## Global Economic Impacts of Physical Climate Risks (WP/2023/183) — Centre for Applied Macroeconomic Analysis ## Source details **Canonical URL:** [Global Economic Impacts of Physical Climate Risks (WP/2023/183) — Centre for Applied Macroeconomic Analysis](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023183-print-pdf.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023183-print-pdf.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023183-print-pdf.pdf.json) --- ### Abstract and high-level results - Evaluates global economic consequences of physical climate risks under SSP 1-2.6 and SSP 2-4.5 using firm-level evidence. - Method: - Estimate historical sectoral productivity changes from chronic climate risks (average temperature and precipitation) and extreme climate conditions (heatwaves, coldwaves, droughts, floods) using firm-level damage functions. - Produce forward-looking sectoral productivity changes for a global multisectoral sample of firms. - For floods, account for persistent productivity changes from damage to firms’ physical capital using Huizinga et al. (2017) damage functions and Jupiter Intelligence flood projections. - Assess macroeconomic impact using the global, multisectoral, intertemporal general equilibrium model G-Cubed (GGG20C_v169). - Key aggregate findings: - Without additional adaptation beyond 2020, all economies experience substantial losses under both SSPs; losses increase with global warming. - SSP 1-2.6: could cost the world 1.2 percent of GDP per annum on average during 2021 to 2100 (2.4 percent of GDP in 2100). - SSP 2-4.5: losses could more than double to 3.2 percent of GDP per annum on average during 2021 to 2100 (6.4 percent of GDP in 2100). ### Data, samples, and sector definitions - Firm samples: - Sample 1: 59,554 firms — top 1,000 non-financial firms from each IMF member nation (where available). - Sample 2: 20,215 firms — cleaned dataset with computed TFP after cleaning. - Sector aggregation (NACE → four broad sectors): Agriculture, Mining, Manufacturing, Services. - Exact sector counts (Sample 1 / Sample 2): - Agriculture: 1,401 / 519 (countries: 91 / 45) - Mining: 1,605 / 464 (countries: 101 / 46) - Manufacturing: 15,849 / 6,945 (countries: 125 / 48) - Services: 40,699 / 12,287 (countries: 143 / 47) - Financial data: Orbis 2000–2018; TFP computed following Ackerberg et al. (2015); cleaned per Kalemli-Ozcan et al. (2015). - Representativeness assumption: Sample 2 assumed representative of Sample 1 for regional and sectoral projection. ### Climate scenarios and temperature outcomes - SSPs used: SSP1-2.6 (low emissions) and SSP2-4.5 (intermediate emissions). Flood data available for SSP1-2.6, SSP2-4.5, and SSP5-8.5 (SSP5-8.5 excluded by authors). - Temperature outcomes (selected exact entries from Table 2 / IPCC (2021)): - SSP1-2.6: Estimated Global Warming 2041-2060 (°C): 1.7; Estimated Global Warming 2081-2100 (°C): 1.8; 2081-2100 (Range in °C): 1.3 – 2.4. - SSP2-4.5: Estimated Global Warming 2041-2060 (°C): 2.0; Estimated Global Warming 2081-2100 (°C): 2.7; 2081-2100 (Range in °C): 2.1 – 3.5. - (Full table includes SSP1-1.9, SSP3-7.0, SSP5-8.5 with corresponding values.) ### Empirical estimation approach and identification - Estimation method: Ridge regression to address collinearity among many climate indicators. - Controls and fixed effects: - Firm- and country-specific climate indicators (manufacturing and services); agriculture and mining estimates exclude firm-specific indicators. - GDP growth, region-specific and year-specific fixed effects. - Location proxy: head office address used for firm location; representativeness differs by sector (manufacturing/services vs agriculture/mining). - Linear relationship assumed between physical climate risks and TFP. ### Sectoral empirical findings (selected exact impacts) - Agriculture (empirical impacts): - A one percent increase in extremely dry conditions reduces agriculture productivity by more than two percent. - A one percent increase in extremely warm night conditions reduces agriculture productivity by more than two percent. - An increase in mean temperature by one degree reduces productivity by more than one percent. - A one percent increase in extremely wet conditions could boost productivity by almost 1.5 percent. - Agriculture projections: - Agriculture productivity reduces by 10 to 20 percentage points in most regions under the two SSPs. - Under SSP 1-2.6: Central Asia, North America, and Western Europe exceed six percentage point reductions by 2100 relative to 2020. - Under SSP 2-4.5: South Asia, North America, and Western Europe exceed 15 percentage point reductions by 2100 relative to 2020. - Australia and New Zealand remain roughly above two percentage points compared to 2020 productivity change. - Mining and Energy (empirical impacts): - Increase in mean temperature by one degree reduces mining productivity by almost three percent. - Extremely wet conditions reduce mining productivity by almost 1.5 percent. - Extremely warm conditions (day and night) and prolonged extremely dry conditions observed to positively affect mining productivity in the sample. - Mining projections: - Under SSP 1-2.6: eight regions experience notable reductions; North America shows almost five percentage point reduction relative to 2020. - Under SSP 2-4.5: almost all regions experience notable reductions; Northern Europe is the only exception with reduction less than one percentage point. - Manufacturing (empirical impacts): - Most climate indicators have lower impacts on manufacturing productivity than on other sectors. - A unit increase in mean temperature at the firm location and at the country level decreases manufacturing productivity; firm-location temperature has stronger effect. - Country-wide extremely warm conditions are detrimental; country-wide extremely wet and extremely dry conditions reduce productivity. - Manufacturing projections: - Under SSP 1-2.6: most regions experience manufacturing productivity contractions compared to 2020; North America and East Asia contractions lie between one and two percentage points by 2100. - Under SSP 2-4.5: more regions experience reductions; North and Latin America exceed 1.75 percentage point reductions by 2100 relative to 2020. - Services (empirical impacts and projections): - Country-level increase in average mean temperature has the strongest adverse effect on service productivity; firm-specific mean temperature increases can boost productivity. - Under SSP 1-2.6: all regions experience service sector productivity reductions; Central Asia reaches almost a one percentage point reduction by 2100 relative to 2020. - Under SSP 2-4.5: all regions experience declines; most regions observe reductions within one to two percentage points by 2100; Sub-Saharan Africa and Western Europe experience almost three percentage point reductions by 2100 relative to 2020. - Sectoral summary: - All sectors experience losses; agriculture most vulnerable (10 to 20 percentage point reductions in most regions), mining losses exceed five percentage points in some regions, manufacturing and services least affected. ### Floods and persistent capital damage - Flood damage methodology: - Damage functions from Huizinga et al. (2017) map water depth to fractional damage by sector and continent. - Jupiter Intelligence provides projected flood depth and fraction flooded for return periods: 1/10, 1/20, 1/50, 1/100, 1/200, 1/500 for 2020–2100 under SSPs. - Firm-level damage probability pd_{c,i,RP} = 휑_{c,i,RP} * f(훿_{c,i,RP}). - Piecewise rule for percentage damage D_{c,i} across depth thresholds (explicit thresholds and averaging rules preserved). - Expected percentage of annual damage E[D_{c,i}] computed as weighted sum across return period intervals (explicit formula preserved). - Flood impact findings: - Under SSP 1-2.6: firms in Asia and Africa exposed to significant increases in flood damage over time across all sectors; other regions show moderate increases, constancy, or slight declines. - Under SSP 2-4.5: damages further increase in Asia and Africa; some regions (example: North America for mining) show further decreases. - Largest damages occur in Asia for all sectors except mining, where largest damages occur in North America. - Conversion to persistent TFP shocks: - Flood damage converted to persistent TFP shocks by adjusting for sector reliance on capital using GTAP 10 Input-Output tables. - Final TFP shock for a G-Cubed sector equals contemporaneous chronic/extreme TFP shock plus persistent TFP shocks from flood damage adjusted for capital reliance. ### G-Cubed model implementation and features - Model: G-Cubed (GGG20C_v169) — global, multisectoral, intertemporal general equilibrium with eleven regions and twenty sectors. - Regions: AUS, CAN, CHI, EUW, IND, JPN, OEC, OPC, ROW, RUS, USA. - Sectors: 1–20 with mapping to broad calibrated TFP shocks for Agriculture, Mining, Manufacturing, Services; renewable generation sectors (16–19, 20) received no shocks. - Key mechanisms: - Heterogeneous households and firms; forward-looking and rule-of-thumb agents. - Distinction between physical and financial capital; financial capital mobile, physical capital subject to quadratic adjustment costs and slow adjustment (stranded assets). - Sectoral trade linkages propagate shocks domestically and internationally. - Baseline: - Starts in 2018 projected to 2100; includes historical climate policies and adaptation up to 2018. - Effective baseline simulation (2020–2100) includes climate shocks equivalent to 2020 levels (results reported relative to this baseline). - Shocks introduced: - Cumulative TFP shocks (chronic and extreme, excluding floods) and persistent TFP shocks from flood damage for selected sectors (Annexure 6 lists cumulative shocks for eight sectors normalized relative to 2020). ### Macroeconomic channels and simulation results (selected outcomes) - Dominant modeled channel: TFP shocks (chronic, extreme, and flood-induced persistent effects). - Common propagation channels noted (immediate capital destruction, TFP decline, reconstruction, unemployment, financial market effects) though the paper focuses on TFP and persistent capital channels. - Consumption: - Changes closely follow Real GDP. - Under SSP 1-2.6: Russia consumption is seven percent below baseline towards 2100; Canada closely follows Russia from 2050 onward. - Under SSP 2-4.5: all regions experience higher consumption contractions; Canada and Russia experience the strongest reductions. - Stronger consumption declines in all regions after 2050 under both scenarios. - Investment: - Investment contractions can be much larger than Real GDP and consumption contractions due to: - Swift relocation of financial capital away from risky sectors/regions. - Physical investment adjustment costs discouraging reinvestment in vulnerable sectors/regions. - Stranded assets feeding back into Real GDP declines. - Capital controls amplify investment reductions. - Under SSP 1-2.6: Canada investment almost 11 percent below baseline by 2100; Russia more than eight percent below baseline by 2100. - Under SSP 2-4.5: larger initial and subsequent investment reductions; Russia experiences notably larger reductions than under SSP 1-2.6. - Broader macro-financial adjustments: - Changes in real interest rates, current account balance, real exchange rate, trade balance, and inflation observed across regions (Annexures 07–09 contain detailed panels). ### Limitations, scope, and interpretation caveats - Does not assume additional adaptation beyond that achieved by 2020. - Does not explicitly account for tropical cyclones and wildfires. - Firm locations proxied by head office only; TFP estimation sample reduced due to data-cleaning constraints and biased toward developed countries. - Empirical estimates and projections should be interpreted as an optimistic upper bound of climate impacts on productivity. - Scenarios are not assigned likelihoods and are not treated as “business as usual”; SSPs used to obtain ranges of economic consequences. ### Policy relevance and applications - Results provide scenario inputs for policymakers, central banks, supervisors, and practitioners conducting climate risk analysis and macro-financial stress testing. - Outputs can be used for standard stress testing approaches for credit and market risks and for assessing sectoral and regional vulnerabilities to physical climate risks. *Source: wpiea2023183-print-pdf — Global Economic Impacts of Physical Climate Risks, Working Paper No. WP/2023/183 (excerpt provided).* ### 1. Centre for Applied Macroeconomic Analysis, ### 1. Centre for Applied Macroeconomic Analysis, ### ABSTRACT - Evaluates the global economic consequences of physical climate risks under two Shared Socioeconomic Pathways (SSP 1-2.6 and SSP 2-4.5) using firm-level evidence. - Methodology: - Estimate historical sectoral productivity changes from chronic climate risks (gradual changes in temperature and precipitation) and extreme climate conditions (heatwaves, coldwaves, droughts, floods). - Produce forward-looking sectoral productivity changes for a global multisectoral sample of firms. - For floods, account for persistent productivity changes from damage to firms’ physical capital. - Assess macroeconomic impact using the global, multisectoral, intertemporal general equilibrium model: G-Cubed. - Key aggregate findings: - Without additional adaptation relative to that already achieved by 2020, all economies experience substantial losses under both climate scenarios, and losses increase with global warming. - Keywords: Climate change, Climate risks, Extreme events, Macroeconomic modeling - JEL Codes: C51, C53, C54, C55, C68, F41, Q51, Q54 ### INTRODUCTION — PURPOSE, GAPS, AND APPROACH - Motivation: - IPCC (2021) documents increases in frequency and intensity of natural hazards from global warming, including extreme (acute) and chronic physical climate risks. - Literature gaps: most studies emphasize chronic risks; fewer estimate extreme risks; analyses combining both within a single framework and with sectoral heterogeneity and cross-country transmission are scarce. - Two main contributions: 1. Estimate sectoral changes in Total Factor Productivity (TFP) due to physical climate risks using firm-level damage functions for four broad sectors: agriculture, mining, manufacturing, services. - Damage functions account for both chronic changes (average temperature and precipitation from a historical baseline) and some extreme risks (extreme conditions related to temperature and precipitation). - Project productivity impacts forward to 2100 for a broader sample covering the largest 1,000 firms by asset size in each IMF member nation. - Scenarios: SSP 1-2.6 (low emissions) and SSP 2-4.5 (intermediate emissions). - Second damage function source: Huizinga et al. (2017) to assess flood impacts (coastal and river) on firms’ physical capital and consequent persistent productivity effects. 