When I multiply my Intercept in a multiple linear regression, what is the effect on t value and Standard error?

113 Views Asked by At

I am running a regresssion on Fama and French 4 factor model in R. I need the Intercept(Alpha) not per month, but per year. When I multiply the Intercept by 12, should I do the same with the t value? Does my Standard error of Alpha changes? Does my p value of Alpha changes?(In this case, I think not)

1

There are 1 best solutions below

0
Morgan Ball On

No you don't need to recalculate the p-stat. While the formulae to calculate the test statistic for alpha depends on the magnitude for the estimated parameter alpha, scaling the alpha by 12 would also scale the SE by 12 because in effect your scaling all your data by 12.

See here for confirmation: http://www2.kobe-u.ac.jp/~kawabat/ch06.pdf