## 1canea2021002 ## Source details **Canonical URL:** [1canea2021002](https://www.imf.org/-/media/files/publications/cr/2021/english/1canea2021002.pdf) ## Other formats - [Markdown version](/-/media/files/publications/cr/2021/english/1canea2021002.pdf.md) - [Structured JSON version](/-/media/files/publications/cr/2021/english/1canea2021002.pdf.json) --- ### Introduction and context - COVID-19 lockdown caused a sharp decline in employment in Canada, with lower-paid workers particularly hard hit. - Government fiscal response provided significant income support; many households received emergency transfers exceeding pre-crisis income levels. - Purpose: describe recent (pre-COVID-19) trends in income inequality across Canada to inform policy discussion and monitoring against government objectives. - Measurement: - Gini coefficient used (range 0–1). - Three income concepts: Market income; Total income (market income plus government transfers); Disposable income (total income less taxes). - Redistribution measured by difference between market-income Gini and disposable-income Gini. *Key contextual points and targets* - Upward trend in income inequality in advanced economies over recent decades. - Poverty reduction targets: 20 percent reduction by 2020 and 50 percent reduction by 2030 (relative to 2015 poverty level). - Correlation between poverty rates and income inequality across OECD countries: reported R² values R² = 0.6872 and R² = 0.606 in referenced figures. *Prepared by Troy Matheson and Dan Pan (WHD).* ### Inequality across provinces and trends - Market income inequality highest in Quebec, Ontario, and Newfoundland and Labrador; lowest in Alberta, British Columbia, and Prince Edward Island. - Market income inequality generally higher nationally than provincially; only Ontario and Newfoundland and Labrador exceed the national market-income inequality. - Disposable income inequality lower than market income inequality in every province, indicating redistributive impact of taxes and transfers. - Historical trends: - Disposable income inequality broadly stable until the second half of the 1990s, then rose and settled at a higher level in the 2000s. - Market income inequality rose beginning in the 1980s; redistribution offset this rise until about 1995. - Between 1995 and 2000, redistributive effect fell while market inequality stayed high, increasing disposable inequality. - Since around 2015, redistributive policies have had greater impact in lowering disposable inequality, especially in Alberta and Newfoundland and Labrador (partly reflecting EI benefits during the oil supply shock starting in 2014). ### Does inequality reduce growth in Canada? — empirical strategy - Panel regressions using provincial data in non-overlapping 5-year periods (τ = 5). - Baseline equation: - ln Y_{i,t} − ln Y_{i,t−τ} = α ln Y_{i,t−τ} + β X_{i,t−τ} + γ Z_{i,t−τ} + u_i + v_t + ε_{i,t} - Y_{i,t} = real GDP per capita. - X = proxies for inequality; Z = controls including trade openness, investment, education, terms of trade, government expenditure. - Estimation methods: fixed effects and system GMM. - Four model specifications: - Model 1: disposable income Gini (G) and redistribution R = G^M − G. - Model 2: non-linearity with G^2 term. - Model 3: interaction with intergenerational mobility index IM_i (from Corak 2017). - Model 4: short- vs long-term effects including G lagged 10 years (G_{i,t−2τ}). ### Model-based findings on inequality and growth (estimation results summary) - Model 1 (System GMM): negative relationship between inequality and growth, not statistically significant. Joint coefficient of inequality and redistribution positive and insignificant. - Model 2 (System GMM): negative relationship, not statistically significant. No evidence of disproportionate negative impact at higher inequality levels. - Model 3 (System GMM): negative relationship mitigated by higher intergenerational mobility (consistent with Aiyar and Ebeke (2019)); both effects not statistically significant. - Model 4: some evidence consistent with Halter and others (2013) — higher inequality can boost growth short term but hamper it long term; relationships not statistically significant. - Overall: mixed evidence; over the sample inequality in Canada has not had a significant impact on growth in either direction. ### Interpretation and implications - Cross-country literature often finds negative inequality–growth links, but the 50-year empirical literature is mixed. - Theoretical models imply offsetting effects; net effects ambiguous. - Political and social mechanisms linking inequality to growth may operate unpredictably and non-linearly. - Policy implication: empirical neutrality in Canada does not reduce normative importance of reducing inequality based on societal preferences and social justice. ### Drivers of rising market income inequality - Market income inequality rising since 1970s driven by: - Hollowing out of the middle class and rising income share to top earners. - Quintile contributions: bottom 40 percent broadly stable; middle quintile and top 40 percent contributed more to rise. - Major drivers identified in literature: - Globalization and skill-biased technical change. - Immigration (evidence mixed for Canada). - Decline of labor unions. - Declining intergenerational income mobility. ### Redistribution measurement and trends - Definitions: - Total redistribution R = GM − G. - Redistribution via transfers Rb = GM − GT. - Redistribution via taxes Rt = GT − G. - R = Rb + Rt. - Empirical patterns: - Transfers have larger impact on provincial inequality than taxes. - Redistribution through transfers rose from 1970s to mid-1990s, fell until roughly 2014, then rose since 2015 (childcare benefits announced in 2015 contributed). - Taxes became more progressive over time while average tax rates fell since 2000, implying maintained redistributive impact with lower average tax rates. - Decline in transfers from mid-1990s driven by reductions in EI benefits and social assistance more than offsetting rising childcare and pension benefits. - Decomposition (Kakwani 1984 framework): - Transfers: Rb ~ b/(1+b) Pb, where Pb is Kakwani progressivity index for transfers and b is average benefit rate. - Taxes: Rt ~ t/(1−t) Pt, where Pt is Kakwani progressivity index for taxes and t is average tax rate. ### Taxes and shares - Increase in tax progressivity due to: - Rise in share of taxes paid by high-income households. - Falling share paid by middle-income households. - Shares paid by lower-income households broadly stable since 1970s. ### Summary on inequality and redistribution - Upward trend in income inequality documented across advanced economies. - Canada’s inequality relatively stable more recently while other G7 countries continued upward trends. - Canada’s tax and transfer systems reduce inequality but by less than many European countries; Canada ranks toward the middle of OECD countries for disposable income inequality. - Provincial patterns broadly similar with recent increase in redistributive impact since 2015, though overall inequality remains high historically. *Source: Excerpt from the Canada chapter (figures, tables, and referenced studies as presented in the supplied text).* --- ### Corporate sector risk analysis and COVID-19 counterfactuals — objectives and data - Objective: quantify how monetary easing during COVID-19 helped prevent firm defaults and adverse macro-financial feedback using an integrated micro-macro simulation covering non-financial firms and banks. - Data: - Balance sheet and P&L data for 850 firms and the seven largest Canadian banks in one analysis. - Micro-simulation uses historical data for 843 listed and non-listed nonfinancial firms from eight industry segments; Capital IQ sample captures 75 percent of total non-financial corporate debt. ### Policy counterfactual (judgmental assumption) - Counterfactual: policy rates assumed +150 bps higher since March 2020 for the remainder of the year. - Rationale: reflect higher debt costs had monetary easing not occurred. ### Counterfactual macro and firm impacts (selected quantitative results) - Industry-aggregate debt-weighted PD: - Observed 2020 average: 3.2 percent. - Counterfactual peak (with second-round feedback): about 5.5 percent; elsewhere text reports could have moved to about 6 percent or beyond 9 percent at the peak when including second-round feedback effects. - Corporate loan pool LGD: - Realized, approximate 2020 LGD: 20–21 percent. - Counterfactual LGD (with feedback): could have moved beyond 30 percent (reported 32 percent inclusive of feedback). - Debt cost increase for firms under counterfactual: close to 300 basis points (text also reports combined 270 bps in one statement). - Macro impacts (text statements and Table 3): - Real GDP growth could have been −3.4pp (YoY) in 2020 relative to baseline. - Unemployment rate: 1.0pp higher (year average). - Risk premium: 1.2pp wider (2020Q1-4). - Another text statement: real GDP growth could have been 3.3 ppts lower and unemployment 1 ppt higher; debt cost increase combined 270 bps. - Banks’ regulatory capital ratios: - Impact range across seven largest banks: −25 bps to −240 bps (min–max). - RWA-weighted banking system impact: about −70 bps inclusive of second-round feedback. ### Sector heterogeneity and vulnerabilities - Energy segment (oil and gas): low profitability, low cash-to-debt ratios, comparatively high PDs pre-pandemic; PDs trended up since 2014 to about three-fold of pre-2014 average. - Consumer cyclical, consumer non-cyclical, and utilities sectors: larger increases in shares of firms with ICRs less than 1 under counterfactual. - Debt to asset ratios move little in simulation due to assumption of no additional debt issuance by firms. ### Modeling framework and estimation elements - Micro-macro feedback channels: - Monetary easing compresses firms’ debt cost (A) → fewer defaults (B) → supports employment and demand (C) → bolsters firms’ income through sales (D) → feeds back to fewer defaults (B); bank solvency module captures fewer loan losses and more interest income (E). - PD bridge equation (industry-level logit) inputs: - cash to debt ratio (CD), debt to asset ratio (DTA), asset volatility (AV), EBIT to total asset ratio (EBITR), risk spread (RS), lags, and industry fixed effects. - Revenue and CGS bridge semi-elasticities (sample 1995–2019): - Communications: Revenue 0.75 ; CGS 0.96 - Consumer cyclical: Revenue 1.19 ; CGS 1.00 - Consumer non-cyclical: Revenue 1.36 ; CGS 0.95 - Energy: Revenue 5.21 ; CGS 0.97 - Industrials: Revenue 1.60 ; CGS 0.98 - Basic materials: Revenue 3.15 ; CGS 0.81 - Technology: Revenue 0.83 ; CGS 0.95 - Utilities: Revenue 1.09 ; CGS 1.00 - Note: CGS elasticities for consumer cyclical and utilities capped at 1. - Asset volatility model (log form) includes NGDPG, term spread, unemployment levels and changes, OIL; sign constraints applied (TS and URX positive; oil negative for energy, positive for utilities in constraints). ### Bank impact module and LGD linkage - Bank parameters reflective of 2019 conditions: risk weight densities, IRB shares, PiT and TTC PDs, PiT and downturn LGDs from bank reports as of 2020Q1. - LGD formula applied at bank corporate portfolio level: - LGD_b,t = Φ[ Φ^−1[PD_b,t] − k_b ] PD_b,t - k_b computed assuming asset correlation ρ = 10 percent: - k_b = Φ^−1[PD_b_TTC] − Φ^−1[PD_b_TTC × LGD_b] / √(1−ρ) - TTC PD inputs smoothed with smoothing parameter λ = 0.3 for base results. - Caveats: counterfactual broad and does not account for specific support policies (e.g., wage subsidies); assumes no additional firm debt accumulation; estimates subject to model and parameter uncertainty. *Source: IMF staff calculations and figures from the supplied IMF Canada chapter content.* --- ### Climate mitigation strategy, carbon pricing architecture, and feebates - Federal mitigation core: phase in carbon price floor rising to CAN $170 by 2030; federal backstop where sub-national pricing falls short (fuel charge and OBPS for EITE industries); sectoral tax credits, standards, building retrofits; public funding and revenue recycling. - Notable targets and numbers: - CAN $170 per ton CO2e by 2030 (proposed price floor). - ZEV sales-share targets: 10 percent by 2025, 30 percent by 2030, 100 percent by 2040. - Coal-based power phaseout by 2030. - NDC target: 511 million tons CO2e in 2030 (30 percent below 2005; 15 percent below 1990). - Total GHG emissions in 2018: 729 million tons; 74 percent from fossil fuel energy. - Fossil fuel CO2 by fuel: coal 26 percent; oil 42 percent; natural gas 32 percent. - LULUCF emissions were negative 13 million tons in 2018. - Pan-Canadian Framework proposed: price rise by CAN$10/ton per year to reach $50 by 2022; from 2023 annual increase of $15 per ton implying $170 by 2030. - Federal backstop components: regulatory fuel charge and OBPS; emitters registered in OBPS exempt from regulatory fuel charges. - Provincial systems status (selected entries): - British Columbia, Northwest Territories, Nova Scotia, Quebec: fully meet federal requirements. - Alberta, New Brunswick, Prince Edward Island, Saskatchewan: meet requirements for covered sources. - Federal backstop in place in Manitoba, Nunavut, Ontario, Yukon. - Price examples in Table A1: Northwest Territories CAN$/ton = 79; Nova Scotia = 80; Quebec = 85; federal backstop jurisdictions = 30 in 2019. *Prepared by Ian Parry, Simon Black, Danielle Minnett, and Victor Mylonas.