<p>Since the 1990s, national household income inequality, measured in the Gini coefficient, has rapidly increased in Japan. Few studies in geography and its related disciplines have to date, however, analyzed household income inequality between regions. The lack of empirical studies stems importantly from the absence of reliable household income data at regional scales in Japan. The purpose of this study is first to estimate prefectural household income using micro data from the Housing and Land Survey and second to examine the changing relationship between prefectural household income and inter-prefectural migration. It is widely known that theories of regional inequality predict markedly different outcomes. The cumulative causation theory predicts continuous divergence, if unchecked by policy interventions, while the orthodox neoclassical theory predicts eventual convergence through functioning market mechanisms. In contemporary Japan, income convergence through the relocation of firms from high-wage to low-wage regions of the country is increasingly unlikely due to the easiness of corporations being able to relocate internationally. Conversely, if labor migrates from low-wage to high-wage regions, regional inequality in per capita income may decline, but this will further intensify a spatially uneven distribution of economic activities. In other words, there is a dilemma of regional inequality in which the equalization of per capita income intensifies regional disparities in economic activities. Geographic studies of regional inequality are typically interested in spatial distribution and temporal evolution of income levels and in how these spatio-temporal patterns are related to specific regional characteristics and structures; such as, relative locations in the country, biophysical environment, natural resources, and major industries. Average disproportionality indices, such as the Gini coefficient which is widely used in measurements of inequality, have only a limited utility in geographic research because they cannot answer the question of where. In this study, we estimate the median household income for each prefecture by using non-aggregated micro data from Housing and Land Surveys in 1993, 1998, 2003, and 2008. The estimation of prefectural household income data must account for spatial and temporal variations in (1) household size (e.g., because single-person and nuclear family households are more prevalent in urban areas), (2) demographic aging (e.g., because household incomes will be affected by pensions), and (3) cost of living (e.g., because it affects real household income). This study offers one of the first attempts that account for these factors by considering regionally differentiated average household sizes, age compositions of the head of households, and consumer price indices to derive estimated prefectural household incomes. Main findings of the study are as follows. The analysis confirms some of the anticipated spatial distributions of household income levels: high household income in the central regions of the country such as the Tokyo metropolitan area, and low household income in the peripheral regions such as Northern Tohoku and Southern Kyushu including Okinawa. These spatial patterns are relatively constant over the study period. The study also finds that regional income inequality tends to decline during recessions, and increase during booms. Contrary to the original expectation, regional inequality, measured by the coefficient of variation of years 1993-2008, shows little increase in Japan. However, there has been a considerable amount of reordering in the ranking of prefectural household income levels, where the Tokai region with its strong auto industry rose and the economically struggling Kinki/ Osaka region fell behind. The study also examines whether prefectural household income level is positively correlated with the rate of population</p><p>(View PDF for the rest of the abstract.)</p>