## The Economic Impacts on Germany of a Potential Russian Gas Shutoff (INTRODUCTION; content as of June 22, 2022) ## Source details **Canonical URL:** [The Economic Impacts on Germany of a Potential Russian Gas Shutoff (INTRODUCTION; content as of June 22, 2022)](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022144-print-pdf.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022144-print-pdf.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022144-print-pdf.pdf.json) --- ### Exposure and recent developments - Gas accounts for "just over a quarter" of Germany’s energy use and is "almost all" imported. - As of end-April 2022, "about 35 percent of Germany’s natural gas imports (33 billion cubic meters (bcm) a year) came from Russia", compared with "12 percent of crude oil, and 8 percent of coal". - Historical shares: "In previous years, 55 percent of gas, 35 percent of oil and 50 percent of coal imports came from Russia." - Russian imports are hard to substitute: limited pipelines to alternative wells and limited domestic regasification infrastructure; LNG must be imported via terminals in neighboring countries. - Regional gas price example range in 2022Q2: "US$ 16.7 per MWh in Louisiana, USA to US$110 per MWh in the Netherlands." - June 2022 supply disruptions: - Russia reduced flows through Nord Stream 1 by "around 60 percent" and flows into Germany at Waidhaus fell "some 69 percent" (cuts would amount to "56 bcm a year" if they persist). - On May 11, 2022, Russia sanctioned Gazprom Germania; Gazprom Germania's imports were "about 4 percent of all German gas imports" (media reports). - Shell's German subsidiary reportedly denied further supplies in May, amounting to "some 1 percent of all German gas imports." - As of June 2022, Russian gas supplies were partially or fully stopped to: Austria, Bulgaria, Czech Republic, Denmark, Finland, France, Germany, Italy, the Netherlands, Poland, and Slovakia. ### Government response and mitigation measures - Target: reduce the share of gas imported from Russia to "10 percent by mid-2024." - Storage and procurement measures: - Operators required to fill leased storage up to "90 percent of capacity by November 1, 2022" or risk forfeiture to government. - Government bought "1 bcm of gas in March 2022" and offered credit lines via KfW to THE for further purchases. - Infrastructure and fuel-switching measures: - Record high imports of LNG through Belgium and the Netherlands and measures limiting reverse pipeline flows. - Proposed laws to enable reactivation of coal-fired power plants until "March 2024". - Leased four FSRUs; envisaged first FSRU operational at Wilhelmshaven "by the end of the year", second in Brunsbüttel "in early 2023", with capacities "7.5 bcm and 5 bcm per year, respectively." Remaining two FSRUs have capacities of "5 bcm per year" and expected operational by "2024". - Proposed legal amendments: - Power to take control of critical energy infrastructure under private control. - Authority to allow gas suppliers to pass on higher prices to customers on fixed-price contracts to preserve supplier solvency. - Government concluded nuclear power cannot help offset gas shortages (reference to March 7, 2022 assessment: extension "is not recommended"). ### Literature review: GDP and inflation impacts (selected outcomes) - Summary: recent studies estimate a wide range; impacts in Germany in 2022 placed at "up to 5 percent of GDP in the first year", and "up to 2.6 percentage points’ higher inflation in the same year." - Selected study outcomes (Gas loss; GDP loss where reported): - ECB (2022): "10"; "0.7" - OECD (2022): "20"; "0.9" - Full model (unspecified): "0.2-0.3" - Simplified model 1 (unspecified): "1.3" - Simplified model 2 (unspecified): "2.2" - Study: "9"; "1.5" - Study: "36"; "3.4" - German Council of Economic Experts (March 2022): "30"; "2.0" - Joint Economic Forecast (2022): "n.a." - Bundesbank (April 2022): "40"; "5.0" - Bundesbank (June 2022): "31 in winter '22/'23, 11 in winter '23/'24"; with frictions "3", without frictions "0.4" - Bachmann et al. (2022): "30" - Schnittker et al. (2022): sequence "1.5 in '22, 6.75 in '23, 4.5 in '24" and alternative "0.8 in '22, 5.3 in '23" - Drivers of divergence across studies: - Comprehensiveness of transmission channels (supply-only vs. supply + demand amplification). - Shock size, composition and duration (examples: 30 percent reduction; 40 percent reduction over 9 months; full cessation scenarios). - Treatment of inventories and replacement (Bundesbank assumes no Russian gas replaced; GCEE assumes one-quarter replaced). - Rationing approaches across households, power sector, and firms (household protection vs. proportionate rationing vs. sectoral targeting). - Substitutability of gas with other inputs (higher substitutability lowers output losses). ### Methodology and scenario framing - Projections use information "as of June 22, 2022" and analyze monthly flows between "June 1, 2022 and end- 2027." - Baseline captures ongoing curtailments; adverse assumes "all remaining flows from Russia to the European Union stop on June 1." - LNG imports via neighboring countries assumed to continue and ramp up with Germany’s new FSRUs "even in the adverse scenario." - In the baseline, Russian imports account for "35 percent" of German consumption in April and May 2022. - To capture curtailment through Nord Stream 1 and Waidhaus in June 2022, these import quantities are scaled down by "45 percent in June and by 60 percent thereafter." - Under the baseline, Germany slowly reduces its Russian import share to zero by "end-2024", in line with the government’s target. --- ### Baseline price, infrastructure, inventories, and demand assumptions ### Prices and imports - Baseline average daily wholesale natural gas prices in Germany assumed: - US$104 a MWh in 2022 - US$83 in 2023 - US$64 in 2024 - US$49 in 2025 - Observed net gas imports from all countries Jan–May 2022: "8.1 bcm a month." - IMF estimate of Russia’s portion of German consumption in that period: "3 bcm a month." - Net inventory accumulation over Jan–May 2022: negligible. ### FSRU commissioning timeline (baseline) - December 2022: "7.5 bcm/year" - February 2023: "5 bcm/year" - September 2023: "5 bcm/year" - February 2024: "5 bcm/year" ### Gas-to-coal switching and generation - Activation assumed from July 2022 of whole grid and capacity reserves of coal-fired power stations in baseline. - Power stations assumed to substitute for around "5 GW" of power capacity, equivalent to "4.5 bcm of gas a year." - Government assumed to extend use of these power stations from "March 2024 to March 2027" in the baseline. - Adverse scenario: same coal-fired generation assumptions as baseline. ### Consumption, demand elasticities, and immediate demand responses - Baseline consumption pattern: average levels and seasonal pattern of last five years ("90.5 bcm over 2017–2021"), adjusted down by about "1.8 bcm a year" to reflect 2022 import price increases. - Temperature assumption: temperatures are in line with their recent averages. - Demand elasticities and estimated immediate effects of 2022 price increases: - Destatis: gas import prices increased by "115 (log) percent" between the 