* ### Feebates — design, sectoral applications, and illustrations - Feebates: revenue-neutral sliding fees on above-average emission rates and rebates for below-average rates. - Advantages: - No fiscal cost to government. - Avoids higher energy prices on average households/firms. - More flexible/cost-effective than regulation; strong mitigation incentives. - Applications and illustrations: - Road transport feebate formula: - CO2 price × {CO2/mile ─ CO2/mile of the new vehicle fleet} × {average lifetime vehicle mileage} - Illustration: US$300/ton CO2 feebate → subsidy US$7,500 for ZEVs; tax US$1,800 for 300 g/mile vehicle; assumes discounted lifetime driving 62,000 miles. - Power feebate formula: - Fee = CO2 price × {CO2/kWh ─ industry-wide average CO2/kWh} × electricity generation - Illustration: CAN$50/ton feebate → subsidy 0.5 cents/kWh for zero-carbon generation; fees 1.2 and 4.5 cents/kWh for natural gas and coal plants respectively. - Industry feebate formula: - CO2 price × {CO2/production ─ industry-wide average CO2/production} × production - Forestry feebate: fee = {CO2 rental price} × {carbon storage baseline ─ current stored carbon}; example: $50 feebate, fully stocking a hectare previously without trees increases land value by about $2,000 assuming 3 tons CO2/ha/yr over 20 years discounted at 5 percent. - Agriculture: feebates or emissions fee with revenues rebated proportional to value of output; demand-side taxes on meat/dairy considered. - Fugitive emissions: tax methane emissions with rebates for validated lower rates; venting accounted for 55 percent of fugitive emissions in 2018 (two-thirds oil, one-third gas); flaring 13 percent; other leaks 30 percent; 70 percent of CO2e from methane. ### Sectoral policy recommendations (summary) - Introduce revenue-neutral feebates at federal level for transport, power, and industry. - Use feebates to promote energy-efficient appliances and clean heating systems in buildings. - Apply fee to fugitive emissions with rebates for verified lower rates. - Consider feebates and logging tax to promote forest carbon sequestration. - Consider feebates or emissions pricing with within-sector revenue recycling for agriculture; consumer-facing fiscal incentives to encourage plant- and poultry-based diets. - Consider transition from OBPS to border carbon adjustment (BCA) for EITE industries in medium term. - Promote dialogue on an ICPF among large emitters using Canada’s model as prototype. ### Distributional incidence and revenue recycling - Household burden of $170 carbon price in 2030 (relative to BAU and prior to recycling): - Average burden = 2 percent of consumption. - Composition: 35 percent from road fuel; 33 percent from natural gas; 6 percent from electricity; 25 percent indirectly via higher consumer prices. - Burdens evenly distributed across population-weighted household per-capita consumption deciles at national level; lower in Quebec. - Returning carbon pricing revenues to households offsets about 80 percent of this burden. - Firm incidence and competitiveness: - Most vulnerable industries listed (iron and steel mills, basic chemicals, petroleum and coal products, etc.). - For same carbon price (US$50) in 2030, average cost increases across 20 percent most vulnerable industries would be: - 1.7 times as large in the United States as in Canada. - 2.7 times as large in India as in Canada. - 4.1 times as large in China as in Canada. ### Border Carbon Adjustments (BCAs) and International Carbon Price Floor (ICPF) - BCA considerations: - Transition from OBPS to BCA could more effectively address competitiveness and leakage; a $170 BCA applied to EITE imports would have raised 0.5 percent of GDP in 2015. - Revenue sources under $170 BCA in 2015 breakdown: 41 percent from China, 23 percent from United States, 6 percent each from India and EU. - Design issues: sectoral coverage, measuring embodied carbon (domestic vs country-specific), firm- vs industry-level measures, rebates for exporters, adjusting for foreign carbon pricing, lower rates for developing countries. - ICPF: - IMF staff recommend an ICPF arrangement to scale up mitigation among large emitters; Canada’s carbon pricing provides a prototype. - Price-floor scenarios and G20 CO2 outcomes (Table 4 summary; % reduction in G20 CO2 emissions below baseline, 2030): - Required for 2o (1.5): 28 (55) - China, India, US join a $75/$50/$25 price floor → 28.4 → Illustration: this trio with $75 (US), $50 (China), $25 (India) would cut G20 emissions about 28 percent below baseline (just consistent with 2oC target). - All G20 join a $50 price floor → 29.9 - All G20 join a $75 price floor → 31.1 - Implementation could initially focus on power and industry sectors. ### Revenue use and macro efficiency - Carbon pricing macro costs can be largely offset by using revenues to lower distortionary taxes or fund efficient investments; lump-sum dividends lack efficiency benefits. - Comparative efficiency example (US): an ETS with free allowances or a carbon tax with lump-sum dividends is about twice as costly (for same emissions reduction) as a combination of feebates; most cost-effective is an ETS or carbon tax with revenues used to cut labor/business taxes. *International Monetary Fund — Canada chapter excerpts (content unit 1canea2021002).* --- *Source: Excerpt from the Canada chapter (figures, tables, and referenced studies as presented in the supplied text).* ### Introduction ________________________________________________________________________ 4 ### Introduction ### A. Introduction - The economic lockdown triggered by COVID-19 led to a sharp decline in employment in Canada, with lower-paid workers being particularly hard hit. - The government’s fiscal response included a range of programs providing significant income support to low-income workers, with many households receiving emergency transfers exceeding their pre-crisis income levels. - This chapter describes recent (pre-COVID-19) trends in income inequality across Canada to inform policy discussion and monitoring against government policy objectives. - Key contextual points: - There has been a well-documented upward trend in income inequality in advanced economies over recent decades. - Concerns about high inequality include rising social tensions, populist and protectionist sentiment, and potential political instability (example cited: Chile and a 4 percent public transport fare increase). - There is evidence that high levels of income inequality can hamper economic growth (see Ostry and Others, 2018). - Whether inequality is "excessive" depends on country-specific factors and societal preferences (IMF Fiscal Monitor, October 2017). - Measurement and definitions: - The Gini coefficient is the measure of income inequality used throughout the chapter; it ranges between 0 (perfect equality) and 1 (perfect inequality). - Three definitions of household income for computing Gini coefficients: - Market income (income from market sources such as wages and interest income). - Total income (market income plus government transfers). - Disposable income (total income less taxes paid). - The difference between market-income Gini and disposable-income Gini reflects the impact of government redistributive policies (transfers and taxes). - Poverty and policy context: - There is a strong correlation between poverty rates and income inequality across OECD countries using both market and disposable income Gini coefficients (figure 2; reported R² values shown in figure: R² = 0.6872 for one series, R² = 0.606 for the other). - The Government of Canada released a poverty reduction strategy in August 2018 targeting a 20 percent reduction in poverty by 2020 and a 50 percent reduction in poverty by 2030 (targets set relative to the 2015 level of poverty). - The current federal administration has emphasized supporting the middle class since coming to power in 2015. ### B. A Closer Look at Inequality Across Canadian Provinces - Cross-provincial patterns (figure 4 and discussion): - Market income inequality is highest in Quebec, Ontario, and Newfoundland and Labrador, and lowest in Alberta, British Columbia, and Prince Edward Island (figure 4). - Market income inequality tends to be higher at the national level than at the provincial level; only Ontario and Newfoundland and Labrador have higher market income inequality than Canada as a whole. - Disposable income inequality is lower than market income inequality in all provinces, indicating government redistributive policies reduce inequality within provinces. - Trends since the mid-1970s: - Disposable income inequality remained broadly stable until the second half of the 1990s, when it rose and settled at a higher level in the 2000s. - Underlying increases in market income inequality began in the 1980s and were broadly offset by increases in redistribution through taxes and transfers until about 1995. - Between 1995 and 2000 the redistributive effect of government programs fell while market income inequality remained high, causing disposable income inequality to rise. - Since around 2015, redistributive policies have had a greater impact on lowering disposable income inequality, particularly in Alberta and Newfoundland and Labrador, though overall levels remain high from a historical standpoint. - Footnote: The greater impact of redistributive policies on Alberta and Newfoundland & Labrador (and the national average) after 2015 is likely to partly reflect that many lost their jobs and received employment insurance benefits during the oil supply shock that began in 2014. ### C. Does Inequality Reduce Growth in Canada? - Empirical approach: - The chapter estimates panel regressions using provincial data to assess the impact of inequality on growth. - Baseline equation (non-overlapping 5-year periods, i.e., τ = 5): - ln Y_{i,t} − ln Y_{i,t−τ} = α ln Y_{i,t−τ} + β X_{i,t−τ} + γ Z_{i,t−τ} + u_i + v_t + ε_{i,t} - Y_{i,t} denotes real GDP per capita in province i and year t. - X_{i,t−τ} consists of proxies for inequality. - Z_{i,t−τ} contains control variables used in previous studies, including trade openness, investment, education, terms of trade, and government expenditure. - Models estimated using both fixed effects and system GMM panel regressions. - Four model specifications are estimated to probe different aspects of the inequality–growth relationship. - Model 1 (in line with Berg and others, 2018): - Focuses on disposable income inequality and redistribution. - Specification: - ln Y_{i,t} − ln Y_{i,t−τ} = α ln Y_{i,t−τ} + β_1 G_{i,t−τ} + β_2 R_{i,t−τ} + γ Z_{i,t−τ} + u_i + v_t + ε_{i,t} - Definitions: - G_{i,t} is disposable income inequality (Gini coefficient measured after taxes and transfers). - R_{i,t} is redistribution defined as G^M_{i,t} − G_{i,t}, where G^M_{i,t} is the market income Gini. - Model 2 (non-linearity test, following Aiyar and Ebeke, 2019): - Incorporates a squared Gini term. - Specification: - ln Y_{i,t} − ln Y_{i,t−τ} = α ln Y_{i,t−τ} + β_3 G_{i,t−τ} + β_4 G_{i,t−τ}^2 + γ Z_{i,t−τ} + u_i + v_t + ε_{i,t} - Model 3 (interaction with intergenerational mobility, following Aiyar and Ebeke, 2019): - Accounts for the possibility that inequality’s effect on growth depends on intergenerational income mobility. - Uses a relative income mobility index IM_i from Corak (2017), an estimate of the intergenerational income elasticity (lower elasticity = more mobility). - Because IM_i has no time variation, it enters as an interaction with the Gini: - ln Y_{i,t} − ln Y_{i,t−τ} = α ln Y_{i,t−τ} + β_5 G_{i,t−τ} + β_6 (G_{i,t−τ} × IM_i) + γ Z_{i,t−τ} + u_i + v_t + ε_{i,t} - Model 4 (short- vs long-term effects, following Halter and others, 2013): - Includes a lagged Gini coefficient 10 years earlier to separate short-term and long-term impacts. - Specification: - ln Y_{i,t} − ln Y_{i,t−τ} = α ln Y_{i,t−τ} + β_7 G_{i,t−τ} + β_8 G_{i,t−2τ} + γ Z_{i,t−τ} + u_i + v_t + ε_{i,t} - Rationale: - Halter and others (2013) propose high income inequality can boost growth in the short term but reduce growth over the long term; positive short- to medium-term effects operate through economic mechanisms, while negative long-term effects operate through political disruptions and changes in educational attainment. - Data and appendices: - Summaries of related studies and the data used are provided in tables A1 and A2 in the appendix. - Estimation results are summarized in Table 1 (Estimation Results). *Prepared by Troy Matheson and Dan Pan (WHD).