2. Feed sectoral productivity shocks into G-Cubed to illustrate global economic consequences under the two SSPs. - High-level results preview: - All sectors incur losses; magnitude heterogeneous across sectors and regions. - Agriculture: productivity reducing by 10 – 20 percent in most regions under both SSPs. - Mining: losses exceed five percent in certain regions under both SSPs. - Manufacturing and services: least affected. - Macroeconomic impacts: - SSP 1-2.6: could cost the world 1.2 percent of GDP per annum on average during 2021 to 2100 (2.4 percent of GDP in 2100). - SSP 2-4.5: losses could more than double, amounting to 3.2 percent of GDP per annum on average during 2021 to 2100 (6.4 percent of GDP in 2100). - Additional modeled adjustments and outcomes: - Changes in consumption and investment patterns; investment contractions may be much larger given physical adjustment costs in G-Cubed. - Shifts in imports, exports, and macro-financial variables: real interest rates, current account balance, real exchange rate, trade balance, inflation. - Use of sectoral disaggregation highlights differing effects across sectors within regions and for the same sector across regions, demonstrating importance of general equilibrium effects. - Scope and limitations: - Does not assume additional adaptation beyond that achieved by 2020. - Does not explicitly account for tropical cyclones and wildfires. - Firm locations represented by head office only; TFP estimation sample reduced due to data-cleaning constraints. - Policy relevance: - Results can serve as scenarios for policymakers and practitioners conducting climate risk analysis, including central banks and supervisors exploring physical risks to economies and financial systems. - Macro-financial outputs can be used for standard stress testing approaches for credit and market risks. ### SECTORAL IMPACTS OF PHYSICAL CLIMATE RISK — DATA AND METHODOLOGY OVERVIEW - Sectional roadmap: - Sections 2.2–2.4: firms’ data, climate scenarios, and climate data. - Section 2.5: constructed climate indicators. - Section 2.6: empirical estimation strategy combining firm-level data with climate indicators to derive productivity impacts. - Section 2.7: results from empirical estimations and projected productivity impacts under SSPs. - Section 2.8: data to calculate persistent productivity changes from flood damage to firms’ physical capital and projected impacts under SSPs. - Section 2.9: summary. ### FIRM DATA (Section 2.2) — SAMPLES, COVERAGE, AND TFP - Samples: - Sample 1: 59,554 firms — top 1,000 non-financial firms from each IMF member nation (where available) by aggregate asset value for latest financial year after 2018; for countries with fewer than 1,000 firms in Orbis, all available firms included. - Sample 2: 20,215 firms — cleaned dataset with computed TFP available for firms after applying cleaning procedures. - Geographic and sectoral coverage: - Sample 1 spans 147 countries; firms aggregated into four broad sectors using NACE: agriculture, mining, manufacturing, services. - Sample 2 spans 48 countries. - Distribution skew: more firms from manufacturing and services than agriculture and mining; Europe provides a higher number of firms. - Data sources and period: - Financial data from Orbis for 2000 to 2018. - Firm addresses from Orbis represent head office locations only. - Data processing: - Financial data cleaned following Kalemli-Ozcan et al. (2015). - TFP computed following Ackerberg et al. (2015). - Due to lack of additional information (e.g., industry-specific deflators), cleaned dataset available for 20,215 firms only. - Representativeness assumption: - Firms in Sample 2 assumed representative of Sample 1 in regional and sectoral distribution for projecting sectoral productivity impacts under SSPs. - Historical financial aggregates: - Supplementary Annexure 2 summarizes historical average annual growth in: operating revenue, operating profit, fixed assets, capital, labor costs, material costs, and TFP across Sample 2 firms for 2001 to 2018, aggregated across sectors and UN regions. - Table: Distribution of Firms across Sectors and Countries (exact figures preserved) - Agriculture: Sample 1 (59,554) — No. of Firms 1,401; No. of Countries 91. Sample 2 (20,215) — No. of Firms 519; No. of Countries 45. - Mining: Sample 1 — No. of Firms 1,605; No. of Countries 101. Sample 2 — No. of Firms 464; No. of Countries 46. - Manufacturing: Sample 1 — No. of Firms 15,849; No. of Countries 125. Sample 2 — No. of Firms 6,945; No. of Countries 48. - Services: Sample 1 — No. of Firms 40,699; No. of Countries 143. Sample 2 — No. of Firms 12,287; No. of Countries 47. *Source: wpiea2023183-print-pdf - 1. Centre for Applied Macroeconomic Analysis,* ### 2.3 Climate Scenarios ### 2.3 Climate Scenarios ### SSP scenarios used and selection - The paper focuses on two of the five Shared Socioeconomic Pathways (SSPs) introduced in the IPCC Sixth Assessment Report: SSP1-2.6 and SSP2-4.5, which represent low and intermediate greenhouse gas emission concentration pathways respectively. - The SSP notation: the first number refers to the SSP narrative and the second number refers to the Representative Concentration Pathway (RCP) indicating approximate global radiative forcing by 2100. - The paper excludes very low, high, and very high greenhouse gas emission concentration pathways mainly due to data availability. - The paper does not attribute likelihoods to scenarios and does not treat any scenario as “business as usual.” ### Temperature outcomes under SSPs (Table 2) - SSP1-1.9 — Very low GHG emissions: CO2 emissions reduced to net zero around 2050 - Estimated Global Warming 2041-2060 (°C): 1.6 - Estimated Global Warming 2081-2100 (°C): 1.4 - 2081-2100 (Range in °C): 1.0 – 1.8 - SSP1-2.6 — Low GHG emissions: CO2 emissions reduced to net zero around 2075 - Estimated Global Warming 2041-2060 (°C): 1.7 - Estimated Global Warming 2081-2100 (°C): 1.8 - 2081-2100 (Range in °C): 1.3 – 2.4 - SSP2-4.5 — Intermediate GHG emissions: CO2 emissions around current levels until 2050, then falling but not reaching net zero by 2100 - Estimated Global Warming 2041-2060 (°C): 2.0 - Estimated Global Warming 2081-2100 (°C): 2.7 - 2081-2100 (Range in °C): 2.1 – 3.5 - SSP3-7.0 — High GHG emissions: CO2 emissions double by 2100 - Estimated Global Warming 2041-2060 (°C): 2.1 - Estimated Global Warming 2081-2100 (°C): 3.6 - 2081-2100 (Range in °C): 2.8 – 4.6 - SSP5-8.5 — Very high GHG emissions: CO2 emissions triple by 2075 - Estimated Global Warming 2041-2060 (°C): 2.4 - Estimated Global Warming 2081-2100 (°C): 4.4 - 2081-2100 (Range in °C): 3.3 – 5.7 - Source for Table 2: IPCC (2021). - Note: projected flood data availability is limited to SSP1-2.6, SSP2-4.5, and SSP5-8.5; SSP5-8.5 is excluded by the authors as an extreme scenario whose viability is heavily debated. ### Interpretation guidance - RCPs provide the warming pathways for SSPs; Hausfather and Peters (2020) discussed interpretation of RCPs in light of recent climate science developments. - The scenarios are used to obtain a range of estimates for economic consequences of physical climate risks, not to assign probabilities. *Source: Constructed from Chapter 2.3 of the provided IMF working paper PDF.