: structural deficit as percent of previous year’s GDP. Alternative measures: structure primary deficit, cyclically adjusted deficit, cyclically adjusted primary deficit.
  - Public debt measured as gross public debt as percent of previous year’s GDP.
  - Controls: one lag of the shock, one and two lags of dependent variable, contemporaneous and one lag of real GDP growth and government debt-to-GDP ratio, monetary policy shocks, country and year fixed effects.
  - Variables winsorized at the 2nd and 98th percentiles to limit outlier influence.
- Debt-level interaction specification:
  - Equation (2) includes a dummy HIGHDebt_i,t equal to 1 when public-debt-to-GDP ratio is at or above 55 percent in baseline; robustness uses 50 or 60 percent.

### Data sources and identification of monetary shocks
- Data sources:
  - Fiscal deficit, public debt, output gap, real GDP growth: IMF’s World Economic Outlook database.
  - Real effective exchange rates: IMF’s International Financial Statistics.
  - Inflation expectations: Consensus Economics.
  - Monetary policy shocks: estimated by Checo et al. (2024).
- Monetary shock identification:
  - Constructed using analysts' forecast errors of policy rate decisions, following high-frequency data approaches.
  - Identification assumption: analysts' forecasts incorporate endogenous monetary responses; forecast errors reflect exogenous variation in monetary policy decisions.

### Implications and interpretation
- Strong policy credibility and institutions mitigate pass-through from fiscal shocks to long-term inflation expectations.
- Fiscal expansions and rising public debt materially affect headline inflation, output gap, sovereign spreads, and the real exchange rate—effects are heterogeneous depending on debt levels and the measure of fiscal shock.
- High public-debt regimes amplify some financial market responses (notably CDS spreads) to public debt shocks according to interaction estimates.

*Source: IMF chapter text and associated figures and tables as provided in the content unit.*

### 8. The proposed transition is taking place in the context of considerable global

### 8. The proposed transition is taking place in the context of considerable global uncertainty regarding future fossil fuel usage

### Global scenarios and methodology
- Staff analysis uses the International Energy Agency (IEA) forecasts of global oil, gas, and coal production and prices under three climate change mitigation scenarios:  
  - Stated Policies Scenario (STEPS) — outlook based on the latest policy settings.  
  - Announced Pledges Scenario (APS) — assumes all national energy and climate targets are met in full and on time.  
  - Net Zero Emissions by 2050 (NZE) — assumes mitigation efforts which limit global warming to 1.5 °C.  
- Project-level (bottom-up) modeling methodology used: FARI Methodology (IMF’s FARI project-level cash flow model), modeling each major coal mine and oil field individually using asset-level data (notably from Rystad Energy for oil) to assess profitability and shutdown timing as prices decline.  
- Historical revenues reflect authorities’ reported royalty and income tax collections for the years 2015–2023.

### Oil production projections (under IEA scenarios)
- Current estimated oil production: 736 thousand barrels per day (bpd).  
- STEPS scenario projection: production declines towards 510 thousand bpd by 2030, 184 thousand bpd by 2040, and 44 thousand bpd by 2050.  
- APS scenario projection: production would be 460 thousand bpd by 2030, 150 thousand bpd by 2040, and 35 thousand bpd by 250, with APS converging to a production level 20 percent lower than STEPS in 2050.  
- NZE scenario projection: production drops to 350 (thousand bpd) by 2030, 18 thousand bpd by 2040, and to only 3 thousand bpd by 2050.

### Fiscal implications of declining oil production
- Under NZE, royalty and income tax revenues from oil production would fall to 0.14 percent of GDP by 2030, and to near zero by 2050.  
- Steep decline in oil production would reduce Colombia’s royalties and income tax revenues by 0.9-1.2 ppts of GDP per year by 2030.  
- Losses would be higher once lower dividends from Ecopetrol are factored in.

### Coal production projections and fiscal impact
- Mining production costs in Colombia range from US$40 to US$70 per ton. Staff analysis assumes an operating cost in the range of US$30-45 per ton, with an additional US$20 per ton in transportation costs.  
- APS scenario accelerates decline with early closure of the Cerrejon mine (4 years earlier than in STEPS) and all other mines by 2040.  
- Under APS, government revenues from coal amount to 0.07 percent of GDP by 2030 and 0.01 percent of GDP in 2039 before mining activity ceases.  
- Under NZE, all Colombian coal production is assumed to cease by 2032, when the European coal price is projected to fall to US$54/ton.