* ### 12. The results show mixed evidence on the impact of inequality on growth among ### 1canea2021002 - 12. The results show mixed evidence on the impact of inequality on growth among ### Model-based findings on inequality and growth - Model 1 (System GMM): shows a negative relationship between inequality and growth, but this relationship is not statistically significant. The joint coefficient of income inequality and redistribution (market income inequality) is estimated to be both positive and insignificant. - Model 2 (System GMM): shows a negative relationship between inequality and growth, but this effect is not statistically significant. No evidence that higher levels of income inequality have a disproportionately large negative impact on growth. - Model 3 (System GMM): suggests a negative relationship between inequality and growth and that this effect is mitigated when intergenerational income mobility is higher (consistent with Aiyar and Ebeke (2019)); however, both effects are not statistically significant. - Model 4: provides some evidence consistent with Halter and others (2013) that higher income inequality can boost growth in the short term but hamper it over the long term; these relationships are not statistically significant. ### Empirical ambiguity and interpretation - Cross-country studies have tended to find a negative relationship between inequality and growth, but the broader 50-year empirical literature is best described as mixed. - Theoretical reasons: Many theories predict offsetting effects of inequality on investment and growth and net effects are ambiguous (Barro (2000) quoted). - Political and social mechanisms that link inequality to growth may operate unpredictably and non-linearly, aligning with empirical ambiguities. - Note: "In addition to the baseline specifications, a large number of alternative specifications were also investigated, including models with different lag lengths, different instruments, and with no controls. The qualitative findings are broadly similar to those described below." ### Role and normative importance of inequality - Most academics agree some level of inequality is desirable to incentivize investment and growth; levels that are too high can lead to social disruptions lowering growth. - Empirical findings for Canada over the sample: inequality has not had a significant impact on growth in either direction — "it has been neither too high nor too low." - Policy implication: This empirical neutrality does not diminish the importance of policymakers’ efforts to reduce inequality based on societal preferences and social justice grounds (as emphasized by the Government of Canada). ### Why has inequality risen across Canada? — Market income drivers - Market income inequality has been rising since the 1970s, largely the result of hollowing out of the middle class and a rising share of income going to richer Canadians. - Quintile contributions: bottom 2 quintiles (bottom 40 percent) broadly stable; middle quintile and top 2 quintiles contributed more to the rise in income inequality. - Major drivers identified (Green and others, 2016; Card and DiNardo, 2002; Card, 2009; Western and Rosenfeld, 2011; Legree and others, 2016; Chorak, 2013; Heisz, 2016; Picot and Hou, 2016; Warman and Worswick, 2016): - Globalization and skill-biased technical change: technological improvements raise demand and wages for educated workers, increasing inequality. - Immigration: importing low-skilled workers can raise inequality by increasing labor supply in low-wage sectors; evidence for Canada is mixed. - Decline of labor unions: decline in unionization rates contributes to rising wage inequality. - Intergenerational income mobility: lower mobility correlates with more persistent inequality; evidence suggests mobility has declined in Canada in recent years. ### Redistribution policies: measurement and trends - Definitions: - Total redistribution R is defined as R = GM − G, where GM is market income Gini and G is disposable income Gini. - Redistribution via transfers: Rb = GM − GT, where GT is Gini for total income (market income plus transfers before taxes). - Redistribution via taxes: Rt = GT − G. - R = Rb + Rt. - Empirical patterns: - Redistribution through transfers has a much larger impact on provincial inequality than taxes. - Declining redistributive effects of transfers are the key reason overall redistribution across provinces fell from the middle of the 1990s until 2014; redistributive effects of taxes remained broadly stable over this period. - Since 2015, redistributive impact of transfers has been rising across all provinces. - Decomposition concepts (Kakwani 1984 framework): - For transfers: Rb ~ b/(1+b) Pb, where Pb is the Kakwani progressivity index for transfers and b is the average benefit rate (total transfers received divided by total market income). - For taxes: Rt ~ t/(1−t) Pt, where Pt is the Kakwani progressivity index for taxes and t is the average tax rate (total taxes paid divided by total income). - Empirical findings from Figures and trends: - Transfers: impact rose from the 1970s to mid-1990s, then fell until roughly 2014; recent rise attributed to increased childcare benefits announced in 2015. - Taxes: impact rose from the 1970s and has been broadly stable since 2000. - Principal factor driving redistribution through transfers across provinces is average benefit rate; small changes in transfers’ progressivity also occurred. - Taxes have become more progressive over time while average tax rates fell since 2000, implying governments maintained redistributive impact from taxes with lower average tax rates. - Specific contributors to the decline in transfers since mid-1990s: reductions in EI benefits and social assistance more than offset rising childcare and pension benefits. ### Taxes and shares - The increase in tax progressivity over time is attributable to: - A rise in the share of taxes paid by high-income households. - Falling shares paid by middle-income households. - Shares paid by lower-income households have remained broadly stable since the 1970s. ### Summary and policy conclusions (section E) - Long-run trends: - Upward trend in income inequality in advanced economies over past several decades is well documented. - Canada’s inequality has been relatively stable more recently, while upward trends continued in other G7 countries. - Role of tax and transfer systems: - Tax and transfer systems reduce income inequality across OECD countries. - Canada’s redistribution policies reduce inequality by less than many European countries. - Overall, Canada ranks towards the middle of OECD countries when considering disposable income inequality after accounting for government redistributive impacts. - Provincial patterns: - Little diversity in market income inequality across provinces; trends are similar over time. - Disposable income inequality rose across Canada in the second half of the 1990s and settled at a higher level in the 2000s. - Market income inequality increases starting in the 1980s were broadly offset by redistributive policies until about 1995; between 1995 and 2000 the redistributive effect of government programs fell while market inequality remained higher, increasing disposable income inequality. - Since 2015, redistributive policies have had a greater impact on lowering disposable income inequality, particularly in Alberta and Newfoundland and Labrador. - Despite recent policy changes, overall levels of inequality remain high from an historical standpoint. *Source: Excerpt from the Canada chapter (figures, tables, and referenced studies as presented in the supplied text).* ### 26. While recent cross-country empirical studies tend to find a negative relationship ### 26. While recent cross-country empirical studies tend to find a negative relationship between inequality and growth, this chapter finds mixed evidence for Canada ### Main findings on inequality and growth - Cross-country empirical literature tends to find a negative relationship between inequality and growth; this chapter finds mixed evidence for Canada. - Over the sample period examined, inequality in Canada has not had a significant impact on growth in either direction—suggesting it has been neither too high nor too low. - Theories linking negative effects of inequality on growth emphasize disruptions in political processes and rising social tensions, which may produce empirically unpredictable and non-linear effects and help explain the mixed Canadian evidence. ### Drivers of rising market income inequality in Canada - Rising market income inequality is attributed to: - globalization and skills-bias technical change, - immigration, - the decline in unionization rates, - intergenerational mobility. ### Role of taxes, transfers, and redistribution - A decline in the impact of redistributive policies between the mid-1990s and 2014 contributed to rising income inequality. - The decline in redistributive impact was most pronounced for transfers, especially reductions in benefit rates for employment insurance and social assistance. - The redistributive impact of taxes has remained broadly stable since 2000, despite a decline in average tax rates across all income groups, reflecting an increase in tax progressivity—a rise in the share of taxes paid by high-income households and a concurrent fall in the share paid by middle-income households. - Since the federal government came into power in 2015, redistribution has lessened inequality largely through a rise in childcare benefits. ### COVID-19 shock and inequality - The economic lockdown triggered by COVID-19 led to a sharp decline in employment in Canada, with lower-paid workers being particularly hard hit. - The long-term impact of the crisis on income inequality is uncertain and needs close monitoring against government policy objectives. - To inform policy, the chapter describes recent (pre-COVID-19) trends in income inequality across Canada. ### Corporate sector risk analysis and COVID-19 policy counterfactuals — key results - Objective: quantify how monetary easing during the COVID-19 pandemic contributed to preventing firm defaults and adverse macroeconomic feedback using an integrated micro-macro simulation model covering non-financial firms and banks. - Data and model: - Balance sheet and P&L data for 850 firms and the seven largest Canadian banks used in one analysis. - A micro-simulation uses historical P&L and balance sheet data for 843 listed and non-listed nonfinancial firms from eight industry segments; the Capital IQ sample captures 75 percent of the total non-financial corporate debt. - Counterfactual scenario (policy rates assumed 150 bps higher since March 2020 for the remainder of the year): - Industry-aggregate debt-weighted probability of default (PD) could have moved to about 5.5 percent at the peak (with second-round feedback), compared to an observed 3.2 percent on average in 2020. - Corporate loan pool loss given default (LGD) could have moved beyond 30 percent across industries, from a realized, approximate 20 percent. - Combined with the assumed base rate shift, the counterfactual entailed a debt cost increase for firms of close to 300 basis points. - Real GDP growth could have been 3.4 ppts lower and the unemployment rate 1 ppt higher (year average). - Counterfactual impact on banks’ regulatory capital ratios ranges between -25 bps and -240 bps (min-max) for the seven largest banks. - Sector heterogeneity and vulnerabilities: - Indebtedness and debt-related risk metrics show notable heterogeneity across firm sectors. - The energy segment (oil and gas) had low profitability, low cash-to-debt ratios, and comparatively high probabilities of default already before the pandemic. - Continued downward pressure on oil prices from the pandemic and structural greening trends imply further consolidation pressure on Canada’s energy segment. - Policy implication on support unwinding: - While monetary support had measurable benefits in containing defaults, sizable policy support beyond the short term raises risks of overleveraging incentivized by low interest rates. - Careful lending practices by banks, encouraged by the supervisor, should aim to preempt and counteract overleveraging to prevent macro-financial imbalances. ### Model caveats and limitations (as stated) - The policy counterfactual is defined broadly as a general debt cost increase and does not account for specific corporate support policies such as wage subsidies. - The analysis assumes no additional debt accumulation by firms. - Estimates are surrounded by non-negligible model and parameter uncertainty. *Italic: IMF staff chapter content provided in the supplied PDF excerpt.