2019–2021 average and Jan–Mar 2022 average. - Elasticity of passthrough to retail gas prices used: "0.05". - Household demand elasticity used: "-0.15". - Estimated household demand fall from observed price increases: about "0.2 bcm." - Passthrough to producer gas prices used: "0.44"; estimated industrial demand fall: about "1.6 bcm." - The analysis does not explicitly allow for further demand compression (beyond gas-to-coal switching) such as efficiency gains. ### Storage capacity and dynamics - Germany gas storage capacity: about "24.8 bcm" (just over two months’ winter consumption of "22 bcm"). - Baseline inventory target outcome: storage levels around "80 percent of capacity by November 1" (short of government target of "90 percent"). - Adverse scenario inventory rule: governments remove all requirements to fill gas storage facilities; inventories drawn down with perfect foresight while always maintaining "10 percent of capacity" for ‘base’ gas. - Projections indicate dangerously low storage levels (just above "10 percent of capacity") in all future winters out to winter "2026/2027" in the baseline. ### Renewables and nuclear effects (baseline & adverse) - Steady increase of renewable power equivalent to about "2.7 bcm a year." - Phasing out of nuclear reactors this year equivalent to about "3.5 bcm a year." - These changes affect both baseline and adverse scenarios equally. --- ### Projected shortages under full Russian shutoff (assume shutoff on June 1, 2022) ### Period-specific shortages (bcm) and percent of national consumption - Shortages (bcm): - Feb–May 2022: "3.9" - Jun–Dec 2022: "3.9" - 2023: "8.9" - 2024: "3.4" - 2025: "0.9" - 2026: "0.4" - 2027: "0.6" - Shortage as percent of national consumption in period: - Feb–May 2022: "9" - Jun–Dec 2022: "10" - 2023: "4" - 2024: "1.0" - 2025: "0.5" - 2026: "0.7" ### Inventories, replacement ratios, and Russian supplies lost (bcm) - Replacement ratio ((net release of inventories + inventories not built up due to shutoff) / Russian supplies lost ×100): - Feb–May 2022: "44" - Subsequent periods: reported as "< 0" in table fragment. - Net release of inventories (bcm) by period: - Feb–May 2022: "2.1" - Jun–Dec 2022: "0.5" - 2023: "-2.4" - 2024: "-0.3" - 2025: "0.2" - 2026: "3.4" - Inventories not built up due to shutoff (bcm): - Feb–May 2022: "0.9" - Jun–Dec 2022: "-1.1" - 2023: "2.0" - 2024: "-0.7" - 2025: "-0.7" - 2026: "-4.0" - Russian supplies lost by period (bcm): - Feb–May 2022: "7" - Jun–Dec 2022: "8" - 2023: "3" - 2024 onward: "0" ### Demand compression scenarios (percent of non-protected consumption in period) - If only households (HHs) are protected: - Feb–May 2022: "13" - Jun–Dec 2022: "14" - 2023: "6" - 2024: "1.5" - 2025: "0.7" - 2026: "1.0" - If households and energy transformation are protected: - Feb–May 2022: "19" - Jun–Dec 2022: "22" - 2023: "8" - 2024: "2.3" - 2025: "1.1" - 2026: "1.6" ### Cumulative demand compression (percent of cumulative non-protected consumption) from Jun 2022 - If only HHs are protected: - Dec 2022: "13" - Dec 2023: "14" - Dec 2024: "11" - Dec 2025: "8" - Dec 2026: "6" - Dec 2027: "5" - If HHs and energy transformation are protected: - Dec 2022: "19" - Dec 2023: "21" - Dec 2024: "16" - Dec 2025: "12" - Dec 2026: "10" - Dec 2027: "8" ### Seasonal concentration and distributional effects - Shortfalls concentrated in winters: national consumption would need to contract by "10–20 percent" each month between November 2022 and March 2023, with similar contractions in winter 2023/24. - If households protected per Germany’s Emergency Plan for Gas, shortages fall on firms: - Firms’ consumption cuts required: "13 percent" in remainder of 2022, "14 percent" in 2023, "6 percent" in 2024. - If energy transformation activities also protected, remaining non-energy firms cuts required: "19 percent" in remainder of 2022, "22 percent" in 2023, "8 percent" in 2024. - Germany would have sufficient supplies for protected customers even in worst winter month, implying Germany would likely not be eligible to request gas supplies from neighbors under existing or future solidarity agreements. ### Uncertainty around shortages - IMF monthly shortages do not exceed "21 percent" of consumption in any month; Joint Economic Forecast (2022) median scenario reaches "60 percent" shortage in October 2022. - Factors that could increase shortages: - Infrastructure bottlenecks and cross-border constraints (ENTSO-G notes significant bottlenecks). - Germany supplying more gas than usual to neighboring countries (estimates suggest neighboring winter shortage could amount to "9 bcm" under certain assumptions). - Colder or longer winters could increase consumption by up to "3.1 bcm a year." - Factors that could reduce shortages: - Further reductions in household demand if higher prices pass through: up to "0.8 bcm a year" additional household demand reduction under extreme scenario considered. --- ### GDP impacts: models, estimated effects, timing, and sensitivity ### Modeling approach (three channels) - Captures: - Production cuts due to lack of gas (direct). - Downstream amplification through intermediate products (second-round; one round). - Increases in uncertainty (uncertainty-growth model using local projections). - Two models combined: - Sectoral firms’ production model (17 sectors; nested production function: Cobb–Douglas converting energy and intermediate inputs; CES combining gas, oil, coal and renewables into energy). - Uncertainty-growth model (local projections) estimating historical relationship between uncertainty measures and GDP deviations from trend. - Second-round impacts assumed to materialize 6 months after the direct impact. - Baseline projections of GDP and inflation from IMF (forthcoming). ### Key parameters and calibrations - 17 sectors; constant returns to scale. - Elasticity of substitution between gas and other energy types (ε) varies with horizon (sequence): "0.10", "0.12", "0.16", "0.20", "0.30", "0.40" (June–Dec 2022 to up to Dec 2027). - Implied national expenditure share on gas: α = "0.025". - One-round downstream amplification only. ### Estimated impacts (direct, second-round, uncertainty, combined) - Direct impact: - A "19 percent" contraction in firms’ gas consumption between June and December 2022 reduces GDP by "0.9 percent" relative to baseline in this period. - Over June 2022–December 2027 cumulative reduction in gas consumption "8 percent" → direct effect reduces GDP by less than "0.05 percent." - Second-round (downstream) effects for June–Dec 2022: reduce GDP by a further "0.9 percent", felt 6 months later (Dec 2022–June 2023). - Annualized direct + second-round (calendar years; percent deviation from baseline GDP levels): - 2022: "-0.7 percent" - 2023: "-1.8 percent" - 2024: "-0.4 percent" - Uncertainty channel: - A one-standard-deviation increase in uncertainty could lower GDP by about "1 percent" relative to trend up to two years after the shock. - Assumed shutoff raises uncertainty to the 95th percentile in 2022Q2 and then decays by half from 2022Q3 to 2023Q1. - This uncertainty shock reduces GDP by "0.8 percent" in 2022, "0.9 percent" in 2023, and "0.1 percent" in 2024. - Combined effects (preferred scenario: households and