* ### 2.5 have historically affected sectoral productivity ### 2.5 have historically affected sectoral productivity ### Methodology and identification - Estimation approach: Ridge regression to address collinearity among a large number of climate indicators while retaining predictors. - Controls included: - Firm- and country-specific climate indicators (manufacturing and services); agriculture and mining estimates exclude firm-specific climate indicators. - GDP growth to account for national economic growth impacts and to capture country and year-specific fixed effects. - Region-specific and year-specific fixed effects to control for unobserved time-invariant regional heterogeneities and time-variant effects, and to implicitly account for local adaptation via indicator thresholds. - Location proxy: single head office address used to represent firm location; assumption of representativeness varies by sector: - Manufacturing and services: firms can choose locations, so head office location assumed representative of domestic production network. - Agriculture and mining: primary operations typically occur in more vulnerable areas than head office location; firm-specific indicators therefore excluded for these sectors. - Linear relationship between physical climate risks and TFP assumed, following existing literature. ### Empirical samples and scope - Agriculture assessment: 519 firms across 45 countries from Sample 2; projected impacts derived for 1,401 firms across 91 countries from Sample 1. - Mining assessment: 464 firms across 46 countries from Sample 2; projected impacts derived for 1,605 firms across 101 countries from Sample 1. - Manufacturing assessment: 6,945 firms across 48 countries from Sample 2; projected impacts derived for 15,849 firms across 125 countries from Sample 1. - Regional aggregation and projections: Annexure 3 summarizes average impacts across 15 UN regions under two SSPs (SSP 1-2.6 and SSP 2-4.5) from 2021 to 2100, with variations normalized relative to 2020 levels. ### Agriculture: empirical findings and projections - Empirical impacts (Figure 3 summary): - A one percent increase in extremely dry conditions reduces agriculture productivity by more than two percent. - A one percent increase in extremely warm conditions during the night reduces agriculture productivity by more than two percent. - An increase in mean temperature by one degree reduces productivity by more than one percent. - A one percent increase in extremely wet conditions could boost productivity by almost 1.5 percent. - Interpretation: impacts operate through soil moisture, growing season timing, water-use efficiency, CO2 concentration effects, and increased vulnerability to pests, diseases, and extreme events. - Projections (SSP scenarios, Annexure 3 summary): - Under SSP 1-2.6: almost all regions experience reductions in agriculture productivity relative to their 2020 productivity changes; Central Asia, North America, and Western Europe experience the highest reductions exceeding six percentage points by 2100 relative to 2020. - Under SSP 2-4.5: all regions experience declining productivity; South Asia, North America, and Western Europe experience reductions by 2100 that exceed 15 percentage points relative to their respective productivity change in 2020. - Australia and New Zealand experience the least reduction, remaining roughly above two percentage points compared to 2020 productivity change. ### Mining and Energy: empirical findings and projections - Empirical impacts (Figure 3 summary): - An increase in mean temperature by one degree reduces mining productivity by almost three percent. - Extremely wet conditions reduce mining productivity by almost 1.5 percent. - Extremely warm conditions (day and night) and prolonged extremely dry conditions are observed to positively affect mining productivity in the sample, possibly reflecting sample composition (leading firms, adaptation/efficiency) and water-management efficiencies. - Interpretation: climate risks increase exploration, extraction, production, transportation, and decommissioning costs; rising water scarcity is a main driver of rising costs; higher temperatures can reduce labor efficiency and affect water-dependent energy generation (cooling needs, hydroelectric disruption). - Projections (SSP scenarios, Annexure 3 summary): - Under SSP 1-2.6: eight regions experience notable mining productivity reductions relative to 2020; North America shows the strongest reduction, with productivity reducing almost by five percentage points relative to 2020. - Under SSP 2-4.5: almost all regions experience notable productivity reductions relative to 2020; Northern Europe is the only exception with a reduction of less than one percentage point. ### Manufacturing: empirical findings and projections - Empirical impacts (Figure 3 summary): - Most climate indicators have lower impacts on manufacturing productivity than on other sectors. - A unit increase in mean temperature at the firm location and at the country level decreases manufacturing productivity; the average temperature increase at the firm location has a stronger effect. - Country-wide extremely warm conditions are detrimental to manufacturing productivity. - Country-wide extremely wet and extremely dry conditions reduce manufacturing productivity, potentially via effects on raw material availability. - Short-term extreme temperature indicators show mixed impacts, likely via effects on working conditions. - Interpretation: channels include direct effects on labor efficiency, increased production costs (substitution of raw materials, retrofits, equipment failures), and indirect effects via impacts on upstream (agriculture, mining, energy) and downstream (transportation) activities in supply chains. - Existing literature context: heterogeneous sectoral responses and prior projections include large potential output reductions in manufacturing absent adaptation. *Source: wpiea2023183-print-pdf - 2.5 have historically affected sectoral productivity* ### 2100. Under SSP 1-2.6, most of the regions experience manufacturing productivity contractions ### 2100. Under SSP 1-2.6, most of the regions experience manufacturing productivity contractions ### Manufacturing productivity projections - Under SSP 1-2.6, most regions experience manufacturing productivity contractions compared to their respective changes in productivity in 2020. - North America and East Asia experience the highest contractions, which lie between one and two percentage points towards 2100. - Some regions, such as Western and Southern