* ### Appendix Figure 1 compares various ratios (for example, cash to debt, debt to assets, etc.) from the ### Appendix Figure 1 compares various ratios (for example, cash to debt, debt to assets, etc.) from the ### Data comparison and key micro-macro correspondence - Appendix Figure 1 compares firm micro data (Capital IQ) with sector aggregate data (Statistics Canada) for ratios such as cash to debt and debt to assets; they suggest a broad correspondence between the averages of such ratios across industries. - Exception: debt to asset ratios from STATCAN aggregates are lower across all industry segments compared to Capital IQ micro data, suggesting smaller firms (not covered by the micro data) may be characterized by lower indebtedness than larger firms. - Note on scope difference: the share reported is an approximation and should be interpreted with caution because of differences in scope—firms’ consolidated debt inclusive of subsidiaries’ debt abroad in the numerator versus a domestic perspective of firms’ debt data from Statistics Canada in the denominator. - Footnote on ratio definitions repeated across figures: Debt in the debt to asset ratios is defined as short-term plus half long-term debt. Debt in the cash to debt ratios comprises total debt. The red whiskers denote the interquartile ranges. ### Historical PDs and asset volatility by industry segments - PDs and asset volatilities are heterogeneous across firm segments historically and at the end-sample position. - Technology firms experienced higher default rates most of the time historically. - Since 2014, PDs of firms in the energy segment started trending up, to an average three-fold level of the average observed in the years before 2014—reflecting consolidation pressure on the segment. - Sources for PDs and asset volatility: Moody’s Analytics and IMF staff calculations. The PDs and asset volatility are medians of the underlying firm data at any point in time historically. ### Model framework overview and impact channels - Expansionary monetary policy support compresses firms’ debt cost (A), implying fewer firm defaults (B); this supports employment and aggregate demand (C), bolsters firms’ income through stronger sales revenues (D), which feeds back to fewer defaults (B) in addition to lower initial debt financing costs. A bank solvency impact module (E) captures fewer loan losses and more interest income from fewer firm defaults. - Table 1 summarizes sub-components and how they are modeled (component list referenced in text). ### Firm revenue, costs, and accounting relationships - Gross revenue (REV) growth linked to nominal GDP growth; firms’ cost of goods sold (CGS) growth linked to gross revenue growth. Both are firm-level fixed effects panel regressions with industry heterogeneity (see Table 2). - Other operating expenses (OOE) held constant. Earnings before interest and taxes (EBIT) computed as: - EBIT_f,t = REV_f,t − CGS_f,t − OOE_f,t ### PD bridge equations (industry-level, econometric Merton-type) - PD bridge equation structure (logit form) includes industry fixed effects and RHS variables: - cash to debt ratio (CD) - debt to asset ratio (DTA) with debt defined as short-term plus half long-term debt - asset volatility (AV, industry median from MKMV model) - EBIT to total asset ratio (EBITR) - risk spread (RS): sectoral aggregate interest expense (4-quarter trailing sum) over debt minus a 2-year sovereign bond yield - Inclusion of DTA and AV reflects a stock-oriented (solvency) perspective; CD and EBITR reflect a flow (liquidity) perspective. - Equation (2) as presented: - logit(PD_i,t) = α1,i + α2,i CD_i,t + α3,i DTA_i,t + α4,i AV_i,t + α5,i EBITR_i,t + α6,i RS_i,t + … (RHS lags) + ε_i,t ### Bridge equations for revenue and CGS (semi-elasticities, Table 2) - Revenue growth to nominal GDP growth; CGS growth to revenue growth — semi-elasticities by industry (sample period 1995-2019 annual): - Communications: Revenue 0.75 ; CGS 0.96 - Consumer cyclical: Revenue 1.19 ; CGS 1.00 - Consumer non-cyclical: Revenue 1.36 ; CGS 0.95 - Energy: Revenue 5.21 ; CGS 0.97 - Industrials: Revenue 1.60 ; CGS 0.98 - Basic materials: Revenue 3.15 ; CGS 0.81 - Technology: Revenue 0.83 ; CGS 0.95 - Utilities: Revenue 1.09 ; CGS 1.00 - Notes: CGS elasticities for consumer cyclical and utilities capped at 1. ### Estimation methodology: Long-run multiplier BMA - A long-run multiplier (LRM) sign-constrained Bayesian model averaging (BMA) methodology considered all conceivable combinations of RHS variables in eq. (2); RHS lag structures optimized (up to two lags beyond contemporaneous). - Models weighted by predictive performance via Bayesian Information Criterion; weights set to zero if not conforming with at least one LRM sign constraint; final posterior weighted average model obtained. - LRM sign constraints for PD equations: positive for RS, AV, and DTA; negative for remaining variables. - Model estimates indicate roles for asset volatility, leverage, and EBIT to total asset ratios with heterogeneity across industries. - Firm-level balance sheet variables (debt to assets, cash to debt, EBIT to assets) are endogenized via micro simulation aggregations feeding into industry-level equations. Asset volatilities endogenized through econometric bridge equations. ### Asset volatility bridge equations - Asset volatility model (log form) estimated via BMA with RHS variables including nominal GDP growth (NGDPG, QoQ and YoY separately), term spread (TS), unemployment rates (URX) in levels and first differences, and OIL (QoQ and YoY) for energy/utilities. - Equation (3) as presented: - ln(AV_i,t) = β1,i + β2,i NGDPG_t + β3,i TS_t + β4,i URX_t + β5,i ΔURX_t + β6,i OIL_t + … (RHS lags) + ε_i,t - LRM sign constraints: positive for TS and URX; negative for remaining variables. Oil price growth assigned negative constraint for energy and positive for utilities. - Many models find asset volatility driven primarily by GDP growth and employment conditions. ### Macro-financial feedback: SVAR specification - A six-variable Structural VAR (SVAR) used as macro-financial feedback engine: nominal GDP (ln QoQ), GDP deflator (ln QoQ), term spread (10Y sovereign minus 3-month treasury bill rate), unemployment rate, risk spread (firms’ implicit interest rate over 2-year sovereign yield), and logit-transformed aggregate NFC sector annual default rate. - Estimated on quarterly data 2000Q4-2019Q4 (77 obs.). - Identification via sign-restricted impulse responses: an initial shock to firm PDs identified as a supply shock constraining on-impact responses (nominal GDP to fall, GDP deflator inflation non-negative, unemployment not rise). - Simulated impulse responses for a +1 pp PD shock (~+0.7 in logit space) presented (Figure 7). The quarterly impulse response used in iterative “looping”, scaling initial PD impulse until convergence. ### Bank impact module and LGD linkage - Structural bank impact module links industry-level PDs and LGDs to banks’ loan losses and RWAs. Starting point: risk parameter data as of 2019 for corporate loan portfolios of the seven largest Canadian banks (Appendix Figure 2). - Bank corporate loan book parameters included risk weight densities, IRB portfolio shares, PiT and regulatory TTC PDs, PiT and regulatory downturn LGDs—sourced from banks’ financial and Pillar 3 reports as of 2020Q1 (reflective of 2019 conditions). - Bank-level PD projections for 2020 linked to firm sector aggregate PD projections from micro-macro simulation engine; LGDs structurally linked to PDs using Frye-Jacobs method. - LGD formula applied at bank corporate portfolio level: - LGD_b,t = Φ[ Φ^−1[PD_b,t] − k_b ] PD_b,t - Parameter k computed assuming asset correlation (ρ) = 10 percent: - k_b = Φ^−1[PD_b_TTC] − Φ^−1[PD_b_TTC × LGD_b] / √(1−ρ) - Long-term average LGDs unavailable; point-in-time LGDs as of 2019 used instead. Appendix Figure 3 reports estimated k values. - RWA impact via Basel risk weight formula for IRB portion. TTC PD input smoothed from point-in-time PD impacts with smoothing parameter λ set judgmentally to 0.3 for base results. - Regulatory downturn LGDs held constant at reported value during simulations. - Caveat: LGD anchor values sourced from banks (corporate loan portfolios) while micro simulation operates on firms’ overall debt inclusive of corporate debt; assumption that loan portfolios’ LGDs are reasonable proxy for corporate bond LGDs. - Bank of Canada data note: corporate debt is composed of about 60 percent (40 percent) by loans (bonds). ### Policy counterfactual simulation setup - Generic policy counterfactual assumption set judgmentally to +150bps—reflective of the move of policy rates and short-term money market rates after the onset of the pandemic in spring 2020 and throughout the year. - Rationale: assume firms’ debt cost would have been higher had monetary policy easing through various means not been considered. ### Counterfactual firm-level impacts (first and second round) - Firms broadly susceptible to variation in debt costs across all industries: rising PDs, rising LGDs, falling ICRs and cash to debt ratios (Figure 9). - Cross-industry debt-weighted aggregate PD counterfactual trajectory: - Could have moved to about 6 percent, or beyond 9 percent at the peak throughout 2020 when considering second-round feedback effects; measured 2020 average was 3.2 percent. - LGDs, all industries combined: measured 2020 LGD 21 percent → 32 percent under counterfactual inclusive of feedback (LGD increase by a factor of about 1.5). - Shares of firms with ICRs less than 1 increase more visibly for consumer cyclical, consumer non-cyclical, and utilities sectors. - Cash to asset and cash to debt ratios tend to fall comparably across industries. - Debt to asset ratios do not move much due to assumption of no additional debt issuance by firms during simulation (debt increased only if cash stocks turned negative). ### Macro feedback estimates from corporate sector debt cost counterfactual - Text statement: absent broad-based monetary support, real GDP growth could have been lower by 3.3 ppts, unemployment higher by 1 ppt (year average), and risk spreads wider by about 120 bps in 2020; the counterfactual entailed a debt cost increase for firms amounting to a combined 270 bps. - Table 3 reported counterfactual deviations from baseline (as presented in the text): - Real GDP, YoY growth in 2020: −3.4pp - Unemployment rate: 1.0pp - Risk premium, 2020Q1-4: 1.2pp - Assessment: model responses in relative terms across GDP growth, employment, and risk premia comparable to past behavior during the Global Financial Crisis. ### Bank capitalization and system impacts - Assuming no deferral/restructuring behavioral reaction by banks for the analysis, impact on seven largest banks’ CET1 capital ratios ranges between −25 bps and −240 bps (min–max). - RWA-weighted banking system-wide impact about −70 bps inclusive of second-round feedback. - Banks’ capital ratio responses vary across banks due to differences in corporate loan portfolio shares and initial risk parameter structures. ### Micro data appendix highlights - Appendix I: Micro Data Characteristics and Figure A1 compare Statistics Canada and Capital IQ (debt definitions reiterated: debt in debt to asset ratios = short-term plus half long-term debt; cash to debt uses total debt). - Appendix Figures A2 and A3 present banks’ capitalization and corporate loan book conditions at the outset and estimates of k, respectively. *Source: IMF staff calculations and figures from the supplied IMF Canada chapter content.