Regression wheater on stock return

[rvv_erasmus]

Hi you guys. I think I have a problem. I want to investigate the effect of the wheater on the stock return. So I have 5 cities covered in my dataset (Amsterdam,Brussel,Frankfurt,New York and Dublin). From these cities I have the index returns from the periods 1995-2014. I also have the skycoverage of each of this cities from the period 1995-2014. Skycoverage ranges from 0 to 8 (0 means clear and 8 overclouded). So what I want to do is to investigate the difference return on days with an skycoverage of 0-1(sunnydays) and overclouded days (7-8). My idea was to use a dummy variable and to run two regressions. First I run an regression with the dummy I called clear and than another with the dummy called overclouded. I shave sent an example of one of the regression containing data of NY. I was wondering if I doing the regressions in a good way or that I maybe need to change something. I hope you guys can help me out!!

[startz]

Just run one regression of the return on an intercept and the dummy.

[rvv_erasmus]

So you mean instead of running the regression: return_ny c skc_ny skc_ny*(dummy) I should run the regression: return_ny c skc_ny*dummy ?
Thanks for your reply anyway ;P

[startz]

If you what you

want to do is to investigate the difference return on days with an skycoverage of 0-1(sunnydays) and overclouded days (7-8).

then

Code: Select all

return_ny c dummy

Doesn't mean your specification is wrong, just answering a slightly different question. However, if you include the interaction term you should almost certainly include the dummy as an independent variable as well.