## wpiea2020253-print-pdf - introduction reduces both the average rate of energy efficiency gains and the static saturation point. ## Source details **Canonical URL:** [wpiea2020253-print-pdf - introduction reduces both the average rate of energy efficiency gains and the static saturation point.](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020253-print-pdf.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020253-print-pdf.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020253-print-pdf.pdf.json) --- ### Robustness and alternative specifications - Results are qualitatively robust to different specifications, including “distance from the equator” and “average winter and summer temperature” (not shown), which serve as proxies for intrinsic demand for cooling and heating energy services, reducing the significance of income. - A dynamic panel (where energy is introduced as lagged dependent variable) and a specification where energy is in per capita terms deliver qualitatively similar results. - The dynamic panel was not chosen as baseline because it introduces a short- and long-term Kuznets curve and makes coefficient interpretation more problematic. - Per capita specification and cross-sectional dependence: - Baseline specification of Section Results using per capita energy (dropping population and replacing land area with density) yields qualitatively similar results; R-squared is somewhat lower than in Table 2’s regressions but still as high as 0.94. - Specification 8 (OLS FE) of Section Results using per capita energy: overall R-squared of 0.79; within R-squared = 0.56; between R-squared = 0.84. - Cross-sectional dependence cannot be ruled out, but its effects are probably modest given the use of year dummies. - 47 cross-section regressions (one for each year of the main panel) corroborate the Kuznets curve and are statistically significant from 2006 onwards; signs of the three polynomial coefficients often point in the right direction in earlier years. - Between estimator (discarding time variation) applied to Table 2 also finds the Kuznets curve remains highly significant. - Longest historical panel: country-level regressions of energy demand per capita on population, the income polynomial, four lags of real oil prices and a year trend: - For 10 out of 11 countries of Group 1, significant evidence of a Kuznets curve. - For 7 out of 11 countries the year trend is negative and statistically significant, indicating energy efficiency gains. ### Main empirical conclusions: Kuznets curve, elasticities, and saturation - A Kuznets curve characterizes the relationship between energy demand and economic development (proxied by per capita income). - Preferred specification explains a large part of energy demand variation over time and across countries once controlling for manufacturing and common factors. - Income elasticity of energy demand: - The highest value for the elasticity of energy to income is slightly above 1 and is reached at middle-income levels. - Based on this elasticity alone, emerging markets would have to forego growth to slow down their energy consumption absent other factors. - Energy efficiency gains: - Historically reduced energy demand by about 1.2 percent per year on average in the past fifty years. - In the longer sample, energy efficiency gains became significant only after WWI. - In the post WWII period, estimated energy efficiency gains are in part due to fast productivity growth in the manufacturing sector. - The decline of the global manufacturing share explains roughly 0.2 percentage points of the annual rate of global energy efficiency gains. - Country-level examples (changes in manufacturing share of GDP between 1990 and 2017): - Can account for as much as a 4.1 percent decline in US energy consumption. - Can account for a 10.7 percent increase of China’s energy consumption. - Looking forward, the past extent of manufacturing re-shuffling across countries is unlikely to be repeated, so this impact on energy demand will diminish. ### Static vs. dynamic saturation and implications for decoupling - Static saturation (peak of the Kuznets curve): - The peak is far into the future, even for AEs, and is above $107,000 per capita income; absolute energy-income decoupling is still far away. - Dynamic saturation (accounting for continuous efficiency gains and a declining but still positive income elasticity): - Income necessary to reach dynamic saturation is about $43,000. - $43,000 per capita income is already reached by 15 countries in North America, Europe, and the Asia-Pacific region. - Dynamic saturation is better estimated and more robust to misspecification because estimates of efficiency gains and static saturation usually move in opposite directions. - Important implication: further declines in energy consumption are not guaranteed and will depend on future energy efficiency gains continuing to outpace the effect of income growth on energy consumption. ### Energy prices, innovation, and cyclical effects - High oil prices have probably affected innovation in energy efficiency but with slow effects. - The decade of high oil and energy prices which ended in 2014 likely helped stimulate energy efficiency gains in recent years. - Prospects of cheap energy will probably reduce the incentive to innovate in energy and end-use energy services in the absence of policy interventions. - Cyclical movements: - Cyclical movements in energy savings are strongly related to the global business cycle and oil price movements. - Historical episodes: - Oil price shocks of the 1970s and consequent US recessions led to cyclical energy savings (1974 and 1980-83). - Cyclical energy savings rose above zero in 1988, two years after the counter oil shock (breakdown of OPEC). - Cyclical energy efficiency component peaked in 2004 amid the mid-2000s China growth spur. - Formal test of responses (local projections on a trivariate VAR; global GDP is log of sum of PPP GDP; real oil prices in logs and deflated US CPI): - A one standard deviation shock to oil prices, which raises oil price by 29 percent, decreases energy demand by about 0.5 percent in the second year and about 1.2 percent after 8 years. - A one standard deviation increase in global GDP, by about 1.5 percent, leads to an increase in energy demand by about 0.7 percent, but estimates are imprecise. - Cross-correlation between (real) oil prices and deviations from the efficiency trends is below -0.5, with oil prices leading by 3 to 5 years. ### Methodology and variable treatment - A trivariate VAR is estimated. - A linear trend is added to the trivariate VAR. - Local projections methods are used to estimate the VAR. - Real oil prices are in logs and deflated US CPI. - GDP and oil prices are ordered first and second, respectively. *Source: wpiea2020253-print-pdf - introduction reduces both the average rate of energy efficiency gains and the static saturation point (IMF).