* ### 2. Canada’s mitigation strategy has several key elements: ### 2. Canada’s mitigation strategy has several key elements: ### Core elements of the federal mitigation strategy - A proposed requirement that provinces and territories phase in an explicit carbon price floor, or an equivalently scaled emissions trading system (ETS), with a proposed price progressively rising to CAN $170 by 2030. - A federally imposed carbon pricing backstop where sub-national carbon pricing falls short, consisting of: - (i) a fuel charge; and - (ii) an output-based performance standard (OBPS) for energy-intensive, trade exposed (EITE) industries. - Reinforcing federal incentives at the sectoral level, including tax credits for ZEVs, emission rate standards for vehicles and power generators, and building retrofit programs. - Public funding to support low-carbon investments and transitional assistance. - Equitable and transparent recycling of carbon pricing revenues to households and (where revenues substitute for distortionary taxes) incentives for work effort and investment. ### Feebates: recommended federal-level sectoral instrument - Feebates: revenue-neutral, sliding-scale fees on products/activities with above-average emission rates and sliding-scale rebates for below-average emission rates. - Key advantages noted: - No fiscal cost to the government (important given pandemic-induced budgetary pressures). - Avoids the burden of higher energy prices on the average household and firm (which can aid acceptability). - More flexible and cost-effective than regulations; can provide powerful mitigation incentives. - Primary application and extensions: - Most appeal for the transportation sector; can be used alongside regulations to promote ZEVs. - Can also be used in power, industry, buildings, and forestry sectors. - Variants might apply to fugitive emissions, logging on public forestland, and agriculture (supported by consumer-level incentives to encourage plant-based diets). ### Acceptability and distributional implications of carbon pricing - Modelling suggests the carbon price floor trajectory is aligned with the 2030 emissions target, but: - Some uncertainty exists over emissions impacts and political acceptability of high carbon prices. - Some sectors (e.g., transportation, forestry, agriculture) are difficult to decarbonize through pricing alone or are not currently covered by pricing. - Federal sectoral policies combined with planned carbon pricing could enhance overall effectiveness and acceptability. - Distributional findings: - The pricing scheme by itself would impose an average burden on households of 2 percent of consumption in 2030 with burdens evenly distributed across household income groups. - Returning carbon pricing revenues to households offsets about 80 percent of this burden. - For the most part, burdens at the provincial level are broadly representative of those at the national level. ### Transition from OBPS to border measures and international approaches - Medium-term consideration: transition away from the OBPS to a border carbon adjustment (BCA) as being considered/introduced in other jurisdictions (e.g., EU, UK, US). - With deeper decarbonization of industry, a BCA could address competitiveness and emissions leakage concerns more effectively than the OBPS by applying carbon charges to imports with high embodied carbon (exempting trading partners with adequate carbon pricing). - A BCA applied to EITE industries would limit administrative burdens and perhaps legal risks while raising revenues of 0.7 percent of GDP in 2030, with 35 percent from charges on imports from China and 28 percent from the US. - International carbon price floor (ICPF): - An ICPF arrangement, based on the Canadian model, is highlighted as potentially far more effective than unilateral BCAs for achieving mitigation among large-emitting countries needed over the next decade to stay on track with climate stabilization goals. ### Background: emissions, targets, and policy coverage - Global context: - Global CO2 and other GHG emissions must be cut 25–50 percent below 2018 levels by 2030 to be on track with containing projected warming to 1.5°–2°C above preindustrial levels (on a linear reduction pathway) with rapid reductions to emissions neutrality thereafter. - Pandemic: global emissions projections for 2020 are about 8 percent below 2019 levels; without strong mitigation policies, emissions likely to rise again in 2021 as economies recover. - Canada’s targets: - Goal of zero net GHG emissions by 2050 (made legally binding by the tabled Canadian Net-Zero Emissions Accountability Act, if passed). - NDC target: reduce GHG emissions to 511 million tons of CO2 equivalent in 2030 (30 percent below the 2005 level and 15 percent below the 1990 level). - ZEV sales-share targets for passenger vehicles: 10 percent by 2025, 30 percent by 2030, and 100 percent by 2040. - Phasing out coal-based power generation by 2030. - Sequestering a net 7-46 million tons of CO2 in forests in 2030 (depending on harvest rates). - Canada’s emissions and sectoral breakdown (2018): - Total GHG emissions: 729 million tons in 2018; 74 percent from fossil fuel energy. - Other sources: 8 percent industrial processes, 8 percent fugitive emissions, 8 percent agriculture, 2 percent waste. - By sector (fossil fuel CO2): energy (power and district heating) 36 percent; industry and construction 12 percent; transport 34 percent; buildings/other 18 percent. - By fuel type (fossil fuel CO2): coal 26 percent; oil 42 percent; natural gas 32 percent. - LULUCF emissions were negative 13 million tons in 2018. - Emissions trends and projections: - GHG emissions peaked at 744 million tons in 2007 (24 percent above the 1990 level); emissions were 2 percent below this peak in 2018. - IMF staff BAU projection: fossil fuel CO2 emissions 7 percent lower in 2030 than in 2018 (with projected GDP 15 percent higher). - Without planned mitigation policies, Canada would be the third largest per-capita emitter among G20 countries in 2030 and the tenth largest emitter in absolute terms. ### Carbon pricing architecture and provincial implementation - Pan-Canadian Framework (2016) and 2020 Climate Plan: - Ensures carbon pricing applies throughout Canada with increasing stringency up to 2022; 2020 Climate Plan proposes extending escalation to 2030. - Jurisdictions may implement an explicit price-based system (carbon tax) or an ETS. - Minimum for explicit price-based systems: price rising by CAN$10 per ton CO2e per year to reach $50 per ton by 2022. - From 2023 the Climate Plan proposes raising the annual increase to $15 per ton, implying a 2030 carbon price of $170 per ton. - ETSs should have a 2030 emissions reduction target at least equal to Canada’s 30 percent reduction target and declining annual caps corresponding to projected emissions reductions from a price-based system. - Federal Greenhouse Gas Pollution Pricing Act backstop: - Applies in provinces/territories that request it or do not have schemes aligned with federal criteria; supplements sub-national schemes if needed. - Backstop components: - A regulatory charge on fuels (tax-like component). - The Output-Based Pricing System (OBPS) for EITE facilities: tradable performance standard where facilities above standards can comply via credits, banked credits, paying a fee equal to the carbon price, or buying offsets from provincial offset schemes. - Emitters registered for the OBPS are exempt from regulatory fuel charges. - Provincial implementation status (as described): - Systems fully meeting federal requirements: British Columbia, Northwest Territories, Nova Scotia, Quebec. - Systems meeting requirements for covered sources (with federal backstop supplementing other sources): Alberta, New Brunswick, Prince Edward Island, Saskatchewan. - Federal backstop in place in Manitoba, Nunavut, Ontario, and Yukon. - Legal challenges: Alberta, Ontario, and Saskatchewan have taken the pricing requirement to the Supreme Court claiming it is not constitutional; Manitoba, New Brunswick, and Quebec also argued the law should be struck down. (Hearing concluded on September 23, 2020; decision could take several months.) - Revenue treatment: - Proceeds from the federal backstop remain, by law, in the jurisdiction of origin. - Provinces/territories with systems meeting the federal benchmark retain revenues. - For provinces that have not committed to carbon pricing, the federal government returns approximately 90 percent of revenues from the backstop directly to households as tax-free Climate Action Incentive payments; the remaining 10 percent is directed back to the jurisdiction through funding for public services and supports. ### Sectoral policy recommendations (summary of Table 1) - Road transport: - Introduce feebates for passenger vehicles: sliding scale of fees/rebates based on CO2/mile to complement emissions regulations; incentives can be set aggressively to promote ZEVs without a new tax burden on the average motorist or fiscal cost to the government. - Power: - Introduce a feebate: sliding scale of fees/rebates on generators with above/below average CO2/kWh to reinforce incentives for zero carbon fuels, without a new tax burden on the average generator. - Industry: - Introduce feebates: sliding scale of fees/rebates on firms with emission rates above/below the industry average; powerful incentives for cleaner production without a large tax burden on the average firm. - Buildings: - Supplement energy efficiency programs with: (i) tax-subsidy scheme promoting shifting from natural gas/oil heating to electric/clean fuel systems; (ii) feebates to promote adoption of more efficient appliances and lighting. - Forestry: - Consider a carbon tax on logging from public forestland. - Introduce a nationwide feebate applied to private landowners equal to an (annualized) CO2 price times the difference between baseline and current forest carbon storage; promotes mitigation responses with no burden on the average landowner or fiscal cost to the government. - Agriculture: - Introduce a charge on farm-level emissions with revenues returned in proportion to the value of output to improve acceptability; emissions can be estimated from inputs and default emissions factors. - A shift from beef and dairy to crop-based production could be reinforced by fiscal incentives at the consumer level. - Fugitive emissions: - Tax methane emissions from extractives based on a default leakage rate with rebates for firms validating lower emission rates. - Trade and international: - Following future EU experience, consider a BCA and full carbon pricing for EITE industries to substitute for OBPSs. - Promote dialogue on an ICPF for scaling up action among large emitting countries using Canada’s carbon pricing scheme as a potential prototype. *Prepared by Ian Parry, Simon Black, Danielle Minnett, and Victor Mylonas.* ### 15. The federal government is pursuing complementary actions to reduce emissions and ### 1canea2021002 - 15. The federal government is pursuing complementary actions to reduce emissions and ### Federal mitigation instruments and standards - Emission rate standards on coal and natural gas-fired power stations (adopted in 2018). - Federal rebates: $5,000 ZEVs and long-range plug-in hybrid vehicles; $2,500 for short-range plug-in hybrid electric vehicles. - Vehicle fleet average emission rate standards progressively declining from 210-275 grams CO2 per mile in 2016 to 130-175 grams CO2 per mile in 2025, depending on the vehicle footprint. - Building retrofit programs to improve energy efficiency. - Incentives for reducing hydrofluorocarbons (HFCs) and methane leaks from the upstream oil and gas sector. - Clean Fuel Standard (CFS) to lower the carbon intensity of all fossil fuels (standards start in 2022 and increase annually until 2030). ### Public funding for green projects (explicit allocations) - Buildings: - $2.6 billion over seven years starting in 2020-21 for residential energy efficiency improvements through grants for upgrades and energy audits. - $1.5 million over three years for refurbishment of community and municipal buildings. - $2 billion for retrofitting commercial and large-scale buildings. - Power: - $964 million over four years for renewable energy and grid modernization projects (e.g., power storage). - $300 million in clean power projects for remote and Indigenous communities. - Transport: - $150 million over three years for charging/refueling stations for ZEVs. - $1.5 billion for adoption of zero emission buses. - Industry: - $3 billion over five years for a Strategic Innovation Fund that expedites decarbonization projects with large emitters. - Agriculture: - $165.7 million over seven years for clean technology development and deployment. - Forestry: - Up to $3.16 billion to partner with local actors to plant two billion trees by 2030. - Clean fuels: - $1.5 billion for production and use of low-carbon fuels (e.g., hydrogen, biocrude, renewable natural gas and diesel, cellulosic ethanol). - Just Transition: - $35 million fund supporting skills development and economic diversification in Canada’s coal regions. - $150 million infrastructure fund for projects in impacted communities. - Net-Zero Advisory Body: - $15.4 million over three years, starting in 2020-21 providing guidance on net zero emission pathways. ### Role of fuel taxation and economic rationale - Road fuels are subject to tax but prior to carbon pricing undercharge for supply costs and non-carbon environmental costs (local air pollution; for road fuels, congestion, accident, and road damage externalities). - Existing fuel taxes do not undermine the case for carbon pricing; accounting for unpriced non-carbon environmental costs enhances the economic case for carbon pricing—at least for coal and road fuels. - Some level of fuel taxation is efficient to reflect local external costs of driving until more efficient instruments (e.g., mileage-based charging) are widely applied. ### Assessment of the pan-Canadian carbon pricing scheme - Design strengths: - Comprehensive application to all provinces and territories and all fossil fuel and industrial process emissions within those jurisdictions. - Clearly specified trajectory of robust and rising prices—recently proposed to be extended to 2030—providing a critical price signal to redirect investment. - Flexibility in the use of revenues; efficient uses include lowering distortionary taxes on work and investment or increasing socially efficient investments. Lump-sum transfers do not increase economic efficiency. - Compatibility with overlapping federal or sub-national instruments without undermining provincial price floors. - Projections and uncertainty: - IMF staff estimates suggest the carbon price would cut nationwide CO2 emissions about 33 percent below BAU levels which result in emissions slightly above the 2030 target. - Projection sensitivity: BAU emissions growth (depends on GDP projections), responsiveness of emissions to pricing (depends on future cost and availability of clean technologies), and political resistance to higher pricing. - Political acceptability may be influenced by relative progress with pricing elsewhere; implicit prices in some G20 countries’ 2030 targets are much lower than Canada’s (e.g., China, India, Russia, and South Africa well below US$25 per ton). ### Complementary sectoral instruments and limitations - Sectoral measures can enhance effectiveness and acceptability of mitigation while avoiding fiscal cost and burden of higher energy prices on households and firms; they reinforce carbon pricing if pricing becomes politically constrained. - Existing federal measures have value but limitations: - Meeting coal generation phaseout reduces nationwide emissions by 3 percent in 2030. - Meeting ZEV sales share requirement reduces nationwide emissions by 1.5 percent in 2030. - Emission rate standards for power generators may not be cost-effective across shifts from coal to gas, to NGCC with carbon capture, and to zero-carbon fuels. - Tax incentives for ZEVs do not encourage shifting among conventional vehicles and impose a fiscal cost. ### Road transportation — feebate proposal - Policy proposal: Generalize ZEV tax credits into a comprehensive feebate to strengthen incentives and avoid fiscal cost. - Feebate formula for each new vehicle: - CO2 price × {CO2/mile ─ CO2/mile of the new vehicle fleet} × {average lifetime vehicle mileage} - Certification data (CO2 per mile by model type) is currently available to administer fees/rebates. - Feebate properties: - Cost-effective promotion of full range of behavioral responses to reduce emission rates. - Maintains approximate revenue neutrality as fees offset rebates when average emission rate is updated. - Illustration: - A feebate with a price of US$300 per ton CO2 would apply a subsidy of US$7,500 for ZEVs and a tax of US$1,800 for a vehicle with a CO2 emission rate of 300 grams per mile. - Assumes discounted lifetime driving of 62,000 miles (used in illustrative calculations). ### Power generation — feebate illustration - Federal-level feebate for power generators: - Fee = CO2 price × {CO2/kWh ─ industry-wide average CO2/kWh} × electricity generation - Feebate effects: cost-effectively and revenue-neutrally promote improved generation efficiency and fuel switching (coal → gas → nuclear, fossil with carbon capture, renewables). - Political advantage: avoids increase in electricity prices under carbon pricing though it does not generate same reduction in electricity demand. - Illustration: - A feebate with a price of CAN$50 per ton would currently apply a subsidy of 0.5 cents per kWh for zero-carbon generation, and fees of 1.2 and 4.5 cents per kWh for natural gas and coal plants respectively. ### Industry — competitiveness, leakage, and feebates - Concerns: - Carbon pricing burden on industry includes costs of cutting emissions and larger tax or allowance purchase payments for remaining emissions. - Estimated leakage rate for carbon pricing for Canada is 19 percent. - Competitiveness concerns have been partly addressed via OBPSs that do not charge for most infra-marginal emissions for the average firm. - Feebate as alternative: - Feebate formula for firms: - CO2 price × {CO2/production ─ industry-wide average CO2/production} × production - Feebates reinforce incentives for reducing emissions intensity with smaller burden than high carbon pricing. - Feebates are fiscal analogs of OBPSs, avoid trading markets, and provide price certainty; they can be harmonized with carbon prices applied to fossil fuels to promote cost effectiveness across EITE and the rest of the economy. - Annex 3 contains illustrative comparisons for steel and cement industries (note: annex content not reproduced here). ### Buildings and appliances — feebates and codes - Energy efficiency improvements in new and existing buildings and appliances reduce both direct and indirect emissions. - Market failures (liquidity constraints, owner-renter mismatches, uncertainty of savings) justify policy intervention even with adequate nationwide pricing. - Building codes implemented at the state level. - Feebate schemes for appliances: - Example fee formula for product sales: - CO2 price × CO2 per unit of energy × {energy consumption per unit ─ industry-wide energy consumption per unit} × number of units - Example: refrigerators measured by kWh per cubic foot cooled; number of units would be cubic feet. - Similar schemes could tax gas- and oil-based heating systems and subsidize electric heat pumps to accelerate zero-carbon heating transitions. - Feebates avoid fiscal cost to the government and allow harmonized carbon prices across product categories for cost effectiveness. _International Monetary Fund_ ### 32. Venting accounted for 55 percent of fugitive emissions in Canada in 2018 (two-thirds ### 32. Venting accounted for 55 percent of fugitive emissions in Canada in 2018 ### Fugitive emissions: key facts and composition - Venting accounted for 55 percent of fugitive emissions in Canada in 2018 (two-thirds from oil, one-third from gas). - Flaring accounted for 13 percent (mostly from oil). - Other leaks accounted for 30 percent (the majority from gas). - 70 percent of the CO2 equivalent emissions were from methane releases and 30 percent from CO2. - One ton of methane is equivalent to about 25 tons of CO2 in terms of warming equivalents over a 100-year horizon (IPCC 2007). ### Mitigation possibilities for fugitive emissions - Options listed: - Reinjection of gas for enhanced oil recovery or storage. - Using methane for on-site or regional power generation. - Compressing the gas, or liquifying it, for sale. - Improved maintenance of infrastructure for gas processing and distribution. - Canada’s targets: - Reduce fugitive methane emissions by 40-45 percent below 2012 levels by 2025. - Reduce fugitive methane emissions by 60-75 percent below 2012 levels by 2030. - Current regulations: targeted interventions (e.g., routing emissions to vents, replacing or controlling individual high-emitting components, inspecting equipment for methane leaks). ### Pricing approaches and illustrative impacts - Pricing schemes can use default emission rates with rebating for firms demonstrating lower emission rates. - Monitoring technologies: satellites, aircraft, drones, remote sensing from vehicles; generally provide discrete measurements at limited sites. CO2 emissions from flaring are measurable. - Pricing rationale: fugitive emissions occur within Canadian borders and thus should be priced regardless of whether fuel is for domestic or overseas markets. - Illustration: an emissions tax of $25 per ton of CO2 equivalent on fugitive emissions would apply charges (prior to mitigation) of approximately: - $0.5 per barrel of oil - $0.1 per thousand cubic feet of natural gas - These charges are equivalent to about 0.6 and 2 percent of current supply prices. - Studies suggest that this modest level of pricing could lower emission rates by around 20 percent. --- ### Forestry: channels, instruments, and illustrative figures - Three channels for increasing forest carbon storage: - Afforestation. - Reducing deforestation. - Enhanced management of tree farms (e.g., planting larger trees, longer rotations, fertilizing, tree thinning). - Ownership and land-use facts: - 90 percent of forestland is owned by provinces and territories, 4 percent by the federal government, and 6 percent by private landholders. - Over the last 30 years, less than 0.5 percent of Canada’s forestlands have been converted to a non-forest land use. - Carbon measurement: satellite monitoring, aerial photography, on-the-ground tree sampling (rudimentary inventories possible). - Sequestration and removals: - Temperate forests can sequester up to about 3 tons of CO2 per hectare a year during the growth cycle (Domke and others 2020). - In 2018, logging in Canada’s managed forests accounted for removals of about 8 million tons of CO2. - Feebate program design and illustration: - A national feebate could apply fees to landowners at the agricultural/forestry boundary that reduce stored carbon relative to a baseline and rebates to those that increase stored carbon. - Feebates: fee = {CO2 rental price} × {carbon storage in a baseline year ─ stored carbon in the current year}. - Feebates can be designed to be revenue-neutral in expected terms by scaling the baseline over time. - Feebates should involve rental payments (an annualized CO2 price times the interest rate) rather than large one-off payments because carbon storage may not be permanent. - Illustration: fully stocking a hectare that previously had no trees would increase the land value by about $2,000 under a $50 feebate (or 25 percent of average agricultural land values in Canada in 2019). Calculation assumes the planting sequesters an additional 3 tons of CO2 each year over a 20-year growth cycle with payments discounted at 5 percent. - Administration note: fees and rebates could be administered based on the registry of landowners used for business tax collection. - Logging taxes: common globally, usually as sales or income taxes; could be modified to link to carbon. Partial exemptions may be warranted for timber harvested for wood products because emissions are delayed until end of product life. --- ### Agriculture: emissions sources and policy options - Composition of agricultural GHGs: - Enteric fermentation (livestock, particularly beef and dairy cattle): 41 percent of Canadian agricultural GHGs. - Manure-related nitrous oxide: 14 percent. - Crops for human and animal consumption: 42 percent (reducing crops reduces nitrous oxide emissions from soils, especially with intensive chemical fertilizer use). - Monitoring and pricing considerations: - Direct farm-level emissions monitoring is not currently practical; emissions can be estimated indirectly using farm-level data and default emissions factors. - Pricing could be based on proxy estimates of emissions but a compensation scheme for the farm sector may be needed to enhance acceptability and limit emissions leakage. - Feebate approaches worth studying: based on GHG-equivalent emission rates per hectare, nutritional value, or per $ of output. - Alternative: emissions fee with revenues recycled to the agricultural sector as a rebate proportional to the value of farm output (operationally equivalent to a feebate based on emissions per $ of output). - Administrative integration: fees and rebates could be integrated into collection procedures for farmer business tax regimes. - Demand-side measures: taxes