## 1.    Transmission Channel

## Source details

**Canonical URL:** [1.    Transmission Channel](https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021261-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2021/english/wpiea2021261-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2021/english/wpiea2021261-print-pdf.pdf.json)

---

### Context and motivation
- Global temperatures have increased by about 1°C since the pre-industrial period; without decisive measures temperatures can increase by an additional 2° to 5°C by the end of this century.
- Limiting warming to 2°C or less requires substantial reductions in CO2 emissions and an increase in carbon taxes globally, including an immediate global carbon tax that should reach the target of US$75 per ton of CO2 by 2030.
- About 50 countries have a carbon pricing scheme; the global average carbon price is about US$2 per ton.
- Colombian carbon tax in 2021: Co$17,660 or about US$4.96 per CO2 ton; reference 1 UVT in 2021: Co$36,308, or about US$10.20.
- Current Colombian carbon tax covers between 24 percent (World Bank, 2021) and 40 percent (IMF 2019b) of greenhouse gas emissions.
- Revenues: carbon tax resulted in US$117 million in revenues in 2020; collections during 2017–2019 were around US$325 million. Legal allocation: 25 percent to environmental conservation fund; 75 percent to finance implementation of the peace agreement.

### Objective of this chapter
- Quantify the macrofinancial transmission of transition risks from an increase in carbon tax to the banking system in Colombia, focusing on the direct channel through banks’ exposures to nonfinancial corporates.
- Central questions:
  - How would a sharp increase in domestic carbon pricing in Colombia affect the banking system in the short-term?
  - How would financial stability implications change if policymakers preferred a gradual increase in the tax to achieve a higher carbon price?

### Framework and methodology
- Focus: transition risks driven by policy change (increase in carbon tax). Physical risks not primary focus.
- Core steps:
  - Estimate CO2 emissions at the sector level.
  - Adjust sector-level emissions by firms’ output to estimate additional burden (new liabilities) on firms’ balance sheets from a carbon tax increase.
  - Identify firms under financial stress after the tax increase by proxying stress with alternative measures (financial liquidity or leverage).
  - Calculate sector shares of financially stressed firms.
  - Merge sector-level shares with bank-level exposures to estimate bank loans at risk.
- Data:
  - ORBIS firm-level dataset (2017), final sample about 178,000 nonfinancial corporates.
  - Banking data from Colombian authorities: outstanding loans to each sector as of end-2017 and bank total assets; 56 individual banks.
- Simplifying assumptions: no pass-through on firms’ side, static stress test, focus on direct transmission channel.

### Firm-level stress measures and implementation
- Two firm-level proxies:
  - Leverage = total liabilities / total assets.
  - (Inverse of) financial liquidity = current liabilities / current assets.
- Identification rule: firm classified as financially stressed if the ratio (leverage or financial liquidity) becomes higher than 1 in the aftermath of the carbon-tax increase (from below 1 initially).
- Implementation of carbon-tax shock:
  - Sector-level emissions per unit output applied to firm output (no within-sector emission heterogeneity).
  - Additional cost imposed as new short-term liabilities: multiply the difference between new and old carbon prices per peso of output with each firm’s output.
  - No pass-through assumption yields an upper-bound estimate of financial risk.

### Banking exposures (reported shares)
- Wholesale and retail trade (G): 19.4 percent of banks’ total outstanding corporate loans; 7 percent of banks’ total assets.
- Manufacturing (C): 19 percent of banks’ total outstanding corporate loans; 6.9 percent of banks’ total assets.
- Construction (F): 16.2 percent of banks’ total outstanding corporate loans; 5.8 percent of banks’ total assets.
- Other notable sectors: Transportation (H), Electricity (D), Agriculture (A), Public administration (O), Real estate (L), Professional/scientific/technical activities (M).
- Heterogeneity: sector shares vary widely across banks (10th, 50th, 90th percentiles). Example: Wholesale and retail trade is above 50 percent of total outstanding corporate loans for the bank at the 90th percentile.

### Scenarios analyzed
- Sharp increase scenario:
  - US$70 increase in the carbon tax (from US$5 to US$75 per ton carbon-equivalent).
- Gradual / smaller increases:
  - US$20 increase.
  - US$15 increase.
  - US$10 increase.
- Results are scenario-based “what if” exercises.

### Main findings — sectoral incidence (Scenario 1: US$70 increase; nonweighted, financial liquidity measure)
- Water supply (E): 66.2 percent of firms newly driven into financial stress.
- Agriculture (A): 46.2 percent.
- Transportation (H): 31.6 percent.
- Electricity (D): 24.4 percent.
- Manufacturing (C): 15.7 percent.
- Accommodation (I): 11.4 percent.
- Leverage measure yields milder effects but similar sector ranking.
- Liability-weighted shares are generally lower than nonweighted shares, indicating stress concentrated among smaller firms.

