The Changing Japanese Retail Market Potential from 1997 to 2007

We augment the literature on retail market potential by combining two research streams, one on customer “out-shopping” and one on structural determinants of consumer expenditure patterns. Using the three most recent Japanese retail trade censuses (1997, 2002 and 2007), we ascertain key factors that determine aggregate retail expenditures per household in over 600 cities comprising ten megalopolises and 75% of the Japanese population. Our key factors include eight socio-economic, environmental and demographic (SEED) variables and five marketing-mix variables. Our research makes five major contributions. First, we introduce two measures of intra-metropolitan out-shopping – individuals living in one part of a metropolitan area shopping in another part of the conurbation. This variable is especially important when workers live in “dormitory suburbs” but shop in central cities. Second, we include a marketing-mix variable (aggregate advertising) that has not been investigated previously, but that is conceptually and statistically important in persuading shoppers to spend. Third, we employ home size as a measure of household wealth; this captures the concept that many home owners treat their residence as a substitute for savings. Fourth, we separately investigate the effect of these variables, and other variables that have been used by previous researchers, on total retail expenditures per household in three census years (¥/H). Few studies have examined changes in retail structure over time; none have done so over three census periods. Using OLS regression, we are able to explain more than 70% of the variation in ¥/H. Fifth, we demonstrate that four of our five marketing-mix variables, and six of our eight SEED variables, have a stable impact over the ten-year period. However, two SEED variables have an impact that varies significantly across these census periods: home size and single-person households lose significance in the last census period. In contrast, the modernity of stores rises from insignificance in the final census period. We link these differences to an aging population and to the decline in small, traditional stores relative to larger, modern retail establishments.