on meat and dairy products (from both domestic and overseas suppliers) could promote household-level demand responses. --- ### Addressing the burden of carbon pricing on households and firms Household incidence - A $170 carbon price in 2030 would, on average (relative to BAU and prior to revenue recycling): - Increase retail electricity prices in Canada 20 percent above BAU levels. - Increase road fuel prices 30 percent above BAU levels. - Increase natural gas prices 200 percent above BAU levels. - Absolute and proportionate price increases: - Natural gas and road fuels: similar across provinces and territories. - Electricity: absolute and proportionate increases differ across provinces due to non-integrated power grids. - On average, the burden on Canadian households from increasing the carbon price to $170 in 2030 (relative to the 2020 price of $30) is 2 percent of consumption prior to revenue use. - Distribution: - Burdens are evenly distributed across population-weighted household per-capita consumption deciles at the national level. - Burdens lower in Quebec where natural gas consumption is limited. - Composition of the household burden: - 35 percent from higher road fuel prices. - 33 percent from higher natural gas prices. - 6 percent from higher electricity prices. - 25 percent indirectly from the impact of higher energy costs on the general consumer price level. - Revenue recycling: about 80 percent of the average household burden in 2030 is offset (at least for the near-to-medium term) when revenues are recycled back into the economy (e.g., broad income tax reductions or general investments). Firm incidence and competitiveness - Most vulnerable industries to carbon pricing include: - Iron and steel mills and other ferrous alloys (stand out as most vulnerable). - Basic chemicals. - Petroleum and coal products (excluding refineries). - Steel products from purchased steel. - Pesticide, fertilizer and other agricultural chemicals. - Resin, synthetic rubber, and artificial fibers and filaments manufacturing. - Other listed sectors (water transportation; motor vehicle metal stamping; forging and stamping; air transportation; water, sewage, other systems; spring/wire products; motor vehicle brake systems; railroad rolling stock; pulp, paper, paperboard mills; other chemical products; architectural/structural metals; boiler, tank, shipping container). - Competitiveness impacts: - For a given carbon price, competitiveness impacts would be less severe at the national level in Canada than in some other large emitters (in part due to Canada’s high share of renewables in electricity). - Averaged across the 20 percent of most vulnerable industries, cost increases from the same level of carbon pricing (US$50) in 2030 (prior to pass-through) would be: - 1.7 times as large in the United States as in Canada. - 2.7 times as large in India as in Canada. - 4.1 times as large in China as in Canada. Border Carbon Adjustments (BCAs) - Context: - The EU intends to announce a proposal for a BCA in June 2021 that would come into force in 2023. - The Biden Administration’s climate plan contains a BCA proposal. - Three main rationales for BCAs: - Help address competitiveness impacts of carbon-price-induced increases in energy prices, enhancing political viability of high carbon prices. - Reduce the risk of emissions leakage (increases in emissions in overseas countries induced by domestic mitigation policy). - Encourage stronger carbon pricing in other countries (through exemptions for those with adequate pricing), though this rationale has little relevance for Canada unless acting in coordination with other large emitters. - Effectiveness and design considerations: - BCAs would be at least as effective as other approaches for addressing competitiveness and leakage, encouraging pricing elsewhere, maintaining mitigation incentives for industry, and mobilizing revenue. - BCAs can be more effective at assisting EITE industries on competitiveness than output-based pricing systems (OBPSs), because OBPSs do not compensate for the costs of reducing emission rates which increase rapidly with the level of abatement. - Concerns about BCAs: administrative complexities and legal risks. - Administrative burden is much reduced if the BCA is limited to EITE industries (reliable aggregated data on embodied carbon in trade flows is publicly available for these industries). - EITE industries account for nearly 90 percent of the emissions from manufacturing in Canada. - Limiting the BCA to EITE industries may enhance legality under trade law (reducing carbon leakage is a potential legal justification under GATT Article 20 for industries with substantial embodied carbon). *International Monetary Fund — Canada chapter excerpts (content unit 1canea2021002).* ### 48. There are various other ### 48. There are various other design issues ### Design issues for Border Carbon Adjustments (BCAs) and harmonization - Other design issues include: - whether to allow rebates for individual overseas exporters that are less carbon intensive than the industry average, - how to adjust charges for carbon pricing or mitigation measures in trading partners, - whether to rebate charges for embodied carbon in exports, - whether to set lower BCA rates for developing country trading partners. - The authorities should consider whether a BCA might be an appropriate instrument for Canada and, at least to some degree, design features that might be harmonized with those adopted by the EU. - EU experience should provide useful guidance. - See Annex 4 for further discussion. ### Revenue implications of a Canadian BCA - A Canadian BCA with a carbon charge of $170 applied to EITE industry imports would have raised revenues of 0.5 percent of GDP in 2015. - Breakdown of the 2015 revenue sources from a $170 BCA: - 41 percent of the revenue would have come from imports from China, - 23 percent from the United States, - 6 percent each from India and the EU. - Using revenues for assisting the domestic clean energy transition (e.g., clean technology infrastructure investments, assistance for vulnerable workers and regions), international climate finance, or rebates to governments of developing country trading partners may enhance the likelihood the BCA is viewed as an environmental measure (and compatible with the WTO) rather than a protectionist measure. - Figure 13: Revenues raised (% GDP) by origin under $170 BCA, 2015 (source: OECD (2021)) — list of reported jurisdictions in the figure includes China, United States, India, EU27, Mexico, Japan, Korea, Russia, Taiwan, Brazil, South Africa, UK, Viet Nam, Indonesia, Thailand, Malaysia, Turkey, Australia, Argentina, Kazakhstan, Rest of world. ### Global mitigation: Canada’s price floor as a prototype for an International Carbon Price Floor (ICPF) - Even if countries achieved their current mitigation commitments for the Paris Agreement, worldwide GHG emissions in 2030 would be reduced less than one-third of the amount consistent with containing projected global warming to 2oC or below. - The main issue: the bulk of low-cost mitigation opportunities is in large emerging market economies, which have relatively lax commitments at present. - Acting unilaterally, countries lack incentives to scale up carbon mitigation due, for example, to concerns about competitiveness impacts. - IMF staff recommend an ICPF arrangement to complement and reinforce the Paris Agreement with the principal aim of increasing near-term mitigation effort in large emerging market economies. - Canada’s carbon pricing scheme provides a valuable prototype for how an ICPF arrangement among large emitting countries might work to scale up global mitigation action. - An ICPF arrangement would be the most efficient approach for addressing countries’ concerns about the competitiveness impacts of carbon mitigation. - The arrangement need only include a small number of large emitting countries facilitating negotiation. - It could be designed equitably with lower requirements for non-advanced economies, and/or transparent transfers, to reflect their lower per-capita income and small contribution to the historical stock of atmospheric GHGs. - The floor could be designed flexibly to accommodate different approaches at the national level including carbon taxes, ETSs, and combinations of pricing, feebates and regulations that achieve the equivalent emissions outcome as implementing the price floor. ### Carbon price floor scenarios and G20 CO2 outcomes (summary of staff estimates) - Table 4: % reduction in G20 CO2 emissions below baseline, 2030. (Assumes CO2 reduced in proportion to total GHGs.) - Required for 2o (1.5) target: 28 (55) - Scenario outcomes (as reported): - Only China, India, and US implement their Paris pledges4.1 → 4.1 - All G20 countries implement their Paris pledges and none join an ICPF10.4 → 10.4 - China, India, US join a $50/25 price floor22.6 → 22.6 - All G20 countries join a $50/25 price floor23.4 → 23.4 - China, India, US join a $50 price floor28.6 → 28.6 - All G20 countries join a $50 price floor29.9 → 29.9 - China, India, US join a $75/50 price floor29.5 → 29.5 - All G20 countries join a $75/50 price floor31.1 → 31.1 - China, India, US join a $75/50/25 price floor28.4 → 28.4 - All G20 countries join a $75/50/25 price floor29.8 → 29.8 - Illustration: if the United States, China, and India were subject to price floors of $75, $50, and $25 per ton respectively in 2030, this would cut G20 emissions about 28 percent below baseline levels, which is just consistent with the 2oC target. - Including all G20 countries would increase G20 emissions reductions, but only moderately, to about 30 percent. - Emissions reductions under the $75/$50/$25 price floor would, broadly speaking, be evenly distributed: - about 20 percent below baseline levels in the EU and India, - about 28 percent in the US, - somewhat over 30 percent in China. ### Implementation issues for an ICPF - Initial focus could be on emissions from the power and industry sectors because: - these emissions are generally the most responsive to pricing and play a key role in early stages of clean energy transitions; - most ETSs currently in place are limited to these sectors; - historically, fuels in these sectors were largely untaxed (or subject to minimal taxation, in terms of CO2 equivalent taxes), making for a clean comparison to a baseline without carbon pricing. - Over time, as the arrangement transitions to broader coverage of fossil fuel emissions, and measuring conventions are developed, focus might move to countries’ ‘effective’ carbon prices which take into account potentially incomplete coverage of formal carbon pricing schemes and changes in pre-existing energy taxes. - Participants could agree to increase their effective carbon prices by a given absolute amount over time, relative to effective prices in a baseline year. ### Summary of Policy Recommendations (as listed) - Introduce a system of revenue-neutral federal feebates to provide strong, reinforcing incentives for reducing emission rates in the transport sector. - Apply feebates to reinforce decarbonization in power generation and industry. - Use feebates to promote adoption of energy-efficient appliances and clean heating systems for buildings. - Apply a fee to fugitive emissions from extractive industries with rebates for firms demonstrating their emission rates are below default rates. - Consider feebates and a logging tax to promote forest carbon sequestration. - Consider feebates or emissions pricing schemes (with within-sector revenue recycling) to promote shifting to less emissions-intensive farming practices, reinforced with fiscal incentives at the consumer level to encourage plant and poultry-based diets. - Consider a medium-term transition from the OBPS to a border carbon adjustment applied to EITE industries. - Promote dialogue on an ICPF arrangement among large emitters to complement the Paris Agreement using Canada’s model as a prototype. ### Appendix I — Further details on the OBPS and provincial/territorial carbon pricing schemes - Facilities covered by the federal OBPS include those that emitted 50,000 tons of CO2 equivalent in 2014 or any subsequent year in the following industries: oil/gas production, minerals, chemicals, pharmaceuticals, iron/steel/metal tubes, mining/ore processing, nitrogen fertilizers, food processing, pulp/paper, automotive and electricity. - Facilities emitting 10,000 tons or more in certain sectors can also apply to participate voluntarily in the OBPS. - Most provincial carbon pricing systems have a version of the OBPS including Alberta, Newfoundland