27.
Summary We apply the Variate-Difference-Method to yearly Austrian data from 1956 to 1975 to test the permanent-income-hypothesis by
M. Fiedman.First we determe the degree of the polynomial, which is sufficient to eliminate the trend (i.e. the permanent part) from time series of disposable income. Then the permanent income variable is constructed in two alternative ways: first, as a polynomial in time and second by
Sheppard's smoothing formulae. Transitory consumption is identified with the residuals of the consumption function. A correlation analysis shows that, transitory income and consumption are correlated, if permanent income is determined by a trend polynomial, while this is not the case if
Sheppard's smoothing formulae is used to estimate permanent income.Consumption functions are specified for total consumption, durables and non-durables and eleven sub-groups of total private consumption. The results indicate that autocorrelation can considerably be reduced by using
Sheppard's smoothing formulae.
相似文献