### Subnational fiscal impacts
- Royalties from coal and oil are distributed to subnational departments and municipalities according to a revenue sharing formula that factors population, poverty, and unemployment.  
- Analysis under APS shows regions like Antioquia, La Guajira and Huila will see more marked declines due to mine closures and oil field depletion.  
- Major oil and coal regions such as Meta, Casanare and Cesar would see decreases in direct allocation of royalties in the range of 38 to 46 percent.

### Employment: direct and indirect impacts
- Estimated jobs in oil and coal sectors in 2023: around 105,000 jobs, accounting for about 1 percent of Colombia’s total labor force.  
- This excludes employment in related upstream and downstream industries.  
- Interlinkages:  
  - For oil: support services for extractive industries, land transportation, and oil refining and fuel blending account for 75 percent of the total value of inputs; downstream, oil refining and fuel blending and gas distribution account for 91 percent of total value of outputs.  
  - For coal: storage and complementary transport services, support services for extractive industries, oil refining and fuel blending, land transport and vehicle repair and maintenance account for 70 percent of inputs; oil refining and electricity generation account for more than 80 percent of total outputs.  
- Preliminary estimates (APS production paths): annual job losses could average over 10,000 over the next decade, with the oil sector most impacted. Indirect job losses account for roughly half of total job losses in any given year. Most indirect job losses occur in land transport and pipelines, support services for extractive industries, storage and complementary transport, oil refining, fuel blending, and gas distribution.

### Electricity sector and renewable energy expansion
- Colombia’s generation matrix: hydroelectric production accounts for about two thirds of total installed capacity; thermal energy accounts for 30 percent of total installed capacity (mainly from natural gas and coal); solar and wind account for the remaining 4 percent.  
- Mining and Energy Planning Unit estimates: by 2037, generation capacity could increase by almost 65 percent with respect to 2023 if renewable projects in the indicative expansion plan (2023-2037) are completed timely. Solar and wind would be the main source of expansion.  
- Staff uses an average of four hypothetical scenarios of electricity generation capacity over the next 10 years to assess impacts.

### Employment opportunities from power sector expansion
- Current employment in power sector: nearly 70,000 workers.  
- Estimates: an additional 40,000 full-time jobs could be created if planned renewable projects are completed (subject to uncertainty).  
  - Only 3.3 percent of these would be permanent employments (operation and maintenance, O&M).  
  - About 65 percent of new job positions would come from construction and installation of plants.  
  - About one third of total new employment would come from indirect employment.  
- Average new job creation in the subsequent decade could reach 22,000 per year, mainly linked to construction and installation.  
- Net impact in year T: staff analysis suggests around 24,000 new full-time jobs when construction and installation jobs are included; excluding construction and installation jobs implies a net loss of 2.5 thousand jobs.  
- Results are driven by ambitious expansion of solar and wind; in absence of such growth, the positive impact would be lower.

### Regional heterogeneity in labor market impacts
- Coal workers concentrated in La Guajira, Cesar, Boyacá, and Cundinamarca.  
- Oil workers concentrated in Santander, Meta, Casanare, Putumayo, Antioquia, and Bogota.  
- Wind potential concentrated in La Guajira; solar potential more distributed across Santander, Tolima, Atlántico, Magdalena, Cesar, and La Guajira (Angel-Sanint et al. (2023)).  
- Subnational analysis (subject to uncertainties and assumption that renewable projects proceed as planned):  
  - La Guajira and Cesar: job losses could be offset by electricity sector jobs if projects proceed.  
  - Antioquia and Tolima: likely to benefit from hydroelectric and solar potential.  
  - Boyacá, Huila, Norte de Santander, and Bogotá: expected to face net job losses without compensation from the electricity sector.  
- Authorities should assess skill gaps to enable workforce participation in new opportunities; regions facing net losses should focus on diversification and strengthening social safety nets.

### Key issues, policy recommendations, and social considerations
- Two key requirements for a positive labor-market outcome from the energy transition: developing future skills and securing local community support.  
- Monetary compensation alone (e.g., 'Electricity Transfer') would be insufficient; targeted skill programs, tailored support addressing community concerns, and effective communication are necessary to maximize local benefits.  
- Short-run measures: leverage social safety net to limit impacts on vulnerable households; set up reskilling programs to ensure medium-term labor force reintegration. Government support would be most effective if conditional on participation in reskilling programs.  
- Targeted policies for workers from oil and coal industries and related sectors: reskilling programs and unemployment insurance as part of a comprehensive active labor market policy.

