The Paris Agreement's goal of limiting global temperature increase necessitates significant reductions in greenhouse gas emissions, particularly CO2 from fossil fuels. Japan, aiming for carbon neutrality by 2050, faces the challenge of reducing its reliance on fossil fuels in electricity generation, a major source of emissions. Previous attempts like the Japan Voluntary Emissions Trading Scheme (JVETS) and a carbon tax have shown limited effectiveness due to low participation and a low tax rate. This research focuses on analyzing the regional impact of Japan's carbon pricing system on electricity prices, considering the structure of Japan's electricity market, which involves vertically integrated utilities (GEUs) and the Japan Electric Power Exchange (JEPX), and the market liberalization that began in 2016.

Existing literature explores the theoretical basis of carbon pricing, highlighting its role in internalizing negative externalities associated with air pollution (Pigou, 1920) and incentivizing the shift to renewable energy sources (Nazifi, 2016). Studies by Bauer and Zink (2005), Sijm and Chen (2006, 2012), and Chen (2008) have employed linear regression models to analyze the carbon price's impact on electricity prices. Other research emphasizes the importance of electricity market liberalization for carbon pricing's effectiveness (Fan, 2014). However, some studies (Barlow, 2002; Kanamura, 2007) suggest non-linear relationships between electricity price and demand, which this paper addresses. Prior work on CPTR in Australia (Nazifi, 2016; Nazifi et al., 2021) and Spain (Fabra, 2014) has shown that the cost of carbon emissions is often borne more by consumers than producers. This research contributes by analyzing hourly data at the regional level in Japan, a relatively under-researched area.

The study uses hourly data from April 2016 to August 2021, a period with a stable carbon tax rate in Japan. The CPTR is calculated as the ratio of changes in electricity price to changes in carbon cost. The main analysis employs a polynomial OLS regression (degree two) to model the relationship between electricity price (P), fuel cost, electricity demand (D), and carbon cost (C). A k-fold cross-validation determines the optimal fitting degree. The model accounts for the non-linear relationship between electricity price and demand, and assumes that fuel costs are the primary driver of electricity generation costs, while other costs (labor, maintenance) are considered constant. Carbon cost is derived from the carbon tax rate (JPY 289 per ton of CO2) multiplied by the carbon intensity. To ensure robustness, the study also employs a generalized additive model (GAM) to capture the non-linear relationship between fuel spread and carbon cost.

The analysis reveals significant regional variations in CPTR. Kansai, Chubu, and Chugoku regions show CPTRs greater than 1, indicating that consumers bear a disproportionate share of the carbon cost increase. Conversely, Hokuriku, Kyushu, Shikoku, Tohoku, and Tokyo exhibit CPTRs less than 1, suggesting that producers absorb a larger portion of the carbon cost. Hokkaido presents an exception with a negative CPTR, possibly due to the influence of feed-in tariffs for renewable energy. Polynomial OLS regression results show statistically significant CPTRs in all regions. The mixed model shows an average CPTR of 0.596, suggesting that a one-unit increase in carbon cost leads to a 0.6-unit increase in electricity price. The GAM analysis confirms the robustness of the polynomial OLS results, with CPTRs remaining statistically significant and similar in magnitude. However, the GAM estimated a much lower CPTR for Kyushu (0.077), compared to the polynomial OLS results (0.531). This is possibly due to the presence of many negative fuel spread values in Kyushu. The table below summarizes the CPTRs across different models: | Regions | Polynomial OLS | Linear OLS | GAM | | :------- | :------------- | :------------- | :-------- | | Kansai | 1.937*** | 1.947*** | 1.962*** | | Chubu | 2.157*** | 2.061*** | 2.153*** | | Chugoku | 2.995*** | 2.895*** | 3.004*** | | Hokkaido | -0.055*** | -0.052*** | -0.051*** | | Hokuriku | 0.446*** | 0.411*** | 0.458*** | | Kyushu | 0.531*** | 0.576*** | 0.077*** | | Shikoku | 0.226*** | 0.230*** | 0.225*** | | Tohoku | 0.585*** | 0.611*** | 0.360*** | | Tokyo | 0.698*** | 0.667*** | 0.728*** |

The findings highlight the uneven impact of carbon pricing on electricity prices across Japan. The variations in CPTR suggest differing levels of cost absorption by producers and consumers in various regions. Regions with higher CPTRs may face increased electricity costs, potentially affecting consumer welfare. Conversely, regions with lower CPTRs might see producers absorbing more of the carbon cost, potentially stimulating investments in renewable energy or energy efficiency improvements. The unique case of Hokkaido underscores the influence of regional policies, such as feed-in tariffs, on mitigating the impact of carbon costs. The difference in CPTR results between the polynomial OLS and GAM models suggests that further investigation into non-linearity and regional specificity is warranted. These results underscore the importance of regionally tailored policies to address the diverse impacts of carbon pricing and accelerate the transition to a low-carbon economy.

This study demonstrates the significant regional heterogeneity in the pass-through of carbon costs to electricity prices in Japan. The findings emphasize the need for differentiated policy approaches across regions, considering the varying capacities of producers and consumers to absorb carbon cost increases. The current carbon tax rate may not sufficiently reflect the social cost of CO2 emissions and might need adjustments. Furthermore, continued electricity market liberalization is crucial to ensure a more competitive market that efficiently reflects the full cost of electricity generation. Future research should explore the long-term impacts of carbon pricing, including its effects on investment in renewable energy and energy efficiency, as well as examining the interplay between carbon pricing and other energy policies.

The study focuses solely on the wholesale electricity market and may not fully capture the retail electricity prices faced by consumers. The analysis assumes that other electricity generation costs remain constant, which might not be entirely true in the long run. The model does not explicitly consider renewable energy sources, which could become increasingly important in the future. The findings might be sensitive to the chosen time period and may not be generalizable to other periods with varying carbon tax rates or other policy interventions.