energy transformation protected) — real GDP level (percent deviation from baseline): - 2022: "-1.5" - 2023: "-2.7" - 2024: "-0.4" - Cumulative GDP loss between 2022 and 2024: "4.8 percent of 2021 GDP." - Alternative protection scenario (only households protected) — real GDP level (percent deviation from baseline): - 2022: "-1.2" - 2023: "-1.9" - 2024: "-0.4" ### Timing, decomposition, and permanence - Losses in 2022 driven mostly by uncertainty effects; downstream amplification effects materialize with lag. - Losses in 2023 roughly evenly split between direct, second-round, and uncertainty channels. - From 2024 onwards, residual losses driven almost entirely by downstream amplification effects. - Losses are permanent relative to baseline (not recovered by deferring production, consumption, or investment to subsequent years). ### Sensitivity to rationing schemes - Allocation of scarce gas across sectors materially affects GDP losses. - Baseline assumption: unprotected sectors face proportional cuts (e.g., "19 percent" cut for each unprotected sector in June–Dec 2022). - Alternative rationing schemes can minimize direct or second-round impacts by prioritizing cuts by gas intensity or downstreamness. - Table summary (Between June and December 2022) — illustrative percent cuts to gas consumption and resulting impacts on output (selected exact entries): - Percent cuts by sector (selected rows across four illustrative scenarios shown as in source): - Non-metallic minerals: "19 | 50 | 50 | 50" - Iron, steel and non-ferrous metals: "19 | 19 | 0 | 0" - Paper, pulp & printing: "19 | 0 | 0 | 0" - Chemical & petrochemical: "19 | 50 | 50 | 23" - Food, beverages & tobacco: "19 | 50 | 5 | 0" - Households: "0 | 0 | 10 | 10" - Energy sector: "0 | 0 | 0 | 10" - Direct impact on output (corresponding columns): "0.88 | 0.31 | 0.21 | 0.12" - Second-round impact on output (corresponding columns): "0.88 | 0.32 | 0.20 | 0.16" - Note: direct and second-round impacts are not additive in the table because second-round impacts can occur with lag. - Household sharing of burden: - If households share in rationing, direct GDP losses can fall by one-third (from "0.31 percent" to "0.21 percent"). - Second-round losses can fall by over one-third (from "0.32 percent" to "0.20 percent"). - Example: a "10 percent" reduction in households’ gas consumption ≈ reducing heating temperature by about "2 degrees" over heating period (BDEW, 2022). --- ### Inflation impacts and passthrough ### Estimation framework and historical passthrough - Uses local projection with state-dependent passthrough; allows larger international energy price shocks to have larger effects. - Defines large international energy price increase following Hamilton (2003); uses Brent crude and Netherlands TTF as international oil and gas prices. - Sample: "1996Q1 to 2022Q1." - Historical passthrough magnitudes: - Passthrough from international to domestic gas prices averages around "0.02" over 2-3 quarters. - Passthrough rises to "0.1" following large increases in international gas prices (fivefold increase). - Contemporaneous passthrough of international oil price averages at "0.2" in normal times but jumps to "0.3" in a price surge. - Passthrough from HICP energy inflation to core inflation: in normal times barely affects core inflation; in large energy price jumps passthrough reaches "0.05" contemporaneously. - International gas passthrough is more lagged than oil due to 12-month fixed gas and electricity contracts. ### Shock scenarios and inflation outcomes (deviation from baseline annual inflation) - International gas prices rise in 2022Q2 by "110", "221" or "367 percent" and last for 12 months. - If international gas prices rise by "110 percent": - German inflation about "¾ of a percentage point" higher in 2022 than without the shutoff. - German inflation about "1¼ percentage points" higher in 2023 than without the shutoff. - If international gas prices rise by "370 percent": - Inflation higher by about "2 percentage points" in 2022 and "3.5 percentage points" in 2023. - If a "370 percent" increase occurs and legal powers permit contemporaneous pass-through by firms (regardless of contracts): - Inflation could be "3¼ percentage points" higher than without the shutoff in both 2022 and 2023. - Caveats: - The international gas price increases considered are without precedent prior to 2021; passthrough and consumer inflation could be greater than estimates. - Estimates larger than Bundesbank (2022b) (Bundesbank: "0.5" in 2022, "1.5" in 2023 and "0.25" in 2024, percentage points). --- ### Policy discussion and recommended measures - Objectives: enhance energy security, build resilience by securing new supplies and accelerating renewables. - Key recommended measures: - Fill gas storage facilities before winter. - Enhance energy efficiency and accelerate renewable deployment. - Ensure financial support for energy companies preserves solvency without exacerbating scarcity. - Maintain international cooperation (REPowerEU referenced) and develop rationing/distribution plans. - Renewable energy acceleration: - "Easter package" proposals aim to raise renewables to "80 percent" of electricity demand by 2030 (from "42 percent" in 2021). - Increase public tenders for solar and onshore wind to "22" and "10 GW a year", respectively. - Financial and solvency measures: - Government measures to ensure energy company solvency are appropriate given rising import prices and volatility. - Monitor financial health of energy companies (regular calls or surveys suggested). - Government plans: loans through KfW, extensions of loan guarantee programs, loans to energy producers to post margins on hedging contracts. - Targeted income support for vulnerable households suggested (one-off energy relief payments or housing benefits). - Energy-saving and compensation mechanisms: - Promote energy-saving campaigns (analogue to Japan’s setsuden). - Programs to exchange gas heaters for heat pumps. - Develop auction mechanism to compensate firms for reducing gas consumption; consider similar schemes for households where feasible. - Smart meters not widely available; provide up-to-date, online gas usage statistics as alternative. - Infrastructure and operational actions: - Publish a clear diagnosis of infrastructure gaps under full shutoff; network development plan consultation was postponed and could be expedited. - Technical cooperation with neighbors (e.g., resolve ‘odorization’ differences with France) and explore additional LNG/FSRU flows (Dunkirk, Eemshaven). - Germany has signed solidarity agreements with Denmark and Austria under EU Regulation 2017/1938; further agreements possible. - Rationing policy design considerations: - Rationing decisions should weigh economic implications alongside social welfare, environmental, health, technical and legal constraints. - Economic analysis can inform but not conclusively determine rationing choices; major decisions require political participation. - The Federal Network Agency has collected firm-level gas usage information and may be well placed to act as central authority for day-to-day rationing. --- ### Annex I — Production model (compact technical summary) - Representative