The Effect of Academic Background on
Household Portfolio Selection:
Evidence from Japanese Repeated Cross Section Data

Yukinobu Kitamura
Taisuke Uchino

August 2010


This paper investigates whether the participation costs play an important role in explaining the households' limited stock market participation in Japan. In order to identify such costs, we focus on the difference of stock holding probability between two household groups whose financial literacy are a priori considered to be different: college-graduate households and non-college-graduate households. Our empirical analysis is utilizing Fairlie (1999, 2005)'s Blinder-Oaxaca decomposition technique in order to decompose the differences that observed in the Japanese repeated cross section survey data. Estimation results show that (1) the participation probability amounts to 35% for college-graduate households, while 18% of non-college-graduate households have stocks, (2) only 30-50% of the differences can be explained by the social and economic attributes, especially by income, wealth and the size of household heads' place of work. These findings imply the existence of participation costs based on the financial literacy.

Full text

PDF Download (PDF: 515KB) (in Japanese)