### Contribution to banking stress (Scenario 1: US$70 increase)
- Sector-level contributions to banks’ total outstanding loans (financial liquidity, nonweighted):
  - Manufacturing (C): 3.0 percent
  - Transportation (H): 2.5 percent
  - Agriculture (A): 2.3 percent
  - Wholesale and retail trade (G): 2.3 percent
  - Electricity (D): 1.6 percent
- Aggregate banking stress (Table 1):
  - Banking loans under stress as percent of total corporate loans:
    - Financial Liquidity, nonweighted: 13.6
    - Financial Liquidity, weighted: 9.3
    - Leverage, nonweighted: 10.5
    - Leverage, weighted: 3.2
  - Banking loans under stress as percent of total assets:
    - Financial Liquidity, nonweighted: 4.9
    - Financial Liquidity, weighted: 3.3
    - Leverage, nonweighted: 3.8
    - Leverage, weighted: 1.2
- Stress in highly exposed banks (average for banks with highest exposure to five key sectors; Table 2):
  - Banking loans under stress as percent of total corporate loans:
    - Financial Liquidity, nonweighted: 24.3
    - Financial Liquidity, weighted: 20.9
    - Leverage, nonweighted: 17.8
    - Leverage, weighted: 8.1
  - Banking loans under stress as percent of total assets:
    - Financial Liquidity, nonweighted: 13.1
    - Financial Liquidity, weighted: 11.3
    - Leverage, nonweighted: 9.6
    - Leverage, weighted: 4.4
- Interpretation: financial liquidity measure produces higher banking stress than leverage measure; estimates are upper bounds given assumptions.

### Alternative scenarios — aggregate banking stress (smaller increases)
- US$20 increase (Panel A):
  - Total corporate loans (percent):
    - Financial Liquidity, nonweighted: 6.9
    - Financial Liquidity, weighted: 4.6
    - Leverage, nonweighted: 5.6
    - Leverage, weighted: 0.9
  - Total assets (percent):
    - Financial Liquidity, nonweighted: 2.5
    - Financial Liquidity, weighted: 1.7
    - Leverage, nonweighted: 2.0
    - Leverage, weighted: 0.3
- US$15 increase (Panel B):
  - Total corporate loans (percent):
    - Financial Liquidity, nonweighted: 6.2
    - Financial Liquidity, weighted: 3.6
    - Leverage, nonweighted: 5.1
    - Leverage, weighted: 0.7
  - Total assets (percent):
    - Financial Liquidity, nonweighted: 2.2
    - Financial Liquidity, weighted: 1.3
    - Leverage, nonweighted: 1.9
    - Leverage, weighted: 0.3
- US$10 increase (Panel C):
  - Total corporate loans (percent):
    - Financial Liquidity, nonweighted: 5.3
    - Financial Liquidity, weighted: 1.4
    - Leverage, nonweighted: 4.6
    - Leverage, weighted: 0.5
  - Total assets (percent):
    - Financial Liquidity, nonweighted: 1.9
    - Financial Liquidity, weighted: 0.5
    - Leverage, nonweighted: 1.7
    - Leverage, weighted: 0.2
- Conclusion: smaller, gradual increases generate much lower banking-system risks than a sharp US$70 increase.

### COVID-19 considerations
- Firm-level balance sheet data are from 2017; major shocks since then may alter vulnerability.
- Sectoral performance 2019 and 2020Q1:
  - Between 2017 and 2019, manufacturing increased real output; COVID-19 negatively affected most sectors thereafter.
  - Hardest hit sectors: construction (F), wholesale and retail trade (G), transportation (H), accommodation (I), arts/recreation (R), other services (S).
  - Wholesale and retail trade and transportation are both high-carbon-emitting and important for transition risk.
  - Agriculture (A) experienced modest growth during the first three quarters of 2020 compared to the previous year.
- Combining pandemic impacts with stress tests suggests heightened vulnerability for wholesale and retail trade and transportation; firm-level pandemic effects should be analyzed when data become available.

### Policy implications and considerations
- Gradual phasing in of carbon pricing, coupled with explicit communication on the use of associated revenue, can balance distributional, efficiency, and political considerations.
- Carbon tax revenues can support a greener and more inclusive recovery; transparent, efficient use of revenues can mitigate distributional concerns.
- For emerging market economies with social challenges and political-economy constraints, managing the transition comprehensively is crucial.
- Policymakers should consider pairing carbon pricing with targeted fiscal measures (e.g., targeted subsidies to affected sectors versus fossil fuel subsidies) and monitor bank-level concentration in high-risk sectors.

### Caveats, limitations, and research needs
- Data vintage: firm-level data are from 2017; scenarios do not fully cover COVID-19 effects.
- Key simplifying assumptions: no pass-through, static stress test, no second- or higher-round effects, no firm/consumer adjustments, no general equilibrium responses.
- Results are scenario-based “what if” estimates and intended to inform policymakers on distributional and financial stability implications rather than provide definitive forecasts.
- Future research: analyze use of carbon tax revenues, other transition-risk triggers (technology adoption, external carbon price changes, value-chain effects, shifts in sentiment), and joint analysis of physical and transition risks.

*Source: wpiea2021261-print-pdf - 1.    Transmission Channel ____________________________________________________8.*

### 1.    Transmission Channel ____________________________________________________8

### 1.    Transmission Channel ____________________________________________________8

### Context and motivation
- Climate crisis described as an existential challenge; global temperatures have increased by about 1°C since the pre-industrial period.
- In the absence of decisive measures, temperatures can increase by an additional 2° to 5°C by the end of this century.
- Limiting warming to 2°C or less requires substantial reductions in CO2 emissions and an increase in carbon taxes globally, including an immediate global carbon tax that should reach the target of US$75 per ton of CO2 by 2030.
- As of now, only about 50 countries have a carbon pricing scheme, and the global average carbon price is about US$2 per ton.
- Carbon taxes are described as “the single most powerful and efficient tool to reduce domestic fossil fuel CO2 emission” and also provide revenues that “should be allocated to reorient public finances in support of sustainable and inclusive growth.”

