Grabbing observations within a date range from long data

[amanjot_42]

Hi,

I want to keep stocks with very high negative beta values also in the respective deciles. The intuition is that both negative and low positive betas are good for portfolio mean-variance.

Thanks

Regards  

[amanjot_42]

Hi,

1. You want to generate CAPM (capital asset pricing model,  i.e. MODEL retx=sprtrn) beta estimates using CRSP monthly data for your sample for a subperiod,  then use the beta estimates to form decile portfolios. 

Ans: Exactly, but my overall objective is Fama-Macbeth.

Then you want rolling 60 month PORTFOLIO returns, meaning that after the regresssion, you also will need a data step or proc to compound monthly returns of  each portfolio component.   (I presume there is no reweighting within portfolios over  the 60 month window).

Ans: You are right, there is no reweighting within portfolios over the 60 month window.

It's time to start a new question.    Don't move the goal posts within a topic.

Ans: I have posted a new question. Thanks for this.

Regards

[mkeintz]

@amanjot_42:

1. You only want companies with data for the complete range, right?
   
   
2. Show us some sample date, including date variable(s), company id vars,  and record order.

[amanjot_42]

Hi,

Thanks for your reply,

This is a sample of my data.

1. You only want companies with data for the complete range, right?

Yes, I want companies' data for complete range specified earlier. 

Regards

[amanjot_42]

Sorry,

file is here