and Labrador, Nova Scotia, Quebec, and Saskatchewan. - Standards for sectors assessed to be at low or medium risk of competitiveness impacts are set at 80 percent of the sector’s average emissions intensity; those assessed to be at high risk are set at 90 or 95 percent of the average. - Effectively, carbon pricing applies to 100 percent of emissions but there is an offsetting output subsidy valued at 80-95 percent of the industry average. - Table A1 provides detail on sub-national pricing schemes (selected entries reported as in the source): - Alberta: ETS on large industrial emitters (carbon tax abolished 2019); Year introduced 2020; Emissions price CAN$/ton CO2 = 48; GHG coverage, percent = 30; Revenue use: Household rebates, green infrastructure, renewable energy, community assistance, small. - Br. Columbia: Carbon tax, ETS on LNG; Year introduced 2008; Emissions price CAN$/ton CO2 = 70; GHG coverage, percent = 40*; Revenue use: Reductions in household and business taxation. - Manitoba: Federal backstop; Year introduced 2019; Emissions price CAN$/ton CO2 = 30; GHG coverage, percent = (blank in table); Revenue use: Rebates to households. - New Brunswick: Carbon tax; Year introduced 2020; Emissions price CAN$/ton CO2 = 39; GHG coverage, percent = 30; Revenue use: Climate Change Fund. - Newfoundland and Labrador: Carbon tax, ETS on large industrial emitters and electricity generation; Year introduced 2019; Emissions price CAN$/ton CO2 = 47; GHG coverage, percent = 20*; Revenue use: (noted in table). - Northwest Territories: Carbon tax; Year introduced 2019; Emissions price CAN$/ton CO2 = 79; GHG coverage, percent = 30; Revenue use: Cost of Living Offset, reductions in reduce GHG emissions. - Nova Scotia: ETS; Year introduced 2019; Emissions price CAN$/ton CO2 = 80; GHG coverage, percent = 20 (floor price); Revenue use: Green Fund. - Nunavut: Federal backstop; Year introduced 2019; Emissions price CAN$/ton CO2 = 30; GHG coverage, percent = (blank); Revenue use: Rebates to households. - Ontario: Federal backstop; Year introduced 2019; Emissions price CAN$/ton CO2 = 30; GHG coverage, percent = (blank); Revenue use: Reductions in household taxation, community assistance, SME support, and GHG reduction. - Pr. Edward Is.: Carbon tax; Year introduced 2019; Emissions price CAN$/ton CO2 = 44; GHG coverage, percent = 30; Revenue use: Reduction in excise tax, transit fees, vehicle registration. - Quebec: Provincial ETS (linked to WCI); Year introduced 2008; Emissions price CAN$/ton CO2 = 85; GHG coverage, percent = 22 (floor price); Revenue use: Green fund. - Saskatchewan: ETS for large industrial emitters; Year introduced 2019; Emissions price CAN$/ton CO2 = 12; GHG coverage, percent = 30; Revenue use: Rebates to households. - Yukon: Federal backstop; Year introduced 2019; Emissions price CAN$/ton CO2 = 30; GHG coverage, percent = (blank); Revenue use: Rebates to households and businesses. - Notes in table: - LNG is liquefied natural gas operations. - WCI is the Western Climate Initiative including the California ETS. - *Increases delayed due to COVID-19. *Source: Excerpt from the IMF staff report (Canada).* ### Appendix II. The Economic Importance of Using Carbon Pricing ### Appendix II. The Economic Importance of Using Carbon Pricing Revenues Productively ### Economic costs and role of revenue recycling - Carbon pricing imposes two sources of costs: - Cost of mitigation responses (e.g., firms using cleaner but more expensive technologies; households using less fuel). - Broader macroeconomic costs: higher energy prices slightly contract overall economic activity by increasing the general price level, reducing real returns to work effort and investment, and compounding distortions in factor markets created by taxes on labor and capital income. - These broader macroeconomic costs can be largely offset (or perhaps more than offset in some cases) by using carbon pricing revenues to increase economic efficiency, for example by lowering taxes on work effort or funding investments warranted on cost-benefit grounds. - Dividends (lump-sum returns of revenue to households) have no efficiency benefits as they do not increase the real return to work effort or investment. ### Comparative efficiency of mitigation instruments (United States example) - An assessment for the United States (policies reduce economywide CO2 emissions 22 percent below BAU) finds: - An ETS with free allowance allocation and emissions price of $50 per ton (or equivalent carbon tax with revenues returned in lump-sum dividends) is about twice as costly—for a given nationwide emissions reduction—as a combination of feebates exploiting major opportunities across the economy for reducing emission rates. - Feebates have smaller impacts on energy prices and therefore smaller macroeconomic costs. - The most cost-effective policy is an ETS with allowance auctions, or a carbon tax, with the bulk of revenues used to cut distortionary taxes on labor and business income, or otherwise increase economic efficiency. - Figure A2: Average cost, $/ton CO2 reduced — shows relative magnitudes for: - Carbon tax: 75% of revenue for cutting labor taxes - Carbon tax and dividend - Feebate/regulatory combination *Italic: Source: IMF (2019a). Note. Policies reduce economywide CO2 emissions 22 percent below BAU.* --- ### Appendix III. Burden of Carbon Pricing on Industries — Conceptual Analysis and Illustrative Impacts ### Conceptual decomposition of unit production cost increases - Two components of the increase in unit production costs for an industry subject to carbon pricing: - Efficiency (resource) cost: induced changes in production methods; indicated by the area under the marginal abatement cost schedule. - Transfer payment: carbon price times the remaining emissions per unit of output (a private cost reflecting a tax payment to government or payment to allowance sellers). - Schemes that avoid (for an average firm) the transfer payment: - Are effective at offsetting most of the increase in unit production costs for modest levels of abatement. - Become progressively less effective at higher levels of abatement (e.g., offset about 95 percent of costs when emissions reductions are 10 percent but only 45 percent of costs when emissions reductions are 70 percent). - Figure A3 illustrates marginal cost, remaining emissions per unit, reduction in emissions per unit, carbon price, efficiency cost, transfer payment, and unit costs. ### Steel — illustrative impacts - Traditional integrated steel process produces about two tons of CO2 per ton of steel. - Alternatives (electrified process using scrap metal; CCS; electric furnace with iron from direct reduction) produce about 0.3–0.4 tons CO2 per ton of steel. - A carbon price of CAN$50/ton of CO2: - Would increase the cost of integrated production by about CAN$100/ton of steel through the first-order transfer payment (about one seventh of recent steel prices). - Would increase the cost under alternative technologies by about CAN$20/ton of steel. - Under a feebate (assumed industry average emission rate of 1 ton CO2 per ton of steel): - Cost increase for integrated production would be CAN$50 per ton of output. - Alternative technologies would receive a subsidy of about CAN$30 per ton of output. - Note: Technology switching is more likely via retrofitting existing plants rather than scrapping and rebuilding, given long-lived existing steel factories. ### Cement — illustrative impacts - Traditional cement production produces about 1 ton of CO2 per one ton of cement; process emissions contribute about 70 percent of these emissions. - Alternatives: - State-of-the-art plants (currently about 10 percent of production) largely eliminate non-process emissions. - CCS: post-combustion reduces emissions about 55 percent; oxy-combustion reduces emissions about 85 percent, with capital costs increasing about 25 and 100 percent respectively. - A carbon price of CAN$50/ton of CO2: - Would increase the cost of traditional production by about CAN$50 per ton of cement (about 30 percent). - Would increase the price of more efficient plants by CAN$30 per ton and CCS-fitted plants by CAN$8–25 per ton through the first-order transfer payment. - Under a feebate with price CAN$50/ton CO2: - Traditional production cost would increase by CAN$5 per ton of cement. - Subsidies to more efficient and CCS-fitted plants of CAN$10 and CAN$18–35 per ton of output respectively. - Note: Cement prices are currently around US$125 per ton of cement. --- ### Appendix IV. Design Issues for Border Carbon Adjustments (BCAs) ### Sectoral coverage - Limiting the BCA to EITE industries: - May make sense on competitiveness, targeted leakage, administrative, and legal grounds. - Competitiveness and leakage concerns are less severe for industries with relatively low energy intensity and trade exposure. - Limits administrative burdens and may limit legal risks (replacement of OBPS; leakage rationale more transparent for EITE products). - Applying a broader BCA to all manufactured products: - More comprehensively addresses leakage and provides a larger incentive for carbon pricing overseas, but effects are modest and administration is more complex. - Administrative burdens rise rapidly; calculating embodied carbon for non-EITE products is more difficult. - Broader BCA might involve higher legal risks; leakage rationale for non-EITE products with low carbon relative to value is more open to question. ### Measuring embodied carbon - Country-specific vs domestic measures: - Country-specific measures appeal on economic efficiency grounds (consumers face correct relative prices across imported products) but increase administrative complexity and raise uncertainties about WTO legal criteria. - Pragmatic approach: use domestic embodied carbon initially to limit legal risks, with progressive transition to country-specific BCAs later. - Industry-wide vs firm-level measures: - Firm-level measures may reflect production heterogeneity but add administrative complexity and are currently limited by data availability (public measures by country and product only for broad classifications). - Industry average benchmarks may be more practical. - A ‘rebuttability’ provision allowing firms to claim rebates, subject to third-party verification, could help WTO compatibility, but risks of gaming exist (e.g., shifting cleaner production away from exports). Rebates could be based on exporter-wide average embodied carbon to avoid plant-level switching. ### Rebates for domestic exporters - A symmetric rebate on exports reflecting the difference between domestic and foreign carbon prices levels the playing field in international markets. - The subsidy element reduces domestic mitigation incentives, makes meeting national mitigation targets harder, and forgoes government revenues. ### Adjusting import charges for foreign carbon pricing or mitigation policy - Lowering BCA rates for exporters with carbon pricing: - Is consistent with reducing carbon leakage and avoids double taxation of overseas emissions. - If BCA motivation is competitiveness or leakage reduction, it might be linked to pricing for power generation and industrial sectors (pricing for residential and transport fuels has little impact on production costs for EITE industries). - Practical complications: - Formal carbon pricing may be partially offset by reductions in pre-existing fuel taxes. - Partial-coverage schemes would need converting to economy-wide equivalents. - Could link BCA to changes in ‘effective’ carbon pricing in overseas countries, though this complicates administration. - Practical compromise: weigh formal pricing schemes by the fraction of emissions covered until widely accepted measures of effective carbon prices are developed. ### Differentiating charges by country income - Applying a lower BCA rate for exporters in low-income countries (LICs): - Would partially undermine the BCA’s ability to address competitiveness and leakage (but only moderately given their modest shares in trade with Canada). - Excluding LICs could be consistent with the principle of common but differentiated responsibilities and potentially with the WTO’s Enabling Clause if exemption criteria are based on objective development indicators. - Country-based exemptions would need design to prevent trans-shipment of goods from covered countries through exempted countries. *Italic: Source: Appendix II–IV, “Appendix II. The Economic Importance of Using Carbon Pricing Revenues Productively” (content unit 1canea2021002).* --- _Source: https://www.imf.org/-/media/files/publications/cr/2021/english/1canea2021002.pdf_