*4.2. Country Clusters and Contextual Factors*

Table 2 summarizes our observations regarding the household economies of scale by various consumption domains and the interaction with population density. Two clusters of countries emerge—one with strong or moderate household economies of scale, and one with lower or no household economies of scale.

**Table 2.** A summary of country clusters with regards to household economies of scale and other contextual differences.


Note: One-sided two-sample unweighted t-test is performed in order to compare the average household sizes between the country clusters under the following hypotheses: *H*0: μcluster2 − μcluster1 = 0, *HA*: μcluster2 − μcluster1 > 0. We estimated separate variances to control for significant differences in sample sizes between the country clusters. Standard errors are presented in parenthesis. T-test significance levels: \* *p* < 0.1, \*\* *p* < 0.05, \*\*\* *p* < 0.01.

The first cluster—with high and moderate household economies of scale—consists of predominantly Northern and Central European countries. An increase in the household size by one member results in a reduction of the total carbon and energy footprint by 3–13% (Figure 5). This cluster is characterized by strong welfare regimes that promote individual independence and female labor market participation [19]—which may explain the lower household sizes in these countries. The cluster includes Belgium, Denmark, Sweden, Finland, France, Germany and the United Kingdom, which are similar in terms of socio-demographic context [56]. The small countries of Malta and Luxembourg are exceptions in terms of welfare regime [47,56]; the regression coefficients of Malta in particular are characterized by relatively high error ranges across most consumption categories, and results should thus be interpreted with caution. Finally, the Czech Republic, Poland, Slovakia, Slovenia and Hungary (and Croatia, which is allocated to the second cluster in terms of household economies of scale in our analysis) are characterized by the Central Europe welfare model, associated with lower income inequality, lower rates of unemployment, higher labor market flexibility and higher social contributions and government expenditure as a share of the Gross Domestic Product compared to the Eastern European countries in the second cluster [56].

The second cluster—with lower or no household economies of scale—consists of predominantly Southern and historically Catholic countries as well as some Eastern European states. An increase in the household size by one member does not change the total per capita carbon and energy footprint, or even increases in the per capita environmental impact (Figure 5). These countries already have higher household sizes, and emphasize the role of the family for mutual support or are more "collectivistic". Greece, Spain, Italy, Cyprus and Portugal stand out from other EU countries in terms of their welfare regimes previously described as the Mediterranean welfare model [56] with stronger influence of Catholicism and traditional family values [57–59]. The Eastern European welfare model—including Lithuania, Estonia and Bulgaria (and also Latvia, which is included in the first

cluster in our analysis)—is associated with strong nuclear family institutions, low social protection expenditure primarily on old-age pensions, high income inequality, rigid and discriminatory labor markets and lower government capacity for generous social policies [56,60]. This might also contribute to higher family dependency for financial and welfare support, and hence higher household sizes. The reliance on extended family for assurance against risks of ill health, unemployment or poverty could be reduced with higher standards of living and the provision of stronger social-security systems in these countries [61].

These clusters show significant differences in terms of the average household sizes, with the second cluster denoting a significantly higher household size (Table 2, Figure 2). Considering the decreasing rate of household economies of scale with rising household sizes, this may partly explain the lower household economies of scale in these countries—where there already is a lot of within-household sharing, thus, there is less to gain by adding a household member.

Heating degree days are positively correlated with the housing-related energy use (and carbon footprints), with dwellings in colder regions requiring more energy to heat over the year [4]. This effect is partly mediated by stricter building standards in northern European countries, which reduces the amount of energy for heating per heating degree day [62]. Nevertheless, colder countries are likely to report higher household economies of scale particularly due to the high importance of home energy for the overall household economies of scale, which is also in line with the country clustering. This might also explain why we find significant positive *HHSIZE* effects in countries such as Spain, Italy, Greece, Portugal and Croatia. Not only are these countries with relatively large average household sizes already (and hence less scope for further within-household sharing), but there is also less of a requirement for heating, which is associated with some of the strongest capacity for household economies of scale.

The positive *HHSIZE* coefficients for some of the categories, where we expect relatively low possibilities for sharing is likely driven by other socio-demographic, infrastructural and economic factors that vary with household size, that we cannot explicitly control for in our model because they are not captured in the HBSs.
