• D12;
  • D30

Abe and Shiotani (2014) are concerned with household-level price differences among households. This type of research has two different roots of investigation.The first is to construct household-level consumer price indices (CPIs) with different household characteristics, namely, the age of the head of household, household income, and household location, for example, Kitamura (2008). The second is to examine the relationship between the household-level actual prices and life-cycle consumption and time allocation behavior, for example, Aguiar and Hurst (2007). This paper follows the second line of research. It is quite sensible to investigate whether Aguiar and Hurst type of observations can be found in Japan.

The advantages of homescan data over the Retail Price Survey conducted by the Japanese Government lie in the former's rich information on the varieties of goods, the purchase prices paid by individual households, and the frequency of shop visits. At the moment, Abe and Shiotani construct shopping behavior-related variables, such as the number of different stores, a Herfindahl–Hirschman index of store shares, the total number of goods bought by a household, a bargain index, a store chain index, and a quality index, and use these indicators as explanatory variables in the household-level price index.

Although Abe and Shiotani's use of this information conveys some interesting aspects of the Japanese household behavior, there could be some alternative ways to use the information from the homescan data. For example, Abe and Shiotani argue that elderly people face higher prices than the young, which is contrary to the results of Aguiar and Hurst (2007). They provide some further analyses of this interesting finding, by finding that the elderly spend more at specialized stores (presumably mom-and-pop retail shops) and supermarkets, and less at drug stores and home_improvement_stores compared to the young. The regression results in Abe and Shiotani's (2014) Table 7 show that store choice can explain part of the price-level differentials between households. However, Abe and Shiotani note from their Table 8 that the elderly face higher prices than the young even within the same store type and that the mechanism behind the higher prices among the elderly needs further investigation.

Another important finding from Abe and Shiotani is that a major determinant of the Japanese household-level price comes from bargain sales. Bargain sales are provided more strategically in supermarkets than in mom-and-pop retail shops. It is worth investigating the items that are the subject of these sales. In so doing, we could identify why the household-level price index is higher for the elderly compared to younger households.

In addition, the elderly's association with specialized stores (that is, mom-and-pop shops) is long term, and young people's shopping at supermarkets is short term (spot trading). There could be an implicit contract between consumers and specialized stores with respect to price stability.

It is also interesting to link the results from Abe and Shiotani with the traditional and existing CPI created by the government. One of the fundamental reasons to study the CPI is to obtain an exact picture of inflation process. Abe and Shiotani focus only on the distribution aspects of price indexes and pay relatively little attention to the inflation process. They may not be able to construct the Laspeyres Price Index from 2004 to 2006, completely comparable to the CPI in the same period because of the data limitations. I believe they can calculate household-level price changes over time and that they can compare this with that of the official CPI with coverage and some other adjustments. As Abe and Shiotani note, the heterogeneity in household-level inflation needs to be investigated. Indeed, it is worth identifying the sources of heterogeneity among households. Abe and Shiotani admit that inflation has different effects depending on the types of households and argue that it is probably that an inflationary policy creates a bigger loss for the elderly and high-income households. These are important policy implications from Abe and Shiotani (2014).


  1. Top of page
  2. References