Japanese carbon emissions patterns shifted following the 2008 financial crisis and the 2011 Tohoku earthquake

The study examines how two major unanticipated events—the 2008 global financial crisis and the 2011 Tohoku earthquake—altered Japan’s energy mix and associated carbon emissions at national and regional levels. Japan’s energy system is highly dependent on imported fossil fuels and historically relied on nuclear power to enhance energy security, with renewables gaining attention mainly after 2011. The financial crisis temporarily reduced economic activity and emissions, while the earthquake and subsequent nuclear shutdown forced a shift in electricity generation toward fossil fuels, especially coal and natural gas. The research aims to: (1) determine how emission trends were affected by the 2008 crisis and the 2011 earthquake; (2) identify the drivers behind emission changes following these events; and (3) assess implications for achieving Japan’s emissions reduction targets (Kyoto commitments, 2030 NDC, and 2050 carbon neutrality). The study addresses gaps in sub-national analyses by decomposing sectoral emissions across ten regions for 2007–2015.

Prior work shows unanticipated shocks (financial crises, disasters) can disrupt emission trajectories despite policy efforts such as the Kyoto Protocol and Paris Agreement. The 2008–2009 crisis reduced emissions globally but rebounds were rapid. Japan’s energy security challenges, low self-sufficiency, and historical reliance on nuclear power shaped its energy mix; post-2011 nuclear suspensions increased fossil fuel dependence. Studies have discussed Japan’s evolving policy environment, renewable energy promotion, nuclear safety debates, and the role of energy diversification in coal resurgence. Decomposition analyses (IDA/LMDI, SDA) have been widely applied to assess drivers of emissions in various contexts; however, few studies provide comprehensive sub-national analyses for Japan’s primary, secondary, and tertiary sectors around these shocks. The paper builds on this literature by quantifying regional drivers and highlighting temporal shifts in dominance among economic effects, energy structure, and energy intensity.

Scope: Emissions from primary, secondary, and tertiary economic sectors across Japan’s 47 prefectures aggregated into 10 regions (aligned with the ten major Electricity Power Suppliers) for 2007–2015. Six major sectors (23 sub-sectors) analyzed: Agriculture/Forestry/Fisheries, Mining, Construction, Manufacturing, Energy supply, and Services. Direct household, transport (passenger and freight outside the defined transport/postal sub-sector), and non-energy-use emissions (industrial processes, waste incineration) are excluded. Emissions accounting: Focus on energy-related CO₂ from three fuels—coal, crude oil, natural gas. Total sectoral emissions are the sum of non-electricity-related emissions and allocated electricity-related emissions. Non-electricity-related emissions are computed by multiplying sectoral-regional fuel consumption by fuel-specific emission factors. Electricity-related emissions are allocated to consuming sectors within each region by: (i) obtaining sectoral electricity consumption (kWh), (ii) identifying the supplying power company for the region, (iii) applying the company’s fuel mix shares to decompose electricity emissions into coal, oil, and gas components, then multiplying by emission factors. Electricity emissions are allocated to end-use sectors based on consumption, not to the Energy sector representing conversion processes. Decomposition analysis: Index Decomposition Analysis using the Logarithmic Mean Divisia Index (LMDI, additive form) decomposes changes in emissions into ten drivers: population; per capita GDP; economic structure (sectoral GDP shares); energy intensity (energy/GDP by sector); energy mix shares for coal, oil, gas; and emission intensity (carbon per unit energy) for coal, oil, gas. Aggregate categories are further used for comparison: population effect, economic effects, energy intensity effect, energy structure effect, emission structure effect. Changes are computed between baseline and target years for pre- (2007–2010) and post-event (2012–2015) periods. Data sources include: Agency for Natural Resources and Energy (prefectural energy consumption by sub-sector/fuel), Cabinet Office Prefectural Accounts (GDP, value added, population; GDP in 2011 constant JPY). Emission dataset available at Figshare (10.6084/m9.figshare.14472534). Assumptions include allocating regions supported by multiple power companies to the one with larger coverage due to data constraints.

