Statistical significance between two egressions

[Rutger]

Hi all,

I use the Fama and French three-factor model of which the intercept measures long-term performance (over three years) of mergers and acquisitions (M&A). I have created monthly portfolios of firms, which engaged in M&A for a period of three year and regressed these monthly portfolios on the Fama and French three-factor model. I have two sub samples namely domestic (D) and cross-border (CB) M&A for which I calculated value weighted (VW) monthly returns. I have regressed (OLS AND WLS) both sub samples on the Fama and French three-factor model, which works fine. My problem is that I want to know whether the domestic M&A outperform the cross-border M&A. In other words, I want to know if the two intercepts differ significantly... Unfortunately, I do not know how to do this. I hope someone can help me!!


the equation = Rp-Rf c HML SMB (Rm-Rf)



See jpg file for the tests


Greets,

Rutger

[startz]

In this particular case all you need to do is regress D_VW-CB-VW on the right hand side variables (noting that the sample sizes differ between your two regressions). The coefficients should be numerically identical to the difference in the coefficients you get in the individual regressions, so you can just do a standard test on the intercept.

[Rutger]

Tnx for your quick answer!! I did the regression as you explaind but the coefficients are not numerically identical to the difference in the coefficients of the individual regressions.

the difference should be 0.001597 (0.005412-0.003815 for the intercept) and I get 0.001449....

Do you know how this is possible??

[startz]

Your first regression on CB used a different sample.

[Rutger]

Ok. I don't know much about statistical tests and eviews but is it oke to reject a hypothesis based on this tests?? I have heard something about anova to tests for differences... I also do not know what a minus sign for the SMB coefficient ?? Does a minus sign for the SMB coefficient menas that smaller firms perform worse than bigger firms (in case the intercept is negative)??

[startz]

Yes, you can reject equal SMB coefficients. You're showing that SMB has a smaller effect on domestic than on cross-border firms. (By the way, only the C should be called an intercept.)

[Rutger]

Tnx I get it. I would like to thank you for your time and effort helping me out!!!

Greets