An Examination of the Effects of Consumption Expenditures on Life Satisfaction in Australia

Abstract

This paper studies how consumption expenditures, especially spending on certain types of goods and services, affect people’s life satisfaction. The results demonstrate that conspicuous (i.e., visible and positional) spending increases life satisfaction. The analysis suggests that it is one’s conspicuous consumption expenditures relative to those of other households in one’s reference group that really matter. In contrast, savings and spending on basic goods and services, the less visible components of income, do not contribute to life satisfaction. This paper also evaluates the often-discussed relationship between income and life satisfaction, finding that the relationship is largely driven by conspicuous spending. Evidence of relationship heterogeneity across income groups is presented: conspicuous consumption expenditures have a positive influence on life satisfaction for individuals in all income groups; basic consumption expenditures, however, have a negative influence for people in the lowest income quartile. The findings of this paper underscore the importance of social comparisons to people’s well-being and imply that interdependence should be modeled in utility functions. The results are based on the Arellano–Bond generalized method of moments estimation, which controls for bias from unobserved individual heterogeneity and endogenous variables. The data come from the Household, Income and Labour Dynamics in Australia Survey (2006–2010).

Appendices

Appendix 1: What is Subjective Well-Being

Subjective well-being is the general expression used to cover a variety of individual self-reports of quality of life (Sachs et al. 2013). There are, in general, two types of measures of SWB: cognitive life evaluations and emotional reports. Emotional reports, such as positive affect (a range of positive emotions) and negative affect (a range of negative emotions), capture feelings at one point in time. The focus of this study is on life evaluations, represented by questions asking how happy or satisfied people are with their lives as a whole. Such measures are considered plausible measures of well-being since they are closely related to life circumstances, consistent over a short period of time, and strongly correlated with both objective and subjective measures of well-being.²⁷

The study of SWB was the domain of psychology for a long time (Bradburn 1969; Andrews and Withey 1974; Campbell et al. 1976; Diener 1984, 1994, 2000; Diener et al. 1985; Frey and Stutzer 2002). The groundbreaking contribution by Easterlin (1974) linked this sychological research to economics, but there were few followers at that time. Since the late 1990s, the literature on empirical analyses of the determinants of life satisfaction or happiness in different regions and periods has started to grow. The important determinants include, but are not limited to, age, gender, unemployment, education, marriage, health, and material living conditions (e.g., Blanchflower and Oswald 2004a; Blanchflower 2009; Clark and Oswald 1994; Easterlin 1995, 2001; Di Tella et al. 2001; Knight et al., 2009; Frey and Stutzer 2010).

Appendix 2: The Quality of Household Expenditure Data from the HILDA

The HILDA survey asks retrospective questions on various expenditures in the past financial year. This is comparatively simple compared with what most official surveys do. Household Expenditure Survey (HES), which is conducted by Australian Bureau of Statistics every 6 years, collects information on over 600 categories of consumption expenditures by making use of a diary method over a period of two weeks.²⁸ A comparison between the HILDA 2010 and HES 2009–2010 on weekly household expenditures in Table 9 in appendix shows considerable consistency in the relative importance of each type of consumption between these two sources. The consistency suggests that the consumption expenditure data from the HILDA can be of good quality.

Appendix 3: Measurement Error in Life Satisfaction

Consider the following dynamic model with unobserved individual fixed effects.

$$LS_{it}^{*} = \beta_{0} LS_{it - 1}^{*} + \beta_{1}^{'} M_{it} + \beta_{2}^{'} X_{1,it} + \beta_{3}^{'} X_{2,it} + \beta_{4}^{'} Z_{i} + \alpha_{i} + \varepsilon_{it}$$

(6)

$LS_{it}^{*}$ is true self-rated life satisfaction. However, we cannot observe $LS_{it}^{*}$. What we can observe is a proxy $LS_{it}$, which is reported by an individual. Individuals may have a systematic tendency to misreport their true life satisfaction scores. For example, more extroverted individuals may always overrate their life satisfaction when asked. Following Powdthavee (2009), I assume that there is a systematic error to how life satisfaction is measured and that this measurement error is driven by a fixed effect and a time-varying component:

$$LS_{it} = LS_{it}^{*} + \theta_{i} + \mu_{it}$$

(7)

Assume the error $\mu_{it}$ is not serially correlated and is uncorrelated with $LS_{it}^{*}$ as well as $M_{it}$, $X_{1, it}$, and $X_{2, it}$. Substituting (7) into (6)

$$LS_{it} = \beta_{0} LS_{it - 1} + \beta_{1}^{'} M_{it} + \beta_{2}^{'} X_{1,it} + \beta_{3}^{'} X_{2,it} + \beta_{4}^{'} Z_{i} + \left[ {\alpha_{i} + \left( {1 - \beta_{0} } \right)\theta_{i} } \right] + \left( {\varepsilon_{it} + \mu_{it} - \beta_{0} \mu_{it - 1} } \right)$$

(8)

The presence of time-varying measurement error would make the regression error follow a first order moving average process. Therefore, the error is serially correlated, and I cannot use the second lag of life satisfaction as an instrument in the first-difference equation. What I can do is to use higher order lags (see Tables 8, 9, 10, 11, 12, 13, 14).

Table 8 Consumption categories

Table 9 Weekly expenditures: household expenditure survey (HES) (2009–2010) and HILDA (2010)

Table 10 Life satisfaction, income, total household consumption expenditures, savings, conspicuous consumption expenditures, and basic consumption expenditures (Blundell and Bond GMM estimation results)

Table 11 Life satisfaction and total consumption expenditures, and savings (OLS and FE results)

Table 12 Life satisfaction, conspicuous consumption expenditures and basic consumption expenditures (OLS and FE results)

Table 13 Robustness check: alternative definitions of conspicuous and basic consumption expenditures

Table 14 Life satisfaction and expenditure on durable goods, nondurable goods, and services