Nonparametric Estimation of Household Gasoline Demand in Japan

2004 Fiscal Year Final Research Report Summary

• Project/Area Number: 15530166
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Applied economics
• Research Institution: Seikei University
• Principal Investigator: INOUE Tomoo Seikei University, Faculty of Economics, Preofesor, 経済学部, 教授 (70307114)
• Co-Investigator(Kenkyū-buntansha): SHIMOJI Makoto Yokohama National University, Inaternational Graduate School of Social Sciences, Associate Professor, 国際社会科学研究科, 助教授 (00323923) ; UESUGI Iichiro Research Institute of Economy, Trade and Industry, Fellow, 研究員 (40371182) ; HIRAO Yukiko Seikei University, Faculty of Economics, Professor, 経済学部, 教授 (40276647)
• Project Period (FY): 2003 – 2004
• Keywords: nonparametric analysis / panel data / gasoline demand / CO2 / Family Income and Expenditure Survey / elasticity

Research Abstract

We estimate the gasoline demand in Japan by using the Family Income and Expenditure Survey conducted by the Ministry of Internal Affairs and Communications. We compare the estimated price and income elasticities with the ones in the previous studies including Hausman and Newey (1995), Schmalensee and Stoker (1999), and Yatchew and No (2001). Compared with these studies, our data set has two major advantages, i.e., (1) the long sample periods of more than 20 years, and (2) availability of the actual income values rather than income classes. These enable us to add further insights on the gasoline demand behavior not examined in the literature.
One of the insights is on the intertemporal aspect of gasoline purchase. In contrast to other non-durables, gasoline is likely to be stored in a sizable amount as inventories. As pointed out by Hendel and Nevo (2002), omitting the stock-aspect variables necessarily biases the estimates of price elasticities. Our stock variables are significant in the estimation, which implies that the elasticities in the previous literature without inventory variables are possibly biased.
Considering the stock effects is important especially in the gasoline market where the price variations reflect the strategic relationship between consumers and gasoline stations (Noel (2003), Inoue and Shimoji (2005)). Our results demonstrate the importance of dynamic purchase behavior, which cannot be captured in the static supply-demand analysis.
Several environmental policy implications are also presented in the paper.