### Objective of this chapter
- Quantify the macrofinancial transmission of transition risks from an increase in carbon tax to the banking system in Colombia, focusing on the direct channel through banks’ exposures to nonfinancial corporates.
- Two central questions:
  - How would a sharp increase in domestic carbon pricing in Colombia affect the banking system in the short-term?
  - How would financial stability implications change if policymakers preferred a gradual increase in the tax to achieve a higher carbon price?

### Framework and methodology
- Focus on transition risks driven by policy change (increase in carbon tax); physical risks are acknowledged but not the primary focus.
- Core steps in the transmission framework:
  - Estimate CO2 emissions at the sector level.
  - Adjust sector-level emissions by firms’ output to estimate the additional burden (new liabilities) imposed on firms’ balance sheets by an increase in the carbon tax.
  - Identify firms under financial stress after the tax increase by proxying financial stress with alternative measures (financial liquidity or leverage).
  - Calculate the share of financially stressed firms in each sector.
  - Merge sector-level shares of stressed firms with bank-level exposures to each sector to estimate bank loans at risk due to the increase in the carbon tax.
- Data used:
  - Firm-level dataset from ORBIS, with data from 2017, covering small, medium-sized, and large firms, private and publicly listed firms, across sectors, including output and balance sheet variables.
  - Bank-level data from national authorities including banks’ outstanding loans to each sector and banks’ total assets.
- Simplifying assumptions noted: no pass-through on firms’ side, a static stress test, and focus on the direct transmission channel.

### Main findings
- Transition risks from a higher carbon tax are concentrated in certain sectors: Agriculture, manufacturing, electricity, wholesale and retail trade, and transportation.
- In Scenario 1 (a sharp US$70 per ton increase in tax from the current US$5 per ton level to meet the US$75 per ton target):
  - Banking loans at risk can be as large as 13.6 percent as a share of banks’ total outstanding corporate loans.
  - Banking loans at risk can be as large as 4.9 percent as a share of banks’ total assets, based on different measures and methodologies.
- Risks are unevenly distributed across individual banks: banks more exposed to the five key sectors are much more affected by an increase in carbon tax.
- Under alternative scenarios with lower increases in carbon tax (US$20, US$15, and US$10 per ton), risks are much lower; in some cases loans at risk are under 1 percent of banks’ assets.
- These results suggest gradual increases over years to meet the US$75 per ton target may be preferable for a smoother adjustment for the financial sector and other stakeholders.

### Policy implications and considerations
- Gradual phasing in of carbon pricing, coupled with explicit communication on the use of associated revenue, can balance distributional, efficiency, and political considerations.
- Carbon taxes generate revenue that can be used to support a greener and more inclusive recovery, especially relevant in the post-COVID-19 context.
- For emerging market economies with social challenges and political-economy constraints, managing the transition comprehensively is crucial.

### Caveats and limitations
- Firm-level data are from 2017; the scenario analyses do not fully cover the COVID-19 pandemic despite some attempt to account for potential effects.
- The analysis relies on simplifying assumptions (e.g., no pass-through, static stress test), and does not model second- and higher-round effects.
- Results are scenario-based “what if” estimates and meant to inform policymakers on distributional and financial stability implications rather than provide definitive forecasts.

### Structure of the paper (as outlined)
- Section IA: literature review.
- Section II: data on carbon emissions and pricing in Colombia; firm- and bank-level data.
- Section III: framework for transmission of stress from nonfinancial firms to the banking system.
- Section IV: results from Scenario 1 (sharp US$70 per ton increase).
- Section V: alternative scenarios (US$20, US$15, and US$10 per ton).
- Section VI: potential influence of the COVID-19 shock on findings.

*Source: wpiea2021261-print-pdf - 1.    Transmission Channel ____________________________________________________8.*

### Section VII concludes.

### Section VII concludes

### A. Literature
- Joint work by the Colombia National Planning Department, the Inter-American Development Bank (IDB), and the Economic Commission for Latin America and the Caribbean (ECLAC) shows climate change can be expected to have an average impact of -0.5 percent on annual GDP growth for every year between 2010 and 2100 (BID-CEPAL-DNP, 2014).
- The cumulated effect of climate change at this horizon (without discounting at the present value) would be equivalent to losing about 3.6 times the total value of the 2010 GDP.
- Sectoral impacts from the same study: agricultural and transport industries expected to lose between 5.9 percent and 7.4 percent of production on average every year between 2010 and 2100.
- World Bank (2014) finds limited macroeconomic impact of “green taxes” when revenues are reinvested to generate jobs, with the energy sector most affected.
- Calderon and others (2016) using partial and general equilibrium models find significant CO2 reductions are achievable but also significant effects of 2 percent to 3 percent decreases in GDP.
- Romero and others (2018) using a sectoral general equilibrium model find a much more limited impact of the Colombian carbon tax, concentrated in the highest-income quintile, concluding the tax is progressive.
- Early modeling of environmental fiscal reforms: Bach and others (2002) find energy consumption and CO2 emissions decrease while employment rises and that “the impact on growth is minimal.”
- Financial stability literature on low-carbon transition:
  - Weyzig and others (2014): exposure of 20 banks and 23 pension funds in the EU to fossil fuel firms exceeds €1 trillion; a slow transition could be more costly and disruptive.
  - Battiston and others (2017): indirect exposures via financial networks can generate second-round losses larger than first-round losses; timing matters—early/stable policy allows smooth adjustments, sudden/late transitions could be severely adverse.
  - Vermeulen and others (2019): comprehensive framework applied to more than 80 Dutch financial institutions finds sizeable financial losses under disruptive transition scenarios and recommends close monitoring by authorities.
  - Grippa and Mann (2020): pioneering firm-level study for Norway finds significant but manageable financial stability impacts from transition risk due to a carbon tax; different firm-level risk measures limit direct comparability.
- Contribution: this study is, to the authors’ knowledge, the first to evaluate financial stability implications of a climate-related transition risk (driven by a higher carbon tax) for an emerging market economy using detailed bank- and firm-level data.

