Date:         Wed, 23 Apr 1997 17:55:28 -0500
Reply-To:     "Robin K. Chou" <rchou@BLUE.WEEG.UIOWA.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         "Robin K. Chou" <rchou@BLUE.WEEG.UIOWA.EDU>
Organization: The University of Iowa
Subject:      Help: another time series question
Content-Type: TEXT/PLAIN; charset=US-ASCII

Hi there,
 
Thanks for all of you that helped me with my question yesterday. Since I
am totally new to SAS, I have yet another one to bother you.
 
I have a huge amount of securities data. For each stock, I have several
variables and each variable is a time series. For example , price series,
return series, volume series ...
It looks like;
 
name    date    price   return  volume
 
IBM     01/94   90      xxx     xxx
IBM     02/94   92      xxx     xxx
IBM     03/94   95      xxx     xxx
IBM     04/94   92      xxx     xxx
IBM     05/94   97      xxx     xxx
IBM     06/94   95      xxx     xxx
IBM     07/94   90      xxx     xxx
.
.
IBM     11/94   98      xxx     xxx
IBM     12/94   95      xxx     xxx
MSFT    01/94   120     xxx     xxx
MSFT    02/94   118     xxx     xxx
MSFT    03/94   115     xxx     xxx
.
.
MSFT    11/94   124     xxx     xxx
3M      01/94   58      xxx     xxx
3M      02/94   59      xxx     xxx
.
.
 
Now, I would like to compute the mean of stock prices by month and call
this "average market price" for month n. I know that we can do this easily
by using "proc means" with "by" statement. But my question is that I want
to use these market prices as independent variables and do a time series
regression for each stock using the stock's prices as dependent variables.
 
That is, I want to create a data set that will look like:
 
name    date    price   return  volume  avg_mkt
 
IBM     01/94   90      xxx     xxx     89.3
IBM     02/94   92      xxx     xxx     89.6
IBM     03/94   95      xxx     xxx     xxx
IBM     04/94   92      xxx     xxx     xxx
IBM     05/94   97      xxx     xxx     xxx
IBM     06/94   95      xxx     xxx     xxx
IBM     07/94   90      xxx     xxx     xxx
.
.
IBM     11/94   98      xxx     xxx     xxx
IBM     12/94   95      xxx     xxx     xxx
MSFT    01/94   120     xxx     xxx     89.3
MSFT    02/94   118     xxx     xxx     89.6
MSFT    03/94   115     xxx     xxx     xxx
.
.
MSFT    11/94   124     xxx     xxx     xxx
3M      01/94   58      xxx     xxx     89.3
3M      02/94   59      xxx     xxx     89.6
.
.
 
Thus, I can regress price on avg_mkt for each stock.
Any suggestion will be greatly appreciated. Thanks again.
 
Robin Chou
Department of Finance
University of Iowa
Iowa City, IA 52242