| Abstract |
This paper reviews the suitability of Japanese municipal tax data for economic research. It outlines procedures for data construction and anonymisation, and discusses key institutional features of Japan’s tax system that affect data completeness. Using pooled cross‑sectional data from multiple municipalities, the paper shows that, despite limitations in capturing the full tails of the income distribution, the dataset exhibits representativeness in both demographic and income distributions. Another limitation is the inability to track inter‑municipal out‑migration, which may introduce bias in panel‑based studies. A longitudinal overview reveals (i) systematic mobility patterns, with younger and lower‑income individuals more likely to move, whereas stayers are older, more often married, and have greater household resources; (ii) transitory income shocks concentrated amongst younger cohorts, permanent heterogeneity generally rising with age, and females exhibiting greater dispersion; (iii) equivalised income generally displaying the lowest transitory variance compared with other income definitions; and (iv) marital status raising zero‑income risk for females but lowering it for males, with non‑linear age profiles and gender‑asymmetric associations with log‑income levels. Collectively, these findings suggest that, despite untracked out‑migrants, the dataset captures key labour market dynamics and retains representativeness in the longitudinal dimension.
|