* The relationship between natural log (LOG) and logarithm to
* the base 10 (LOG10) is: LOG(x) = 2.3026 LOG10(x)
* This example compares the results of:
* (1) a regression with all variables transformed to
* logarithm to the base 10.
* (1) a regression with all variables transformed to
* natural logarithms.
* Load the Theil textile test data set
* Henri Theil, Principles of Econometrics, Wiley, 1971.
DEMO
* Generate common logarithms - logarithms to the base 10.
GENR LCON=LOG(CONSUME)/2.3026
GENR LINC=LOG(INCOME)/2.3026
GENR LPRICE=LOG(PRICE)/2.3026
* OLS regression - Theil, Equation (3.18), p. 116.
OLS LCON LINC LPRICE
* The estimates of the slope coefficients are identical
* to using natural logarithms.
* Generate natural logarithms
GENR LCON=LOG(CONSUME)
GENR LINC=LOG(INCOME)
GENR LPRICE=LOG(PRICE)
OLS LCON LINC LPRICE
STOP