*Key numeric findings and estimates reported in the literature section above are preserved verbatim from the source.*

### II. DATA AND FACTS

#### A. Carbon Emissions — Context and composition
- Colombia’s gross greenhouse gas emissions for 2014 equaled 236 million tons of CO2-equivalent (BUR, 2018).
- Colombia’s CO2 emissions account for close to 70 percent of the country’s gross emissions in equivalent units.
- CO2 absorption offsets only 13 percent of the CO2 emitted.
- Greenhouse gas emissions by activity (net, 2014) shares:
  - Land Use: 39%
  - Transport: 14%
  - Manufacturing, Construction and Industrial Processes: 11%
  - Energy Industries: 11%
  - Cattle Ranching: 11%
  - Waste and Residual Management: 7%
  - Fuel Manufacturing: 4%
  - Other: 3%
- BAU and NDC commitments:
  - Projected BAU emissions by 2030: 345 megaton (Mt) CO2-equivalent.
  - Revised 2030 target: cap emissions at 169 Mt CO2-equivalent.
  - The revised commitment implies a reduction of 28 percent with respect to the 2014 emissions reported in the BUR.
- Gas equivalency example preserved as stated: “A ton of methane, for instance, has the equivalent effect on heat trapping in the atmosphere as 84 tons of carbon dioxide over a 20-year horizon.”
- Sectoral emissions per output: methodology matches activities in BUR (2018) to 1-digit NACE Revision 2 sectors A-F and H; sectors G and I-T estimated from Brazilian Exiobase data and converted to Colombian pesos. (Figure notes: tertiary sectors from Exiobase are all well below 0.01.)

#### B. Carbon Pricing in Colombia — Coverage, level, schedule, and social/revenue considerations
- Carbon tax definition and scope:
  - The Colombian carbon tax (“Impuesto Nacional al Carbono”) introduced in 2017 covers the carbon content of all fossil fuels used for combustion (including gasoline, kerosene, fuel oil).
  - The tax burden is generated at the point of sale, withdrawal or import inside the country for consumption or sale, and is only taxed once.
  - The existing carbon tax is limited to emissions from oil and gas fuel combustion; N2O and CH4 and CO2 from other processes (e.g., coal burning, agricultural burning, residue management, land transformation) are not included.
  - Exemptions include exports and carbon neutral companies (carbon offset system introduced in 2017).
- Estimates of coverage of the Colombian carbon tax: between 24 percent (World Bank, 2021) and 40 percent (IMF 2019b) of all the country’s greenhouse gas emissions.
- Tax level and evolution:
  - Initial tax in 2017: Co$15,000 per CO2 ton (adjusted individually to each fossil fuel).
  - By law, tax to be revised each February increasing by previous year’s inflation plus one percentage point until it equals 1 UVT.
  - Tax for 2021: Co$17,660 or about US$4.96 per CO2 ton.
  - Reference 1 UVT in 2021: Co$36,308, or about US$10.20.
  - Projected schedule for reaching 1 UVT is very slow; the target won’t be reached until 2094.
  - Comparison from Calderon and others (2016): a 50 percent abatement scenario would correspond to US$114 per CO2 ton; they proposed starting schedules at US$10 or US$50, growing at an annual rate of 4 percent afterward.
- Social considerations:
  - Concerns about regressivity of carbon taxes; Romero and others (2018) find the highest welfare-reducing impacts on the highest quintile under an aggressive carbon tax, with the middle class impacted the least; regional variability exists (highest decreases in Bogota; muted effects in Arauca, Caqueta, Vichada).
  - The use of carbon tax revenues can mitigate distributional concerns.
- Revenues:
  - The carbon tax resulted in US$117 million in revenues for the Colombian government in 2020 (World Bank).
  - Collections for carbon tax during 2017–2019 were around US$325 million.
  - Legal allocation: 25 percent of this revenue to a fund for environmental conservation and protection projects; 75 percent to finance implementation of the peace agreement.

#### B (Methodology note)
- For the study’s exercise, a more comprehensive carbon tax than currently implemented is considered:
  - The tax is imposed on the basis of all greenhouse emissions (including N2O and CH4, CO2 from land use changes, residue management, coal combustion).
  - The tax burden is assumed to fall on the final emitter rather than on fuel producers.
- Rationale: achieving Colombia’s revised NDC target is “likely to require the widespread adoption of aggressive carbon taxes.”