firm sectoral production function: - Y_k = Z_k E_k^{α_Ek} Π_{j=1}^J X_{jk}^{α_{jk}}. - Energy aggregate (CES) combining gas and other energy: - E_k = (ω e_gas_k^{(ε−1)/ε} + (1−ω) e_oth_k^{(ε−1)/ε})^{ε/(ε−1)}. - Kmenta (1967) second-order approximation used for CES to derive: - dln Y_k^d ≈ α_Ek ( ω dln e_gas_k + (1/2)(1−1/ε) ω(1−ω) dln e_gas_k dln e_gas_k ). - Calibration: - α_E from energy expenditure share; ω from gas share in energy inputs. - ε (elasticity of substitution) central to sensitivity: higher ε → easier substitution → smaller impact. - Second-round (input-output) propagation: - dln Y_k ≈ α_Ek dln E_k + Σ_j α_{jk} dln Y_j. - Matrix form: dln Y_s⃗ = (I − A')^{−1} dln Y_d⃗ (Leontief inverse L = (I − A)^{−1}). - Analysis focuses on first-round network propagation Σ_j a_{jk} dln Y_j^d given timing and substitutability considerations. - Aggregation: - dln GDP^d = Σ_k λ_k dln Y_k^d. - dln GDP^s = Σ_k λ_k dln Y_k^s. - λ_k = V_k / Σ_k V_k calibrated from input-output table. *Italic: IMF Working Paper — The Economic Impacts on Germany of a Potential Russian Gas Shutoff (INTRODUCTION; content as of June 22, 2022).* ### INTRODUCTION __________________________________________________________________________________ 4 ### INTRODUCTION ### Exposure and risk - Gas accounts for "just over a quarter" of Germany’s energy use and is "almost all" imported. - As of end-April 2022, "about 35 percent of Germany’s natural gas imports (33 billion cubic meters (bcm) a year) came from Russia", compared with "12 percent of crude oil, and 8 percent of coal". - Russian gas imports are difficult to substitute due to lack of pipelines to alternative wells and limited domestic regasification infrastructure; LNG must be imported via terminals in neighboring countries. - 2022Q2 regional gas price example range: "US$ 16.7 per MWh in Louisiana, USA to US$110 per MWh in the Netherlands." ### Recent developments (as of June 2022) - In June 2022, Russia reduced flows through Nord Stream 1 by "around 60 percent" and flows into Germany at Waidhaus fell "some 69 percent" (cuts would amount to "56 bcm a year" if they persist). - On May 11, 2022, Russia sanctioned Gazprom Germania; Gazprom Germania's imports were "about 4 percent of all German gas imports" (media reports). - Shell's German subsidiary reportedly denied further supplies in May, amounting to "some 1 percent of all German gas imports." - As of June 2022, Russian gas supplies were partially or fully stopped to: Austria, Bulgaria, Czech Republic, Denmark, Finland, France, Germany, Italy, the Netherlands, Poland, and Slovakia. ### Government response and mitigation measures - Government target: reduce the share of gas imported from Russia to "10 percent by mid-2024." - Storage and procurement measures: - Operators required to fill leased storage up to "90 percent of capacity by November 1, 2022" or risk forfeiture to government. - Government bought "1 bcm of gas in March 2022" and offered credit lines via KfW to THE for further purchases. - Infrastructure and fuel-switching measures: - Record high imports of LNG through Belgium and the Netherlands and measures limiting reverse pipeline flows. - Proposed laws to enable reactivation of coal-fired power plants until "March 2024". - Leased four floating storage and regasification units (FSRUs); envisaged first FSRU operational at Wilhelmshaven "by the end of the year", second in Brunsbüttel "in early 2023", with capacities "7.5 bcm and 5 bcm per year, respectively." Remaining two FSRUs have capacities of "5 bcm per year" and expected operational by "2024". - Proposed legal amendments: - Power to take control of critical energy infrastructure under private control. - Authority to allow gas suppliers to pass on higher prices to customers on fixed-price contracts to preserve supplier solvency. - Assessment: government concluded nuclear power cannot help offset gas shortages. ### Literature review: estimated GDP and inflation impacts - Summary statement: recent studies estimate a wide range of economic impacts; results put the impact in Germany in 2022 at "up to 5 percent of GDP in the first year", and "up to 2.6 percentage points’ higher inflation in the same year." - Selected study outcomes (percent gas loss, GDP loss where reported): - ECB (2022): Gas loss "10", GDP loss "0.7" - OECD (2022): Gas loss "20", GDP loss "0.9" - Full model (unspecified study row): Gas loss "0.2-0.3" - Simplified model 1 (unspecified): Gas loss "1.3" - Simplified model 2 (unspecified): Gas loss "2.2" - Study with gas loss "9": GDP loss "1.5" - Study with gas loss "36": GDP loss "3.4" - German Council of Economic Experts (March 2022): Gas loss "30", GDP loss "2.0" - Joint Economic Forecast (2022): Gas loss "n.a." - Bundesbank (April 2022): Gas loss "40", GDP loss "5.0" - Bundesbank (June 2022): Gas loss "31 in winter '22/'23, 11 in winter '23/'24"; with frictions "3", without frictions "0.4" - Bachmann et al. (2022): Gas loss "30" (GDP loss not shown in table fragment) - Schnittker et al. (2022): GDP loss sequence "1.5 in '22, 6.75 in '23, 4.5 in '24" and alternative "0.8 in '22, 5.3 in '23" - Di Bella et al. (2022): (entries shown without numeric detail in table fragment) - Cross-study drivers of divergence: - Comprehensiveness of transmission channels: supply-only models versus models combining supply and demand-side amplification (sticky prices, unemployment, tighter financial conditions). - Shock size, composition and duration: assumptions range (e.g., 30 percent reduction, 40 percent reduction over 9 months, full cessation of energy and non-energy trade in some scenarios). - Treatment of inventories and replacement: examples include Bundesbank (2022a) assuming no Russian gas replaced; GCEE assumes one-quarter replaced; Joint Economic Forecast expects shortages to emerge in winter after inventories exhausted. - Approach to rationing across households, power sector, and firms: some studies assume households protected (burden on firms), others proportionate rationing; sectoral targeting (stricter on energy-intensive sectors) can change outcomes. - Substitutability of gas with other inputs: higher substitutability lowers output losses; short-term assumptions of low substitutability yield larger impacts. ### Paper structure - Section 2: estimates of potential gas shortages from a Russian shutoff. - Section 3: potential impacts of shortages on economic activity. - Section 4: importance of gas rationing policy for economic estimates. - Section 5: estimates of effects on German inflation. - Annex I: Production Model details. ### Key methodological note - Projections use information "as of June 22, 2022" and analyze monthly flows between "June 1, 2022 and end- 2027." ### Notable findings reported in literature discussion - Di Bella et al. (forthcoming): complementarities between gas and capital can magnify impact from "1.3 to 2.1 percent of GDP" for Germany. - Albrizio et al. (forthcoming): global LNG market sharing the burden could make economic impacts on Europe "five times smaller." ### Footnotes and contextual figures (preserved) - Historical shares: "In previous years, 55 percent of gas, 35 percent of oil and 50 percent of coal imports came from Russia." - Government