The Joneses in Japan: income comparisons and financial satisfaction

This paper uses relatively large-scale internet survey data from Japan to analyse income comparisons and income satisfaction. In contrast to the vast majority of empirical work in the area of subjective well-being, we are able to measure both the direction (to whom?) and intensity (how much?) dimensions of income comparisons. Relative to Europeans, the Japanese compare more to friends and less to colleagues, and compare their incomes more. The relationship between satisfaction and reference-group income is negative and more negative for those who say that they compare their incomes more. Our main finding concerns the measure of the relevant reference-group income. It is common in non-experimental work to calculate "others' income" as some conditional or unconditional cell-mean, with the cells being defined by neighbourhood, workplace or demographic type. We show that two such cell-mean measures (one from within the dataset, the other matched in from external sources) fit the well-being data worse than does a simple self-reported measure of what relevant others earn. The self-reported measure of others' income would arguably make a useful addition to many existing surveys.