OBJECTIVE: International comparisons of horizontal inequity in health have recently become one of the most pertinent issues in health economics. Japan has not been included in these international comparisons. This paper, focusing on Japan, rectifies this and considers its dynamics over six years from 1992 to 1998, which has never been considered in this field. METHODS: We use the Comprehensive Survey of Living Standards in Japan (CSLSJ) for 1992, 1995, and 1998 so as to perform international comparison following Doorslaer and Wagstaff et al (JHE:2000). The sample size for each year exceeds 60,000. First of all, we regress outpatient utilization on age, gender, self-evaluation of health, and/or list of symptoms and define "Needs" as the predicted. Then, Kakuwani index between actual utilization and "Needs" is calculated. RESULTS: In a rigorous international comparison, we cannot find any horizontal inequity in health in Japan. The point estimator is larger than Belgium and less than Canada, which means the smallest horizontal inequity in OECD countries. Moreover, it gradually changes from pro-rich to pro-poor, though this movement is not significant. CONCLUSION: As a reminder, the demand for outpatient services is defined as current visits to physicians currently, whereas in the previous OECD studies, it was defined by the number of visit to physicians (general practioner) in a certain time period (typically, one year). Since there is no closer data to the other OECD countries than CSLSJ in Japan, we have to survey as originally so as to obtain the completely comparable data even though it is expensive to do so.