assessment on nuclear extension: reference to March 7, 2022 assessment concluded that "an extension of the service life of the three remaining nuclear power plants is not recommended." ### Transition to scenario analysis - The paper calculates shortages as the difference between gas consumption under baseline and adverse scenarios, projecting monthly flows from "June 1, 2022" to "end- 2027", with baseline capturing ongoing curtailments and adverse assuming "all remaining flows from Russia to the European Union stop on June 1." ### Key assumptions for scenario projections (excerpt) - LNG and Russian imports: - LNG imports via neighboring countries rose sharply in early 2022 and are assumed to continue and to ramp up with Germany’s new FSRUs throughout the forecast horizon, "even in the adverse scenario." - In the baseline, Russian imports account for "35 percent" of German consumption in April and May 2022. - To capture curtailment through Nord Stream 1 and Waidhaus in June 2022, these import quantities are scaled down by "45 percent in June and by 60 percent thereafter." - Under the baseline, Germany slowly reduces its Russian import share to zero by "end-2024", in line with the government’s target. *Source: IMF Working Papers — The Economic Impacts on Germany of a Potential Russian Gas Shutoff (INTRODUCTION; content as of June 22, 2022).* ### 5.1 bcm a month, which is the difference between observed net gas imports from all countries (8.1 bcm a ### The Economic Impacts on Germany of a Potential Russian Gas Shutoff ### Baseline price and import assumptions - Average daily wholesale natural gas prices in Germany assumed in the baseline: US$104 a MWh in 2022, US$83 in 2023, US$64 in 2024 and US$49 in 2025. - Observed net gas imports from all countries between January and May 2022: 8.1 bcm a month. - IMF estimate of Russia’s portion of German consumption in that period: 3 bcm a month. - Net inventory accumulation over January–May 2022: negligible. - Modeling approach: net gas imports are modeled rather than gross imports and exports; assumption that a 60 percent curtailment in Russian supplies to Germany in June can be passed on as a 60 percent decline in German re-exports of Russian gas. - June Russian supply cut rounded to 60 percent (dashboard data indicated Nord Stream 1 and Waidhaus fell by 60 and 69 percent respectively, or 63 percent in total). ### Infrastructure expansion and fuel switching measures - FSRU (floating storage regasification units) commissioning timeline in baseline: - December 2022: 7.5 bcm/year - February 2023: 5 bcm/year - September 2023: 5 bcm/year - February 2024: 5 bcm/year - Gas-to-coal switching: - Assumed activation of whole grid and capacity reserves of coal-fired power stations from July 2022 in the baseline. - Power stations assumed to substitute for around 5 GW of power capacity, equivalent to 4.5 bcm of gas a year. - Government assumed to extend use of these power stations from March 2024 to March 2027 in the baseline to prevent gas shortages. - Adverse scenario makes same assumptions regarding coal-fired power generation as baseline. ### Gas consumption and demand adjustments - Baseline consumption pattern: average levels and seasonal pattern of last five years (90.5 bcm over 2017–2021), adjusted down by about 1.8 bcm a year to reflect import price increases seen so far in 2022. - Temperature assumption: temperatures are in line with their recent averages. - Adverse scenario: consumption calculated as a residual. - Demand elasticities and estimated immediate effects of 2022 price increases: - Destatis: gas import prices increased by 115 (log) percent between the 2019–2021 average and Jan–Mar 2022 average. - Elasticity of passthrough to retail gas prices used: 0.05. - Household demand elasticity used: -0.15. - Estimated household demand fall from observed price increases: about 0.2 bcm. - Passthrough to producer gas prices used: 0.44; estimated industrial demand fall: about 1.6 bcm. - The analysis does not explicitly allow for further demand compression (beyond gas-to-coal switching) such as efficiency gains, because such reductions could come with costs to economic output which the study seeks to quantify. ### Inventories and storage capacity - Germany gas storage capacity: about 24.8 bcm (just over two months’ winter consumption of 22 bcm). - Baseline inventory dynamics: accumulate in summers and drawn down in winters following seasonal patterns. - Adverse scenario inventory rule: government assumed to remove all requirements to fill gas storage facilities; inventories drawn down with perfect foresight while always maintaining 10 percent of capacity to allow for ‘base’ gas. - Intermediate storage target outcome in baseline: storage levels around 80 percent of capacity by November 1 (short of government target of 90 percent). - Projections indicate dangerously low gas inventory levels (just above 10 percent of capacity) in all future winters out to winter 2026/2027 in the baseline. ### Renewable and nuclear power effects - Steady increase of renewable power equivalent to about 2.7 bcm a year. - Phasing out of nuclear reactors this year equivalent to about 3.5 bcm a year. - These changes affect both baseline and adverse scenarios equally and are not specially adjusted because shortages are measured as deviations of the adverse scenario from the baseline. ### Projected gas shortages under a full Russian shutoff (scenario: Russian gas shutoff to Germany on June 1, 2022) - Period-specific shortages (bcm): - Feb–May 2022: 3.9 (shortage for Feb–May row shown in table) - Jun–Dec 2022: 3.9 - 2023: 8.9 - 2024: 3.4 - 2025: 0.9 - 2026: 0.4 - 2027: 0.6 - Shortage as percent of national consumption in period: - Feb–May 2022: 9 - Jun–Dec 2022: 10 - 2023: 4 - 2024: 1.0 - 2025: 0.5 - 2026: 0.7 - Replacement ratio ((net release of inventories + inventories not built up due to shutoff) / Russian supplies lost ×100): 44 for Feb–May 2022; reported as < 0 for subsequent periods in table. - Net release of inventories by period (bcm): 2.1 (Feb–May 2022), 0.5 (Jun–Dec 2022), -2.4 (2023), -0.3 (2024), 0.2 (2025), 3.4 (2026). - Inventories not built up due to shutoff (bcm): 0.9 (Feb–May 2022), -1.1 (Jun–Dec 2022), 2.0 (2023), -0.7 (2024), -0.7 (2025), -4.0 (2026). - Russian supplies lost by period (bcm): 7 (Feb–May 2022), 8 (Jun–Dec 2022), 3 (2023), 0 (2024 onward). ### Demand compression scenarios (percent of non-protected consumption in period) - If only households (HHs) are protected: - Feb–May 2022: 13 - Jun–Dec 2022: 14 - 2023: 6 - 2024: 1.5 - 2025: 0.7 - 2026: 1.0 - If households and energy transformation are protected: - Feb–May 2022: 19 - Jun–Dec 2022: 22 - 2023: 8 - 2024: 2.3 - 2025: 1.1 - 2026: 1.6 ### Cumulative demand compression (percent of cumulative non-protected consumption) from Jun 2022 until listed dates - If only HHs are protected: - Dec 2022: 13 - Dec 2023: 14 - Dec 2024: 11 - Dec 2025: 8 - Dec 2026: 6 - Dec 2027: 5 - If HHs and energy transformation are protected: - Dec 2022: 19 - Dec 2023: 21 - Dec 2024: 16 - Dec 2025: 12 - Dec 2026: 10 - Dec 2027: 8 ### Seasonal concentration and distributional effects - Shortfalls concentrated in winters: national consumption would need to