Europe, experience minimal productivity changes. - Under SSP 2-4.5, more regions experience productivity reductions. - North and Latin America are the most vulnerable regions with their productivity reductions exceeding 1.75 percentage points by 2100 relative to their respective changes in 2020. ### Services sector impacts and empirical evidence - The service sector definition used: utilities (linked to energy), trade, accommodation and food services, transportation and warehousing, communication, financial and business services, construction and real estate activities, recreation, education and health, public administration and defense (GTAP 2022). - Physical climate risks affect services both directly and indirectly (example: construction directly; food services indirectly via agriculture). - Empirical assessment uses 12,287 firms across 47 countries from Sample 2. - Country-level increase in average mean temperature has the strongest adverse effect on service sector productivity. - Firm-specific increase in mean temperature can boost productivity. - Extreme short-term temperature conditions have mixed impacts; country-level indicators have more pronounced effects than firm-level indicators. - Extremely dry and wet country conditions reduce productivity, potentially via indirect effects through agriculture and wholesale and retail trade. - Projections derived for 40,699 firms across 143 countries from Sample 1. - Annexure 3 summarizes average impacts across 15 UN regions from 2021 to 2100 under the two SSPs. - Under SSP 1-2.6, all regions experience service sector productivity reductions. - Western Europe experiences the least change from its change in 2020. - Central Asia is most vulnerable, reaching almost a one percentage point reduction by 2100 compared to its 2020 productivity change. - Under SSP 2-4.5, all regions experience declining productivity compared to 2020; most regions observe reductions within one to two percentage points towards 2100. - Notably, Sub-Saharan Africa and Western Europe experience almost three percentage point reductions towards 2100 compared to their 2020 productivity changes. ### Summary of sectoral TFP impacts - All sectors experience losses from changes in physical climate risks; magnitude is heterogeneous across time and regions. - Agriculture is the most vulnerable sector: - Agriculture productivity reduces by 10 to 20 percentage points in most regions under the two SSPs. - Mining experiences losses exceeding five percentage points in certain regions under the SSPs. - Manufacturing and services are the least affected, potentially due to greater flexibility in location and operations to reduce exposure. - Sample composition and limitations: - Top 1,000 (or fewer) firms in each IMF member nation by asset value -> sample of 59,554 firms from 147 countries. - Estimation sample of 20,215 firms is biased toward developed countries (data limitations for TFP measures). - Empirical estimates and projections should be interpreted as an optimistic upper bound of climate impacts on productivity. - Normalizing changes relative to 2020 is used to reduce region-specific biases (such as historical adaptation measures). - Strengths of the exercise: - Incorporates both chronic and extreme climate indicators at firm and country levels within a single framework. - Covers a global sample of multisectoral firms. - Utilizes improvements to conventional estimation approaches. ### Impact of floods on sectoral capital - Extreme events (storms and floods) can cause instantaneous damage to firms by destroying physical capital stocks; effects can be non-linear and disruptive. - Damage functions linking hazard intensity to damage rates are available for riverine and coastal floods (Huizinga et al. 2017). - Huizinga et al. (2017) provide global and regional damage functions depicting fractional damage as a function of water depth and distinguish damage across agriculture, infrastructure, transport, residential buildings, commercial buildings, and industrial buildings. - Figure 4 illustrates sectoral damage functions by continent where 0 represents no damage and 1 represents complete destruction. - Mapping and data application: - Damage functions mapped to 21 NACE sectors covering the 59,554 firms in Sample 1. - Projected flood severity data from Jupiter Intelligence in terms of flood depth and fraction of land flooded under five different return periods: 1/10, 1/20, 1/50, 1/100, 1/200, 1/500 for the two SSPs from 2020 to 2100. - Continent- and sector-specific damage functions applied to firm-location flood depth and fraction of land flooded to derive average annual damage to firms’ physical capital stock under the two SSPs from 2020 to 2100. - Damage estimates averaged for the four broad sectors (Agriculture, Mining, Manufacturing, Services) and 15 UN regions to obtain loss of physical capital stock for each sector-region under the SSPs. - Findings on flood impacts (Annexure 5): - Flood impacts are unevenly distributed across regions. - Under SSP 1-2.6, firms in Asia and Africa are exposed to significant increases in flood damage over time across all sectors; firms in other regions experience moderately increasing, constant, or slightly declining damages. - Under SSP 2-4.5, patterns are similar but damages further increase over time in Asia and Africa and further decrease in some regions (example: North America for mining). - Under both scenarios, largest damages occur in Asia for all sectors except mining, for which the largest damages occur in North America. ### Modeling physical climate risks in the G-Cubed and economic channels - Section 3 outlines simulation of sectoral productivity impacts within the G-Cubed model using estimates from Section 2. - Background on modeling approaches: - Integrated Assessment Models (IAMs) have been widely used; classes include cost-benefit analysis models, biophysical models, and policy guidance models. - Alternative approaches include econometric and economic modeling: cross-sectional and panel regressions, SVAR, DSGE, CGE, ABM, and hybrid DSGE-CGE models. - Prior CGE and hybrid DSGE-CGE studies typically develop shocks using primary-sector damage functions and derive secondary and tertiary sector shocks via reliance on primary sectors; firm-level estimations are largely absent historically. - Common economic channels of propagation of physical climate risks (listed in the source): 1. Immediate physical capital destruction: modeled as a one-time increase in capital depreciation calibrated using damage rates; impacts vary across capital types. 2. Decline in TFP: chronic physical risks reduce firms’ ability to transform inputs into outputs; extreme-event damage can cause misallocation of remaining capital and persistent effects via cost of capital changes. 3. Reconstruction after extreme events: household consumption and firm investment adjustments; reconstruction investment choices differ by private vs public capital and can yield favorable post-disaster firm performance. 4. Impact on unemployment: concurrent shocks to employment can arise from large/immediate destruction of capital stock. 