Volatility in Panel Data of Household Expenditure
The variance of growth rates of recall food expenditure is usually greater than the variance of the income growth rate. Many researchers regard the strong volatility as a symptom of measurement error. Comparing two data sets, diary data and recall data, I find that the measurement error cannot account for the observed large variance in the consumption data. Variance decomposition of the consumption and income growth rates reveals that the permanent component of the consumption variance is smaller than that of the income variance, suggesting that the consumption smoothing holds in the long run. Short run fluctuation in consumption, however, is not caused by measurement error. The implication of this finding for the modeling of household behavior is also discussed.
|Date of creation:||Feb 2008|
|Date of revision:|
|Contact details of provider:|| Postal: 2-1 Naka, Kunitachi City, Tokyo 186|
Web page: http://www.ier.hit-u.ac.jp/
More information through EDIRC
When requesting a correction, please mention this item's handle: RePEc:hst:hstdps:d07-237. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Tatsuji Makino)
If references are entirely missing, you can add them using this form.