#### C. Firm-level Data, Sector Distribution, and Variables
- Data source: ORBIS database compiled by Bureau van Dijk (BvD); 2017 data used (average reporting lag around two years).
- Dataset coverage and cleaning: methodology follows Gal (2013), Gal and Hijzen (2016), Kalemli-Özcan and others (2015); final dataset provided with assistance from IMF Research Department staff.
- Final sample: about 178,000 nonfinancial corporates with variables of interest available.
- Sector coverage: nonfinancial sectors corresponding to NACE 1-digit A–S and NACE 2-digit 1–96, excluding financial services; distribution shows highest number of firms in wholesale and retail trade, followed by professional/scientific/technical activities, manufacturing, and construction.
- Firm-level proxy for financial distress:
  - Standard proxy is interest coverage ratio (ICR)—ability to cover interest expenses by earnings (ICR below 1 indicates inability).
  - ORBIS lacks interest expense data for Colombian firms; therefore this study uses alternative proxies: leverage and financial liquidity (to be explained in later sections).
- Stress-test approach: emissions per output estimates (from Section II A) are applied to sectors and mapped to firms by assuming firms are representative of average emissions per output of their sector; new liabilities due to an increase in carbon tax are estimated and their effects on firm balance sheets are analyzed.

_Italic source attribution: Section VII concludes. Source PDF filename: wpiea2021261-print-pdf - Section VII concludes._

### 1. Percentage of firms                                                    2. Number of firms

### wpiea2021261-print-pdf - 1. Percentage of firms                                                    2. Number of firms

### Methodology: firm-level stress measures and no-pass-through assumption
- Two firm-level proxies for balance-sheet stress:
  - Leverage = total liabilities / total assets. Higher values imply higher stress.
  - (Inverse of) financial liquidity = current liabilities / current assets. Higher values imply higher stress.
- Identification rule: a firm is classified as financially stressed if the ratio (leverage or financial liquidity) becomes higher than 1 in the aftermath of an increase in the carbon tax (from below 1 initially).
- Implementation of carbon-tax shock on firm balance sheets:
  - Sector-level carbon emissions (per unit output) are applied to firm output (firm-level emission heterogeneity within sectors is not observed).
  - The additional cost from an increase in carbon tax is imposed as new short-term liabilities (no-pass-through and no further adjustment): multiply the difference between new and old carbon prices per peso of output with each firm’s output.
  - Both leverage and financial liquidity increase because of higher liabilities.
- Interpretation of the no-pass-through assumption:
  - Firms absorbing the full tax increase is equivalent to an upper-bound estimate of financial risk.
  - The assumption abstracts from firm/consumer adjustments, pass-through to prices, demand responses, changes in production chains, or green investments.

### Banking data and exposures to sectors
- Banking dataset:
  - Provided by Colombian authorities.
  - Contains each bank’s outstanding loans to each sector (NACE 1 Revision 2) as of end-2017, and bank total assets.
  - 56 individual banks in the dataset.
- Banking system exposures (as reported):
  - Wholesale and retail trade (G): 19.4 percent of banks’ total outstanding corporate loans; 7 percent of banks’ total assets.
  - Manufacturing (C): 19 percent of banks’ total outstanding corporate loans; 6.9 percent of banks’ total assets.
  - Construction (F): 16.2 percent of banks’ total outstanding corporate loans; 5.8 percent of banks’ total assets.
  - Other sectors with relatively high loans: Transportation (H), Electricity (D), Agriculture (A), Public administration (O), Real estate (L), Professional/scientific/technical activities (M).
- Heterogeneity across individual banks:
  - Exposures vary substantially by bank (10th, 50th, 90th percentiles reported in analyses).
  - Example: Wholesale and retail trade accounts for 19 percent of total outstanding corporate loans in the banking system overall, but for the bank at the 90th percentile it is above 50 percent of that bank’s total outstanding corporate loans.
  - Example: Manufacturing loans are about 6.9 percent of total assets for the banking system but exceed 13 percent of total assets for some banks.

### Transmission framework from firms to banks
- Steps to quantify transmission:
  1. Compute share of firms in each sector that move into financial distress after the carbon-tax shock (based on leverage or financial liquidity).
  2. Aggregate firm-level stress to sector level using two approaches:
     - Nonweighted: share of firms under stress (number of firms).
     - Weighted: share of firms under stress using existing liabilities as weights.
  3. Combine sector-level share of stressed firms with banking-system exposure to that sector to obtain sector-specific contributions to banking stress.
- Two banking-stress metrics produced:
  - Share of banks’ total outstanding corporate loans under stress (sector share of outstanding loans × sector share of stressed firms).
  - Share of banks’ total assets under stress (sector share of total assets × sector share of stressed firms).
- Four alternative calculations reported (combinations):
  - Financial liquidity measure, nonweighted approach.
  - Financial liquidity measure, weighted approach.
  - Leverage measure, nonweighted approach.
  - Leverage measure, weighted approach.
- Illustration from text:
  - If 50 percent of firms in a sector are under stress and outstanding loans to that sector are 10 percent of total outstanding corporate loans, then 5 percent of total outstanding corporate loans are under stress from that sector. If loans to that sector are 4 percent of banks’ total assets, then 2 percent of banks’ total assets are under stress from that sector.