- National trends: Emissions decreased in 2008 and 2011 relative to prior years, with rebounds thereafter. For 2007–2010, emissions declined by approximately 35.58 MtCO₂; for 2012–2015, emissions increased by 27.06 MtCO₂. Coal-related emissions surged after the 2011 earthquake and stabilized from about 2013 onward, contributing more to emissions than oil and gas; manufacturing remained the dominant sectoral emitter. - Drivers at national level: 2007–2010 decline mainly driven by decreases in per capita GDP (economic downturn) and reduced oil share in the energy mix. 2012–2015 increase driven by per capita GDP growth, industrial value added, and higher coal share in the energy mix. Energy intensity improvements consistently reduced emissions across periods. - Regional patterns: Largest absolute emitters include Kanto, Chubu, Kansai, and Chugoku, reflecting industrial concentration. Manufacturing is the main source of emissions in all regions except Okinawa (services dominate there). Post-2011, coal use increased markedly in Kyushu, Tohoku, and Kansai; Kanto maintained highest oil-related emissions; gas penetration varied, highest in Hokuriku and lowest in Hokkaido. - Regional drivers: In high-emitting regions (Kanto, Chubu, Hokuriku, Kansai, Kyushu), increased coal use was the most significant post-2011 driver of emission increases. In Chugoku and Hokkaido, GDP structure and energy intensity played larger roles (GDP structure increased emissions in Chugoku; in Hokkaido it contributed to decreases). Energy intensity was the strongest driver of emission reductions in all regions except Chugoku and Shikoku. - Temporal dominance of drivers: Around 2009–2010, economic effects dominated; 2012–2013, energy structure (fuel mix) dominated, reflecting the nuclear-to-fossil shift; from 2014 onward, energy intensity and economic effects were more influential. Almost all regions saw emission declines after both events, with smaller variations in Hokkaido and Okinawa. - Notably, signals of increased coal share influencing emissions existed even before 2011, indicating pre-existing trends in energy diversification and security shaping coal use.

The analysis demonstrates that Japan’s emission pathways around the 2008 financial crisis and 2011 earthquake were substantially altered through different mechanisms: economic contraction and recovery primarily drove short-term decreases and rebounds around 2008–2010, while the post-2011 suspension of nuclear power drove a fuel mix shift toward coal (and gas), elevating emissions nationally and in most regions. Energy intensity improvements consistently mitigated emissions and became more prominent later in the period. The regional heterogeneity in levels and drivers underscores the importance of tailoring mitigation strategies to local industrial structures and energy mixes. The findings inform the feasibility of meeting national targets (Kyoto, NDC 2030, and 2050 neutrality), highlighting coal’s central role in post-2011 emissions and the need for differentiated regional responses, especially in industrial hubs. The observed pre-2011 trend of rising coal shares suggests underlying geopolitical and energy security drivers, with the earthquake accelerating rather than initiating coal dependence. This nuanced understanding of driver dominance over time (economic effects → energy structure → energy intensity/economic effects) provides guidance for sequencing policy levers across time and space.

The study quantifies how the 2008 financial crisis and 2011 Tohoku earthquake reshaped Japan’s national and regional carbon emissions across major sectors. It finds: (i) a national emissions decline of ~35.6 MtCO₂ in 2007–2010 and an increase of ~27.1 MtCO₂ in 2012–2015; (ii) manufacturing dominance in emissions; (iii) coal use as the principal post-2011 driver of emission increases in most regions; and (iv) energy intensity improvements as the consistent driver of emission reductions. The paper contributes a disaggregated regional decomposition of drivers, revealing temporal shifts in dominance among economic factors, energy structure, and energy intensity. Policy implications include prioritizing industrial decarbonization—especially coal-related emissions in key regions (Kanto, Chugoku, Chubu, Kansai, Kyushu)—and leveraging energy efficiency. Future work should produce finer spatial resolutions (prefecture/city, industrial clusters) and incorporate technology-specific data and post-COVID-19 dynamics to inform sub-national mitigation pathways and energy mix decisions aligned with Japan’s NDC and 2050 neutrality goals.

- Electricity emissions allocation uncertainty: Regional allocation based on power company coverage; where multiple companies serve a prefecture, the larger coverage area was used—affecting a limited set of cities but not expected to change overall conclusions. - Technology granularity: Lack of plant-level technology and fuel consumption data prevents detailed bottom-up analysis or differentiation of emission intensities among coal technologies; a top-down approach was used assuming limited inter-company intensity differences (except Okinawa vs. others in 2019). - Spatial resolution: Analysis uses ten regions (power supplier boundaries) rather than prefectures or cities; this coarser resolution limits direct applicability to sub-national planning that requires finer granularity. - Scope exclusions: Direct household, comprehensive transport emissions (outside the defined transport/postal sub-sector), and non-energy use/process emissions are excluded, potentially understating total regional emission dynamics. - Temporal constraints: Study period ends in 2015; does not capture later policy shifts or impacts from COVID-19, recent market changes, or nuclear restarts.