### Macroeconomic modeling of oil depletion and role of reforms (GIMF-GVC)
- Model: Global Integrated Monetary and Fiscal Model (GIMF-GVC) used to model macroeconomic impacts of oil depletion over the next 30 years and policies to reduce losses.  
- Baseline scenario: oil production declines by 95 percent over 30 years; depletion is fully anticipated by forward-looking agents; impulse responses are calculated as deviations from a counterfactual where oil production remains constant at current levels.  
- Key model results:  
  - Real GDP would be 12 percent lower compared to the counterfactual scenario in which oil production remains constant at current production levels.  
  - Potential growth would decline by 0.4 percentage point annually over 30 years.  
  - Declines in exports, consumption, and investment lower GDP by 6 percent, 5 percent, and 2 percent, respectively.  
  - Capital in the oil, coal, and energy sector would decline by 20 percent of GDP. With a capital income share of 23 percent, this explains around half of the long-term losses in GDP.  
  - Employment in oil, coal, and energy sectors projected to fall sharply, reducing total labor supply by 0.8 percent as workers shift to tradables and nontradables.  
  - GDP losses of 9 percent are attributable to lower labor productivity overall while 3 percent are attributable to labor reallocation away from oil, coal, and energy.

*Source: IMF staff estimates and analysis as presented in the supplied content.*

### 24. In addition, the real exchange rate would weaken by about 12 percent over the same

### 24. In addition, the real exchange rate would weaken by about 12 percent over the same

### Macroeconomic impact of oil depletion and baseline dynamics
- Real exchange rate would weaken by about 12 percent over the same period.
- When including the “news” effects of future depletion, the total depreciation is 21 percent.
- Export losses from oil, coal, and energy dominate; depreciation helps boost tradables exports, but the effect is small compared to those losses.
- Government revenue losses are about 3 percent of GDP; under the fiscal rule government debt remains broadly stable but reduced transfers lower consumption among liquidity-constrained households.

### Policy reform scenarios considered
- Three reform packages considered:
  - Domestic reform package implemented over a 30-year period that assumes higher TFP (5 percent) and labor productivity and lower markups in tradables and non-tradables.
  - Financial reform package implemented over a 15-year period aimed at lowering verification costs and borrower riskiness (by 100 bps) in tradables and non-tradables.
  - Export reform package implemented over a 30-year period that increases Colombia’s trade weights in partners’ import demand for tradables by 50 percent.

### Impact of individual reforms (key outcomes)
- Domestic reforms (30 years, higher productivity and lower markups):
  - Instead of losses of 12 percent, GDP would decline by only 4 percent over the next 30 years.
  - Higher consumption and investment due to positive income effects and higher returns to capital in tradables and non-tradables.
  - Government debt declines from around 53 percent of GDP to around 49 percent of GDP.
  - Higher productivity and lower markups in non-tradables push domestic inflation down, further depreciating the real exchange rate and supporting the current account balance over much of the reform period.
- Financial reforms (15 years, lower verification costs and borrower riskiness by 100 bps):
  - Compared to the oil depletion with no policy scenario, GDP would be around 1.5 percent higher after 30 years, at around 10.5 percent below baseline GDP.
  - Improvement arises from higher consumption and investment in broadly equal proportion; relatively stronger investment than under domestic reforms.
  - Impact on the real exchange rate is muted compared to domestic reform scenario.
- Export reforms (30 years, increase trade weights by 50 percent):
  - GDP is 9.5 percent smaller than baseline after 30 years, with relatively larger contributions from consumption and exports.
  - Reforms increase imports due to stronger demand for intermediate goods in tradables.
  - Exchange rate support: instead of falling by 12 percent, declines by 4.5 percent after 30 years.

### Combined reforms
- When all three reform packages are implemented together:
  - Real GDP recovers beyond the losses from oil depletion—instead of 12 percent losses, real GDP would be 0.8 percent larger than the baseline.
  - Strong responses of consumption and investment, both exceeding their baseline levels.
  - About 40 percent of export losses from oil depletion are compensated by higher tradables exports.
  - The peso depreciates by around 13 percent, broadly similar to the depreciation in the no-reform scenario.

### Financing the energy transition (Annex I highlights)
- Colombia would need about US$92 billion in additional investment by 2050 to finance the energy transition.
- This amount represents 1.2 percent of the country’s discounted cumulative GDP over 2030-2050.
- Public sector resources would be limited; staff estimates that the private sector would need to contribute at least 80 percent of Colombia’s additional investment needs.