contract by 10–20 percent each month between November 2022 and March 2023, with similar contractions in winter 2023/24. - If households are protected per Germany’s Emergency Plan for Gas, shortages would fall on firms: - Firms’ consumption cuts required: 13 percent in remainder of 2022, 14 percent in 2023, 6 percent in 2024. - If energy transformation activities are also protected, remaining non-energy firms would have to cut consumption by 19 percent in remainder of 2022, 22 percent in 2023, and 8 percent in 2024. - Calculations suggest Germany would have sufficient supplies for protected customers like households even in the worst winter month, implying Germany would likely not be eligible to request gas supplies from neighbors under existing or future solidarity agreements. ### Comparison with other studies and key uncertainty drivers - The IMF estimates of shortages are similar to Bundesbank (2022b) for winter 2022/23 (Bundesbank reports 31 percent of unprotected sectors’ consumption) and larger in winter 2023/24 (Bundesbank: 11 percent; IMF: around 20 percent). - IMF monthly shortages do not exceed 21 percent of consumption in any month; Joint Economic Forecast (2022) median scenario reaches 60 percent shortage in October 2022. - Factors that could increase shortages: - Infrastructure bottlenecks and cross-border constraints (ENTSO-G notes significant bottlenecks). - Germany supplying more gas than usual to neighboring countries (estimates suggest neighboring winter shortage could amount to 9 bcm under certain assumptions). - Colder or longer winters could increase consumption by up to 3.1 bcm a year. - Factors that could reduce shortages: - Further reductions in household demand if higher prices pass through: up to 0.8 bcm a year additional household demand reduction under the most extreme scenario considered (370 percent increase in Netherlands TTF prices, full pass-through to German import prices, passthrough of 0.05 to retail and household elasticity -0.15). *IMF Working Paper: The Economic Impacts on Germany of a Potential Russian Gas Shutoff* ### 3. GDP impacts ### 3. GDP impacts ### Approach to estimating GDP impacts - Analysis captures three mechanisms: production cuts due to a lack of gas, downstream amplification through intermediate products, and increases in uncertainty. - Estimates add results from two models: - A sectoral firms’ production model (17 sectors, representative firm per sector) that: - uses a nested production function (Cobb–Douglas converting energy and intermediate inputs into output; CES converting gas, oil, coal and renewables into energy); - captures direct production losses from reduced gas and downstream amplification via input—output links (one round of downstream propagation only); - is static and applied over different scenario horizons to produce time-path impacts. - An uncertainty-growth model (local projections) estimating the historical relationship between uncertainty and deviations of GDP from trend, using two uncertainty measures (frequency of the word “uncertainty” in Economist Intelligence Unit reports for Germany; dispersion of economic forecasts). Regression specification: - gt+h = α + βut + ηut−1 + λg̃t−1 + ∑k Γk Xt−k + εt+h - Sample runs from 1991Q2 to 2019Q3. - Second-round impacts are assumed to materialize 6 months after the direct impact. - Baseline projections of GDP and inflation used are from IMF (forthcoming). ### Model specification and key parameters - 17 sectors; constant returns to scale; nested production function (Cobb–Douglas + CES). - Elasticity of substitution between gas and other energy types is a key parameter and varies with horizon: - June–December 2022: 0.1 - June 2022–December 2023: 0.12 - Between June 2022 and December 2027: up to 0.4 (sequence shown in Table 3: 0.10, 0.12, 0.16, 0.20, 0.30, 0.40) - Implied national expenditure share on gas: α = 0.025 (used in formula in Table 3 notes). - Implied change in expenditure share (Panel (a), expanding windows): 2.0, 1.8, 1.0, 0.5, 0.3, 0.1 (matching the elasticity sequence). - One-round downstream amplification only (motivated by uncertainties in timing and increasing substitutability over time). ### Estimated GDP impacts (direct, second-round, and uncertainty channels) - Direct impact of a 19 percent contraction in firms’ gas consumption between June and December 2022: reduce GDP by 0.9 percent relative to baseline in this period. - Direct impact between June 2022 and December 2023 (21 vs. 19 percent shock with elasticity 0.12 vs. 0.1): 0.9 percent. - Over longer horizons, as cumulative gas shortage shrinks and elasticity increases, direct impact falls; e.g., June 2022–December 2027: gas consumption reduced by 8 percent and direct effect reduces GDP by less than 0.05 percent. - Second-round (downstream) effects for shortages June–December 2022: reduce GDP by a further 0.9 percent, felt 6 months later (Dec 2022–June 2023). - Annualized direct + second-round (converting expanding windows to calendar years): - GDP lower than baseline by 0.7 percent in 2022 - GDP lower than baseline by 1.8 percent in 2023 - GDP lower than baseline by 0.4 percent in 2024 - (These are percentage deviations from baseline GDP levels, not growth rates.) - Uncertainty channel: - Estimated β implies a one-standard-deviation increase in uncertainty could lower GDP by about 1 percent relative to trend up to two years after the shock. - Assumed gas shutoff raises uncertainty to the 95th percentile in 2022Q2 and then decays by half from 2022Q3 to 2023Q1. - This uncertainty shock reduces GDP by 0.8 percent in 2022, 0.9 percent in 2023, and 0.1 percent in 2024; no clear effect thereafter. - Combined effects (direct + second-round + uncertainty) for the preferred scenario (households and energy transformation protected): - Real GDP level (percent deviation from baseline): - 2022: -1.5 - 2023: -2.7 - 2024: -0.4 - Cumulative GDP loss between 2022 and 2024: 4.8 percent of 2021 GDP. - Alternative protection scenario (only households protected): - Real GDP level (percent deviation from baseline): - 2022: -1.2 - 2023: -1.9 - 2024: -0.4 ### Timing and decomposition of losses - Losses in 2022 are driven mostly by uncertainty effects (downstream amplification effects materialize with lag). - Losses in 2023 are approximately evenly split between direct, second-round, and uncertainty channels. - From 2024 onwards, residual losses are driven almost entirely by downstream amplification effects. - Losses are permanent in the sense they are not “caught up” by deferring production, consumption, or investment to subsequent years. ### Sensitivity to rationing schemes (overview) - The way scarce gas supplies are allocated across sectors materially affects GDP losses. - Baseline assumption: unprotected sectors face proportional cuts to their gas consumption (e.g., 19 percent cut for each unprotected sector in June–Dec 2022). - Alternative rationing schemes can minimize direct or second-round impacts by prioritizing cuts to sectors in order of gas intensity or “downstreamness”. - Table 3 (Panel (a)) summary figures for expanding windows (households and transformation protected): - Required demand compression (percent of consumption): Dec '22: 