"Factor Decomposition of Inter-prefectural Health Care Expenditure Disparities in Japan"

Hayashi, Masayoshi and Akiko Oyama

December 2014

Despite frequent discussions on regional variations in health care expenditure (HCE), few studies account for the sources of such regional disparities. This study bridges this gap in the literature by taking the following two steps. First, we explore the determinants of regional HCE in Japan, covering a data period that expands the scope of previous studies (i.e., the 2000s). Second, we decompose the variations in regional HCE into contributions explained by the HCE determinants examined in the first step, utilizing a regression-based decomposition method. In the regression analysis, we find that the effect of the number of hospital beds on per capita HCE is larger than that of the other determinants, except the proportion of the elderly population. In particular, a 1% increase in the number of hospital beds induces a .22−.43% increase in HCE, in line with Roemer’s Law. The decomposition analysis also finds the salient effect of the number of hospital beds. In particular, this variable accounts for a large proportion of inequality (between 37.6% and 83.9%). This finding also corroborates Roemer’s Law. Our results strongly suggest that the national policy in Japan of reducing hospital beds regionally has been an effective instrument for containing rapidly increasing HCE.