*2.3. Data*

This study uses the time-series Japan input–output tables (TJIO) consisting of the economic transaction (*Lij*) and household final demand structures ( *b fib*) for 1990, 1995, 2000, and 2005 based on the 2005 price with 397 common commodities. Hence, the data are comparable among periods. Further, we disaggregated the commodities within petroleum products into six detailed commodities including gasoline, light oil, kerosene, liquefied petroleum gas (LPG), and other petroleum products (i.e., lubricants) by using the Comprehensive Energy Statistics [39]. To identify the relationship between household CF and demographic trends, the final household demands from the TJIO were divided into consumption expenditures for six age groups of the highest income earner in the household (29≤, 30–39, 40–49, 50–59, 60–69, ≥70) for each year using the national survey of family income and expenditure (NSFIE) [40]. Then, the consumption share by attribute, summation of each final consumption type by commodity and the di fference between the producer-based price and consumer-based price were considered as detailed in a previous study [14]. *eind i* was obtained based on Equation (3). *qi* was calculated by dividing sectoral CO2 emissions obtained from the 3EID [41] by total output utilizing the TJIO. *b* denotes the age of the highest income earner (1: ≤29, 2: 30–39, 3: 40–49, 4: 50–59, 5: 60–69, 6: ≥70).