5. Effects from financial markets: physical shocks can reduce investor attraction to vulnerable countries/sectors, be reflected in equity or corporate bond markets, and weaken sovereign credit ratings increasing real interest on external debt. - This paper’s focus within the G-Cubed: - Focuses on the TFP channel. - Accounts for TFP shocks from both chronic and extreme climate risks (excluding floods) as estimated in Section 2.7. - Computes persistent TFP effects due to floods’ damage to firms’ physical capital. - Recognizes other channels may be important and are left to future studies. *Italic: Source: wpiea2023183-print-pdf (excerpt provided)* ### 3.3 The G-Cubed Model ### 3.3 The G-Cubed Model ### Overview - The G-Cubed is a global, multisectoral, intertemporal general equilibrium model developed by McKibbin and Wilcoxen (2013; 1999). - The model bridges econometric general-equilibrium modeling, international trade theory, and modern macroeconomics and is particularly well suited to capture climate risks due to its regional and sectoral representation. - Version used: GGG20C_v169 with eleven regions and twenty sectors. - The first twelve sectors are aggregated from the 65 sectors in the GTAP 10 database; electricity is disaggregated into electricity delivery (Sector 1) and eight electricity generation sectors (Sectors 13–20). - The model has been used in studies of climate change and transition pathways (Fernando 2023; Fernando et al. 2021; Jaumotte et al. 2021; Liu et al. 2020) and for NGFS scenario work (NGFS 2022a). ### Regions and Sectors (model configuration) - Regions (Table 4): - AUS: Australia - CAN: Canada - CHI: China - EUW: Europe - IND: India - JPN: Japan - OEC: Rest of the OECD - OPC: Oil-Exporting developing countries - ROW: Rest of the World - RUS: Russian Federation - USA: United States - Sectors (Table 5) — mapping and notes: - 1 Electricity delivery — Energy Sectors excluding Electricity Generation — Broad mapping: Manufacturing - 2 Gas extraction and utilities — Manufacturing - 3 Petroleum refining — Manufacturing - 4 Coal mining — Mining - 5 Crude oil extraction — Mining - 6 Construction — Goods and Services — Manufacturing - 7 Other mining — Mining - 8 Agriculture and forestry — Agriculture - 9 Durable goods — Manufacturing - 10 Non-durable goods — Manufacturing - 11 Transportation — Services - 12 Services — Services - 13 Coal generation — Electricity Generation Sectors — Mining - 14 Natural gas generation — Mining - 15 Petroleum generation — Mining - 16 Nuclear generation — No Shocks were Applied. - 17 Wind generation — No Shocks were Applied. - 18 Solar generation — No Shocks were Applied. - 19 Hydroelectric generation — No Shocks were Applied. - 20 Other generation — No Shocks were Applied. - Mapping note: four broad sectors used for calibrated TFP shocks are Agriculture, Mining, Manufacturing, and Services. Renewable electricity generation sectors (Sectors 16–19 and 20) were excluded from shocks because the firms in the sample do not include firms engaged in renewable electricity generation and the broad sectors do not reflect exposures to renewables in this paper. ### Model Structure and Features - Agent types and expectations: - Heterogeneous households and firms, a government, and a central bank in each region. - Representative households and firms in each sector may have forward-looking expectations or follow rules of thumb (optimal in the long run but not necessarily in the short run). Forward-looking agents smooth consumption and investment in the presence of continuing shocks. - Domestic and international linkages: - Sectoral trade linkages allow shocks in one sector to spill over to domestic and foreign sectors that rely on it. - Final impact on a sector depends on that sector’s influence on world prices. - Capital flows and adjustment: - Distinguishes physical capital from financial capital. - Financial capital can move immediately across industries and regions. - Physical capital faces sector-specific quadratic adjustment costs and adjusts sluggishly, creating stranded assets. ### Baseline Construction - Baseline starts in 2018 and is projected to 2100. - 2018 chosen as latest year with comprehensive data for calibration. - Baseline driver: region-specific sectoral productivity growth rates as a function of labor force growth and labor productivity growth. - Labor force growth rates derived from United Nations Population Prospects (2019). - Sectoral labor productivity growth (labor-augmenting technological progress) determined using a Barro-style catch-up model assuming the average annual catch-up rate to the global frontier is two percent. - Initial sectoral productivity data from Groningen Growth and Development database (2022); US sectors assumed as the frontier. - Catch-up rates varied by economy based on recent growth experiences. - Two-layer baseline interpretation: 1. Model baseline: assumes historical climate policies and adaptation measures implemented by 2018 and no future climate shocks. 2. Effective baseline simulation (2020–2100): includes climate shocks equivalent to 2020 levels and assumes adaptation for those shocks. - Climate scenarios: normalized shocks under two SSPs from 2020 to 2100 are introduced as unanticipated shocks to the baseline; simulations show how economies attempt to return to the baseline. ### Treatment of Climate-Related Shocks (link to G-Cubed sectors) - Two main types of TFP shocks introduced to replicate physical climate risks under two SSPs: 1. Sectoral TFP changes due to chronic and extreme climate risks (excluding floods), estimated and projected in Section 2.7. 2. Persistent TFP effects due to floods’ damage to firms’ physical capital (Section 2.8). - Conversion of flood damage to TFP: - Consider sector reliance on capital as proportion of total inputs using GTAP 10 Input-Output tables (Aguiar et al. 2019). - Ultimate TFP shock on a given G-Cubed sector (e.g., Coal Mining) equals the sum of: - (1) contemporaneous TFP shock on corresponding broad sector (e.g., Mining) due to chronic and extreme risks, and - (2) persistent TFP shocks from flood damage to physical capital adjusted for sectoral reliance on capital. - Because reliance on capital varies across sectors via GTAP IO linkages, two G-Cubed sectors mapped to the same broad sector can receive different final TFP shocks. - Empirical inputs: - Annexure 6: cumulative TFP shocks introduced to eight selected sectors under the two SSPs (Coal Mining, Crude Oil Extraction, Construction, Other Mining, Agriculture, Durable Manufacturing, Non-durable Manufacturing, Services); shocks normalized relative to 2020. - Supplementary Annexure 11: descriptive statistics for shocks grouped by region and sector under the two SSPs. ### Role in Simulations and Scope Limitations - The G-Cubed baseline includes historical climate policies and adaptation up to 2018 and effective shocks up to 2020 levels; reported results are relative to this baseline, i.e., additional climate shocks relative to 2020 levels. - Physical risk macro-financial channels beyond TFP shocks (e.g., broader financial sector effects) are noted as not included in this paper (see Section 3.2). *Source: wpiea2023183-print-pdf - 3.3 The G-Cubed Model* ### 4.3 Changes in Consumption and Investment ### 4.3 Changes in Consumption and Investment ### Consumption: patterns, magnitudes, and regional outcomes - Consumption changes closely follow changes in Real GDP due to income effects on consumption. - Under SSP 1-2.6: - European consumers observe the lowest consumption adjustment as they experience the lowest Real GDP contraction. - Russia experiences the highest consumption contraction which is seven percent below the baseline towards 2100. - Canada closely follows Russia from 2050 onwards. - China, the Oil Producing Countries, and Japan also experience higher contractions although they are lower than those of Canada and Russia. - Under SSP 2-4.5: - All regions experience higher consumption contractions compared to SSP 1-2.6. - Consumption reduction patterns remain similar across scenarios. - Canada and Russia experience the strongest consumption reductions under SSP 2-4.5. - Timing: - In all regions, under both scenarios, there are stronger consumption declines after 2050, reflecting substantial changes in physical risks in the medium to long term. ### Investment: mechanisms, model-specific amplifications, and regional outcomes - Investment changes follow a similar pattern to Real GDP and consumption but can be much larger than Real GDP and consumption. - Model mechanisms in G-Cubed that amplify investment changes: - Explicit distinction between physical and financial capital allows swift relocation of financial capital to sectors and regions experiencing lower risks. - Physical investment adjustment costs discourage reinvestment in sectors and regions more vulnerable to physical climate risks. - Stranded assets or idling stock of capital without productive use due to vulnerability to climate risks have feedback effects on Real GDP. - Structural features such as capital controls: economies with capital controls experience much larger investment reductions when investors respond to physical climate risks. - Comparative note: - Investment reductions in the G-Cubed are larger than in similar CGE/DSGE models and especially IAMs, partly due to the above capital distinctions and adjustment dynamics (see referenced NGFS discussion on investment adjustment costs). - Under SSP 1-2.6: - All regions experience substantial investment reductions. - Canada experiences the highest contraction which is almost 11 percent below the baseline by 2100. - Investment in Russia is more than eight percent below the baseline by 2100. - Under SSP 2-4.5: - Both the initial and subsequent investment reductions are larger than SSP 1-2.6. - Russia experiences notably larger investment reductions under SSP 2-4.5 compared to SSP 1-2.6. ### Links between consumption, investment, and wider macroeconomic dynamics - Investment contractions amplify reductions in demand for capital goods, feeding back into durable manufacturing and broader economic activity. - Financial market responses and cross-border capital flows (affected by differential productivity impacts) influence real exchange rates, trade flows, and current account balances, further affecting investment and consumption patterns across regions. - The combination of climate shocks, characteristics of agents, and structural features of economies drives the differential investment outcomes observed across regions and scenarios. *Excerpted from "4.3 Changes in Consumption and Investment" in the provided IMF content unit.* ### References ### wpiea2023183-print-pdf - References (Annexes and Supplementary Material) ### Major methodological elements and data sources - Key climate and damage data sources referenced or used: - CRU (2022) - ISIMIP (2022) - Jupiter Intelligence (flood depth and fraction flooded) - Huizinga et al. (2017) damage functions and database - Orbis Data (2022) for firm-level financials - GTAP 10 database (Aguiar et al. 2019) and GTAP Data Bases: GTAP 10 Data Base Sectors - G-Cubed Model (version GGG20C_v169) - Modeling frameworks and literature referenced include: - G-Cubed (McKibbin & Wilcoxen) - DSGE models for transmission (Hashimoto & Sudo, Xu) - Integrated assessment and impact model intercomparison (ISIMIP, IPCC, O'Neill et al. 2016) - Empirical micro- and firm-level studies on disasters, productivity, and climate impacts (Ackerberg et al. 2015; Hsiang et al. 2017; Zhang et al. 2018; Dasgupta et al. 2021; Fornino et al. 2023; others) ### Flood damage computation (firm-level) - Notation and inputs for each firm i in country c and return period RP: - Flood depth: 훿_{c,i,RP} (meters) - Fraction of land flooded: 휑_{c,i,RP} - Available return periods from Jupiter: 1-in-10 years (RP_{10}), 1-in-20 years (RP_{20}), 1-in-50 years (RP_{50}), 1-in-100 years (RP_{100}), 1-in-200 years (RP_{200}), 1-in-500 years (RP_{500}) - Damage probability for a given return period RP: - pd_{c,i,RP} = 휑_{c,i,RP} * f(훿_{c,i,RP}), where f(⋅) is the damage function from Huizinga et al. (2017) for the firm’s sector and continent. - Percentage damage for a firm D_{c,i} defined by flood depth thresholds (piecewise rule): - 0, if 훿_{c,i,RP} < 훿_{c,i,RP_{10}} - (pd_{c,i,RP_{10}} + pd_{c,i,RP_{20}})/2, if 훿_{c,i,RP_{10}} < 훿_{c,i,RP} < 훿_{c,i,RP_{20}} - (pd_{c,i,RP_{20}} + pd_{c,i,RP_{50}})/2, if 훿_{c,i,RP_{20}} < 훿_{c,i,RP} < 훿_{c,i,RP_{50}} - (pd_{c,i,RP_{50}} + pd_{c,i,RP_{100}})/2, if 훿_{c,i,RP_{50}} < 훿_{c,i,RP} < 훿_{c,i,RP_{100}} - (pd_{c,i,RP_{100}} + pd_{c,i,RP_{200}})/2, if 훿_{c,i,RP_{100}} < 훿_{c,i,RP} < 훿_{c,i,RP_{200}} - pd_{c,i,RP_{500}}, if 훿_{c,i,RP} ≥ 훿_{c,i,RP_{500}} - Expected percentage of annual damage E[D_{c,i}]: - E[D_{c,i}] = 0 * 1/10 + (pd_{c,i,RP_{10}} + pd_{c,i,RP_{20}})/2 * (1/10 − 1/20) + (pd_{c,i,RP_{20}} + pd_{c,i,RP_{50}})/2 * (1/20 − 1/50) + (pd_{c,i,RP_{50}} + pd_{c,i,RP_{100}})/2 * (1/50 − 1/100) + (pd_{c,i,RP_{100}} + pd_{c,i,RP_{200}})/2 * (1/100 − 1/200) + pd_{c,i,RP_{500}} * 1/500 ### SSP scenario narratives included in the annexes - SSP1: Sustainability – Taking the Green Road (Low challenges to mitigation and adaptation) - Emphasis on inclusive development, improved global commons management, accelerated demographic transition, reduced inequality, low material growth and lower resource and energy intensity. - SSP2: Middle of the Road (Medium challenges to mitigation and adaptation) - Social, economic, technological trends continue historically; uneven development; slow progress on sustainable development; moderate global population growth. - SSP3: Regional Rivalry – A Rocky Road (High challenges to mitigation and adaptation) - Resurgent nationalism, regional focus on energy and food security, declining investments in education and technology, strong environmental degradation in some regions. - SSP4: Inequality – A Road Divided (Low challenges to mitigation, high challenges to adaptation) - Increasing inequalities and stratification; high-tech internationally connected economy vs. fragmented low-income economies; mixed energy investments. - SSP5: Fossil-fueled Development – Taking the Highway (High challenges to mitigation, low challenges to adaptation) - Rapid technological progress and integrated markets, strong investments in human capital, exploitation of fossil fuels, resource- and energy-intensive lifestyles. ### Sample sizes and firm