**Appendix A**

Figure A1 depicts direct and indirect CO2 emissions per household during the period 1990–2005. Among age groups, both direct and indirect CO2 emissions per household for those in their 50s were estimated to be the largest, followed by those in their 40s during all periods. This is mainly because their average household income and family size were larger than for other households (e.g., in 2005, the average household income and family size in their 50s and 40s are 7.83 and 7.25 M-JPY/y and 2.87 and 3.17 people/household, respectively). While indirect CO2 emissions for those in their 60s were the third largest for the 15 years investigated, their direct CO2 emissions were fourth—smaller than for those in their 30s from 1990 to 2000. In 2005, emissions in the 60s group increased to become larger than for the 30s. Comparing the per household results to the per capita results, as presented in Figures 1b and 2b, the largest emissions were seen from those in their 50s. However, the orders of magnitude of indirect CO2 emissions for those in their 40s and 60s di ffers between per household and per capita results, as seen in the emissions levels for 1990 and 2005. This is also observed for direct CO2 emissions for households aged between their 20s and 70s. These di fferences between per household and per capita are most a ffected by the average family size and composition of households (as described in the body of this study).

**Figure A1.** Sectoral composition of per household CO2 emissions by age group (t-CO2/y) from 1990 to 2005. (**a**) Direct and (**b**) indirect.

Figure A2 depicts a sectoral-level decomposition analysis for indirect CO2 emissions, except for the sectors discussed in Section 3.3.2. Note that the highest and lowest growth rates of indirect emissions were shown in the information and communication and clothing and footwear sectors, respectively. The growth rate in the information and communication sector was 267%, while that in the clothing and footwear sector was −31% during the studied period. The reasons behind these changes can be explained as follows. In the information and communication sector, the pattern e ffect was the major factor underpinning an increase in emissions (Figure A2e). From the end of the 20th century, the world, including Japan, began to enter into the information technology age, and computers and the internet began to spread. At the same time, mobile phones also became more common. This means

that information and communication technology penetrated into daily life, and the consumption of such technologies has rapidly increased, having a commensurate effect on indirect emissions.

In the clothing and footwear sector, the pattern effect was also the main factor responsible for reducing indirect emissions (Figure A2c). During the studied period, the outsourcing of production in this sector to developing Asian countries reduced production costs (as a result, the supply chain effect became positive during this period). At the same time, fast fashion became popular (Uniqlo, etc.). These complementary phenomena resulted in a lower consumption of apparel and led to a reduction in indirect emissions.

**Figure A2.** *Cont*.

**Figure A2.** SDA results for indirect CO2 emissions between 1990 and 2005 for ten sectors not shown in Section 3.3.2. Δintensity is intensity effect, <sup>Δ</sup>supply chain is the supply chain structure effect, Δpattern is the consumption pattern effect, Δvolume is the consumption volume effect, Δsize is the household size effect, Δdistribution is the household distribution effect, and Δhousehold is the household number effect. (**a**) Food and non-alcoholic beverages, (**b**) alcoholic beverages and tobacco, (**c**) clothing and footwear, (**d**) furnishings, (**e**) information and communication, (**f**) recreation and culture, (**g**) education, (**h**) restaurants and hotels, (**i**) consumable goods, and (**j**) margins, religions and other services.
