Osaka University of Economics

How Much Information Can We Get?

Email address: yhayashi@osaka-ue.ac.jp

Keywords: expected income, consumption

JEL Classifications: C22 D12 D91

Abstract:
Assuming that real return rate is equal to discount rates, household consumption, which follows PIH hypothesis, is varied by only difference between expectations of periods t and t-1. In this paper, we investigate how long the difference is significant. For example, when our revised period is 3-month, we can get more information on next period's income than 10 years after income during the 3-month. We investigate how much information we get. Our estimation model is extremely simple as following. Ct=a0 + a1 Ct-1 + u't (1) Ct=b0+ b1 YDt + b2 Ct-1 + ut (2) �@Where YD and C are real income and real consumption, respectively. In equation (1), u't implies unpredictable income shock in time t. In equation (2), b1 means unpredictable income effect for present consumption and ut means unpredictable income shock for after period t+1. And b2 implies the influence of a prior consumption, which dose not include predictable income in period t, on present consumption. Comparing equation (1) and (2), we get the ratio of permanent income to predictable present income as follow: The percentage of consumption which dose not include predictable income in period t is b2/a2. Then the predictable income can explain 1-b2/a2 of consumption. Therefore we can estimate the degree of influence in each period using equation (3). Ct=b0+ b1j Ydt+j + b2 Ct-1 + ut (3) And calculate ID=1-b1j/a1. (4) We compare several revised periods of ID and significance of b1j.

PDF file of paper: hayashi.pdf

Session: Consumer Demand Behavior

Time: Sunday, 8 July, 2:15pm - 3:45pm

Room: D