distributions (exact counts) - Firms used for empirical analysis: Total = 20,215 firms - Sector totals: AGR = 519, MAN = 6,945, MIN = 464, SVC = 12,287 - Region breakdown (selected examples with exact counts and proportions): - ANZ: Count_AGR 25, Count_MAN 251, Count_MIN 57, Count_SVC 506; Proportion_AGR 4.82, Proportion_MAN 3.61, Proportion_MIN 12.28, Proportion_SVC 4.12 - EEU: Count_AGR 111, Count_MAN 1,181, Count_MIN 71, Count_SVC 1,870; Proportion_AGR 21.39, Proportion_MAN 17.01, Proportion_MIN 15.30, Proportion_SVC 15.22 - NEU: Count_AGR 174, Count_MAN 831, Count_MIN 72, Count_SVC 2,654; Proportion_AGR 33.53, Proportion_MAN 11.97, Proportion_MIN 15.52, Proportion_SVC 21.60 - Country-level examples (exact counts and proportions): - AT: Count_AGR 0, Count_MAN 67, Count_MIN 1, Count_SVC 284; Proportion_MAN 0.96, Proportion_MIN 0.22, Proportion_SVC 2.31 - US: Count_AGR 0, Count_MAN 256, Count_MIN 24, Count_SVC 465; Proportion_MAN 3.69, Proportion_MIN 5.17, Proportion_SVC 3.78 - VN: Count_AGR 19, Count_MAN 264, Count_MIN 32, Count_SVC 480; Proportion_AGR 3.66, Proportion_MAN 3.80, Proportion_MIN 6.90, Proportion_SVC 3.91 - Firms used for projections: Total = 59,554 firms - Sector totals: AGR = 1,401, MAN = 15,849, MIN = 1,605, SVC = 40,699 - Region breakdown (selected examples with exact counts and proportions): - EEU: Count_AGR 341, Count_MAN 2,339, Count_MIN 165, Count_SVC 5,578; Proportion_AGR 24.34, Proportion_MAN 14.76, Proportion_MIN 10.28, Proportion_SVC 13.71 - NEU: Count_AGR 308, Count_MAN 1,596, Count_MIN 171, Count_SVC 7,456; Proportion_AGR 21.98, Proportion_MAN 10.07, Proportion_MIN 10.65, Proportion_SVC 18.32 - Country-level examples (exact counts and proportions): - CN: Count_AGR 4, Count_MAN 483, Count_MIN 32, Count_SVC 329; Proportion_AGR 0.29, Proportion_MAN 3.05, Proportion_MIN 1.99, Proportion_SVC 0.81 - US: Count_AGR 0, Count_MAN 298, Count_MIN 32, Count_SVC 669; Proportion_MAN 1.88, Proportion_MIN 1.99, Proportion_SVC 1.64 - MD (Moldova): Count_AGR 93, Count_MAN 239, Count_MIN 9, Count_SVC 612; Proportion_AGR 6.64, Proportion_MAN 1.51, Proportion_MIN 0.56, Proportion_SVC 1.50 ### Mappings, sectors, and model structure - Mapping of NACE sectors to Huizinga et al. (2017) flood-risk sector categories provided (selected examples): - A - Agriculture, forestry, and fishing → Agriculture - B - Mining and quarrying → Industrial - C - Manufacturing → Industrial - F - Construction → Infrastructure - S - Other service activities → Commercial - T - Activities of households as employers; undifferentiated goods- and services-producing activities of households for own use → Residential - Mapping of United Nations countries to UN regions and G-Cubed regions provided in detailed ISO tables (full ISO-to-GTAP10/UN/GGG20C mapping included). - Mapping of GTAP sectors to G-Cubed sectors included (examples preserving exact mapping): - 15 COA → Coal → Coal Mining - 16 OIL → Oil → Crude Oil Extraction - 32 P_C → Petroleum, coal products → Petroleum Refining - 46 ELY → Electricity → Electric Utilities - 49 CNS → Construction → Construction - 59 RSA → Real estate activities → Services - Production structure reference: G-Cubed Model (version GGG20C_v169) listed as source. ### Representative projected climate and economic results (figure and annex references) - Annexes present projected variation in climate indicators under SSP scenarios and time horizons (figures constructed from ISIMIP (2022)): - Annexure 02: Projected Variation in Climate Indicators under SSP 1−2.6 and SSP 2−4.5 (temperature and precipitation deviations from 1961−90 baseline; multi-decadal series from 2025 to 2100) - Supplementary Annexure 05 and 06: Projected Average Variation in Climate Indicators at Firm Locations under SSP 1−2.6 and SSP 2−4.5 (temperature, precipitation, extremely warm/cold/wet/dry conditions) - Annexure 03: Projected Variation in Productivity under SSP 1−2.6 and SSP 2−4.5 (sectoral productivity deviations plotted from 2025 to 2100; "MINSVC" and "AGRMAN" headings indicate aggregated sector groups) - Annexure 05: Mean Flood Damage (regional panels) under: - SSP 1−2.6 (panel showing AF R, ASA, EUR, NAM, OCN, SAM columns) - SSP 2−4.5 (panel showing AF R, ASA, EUR, NAM, OCN, SAM columns) - Annexure 06: TFP Shocks under SSP 1−2.6 and SSP 2−4.5 (country and sector-level percentage changes in productivity displayed for AUS, CAN, CHN, EUR, IND, JPN, OEC, OPC, ROW, RUS, USA) - Annexure 07: Macroeconomic Results (under SSP 1−2.6 and SSP 2−4.5) using G−Cubed Simulation Results (GGG20C_v169): - Variables displayed as Percentage Deviation from Baseline: Real GDP, Imports, Investment, Consumption, Exports (panels for AUS, CAN, CHN, EUR, IND, JPN, OEC, OPC, ROW, RUS, USA) - Annexure 08: Macro−financial Results (under SSP 1−2.6 and SSP 2−4.5) using G−Cubed Simulation Results (GGG20C_v169): - Variables include Short−term Real Interest Rate (Percentage Points from Baseline), Long−term Real Interest Rate (Percentage Points from Baseline), Trade Balance (Percentage GDP Deviation from Baseline), Real Exchange Rate (Percentage Points from Baseline), Current Account Balance (Percentage GDP Deviation from Baseline), Inflation (Percentage Points from Baseline) - Annexure 09: Sectoral Results (under SSP 1−2.6 and SSP 2−4.5) using G−Cubed Simulation Results (GGG20C_v169): - Sectoral percentage deviations from baseline for Other Mining, Services, Durable Manufacturing, Non−durable Manufacturing, Construction, Crude Oil Extraction, Agriculture, Coal Mining across country groups. ### Descriptive statistics for shocks (selected exact entries) - SSP 1-2.6 — sample of sector-region means reported (Sector / Region / Mean / Median / SD / Min / Max): - Agriculture AUS -0.68 -0.66 0.42 -1.47 0.00 - Agriculture CAN -3.84 -3.84 2.26 -7.69 0.00 - Agriculture CHN -1.91 -1.90 1.14 -3.86 0.00 - Agriculture EUR -0.33 -0.27 0.27 -0.91 0.00 - Agriculture IND -1.08 -1.05 0.66 -2.26 0.00 - Agriculture JPN -1.73 -1.72 1.02 -3.46 0.00 - Agriculture USA -3.84 -3.84 2.26 -7.69 0.00 - SSP 2-4.5 — sample of sector-region means reported (Sector / Region / Mean / Median / SD / Min / Max): - Agriculture AUS -5.06 -5.06 2.97 -10.11 0.00 - Agriculture CAN -8.58 -8.58 5.04 -17.16 0.00 - Agriculture CHN -3.15 -2.76 2.33 -7.82 0.00 - Agriculture EUR -4.17 -4.06 2.58 -8.77 0.00 - Agriculture IND -5.17 -5.16 3.06 -10.40 0.00 - Agriculture USA -8.58 -8.58 5.04 -17.16 0.00 - Durable Manufacturing USA -0.96 -0.96 0.56 -1.92 0.00 - (Full descriptive tables contained in Supplementary Annexure 11 for many sector-region combinations.) ### Supplementary historical firm performance indicators - Supplementary Annexure 02 provides historical financial performance indicators (Orbis 2022) time series panels (2001–2018) showing growth in: - Operating Revenue, Operating Profit, Fixed Assets, Capital, Labor, Material, Total Factor Productivity - Panels are presented as Annual Average Percentage Growth by region and sector grouping (ANZ, EAS, EEU, LAM, NAF, NEU, SEA, SEU, SSA, WAS, WEU). *Source: Extracted from the References and Annexes of "Global Economic Impacts of Physical Climate Risks, Working Paper No. WP/2023/183".* --- _Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023183-print-pdf.pdf_