19; Dec '23: 21; Dec '24: 16; Dec '25: 12; Dec '26: 10; Dec '27: 8 - First-round impact on cumulative GDP (percent deviation from baseline) when households and transformation protected: Dec '22: -0.9; Dec '23: -0.9; Dec '24: -0.4; Dec '25: -0.2; Dec '26: -0.1; Dec '27: -0.1 - Second-round impact on GDP in 6 months' time (percent deviation from baseline) when households and transformation protected: Dec '22: -0.9; Dec '23: -0.9; Dec '24: -0.4; Dec '25: -0.2; Dec '26: -0.1; Dec '27: -0.1 ### Comparisons with other studies and caveats - Estimated impact of -1.5 percent in 2022 matches Bundesbank (2022b) but is smaller than Bundesbank (2022a) which estimated -5 percent (differences due to narrower scope and smaller energy shortages here). - Over twelve months, GDP losses align with the 2–3 percent impact in Di Bella et al. (forthcoming) with adjustment frictions. - Over two years, the cumulative loss of 4.3 percent of 2021 GDP for 2022–2023 here is smaller than the 6.5 percent reported by the Joint Economic Forecast (2022) (differences driven by assumptions about substitutability, additional channels, and larger gas shortages in that forecast). - Model limitations noted: - Does not differentiate between domestic vs. export production or domestic vs. imported inputs — explicitly modeling trade could reduce estimated effects. - Does not model the energy transformation sector in full detail; results focus on scenario where energy transformation activity is protected from rationing to avoid underestimation of welfare and economic impacts. - One-round downstream propagation only; further rounds would shrink rapidly but timing and substitutability uncertainties increase with rounds. _Italic: IMF Working Paper — 3. GDP impacts (from wpiea2022144-print-pdf)_ ### 0.31 percent), and the second-round impact by a similar magnitude (from 0.88 to 0.32), if gas supplies to some ### wpiea2022144-print-pdf - 0.31 percent), and the second-round impact by a similar magnitude (from 0.88 to 0.32), if gas supplies to some ### Rationing: sectoral ordering, identified sectors, and GDP impacts - Sectors identified by the analysis for rationing (in various orders and combinations): non-metallic minerals; iron, steel and non-ferrous metals; paper, pulp and printing; chemicals and petrochemicals; and food, beverages and tobacco. - Rationing algorithm criteria: - Chooses sectors because they are relatively gas-intensive or appear to have fewer downstream linkages than other sectors. - Balances both gas-intensity and downstreamness of sectors. - Assumes no sector can have its gas consumption cut by more than 50 percent. - Key pattern: GDP losses are generally minimized when sectors higher in the gas-intensity ordering are rationed first, but exceptions occur when downstream linkages outweigh gas intensity (e.g., paper or iron and steel may be skipped). - Table 4 (Between June and December 2022) — illustrative percent cuts to gas consumption by scenario and resulting impacts on output: - Percent cuts to gas consumption by sector (selected rows shown exactly as in source): - Non-metallic minerals: 19 | 50 | 50 | 50 - Iron, steel and non-ferrous metals: 19 | 19 | 0 | 0 - Paper, pulp & printing: 19 | 0 | 0 | 0 - Chemical & petrochemical: 19 | 50 | 50 | 23 - Food, beverages & tobacco: 19 | 50 | 5 | 0 - Textile & leather: 19 | 0 | 0 | 0 - Mining & quarrying: 19 | 0 | 0 | 0 - Wood & wood products: 19 | 0 | 0 | 0 - Not elsewhere specified (industry): 19 | 0 | 0 | 0 - Machinery: 19 | 0 | 0 | 0 - Transport equipment: 19 | 0 | 0 | 0 - Construction: 19 | 0 | 0 | 0 - Transportation: 19 | 0 | 0 | 0 - Service: 19 | 0 | 0 | 0 - Agriculture: 19 | 0 | 0 | 0 - Households: 0 | 0 | 10 | 10 - Energy sector: 0 | 0 | 0 | 10 - Direct impact on output (corresponding columns): 0.88 | 0.31 | 0.21 | 0.12 - Second-round impact on output (corresponding columns): 0.88 | 0.32 | 0.20 | 0.16 - Note: The direct and second-round impacts on output shown are not added together because second-round impacts might occur with a lag. ### Household sharing of the burden and quantitative effects - If households share in gas rationing, GDP losses fall: - Direct GDP losses can be reduced by one-third: from 0.31 percent to 0.21 percent. - Second-round losses can be reduced by over one-third: from 0.32 percent to 0.20 percent. - Specific household reduction example: a 10 percent reduction in households’ gas consumption is equivalent to reducing heating temperature by about 2 degrees over the period where heating is needed (BDEW, 2022). - Means to achieve household reductions: voluntary behavioral changes (e.g., Japan’s setsuden movement after the 2011 earthquake) or responses to rising prices. - Policy note: Even customers on fixed-price contracts could see increases in gas prices in a crisis due to legal powers permitting pass-through of wholesale price increases; households’ purchasing power could be supported via income transfers or other policy measures. ### Inflation impacts: passthrough estimates and scenario outcomes - Estimation framework: - Uses local projection specification with state-dependent passthrough; allows larger international energy price shocks to have larger effects. - Defines a large increase in international energy prices in quarter t as 휋̃_t^E* = max(0, p̃_t / p̃_{t,3yr max} − 1), following Hamilton (2003). - Uses Brent crude and Netherlands TTF as international oil and gas prices, respectively. - Sample: 1996Q1 to 2022Q1. - Historical passthrough magnitudes: - Passthrough from international to domestic gas prices averages around 0.02 over 2-3 quarters. - Passthrough rises to 0.1 following large increases in international gas prices (fivefold increase). - Contemporaneous passthrough of international oil price averages at 0.2 in normal times but jumps to 0.3 in a price surge. - Passthrough from HICP energy inflation to core inflation: in normal times barely affects core inflation; in large energy price jumps passthrough reaches 0.05 contemporaneously. - International gas passthrough is more lagged than oil due to customers’ gas (and electricity) prices usually being fixed in 12-month contracts. - Assumed shock scenarios for international gas prices: - International gas prices rise in 2022Q2 by 110, 221 or 367 percent and last for 12 months. - Estimated impacts on German inflation (deviation from baseline annual inflation): - If international gas prices rise by 110 percent: - German inflation about ¾ of a percentage point higher in 2022 than without the shutoff. - German inflation about 1¼ percentage points higher in 2023 than without the shutoff. - If international gas prices rise by 370 percent: - Inflation higher by about 2 percentage points in 2022 and 3.5 percentage points in 2023. - If a 370 percent increase occurs and the government activates powers permitting firms to pass on price increases regardless of contracts (leading to contemporaneous passthrough): - Inflation could be 3¼ percentage points higher than without the shutoff in both 2022 and 2023. - Caveats: - The international gas price increases considered are without precedent prior to 2021; passthrough and resultant consumer price inflation rates could be greater than these estimates. - Estimates are larger