### ESG finance landscape and market facts
- Institutional strengthening: National Climate Change System, Regional Climate Change Nodes, and SFC’s biennial ESG and climate risks and opportunity survey (since 2019) have advanced ESG finance capacity.
- As of 2023, the green portfolio accounted for 2.4 percent of total loans.
- Between 2020 and 2024, Colombia issued US$15.3 billion in ESG bonds—nearly 60 percent in social bonds.
- Around 80 percent of ESG bonds were issued by the government; the private sector accounted for 18 percent of total issuances.
- Colombia’s ESG bond market is underdeveloped compared to regional peers, accounting for only 6 percent of total ESG issuance in LAC.

### Effects of ESG labeling and disclosure on bond yields
- Matched-sample analysis of 2,105 bonds (2016-2024) shows ESG bonds have yields to maturity (YTM) that are, on average, 100 basis points lower than matched non-ESG bonds (ATT = -1.0 in Annex I Table 1, Full Sample).
- Temporal patterns:
  - Before Covid: no significant difference (ATT = 0.24).
  - After Covid: ESG bonds yield 190 basis points less (ATT = -1.9).
- By issuer type: effect strongest among corporate issuers (ATT = -0.97, P-Value 0.04); no significant difference for government or supranational bonds in the full-sample table.
- Annex I Table 1 (selected entries):
  - Full Sample: ESG bond 6.38; Non-ESG bond 7.38; ATT -1.0; T-Statistics -2.51; P-Value 0.01.
  - After Covid: ESG bond 6.05; Non-ESG bond 7.96; ATT -1.9; T-Statistics -3.65; P-Value 0.0.
  - Corporate: ESG bond 6.77; Non-ESG bond 7.73; ATT -0.97; T-Statistics -2.02; P-Value 0.04.

### Role of ESG disclosure frameworks
- A composite disclosure framework score (Principal Component Analysis of four dimensions: disclosure approach, applicability, assurance requirements, materiality perspective) is associated with lower ESG bond spreads.
- Regression results show stronger ESG disclosure frameworks are significantly associated with an average reduction of 33 basis points in ESG bond spreads (Panel A: Disclosure -0.33***).
- Subsample analyses indicate the reduction is more pronounced post-COVID, and especially for agency issuers, domestic currency bonds, and high-yield instruments.
- Annex I Table 2 (selected coefficients):
  - Panel A (Full Sample): Disclosure -0.33*** (t-stat -6.19); Observations 1794; R-Squared 0.41.
  - Panel A (After Covid): Disclosure -0.6*** (t-stat -8.35); Observations 1017; R-Squared 0.47.
  - Panel B (Domestic Currency): Disclosure -0.65*** (t-stat -7.16); Observations 494; R-Squared 0.52.
  - Panel B (High Yield): Disclosure -0.56*** (t-stat -3.92); Observations 375; R-Squared 0.33.

### Policy recommendations to strengthen ESG financial market
- Ensure macro-financial stability to reduce investor risk premia.
- Improve regulatory framework: standardized reporting and third-party verification to enhance transparency.
- Expand issuer participation: support municipalities and SMEs through credit enhancement and technical assistance to lower borrowing costs.
- Foster regional and international cooperation: integrate with Latin American sustainable finance platforms and partnerships with multilateral institutions for technical and financial support.

### Labor force participation (summary highlights)
- Colombia’s labor force participation (LFP) rose to 68 percent in 2010-12 and declined to around 63 percent thereafter.
- Male LFP oscillated around 80 percent; female LFP rose to 55 percent mid-decade and fell to 52 percent thereafter.
- Female LFP starts to drop sharply at 45-49 years old—10 years earlier than male LFP.
- Migration explains only a small fraction of the LFP decline; literature finds small or no significant effects (e.g., Bonilla and others (2024) find a reduction of 0.3 percentage points).
- Colombia’s gender gap in LFP is almost 25 percent, second only to Mexico among peers; peers like Peru and Uruguay have gender gaps around 14-15 ppts.