### Scenarios explored
- Sharp increase scenario:
  - US$70 increase in the carbon tax (from US$5 to US$75 per ton carbon-equivalent).
  - Rationale: US$75 per ton is the mid-point of the range estimated by the High-Level Commission on Carbon Prices (2017) and reported by the IMF; associated with Paris targets in referenced literature.
- Gradual / smaller increases also analyzed:
  - US$20 increase.
  - US$15 increase.
  - US$10 increase.
- Note: scenarios are presented as “what if” exercises and not as policy recommendations per se.

### Scenario 1 (US$70 increase to US$75): firms driven into stress — key sector-level impacts (nonweighted shares)
- Financial liquidity measure, nonweighted (Panel 1 highlights):
  - Water supply (E): 66.2 percent of firms newly driven into financial stress.
  - Agriculture (A): 46.2 percent.
  - Transportation (H): 31.6 percent.
  - Electricity (D): 24.4 percent.
  - Manufacturing (C): 15.7 percent.
  - Accommodation (I): 11.4 percent.
- Leverage measure, nonweighted (Panel 2):
  - Effects are milder relative to the financial liquidity measure, but the ranking of most affected sectors is similar.
- Weighted vs nonweighted comparisons (Panels 3 and 4):
  - Weighted (liability-weighted) shares are lower in many sectors than nonweighted shares.
  - This suggests stress is more concentrated among small firms whose liabilities are large relative to their asset size.

### Analytical caveats highlighted in the source
- Firm-level emission heterogeneity within sectors is not captured because emissions are available only at the sector level.
- Lack of firm-level bank-loan linkage: banking loans are available only at sector level for the overall banking system, so aggregation is done at 1-digit sector level.
- No general equilibrium adjustments are made (no pass-through, no demand-side responses, no second-round effects, no investment responses).

*Source: ORBIS, IMF staff calculations; banking dataset provided by Colombian authorities; figures and notes as described in the supplied content.*

### 1. Based on number of firms, financial liqudity         2. Based on number of firms, leverage

### wpiea2021261-print-pdf - 1. Based on number of firms, financial liqudity         2. Based on number of firms, leverage

### Contribution to the Banking Stress
- Framework: Sector shares of firms under stress following a US$70 increase in the carbon tax are computed using (i) firm-level financial stress measures (financial liquidity versus leverage) and (ii) two approaches to aggregate sector shares (nonweighted based on number of firms and weighted based on liability weights).
- Key sector contributions to banks’ total outstanding loans to nonfinancial corporates (financial liquidity measure, nonweighted):
  - Manufacturing (C): 3.0 percent
  - Transportation (H): 2.5 percent
  - Agriculture (A): 2.3 percent
  - Wholesale and retail trade (G): 2.3 percent
  - Electricity (D): 1.6 percent
- Key sector contributions (leverage measure, nonweighted):
  - Manufacturing (C): 2.1 percent
  - Wholesale and retail trade (G): 2.0 percent
  - Agriculture (A): 1.9 percent
  - Transportation (H): 1.8 percent
  - Electricity (D): 1.0 percent
- Weighted calculations (liability weights) yield generally lower risk contributions for each measure relative to nonweighted results, but the five highest-risk sectors remain: agriculture, manufacturing, electricity, wholesale and retail trade, and transportation.

_Source: IMF staff calculations._

### Aggregate Banking Stress and Individual Exposures
- Aggregate banking stress following a US$70 increase in the carbon tax (Table 1):
  - Banking loans under stress as percent of total corporate loans:
    - Financial Liquidity, nonweighted: 13.6
    - Financial Liquidity, weighted: 9.3
    - Leverage, nonweighted: 10.5
    - Leverage, weighted: 3.2
  - Banking loans under stress as percent of total assets:
    - Financial Liquidity, nonweighted: 4.9
    - Financial Liquidity, weighted: 3.3
    - Leverage, nonweighted: 3.8
    - Leverage, weighted: 1.2
- Interpretation:
  - Financial liquidity measure produces higher banking stress than leverage measure for both aggregation approaches.
  - Estimates are characterised as an upper bound given simplifying assumptions (e.g., no-pass-through). Footnote: amounts of loans under risk do not correspond to nonperforming loans (NPLs) one to one; they can be viewed as an upper bound for NPLs.
- Stress in highly exposed banks (average for banks with highest exposure to five key sectors; Table 2):
  - Banking loans under stress as percent of total corporate loans:
    - Financial Liquidity, nonweighted: 24.3
    - Financial Liquidity, weighted: 20.9
    - Leverage, nonweighted: 17.8
    - Leverage, weighted: 8.1
  - Banking loans under stress as percent of total assets:
    - Financial Liquidity, nonweighted: 13.1
    - Financial Liquidity, weighted: 11.3
    - Leverage, nonweighted: 9.6
    - Leverage, weighted: 4.4
- Implication: Individual bank exposures to the five high-risk sectors can yield much higher stress than aggregate averages, warranting attention to bank-level concentration.