than Bundesbank (2022b) estimates (0.5 in 2022, 1.5 in 2023 and 0.25 in 2024, in percentage points), potentially due to demand-side differences and model non-linearity. ### Policy discussion and recommended measures - Key policy objectives: - Enhance energy security and build resilience by securing new gas supplies and accelerating renewable energy. - Fill gas storage facilities before winter, enhance energy efficiency, ensure financial support does not exacerbate scarcity, maintain international cooperation, and develop rationing/distribution plans. - Renewable energy acceleration: - "Easter package" legislative proposals aim to raise the target to 80 percent of electricity demand met with renewables by 2030, from 42 percent observed in 2021. - Increase volume of public tenders for solar and onshore wind power to 22 and 10 GW a year, respectively. - Financial and solvency measures: - Government measures to ensure solvency of energy companies are appropriate given rising import prices and volatility. - Monitor financial health of energy companies (regular calls or surveys suggested). - Recent legal amendments allow more flexibility in retail prices which could help maintain solvency; such measures may need targeted income support for vulnerable households (e.g., one-off energy relief payments or housing benefits). - Government plans to support firms’ liquidity with loans through KfW, extensions of loan guarantee programs, and loans to energy producers to post margins on hedging contracts. - Energy-saving and compensation mechanisms: - Promote energy-saving campaigns (analogue to Japan’s setsuden) and programs to exchange gas heaters for heat pumps. - Government announced intention to develop an auction mechanism to compensate firms for reducing gas consumption; consider similar schemes for households (rebates or discounts) depending on technical feasibility. - Smart meters are not widely available in Germany; provide up-to-date, online gas usage statistics as an alternative. - Infrastructure and investment: - Authorities should provide a clear diagnosis of infrastructure gaps under a full shutoff scenario; publication of the network development plan consultation was postponed and a diagnosis could be published sooner given urgency. - International cooperation and operational issues: - Cooperation with the EU is key; REPowerEU is an important step. - Germany could work with France to solve the ‘odorization’ issue (different practices currently impede flows). - Potential additional gas flows from Dunkirk LNG terminal and planned FSRU at Eemshaven. - Germany has signed solidarity agreements with Denmark and Austria under EU Regulation 2017/1938; further agreements possible. - Rationing policy design considerations: - Rationing decisions should take economic implications into account alongside social welfare, environmental, health, technical and legal constraints. - Economic analysis (such as presented here) can inform but not be conclusive for rationing choices. - Major rationing decisions require political participation; delegating entirely to a technical agency could exceed mandates and undermine credibility. - Federal Network Agency has collected crucial firm-level gas usage information and might be well placed to act as a central authority to manage day-to-day rationing. *IMF Working Paper — The Economic Impacts on Germany of a Potential Russian Gas Shutoff* ### Annex I. Production model ### Annex I. Production model ### Nested Production Model and Extensions - Representative firm in sector k operates a constant return to scale production function: - Y_k = Z_k E_k^{α_Ek} Π_{j=1}^J X_{jk}^{α_{jk}}. (1) - Energy aggregate E_k combines gas and other energy inputs: - E_k = (ω e_gas_k^{(ε−1)/ε} + (1−ω) e_oth_k^{(ε−1)/ε})^{ε/(ε−1)}. (2) - ω>0 captures importance of gas; ε is elasticity of substitution between gas and other energy inputs. - Kmenta (1967) approximation (second-order) for the CES energy aggregate: - E_k ≈ e_gas_k^{ω} e_oth_k^{1−ω} (e_gas_k / e_oth_k)^{(1/2)(1−1/ε) ω(1−ω) ln(e_gas_k / e_oth_k)}. - Differential form (second-order approximation): - dln E_k = ω dln e_gas_k + (1/2)(1−1/ε) ω(1−ω) dln e_gas_k dln e_gas_k. (4) - Log-linear approximation of sectoral production function: - dln Y_k ≈ α_Ek dln E_k + Σ_j α_{jk} dln X_{jk}. (5) - Combining (4) and (5) yields direct-impact relationship of gas usage on gross output: - dln Y_k^d ≈ α_Ek ( ω dln e_gas_k + (1/2)(1−1/ε) ω(1−ω) dln e_gas_k dln e_gas_k ). (6) - dln Y_k^d denotes the direct impact on gross output. ### Direct impact (calibration and interpretation) - Calibration: - α_E calibrated to energy expenditure share P_E e_gas_k / P_k Y_k in total output. - ω calibrated to gas share in energy inputs P_gas e_gas_k / P_E e_gas_k. - Rewriting (6) in expenditure-share terms: - dln Y_k^d ≈ p_gas e_gas_k / P_k Y_k dln e_gas_k + (1/2) Δ p_gas e_gas_k / P_k Y_k dln e_gas_k, (7) - where changes in gas expenditure shares Δ p_gas e_gas_k / P_k Y_k = α_E (1−1/ε) ω(1−ω) dln e_gas_k. - Interpretation: - ε reflects ease of substituting gas with other inputs; higher ε → easier substitution → smaller implied changes in gas expenditure shares for a given price shock. - Sectoral production function approach is similar to Bachmann et al. (2022). ### Second-round impact (input-output propagation) - Cobb-Douglas property: sectoral expenditure share on intermediate goods is invariant to supply shocks → changes in intermediate inputs from sector j into k equal changes in output of sector j: - dln X_{jk} = dln Y_j. (footnotes 42, 43) - Using (5) and this invariance: - dln Y_k ≈ α_Ek dln E_k + Σ_j α_{jk} dln Y_j. - Matrix form for second-round (network) effects: - dln Y_s⃗ = (I − A')^{−1} α_E dln E⃗ = (I − A')^{−1} dln Y_d⃗, - where A = [a_{jk}] is the input-output matrix and α_{jk} summarizes inputs from j required for one unit of output k (equivalent to sector k’s expenditure share on inputs from sector j). - Leontief inverse L = (I − A)^{−1} = Σ_{i=0}^∞ A^i. - Relationship between direct and second-round effects at sector level: - dln Y_s⃗ = (I − A')^{−1} dln Y_d⃗. - Element-wise second-round impact for sector k: - dln Y_k^s = Σ_j l_{jk} dln Y_j^d = Σ_j (a_{jk} + Σ_r a_{jr} a_{rk} + ⋯ ) dln Y_j^d. - Propagation interpretation: - Second-round impacts propagate directly from input suppliers and indirectly via downstream suppliers. - Analysis focuses on second-round impact captured by Σ_j a_{jk} dln Y_j^d because higher-order effects take time to materialize and intermediate inputs become more substitutable over time, easing supply bottlenecks. ### Aggregation - Aggregate log change in GDP (using value-added shares λ_k): - dln GDP = Σ_k dln VA_k. (aggregation identity) - Assuming changes in value added equal changes in output at sectoral level (footnote 44), aggregate direct and second-round impacts: - dln GDP^d = Σ_k λ_k dln Y_k^d_k. - dln GDP^s = Σ_k λ_k dln Y_k^s_k. - λ_k = V_k / Σ_k V_k is the value-added share of sector k, calibrated from the input-output table. *Source: Annex I. Production model — The Economic Impacts on Germany of a Potential Russian Gas Shutoff, Working Paper No. WP/20220/144.* --- _Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022144-print-pdf.pdf_