Micro Data Studies on Japanese Household Consumption

Part I : New Evidence on the Response of Expenditures to Anticipated Income Changes: Panel Data Estimates on Japanese Bonus System This paper exploits a notable institutional feature of the Japanese economy, the bonus system, to test whether households in Japan smooth consumption under large and predictable income changes. We use Japanese micro-data on individual household consumption and compare expenditure seasonality between the households that are in a bonus scheme and those are not. Contrary to the earlier studies on other countries, our empirical evidence indicates that the changes in income that originate from bonus payments undoubtedly affect the seasonal pattern of consumption expenditures, despite the fact that the bonus payments are large and fairly predictable in Japan. Euler equation estimates suggest that Japanese consumption is "excessively sensitive" to bonus income with an instantaneous marginal propensity to consume of less than 0.1. Our micro data analysis is also confirmed by aggregate data. Part II : Are Japanese Consumers Prudent in the Late 1990s? Evidence From Japanese Micro Data This paper explores empirically whether Japanese consumers are prudent in the 1990s, a decade with historically low growth rate. Using the methodology of Dynan (1993), this study uses micro-level data from the Family Income and Expenditure Survey and Family Savings Survey to estimate the coefficient of prudence of Japanese households in the second half of the 1990s. The estimates indicate that the coefficient of prudence is positive and statistically significant in 1998-1999 with an implied coefficient of prudence of 4, much higher than those estimated for U.S. households. The estimates are higher for worker's households and for younger households whose head of household age is under 40. There is not much difference between manufacturing and non-manufacturing or households with different liquidity conditions. Part III : Income Variability and Consumption: A Full Consumption Insurance Test Using Japanese Micro Data in the 1990s This paper explores the relationship between income variability and consumption in the 1990s in Japan. Using micro-level data from the Family Income and Expenditure Survey, this study tests the full consumption insurance hypothesis and quantifies the sources of insurance. The consumption full insurance hypothesis is strongly rejected for both total consumption and consumption by type of goods, which implies that Japanese households fail to smooth consumption. However, the deviation from full insurance is small: there is only a 1 yen to 6 yen change in total consumption per 100 yen change in head of household income, which is comparable to previous studies in the U.S. The main source of insurance to compensate for changes in head of household income is tax payment and the change in tax paid is 12-13 yen in response to a 100 yen change in head of household income. Other sources such as spouse's income, social security transfers and savings a play smaller role in compensating for changes in head of household income. These findings hold even after controlling for different liquidity constraints.