* ### introduction reduces both the average rate of energy efficiency gains and the static saturation ### wpiea2020253-print-pdf - introduction reduces both the average rate of energy efficiency gains and the static saturation point. ### Robustness and alternative specifications - Results are qualitatively robust to different specifications, including “distance from the equator” and “average winter and summer temperature” (not shown), which serve as proxies for intrinsic demand for cooling and heating energy services, reducing the significance of income. - A dynamic panel (where energy is introduced as lagged dependent variable) and a specification where energy is in per capita terms deliver qualitatively similar results. - The dynamic panel was not chosen as baseline because it introduces a short- and long-term Kuznets curve and makes coefficient interpretation more problematic. - Per capita specification and cross-sectional dependence: - Baseline specification of Section Results using per capita energy (dropping population and replacing land area with density) yields qualitatively similar results; R-squared is somewhat lower than in Table 2’s regressions but still as high as 0.94. - Specification 8 (OLS FE) of Section Results using per capita energy: overall R-squared of 0.79; within R-squared = 0.56; between R-squared = 0.84. - Cross-sectional dependence cannot be ruled out, but its effects are probably modest given the use of year dummies. - 47 cross-section regressions (one for each year of the main panel) corroborate the Kuznets curve and are statistically significant from 2006 onwards; signs of the three polynomial coefficients often point in the right direction in earlier years. - Between estimator (discarding time variation) applied to Table 2 also finds the Kuznets curve remains highly significant. - Longest historical panel: country-level regressions of energy demand per capita on population, the income polynomial, four lags of real oil prices and a year trend: - For 10 out of 11 countries of Group 1, significant evidence of a Kuznets curve. - For 7 out of 11 countries the year trend is negative and statistically significant, indicating energy efficiency gains. ### Main empirical conclusions: Kuznets curve, elasticities, and saturation - A Kuznets curve characterizes the relationship between energy demand and economic development (proxied by per capita income). - Preferred specification explains a large part of energy demand variation over time and across countries once controlling for manufacturing and common factors. - Income elasticity of energy demand: - The highest value for the elasticity of energy to income is slightly above 1 and is reached at middle-income levels. - Based on this elasticity alone, emerging markets would have to forego growth to slow down their energy consumption absent other factors. - Energy efficiency gains: - Historically reduced energy demand by about 1.2 percent per year on average in the past fifty years. - In the longer sample, energy efficiency gains became significant only after WWI. - In the post WWII period, estimated energy efficiency gains are in part due to fast productivity growth in the manufacturing sector. - The decline of the global manufacturing share explains roughly 0.2 percentage points of the annual rate of global energy efficiency gains. - Country-level examples (changes in manufacturing share of GDP between 1990 and 2017): - Can account for as much as a 4.1 percent decline in US energy consumption. - Can account for a 10.7 percent increase of China’s energy consumption. - Looking forward, the past extent of manufacturing re-shuffling across countries is unlikely to be repeated, so this impact on energy demand will diminish. ### Static vs. dynamic saturation and implications for decoupling - Static saturation (peak of the Kuznets curve): - The peak is far into the future, even for AEs, and is above $107,000 per capita income; absolute energy-income decoupling is still far away. - Dynamic saturation (accounting for continuous efficiency gains and a declining but still positive income elasticity): - Income necessary to reach dynamic saturation is about $43,000. - $43,000 per capita income is already reached by 15 countries in North America, Europe, and the Asia-Pacific region. - Dynamic saturation is better estimated and more robust to misspecification because estimates of efficiency gains and static saturation usually move in opposite directions. - Important implication: further declines in energy consumption are not guaranteed and will depend on future energy efficiency gains continuing to outpace the effect of income growth on energy consumption. ### Energy prices, innovation, and cyclical effects - High oil prices have probably affected innovation in energy efficiency but with slow effects. - The decade of high oil and energy prices which ended in 2014 likely helped stimulate energy efficiency gains in recent years. - Prospects of cheap energy will probably reduce the incentive to innovate in energy and end-use energy services in the absence of policy interventions. - Cyclical movements: - Cyclical movements in energy savings are strongly related to the global business cycle and oil price movements. - Historical episodes: - Oil price shocks of the 1970s and consequent US recessions led to cyclical energy savings (1974 and 1980-83). - Cyclical energy savings rose above zero in 1988, two years after the counter oil shock (breakdown of OPEC). - Cyclical energy efficiency component peaked in 2004 amid the mid-2000s China growth spur. - Formal test of responses (local projections on a trivariate VAR; global GDP is log of sum of PPP GDP; real oil prices in logs and deflated US CPI): - A one standard deviation shock to oil prices, which raises oil price by 29 percent, decreases energy demand by about 0.5 percent in the second year and about 1.2 percent after 8 years. - A one standard deviation increase in global GDP, by about 1.5 percent, leads to an increase in energy demand by about 0.7 percent, but estimates are imprecise. - Cross-correlation between (real) oil prices and deviations from the efficiency trends is below -0.5, with oil prices leading by 3 to 5 years. *Italic: Source: wpiea2020253-print-pdf - introduction reduces both the average rate of energy efficiency gains and the static saturation point (IMF).* ### 1. Real oil prices are in logs and deflated US CPI. A linear trend is added to the trivariate VAR where GDP and oil ### Real oil prices are in logs and deflated US CPI. A linear trend is added to the trivariate VAR where GDP and oil ### Methodology - A trivariate VAR is estimated. - A linear trend is added to the trivariate VAR. - Local projections methods are used to estimate the VAR. ### Variable treatment and ordering - Real oil prices are in logs and deflated US CPI. - GDP and oil prices are ordered first and second, respectively. *Source: wpiea2020253-print-pdf - 1. Real oil prices are in logs and deflated US CPI. A linear trend is added to the trivariate VAR where GDP and oil* --- _Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020253-print-pdf.pdf_