_Source: IMF staff calculations._

### Other Scenarios (gradual carbon tax increases)
- Purpose: Assess banking-system risk under smaller annual increases to reach US$75 over time instead of a single US$70 shock.
- Panel A: A 20 USD increase in tax
  - Total corporate loans (percent):
    - Financial Liquidity, nonweighted: 6.9
    - Financial Liquidity, weighted: 4.6
    - Leverage, nonweighted: 5.6
    - Leverage, weighted: 0.9
  - Total assets (percent):
    - Financial Liquidity, nonweighted: 2.5
    - Financial Liquidity, weighted: 1.7
    - Leverage, nonweighted: 2.0
    - Leverage, weighted: 0.3
- Panel B: A 15 USD increase in tax
  - Total corporate loans (percent):
    - Financial Liquidity, nonweighted: 6.2
    - Financial Liquidity, weighted: 3.6
    - Leverage, nonweighted: 5.1
    - Leverage, weighted: 0.7
  - Total assets (percent):
    - Financial Liquidity, nonweighted: 2.2
    - Financial Liquidity, weighted: 1.3
    - Leverage, nonweighted: 1.9
    - Leverage, weighted: 0.3
- Panel C: A 10 USD increase in tax
  - Total corporate loans (percent):
    - Financial Liquidity, nonweighted: 5.3
    - Financial Liquidity, weighted: 1.4
    - Leverage, nonweighted: 4.6
    - Leverage, weighted: 0.5
  - Total assets (percent):
    - Financial Liquidity, nonweighted: 1.9
    - Financial Liquidity, weighted: 0.5
    - Leverage, nonweighted: 1.7
    - Leverage, weighted: 0.2
- Conclusion: Smaller, gradual increases (US$20, US$15, US$10) generate much lower banking-system risks than a sharp US$70 increase; a gradual approach could be preferable.

_Source: IMF staff calculations._

### Considerations on the Effects of the COVID-19 Pandemic
- Data vintage: Firm-level balance sheet data correspond to 2017; major shocks occurred since then.
- Sectoral performance 2019 and 2020Q1 (Banco de la Republica and IMF staff calculations):
  - Between 2017 and 2019, manufacturing increased real output; the COVID-19 crisis negatively affected most sectors thereafter.
  - Hardest hit sectors: construction (F), wholesale and retail trade (G), transportation (H), accommodation (I), arts/recreation (R), and other services (S).
  - Wholesale and retail trade and transportation are both high-carbon-emitting and important for transition risk in the analysis.
  - Agriculture (A) experienced modest growth during the first three quarters of 2020 compared to the previous year.
- Combining pandemic impacts with stress tests suggests wholesale and retail trade and transportation face heightened vulnerability to a sharp carbon tax increase; firm-level balance sheet impact of the pandemic should be analyzed once data become available.

_Source: Banco de la Republica and IMF staff calculations._

### Concluding Remarks
- Framework summary: The study presents a firm- and bank-level framework to quantify transmission of climate-related transition risk (via a higher carbon price) from nonfinancial corporates to the banking system for Colombia.
- Main findings:
  - Five sectors—agriculture, manufacturing, electricity, wholesale and retail trade, and transportation—drive most of the transmission risk.
  - A sharp US$70 increase in the carbon price leads to significant but potentially manageable banking-system risks; stresses are uneven across banks depending on exposures.
  - Gradual increases in the carbon price reduce short-term financial-stability risks and may be preferable; policymakers should nevertheless act sooner rather than later.
- Caveats and research needs:
  - Analysis relies on strong simplifying assumptions and a static stress-test perspective; it accounts only for the direct channel.
  - The study is the first detailed firm-level analysis of climate-related transition risks for an emerging market economy (Colombia), to the authors' knowledge.
  - Future research should analyze use of new fiscal revenues from higher carbon taxes (e.g., targeted subsidies to affected sectors versus fossil fuel subsidies), other transition-risk triggers (technology adoption, external carbon price changes, value-chain effects, shifts in public opinion/market sentiment), and joint analysis of physical and transition risks.
- Policy implication: Consider holistic policies that pair carbon pricing with transparent, efficient use of revenues to support affected sectors and mitigate short-term financial risks while preserving mitigation effectiveness.