*Source: IMF staff estimates.*

### 4. The decline in LFP has hampered Colombia’s labor and real GDP growth over the past

### 4. The decline in LFP has hampered Colombia’s labor and real GDP growth over the past

### Summary of trends
- During 2000-19, Colombia’s labor has grown on average 2.3 percent per year—0.3 ppts faster than in its Latin American peers (LA8 countries).
- Components of labor growth (2000-19):
  - Population growth: 1.3 percent in Colombia; 1.1 percent on average in LA8.
  - Working-age population (“demographics”): 0.7 percent in Colombia; 0.6 percent in LA8.
  - Employment growth: 0.5 percent in Colombia; 0.3 percent in LA8.
  - Labor force participation growth: -0.3 percent in Colombia (the lowest among LA8 countries).
- In the two decades up to the pandemic, labor dynamics contributed on average with 0.7 percentage points to Colombia’s annual real growth rate; going forward negative population changes are expected to reduce the labor contribution to long-term growth by 0.1-0.2 ppts.

### Data, methodology, and robustness
- Data: Microdata from the Colombian household survey (Gran Encuesta Integrada de Hogares) with migration modules for individuals 15-64 years old; over 576 thousand observations for 2023; 54 percent of surveyed individuals were women.
- Method: Probit regressions of labor force participation (LFP) for women and men including household characteristics (head, married), young children, elders, education, Venezuelan migrant status, age, rural/urban domain, and per capita homicide rate; robustness checks include regressions for 2019 and 2024 and pre-pandemic years 2013–2019 (with some variable exclusions where data limited).
- For the pre-pandemic period (2013–2019), most explanatory variables have similar marginal effects; the marginal effect of “children” declined from 0.0440 in 2013 to 0.0287 in 2019.

### Key quantitative findings (marginal contributions and effects)
- Female drivers of probability of participating in the labor force (2023, marginal contribution to an individual’s probability):
  - head: 19.0%
  - education: 9.7%
  - ven_migrant: 8.8%
  - children: 3.8%
  - age: 0.2%
  - married: -1.9%
  - elders: -2.4%
  - rural: -5.6%
  - young_children: -6.4%
- Male drivers of probability of participating in the labor force (2023, marginal contribution):
  - married: 16.1%
  - ven_migrant: 12.8%
  - head: 8.3%
  - young_children: 6.7%
  - rural: 5.8%
  - education: 3.5%
  - age: 0.4%
  - elders: -1.4%
  - children: -1.7%
- Childcare and eldercare effects:
  - Having young children (less than 5 years old) reduces women’s probability of staying in the Colombian labor force by around 6 percentage points.
  - Having elders living at home reduces women’s probability by an additional 2 percentage points.
- Venezuelan migrants:
  - Venezuelan migrants appear more educated and younger (28 years old on average) than Colombians (32 years old on average).
  - About 74 percent of Venezuelan migrants actively sought a job, compared with 63 percent of locals.
  - After controlling for characteristics, being a Venezuelan migrant boosts the probability of participating in the Colombian labor force by about 10 percentage points on average—9 ppts for women and 12 ppts for men.
  - Migration could have heterogeneous impacts (e.g., possible negative effects on less-skilled Colombian female LFP due to competition; positive effects among high-skilled Colombian women, especially those with young children).
- Crime measure:
  - Homicide rates in levels are statistically significant but the coefficient is small; interpretation limited because other forms of crime/insecurity may not be captured.
- Demographic note:
  - According to DANE, the number of births per year has declined from more than 200 thousand in 2019 to less than 90 thousand last year.

### Counterfactuals and macro impact
- Counterfactuals on long-term growth (labor contribution):
  - If LFP returned to its peak, Colombia’s long-term growth rate could increase by 0.3 ppts.
  - If the gender gap in labor participation were narrowed to 10 percent (the average gap in advanced economies), the contribution from labor could boost long-term growth by 0.6 ppts.
  - Further integrating migrants in the labor force would add to these gains.
- Increasing labor force participation would reduce funding pressures on Colombia’s pension system.

### Policy recommendations
- Enhance work flexibility to facilitate higher participation, especially for women with caregiving responsibilities.
- Expand childcare and eldercare services to mitigate the negative marginal effects of young children and elders on female LFP.
- Better integrate migrants in the labor force to capture their positive participation propensity:
  - Expand access to formal labor markets through more integrated approaches.
  - Deepen active labor market policies including job search assistance and labor protections.
  - Formalize migrant workers and ease access to formal employment to reduce misallocation and mismatching.
- Reap economic gains from migration by accelerating and strengthening labor market integration measures.

*Source: IMF staff calculations and analysis from the chapter "The decline in LFP has hampered Colombia’s labor and real GDP growth over the past."*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2025/english/1colea2025002-source-pdf.pdf_