*Source: IMF staff calculations.*

### REFERENCES

### REFERENCES

### References cited
- Adrian, Tobias, James Morsink, and Liliana Schumacher. “Stress Testing at the IMF.” IMF Monetary and Capital Markets Department (2020).
- Ahn, JaeBin, Romain Duval, and Can Sever. “Macroeconomic policy, product market competition, and growth: the intangible investment channel.” IMF Working Paper No 20/25 (2020).
- Bach, Stefan, Michael Kohlhaas, Bernd Meyer, Barbara Praetorius, and Heinz Welsch. “The effects of environmental fiscal reform in Germany: a simulation study.” Energy policy 30, no. 9 (2002): pp. 803–811.
- Battiston, Stefano, Antoine Mandel, Irene Monasterolo, Franziska Schütze, and Gabriele Visentin. “A climate stress-test of the financial system.” Nature Climate Change 7, no. 4 (2017): pp. 283–288.
- BID-CEPAL-DNP. 2014. Silvia Calderón, Germán Romero, Alejandro Ordoñez, Andrés Álvarez, C. Ludeña, L. Sánchez, Carlos De Miguel, Karina Martínez, and Mauricio Pereira. “Impactos económicos del cambio climático en Colombia.” Washington DC: Banco Interamericano de desarrollo. Recuperada de https://repositorio. cepal. org/bitstream/11362/37879/1/S1500268_es. pdf (2014).
- Biennial Update Report to the UNFCCC (BUR). 2018. “Inventario nacional de emisiones de gases efecto invernadero. República de Colombia segunda comunicación nacional ante la convención marco de las naciones unidas sobre cambio climático” IDEAM, MAVDT y PNUD (Instituto de Hidrología, Meteorología y Estudios Ambientales), (Ministerio de Ambiente, Vivienda y Desarrollo Territorial) y (Programa de las Naciones Unidas para el Desarrollo).
- Calderón, Silvia, Andrés Camilo Alvarez, Ana María Loboguerrero, Santiago Arango, Katherine Calvin, Tom Kober, Kathryn Daenzer, and Karen Fisher-Vanden. “Achieving CO2 reductions in Colombia: Effects of carbon taxes and abatement targets.” Energy economics 56 (2016): pp. 575–586.
- Dell, Melissa, Benjamin F. Jones, and Benjamin A. Olken. “What do we learn from the weather? The new climate-economy literature.” Journal of Economic Literature 52, no. 3 (2014): pp. 740–798.
- Duval, Romain, Gee Hee Hong, and Yannick Timmer. “Financial frictions and the great productivity slowdown.” The Review of Financial Studies 33, no. 2 (2020): pp. 475–503.
- Gal, Peter, and A. Hijzen, 2016. “The Short-Term Impact of Product Market Reforms: A CrossCountry Firm-Level Analysis,” IMF Working Paper Series, WP/16/116. International Monetary Fund.
- ———, 2013. “Measuring Total Factor Productivity at the Firm Level using OECDOrbis,” OECD Economics Department Working Paper 1049, OECD Publishing.
- Gaspar, Vitor, Paolo Mauro, Ian Parry and Catherine Pattillo. 2019. “Fiscal Policies to Curb Climate Change.” IMF Blog. Available at: https://blogs.imf.org/2019/10/10/fiscal-policies-to-curb-climate-change/
- Gopinath, Gita, Şebnem Kalemli-Özcan, Loukas Karabarbounis, and Carolina Villegas-Sanchez. “Capital allocation and productivity in South Europe.” The Quarterly Journal of Economics 132, no. 4 (2017): pp. 1915–1967.
- Grippa, Pierpaolo, and Samuel Mann. “Climate-Related Stress Testing: Transition Risks in Norway.” IMF Working Paper No. 2020/232 (2020).
- High-Level Commission on Carbon Prices (2017) Report of the High-Level Commission on Carbon Prices. Washington, DC: World Bank.
- IMF 2019a. “Fiscal Policies for Paris Climate Strategies—from Principle to Practice.” IMF Policy Paper No. 19/010
- ———2019b. Fiscal Monitor, October 2019: How to Mitigate Climate Change. October 2019. Available at: https://www.imf.org/en/Publications/FM/Issues/2019/09/12/fiscal-monitor-october-2019
- ———2020a. Finding the Right Policy Mix to Safeguard our Climate. IMF Blog. Available at: https://blogs.imf.org/2020/10/07/finding-the-right-policy-mix-to-safeguard-our-climate/
- ———2020b. World Economic Outlook, October 2020: A Long and Difficult Ascent. Available at: https://www.imf.org/en/Publications/WEO/Issues/2020/09/30/world-economic-outlook-october-2020
- Kalemli-Ozcan, Sebnem, Bent Sorensen, Carolina Villegas-Sanchez, Vadym Volosovych, and Sevcan Yesiltas. How to construct nationally representative firm level data from the Orbis global database: New facts and aggregate implications. No. w21558. National Bureau of Economic Research, 2015.
- Lagarde, Christine and Vitor Gaspar. 2019. “Getting Real on Meeting Paris Climate Change Commitments.” IMF Blog. Available at: https://blogs.imf.org/2019/05/03/getting-real-on-meeting-paris-climate-change-commitments/
- Romero, Germán, Andrés Álvarez-Espinosa, Silvia Calderón, and Alejandro Ordóñez. “Impactos distributivos de un impuesto al carbono en Colombia: vínculo entre modelos de microsimulaciones y equilibrio general.” Lecturas de Economía 89 (2018): pp. 163–198.
- Uras, Burak R. “Corporate financial structure, misallocation and total factor productivity.” Journal of Banking & Finance 39 (2014): pp. 177–191.
- Vermeulen, Robert, Edo Schets, Melanie Lohuis, Barbara Kölbl, David-Jan Jansen, and Willem Heeringa. “The heat is on: A framework for measuring financial stress under disruptive energy transition scenarios.” (2019).
- Weyzig, Francis, Barbara Kuepper, Jan Willem Van Gelder, and Rens Van Tilburg. “The price of doing too little too late; the impact of the carbon bubble on the European financial system.” Green New Deal Series 11 (2014).
- World Bank Group, 2019. Report of the High-Level Commission on Carbon Pricing and Competitiveness. World Bank Group, Washington, D.C
- ———, 2021. “Carbon Pricing Dashboard.” Updated April 1, 2021. Available at: https://carbonpricingdashboard.worldbank.org/map_data Accessed: June 16, 2021.

*Source: wpiea2021261-print-pdf - REFERENCES*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021261-print-pdf.pdf_