Hi, I'm a novice in SAS. I have downloaded monthly stock data for AMEX from CRSP. The queries I selected are PERMNO, DATE, EXCHCD, PRC, RET, SHROUT and EWRETD. I want to perform regression analysis of a single stock return on market return using Model statement. I just don't know which variable out of these is the market return. Is it EWRETD or the average of EWRETD of all stocks in AMEX? or is there another formula? Please help. Thanks
When it comes to what variables in a data set mean you need to talk to the source of the data. Perhaps where you downloaded the data they had something labeled with a name similar to "data dictionary" that you should also download that has the details.
Topics like how to use specific sources of data, such as stock market data, may not get answers unless someone with interest or understanding of the topic comes along (not me, finance gives me headaches because so many o f the "rules" have so many exceptions).
The normal Capital Asset Pricing Model requires (1) return of the asset in question (i.e. return for the given PERMNO, (2) risk-free return (typically government bonds), and (3) the general market return (usually S&P500).
You have item 1, in the variable RET (make sure it is returns with dividends)
I don't see a variable name that suggest risk-free returns. You need to examine the literature to see what is commonly used. CRSP may provide that data as well, although you might get away with choosing a mutual fund traded on the market that has a consistent mix of government bonds as its portfolio. If that is satisfactory, you could download monthly returns for that mutual fund.
As for market returns, you apparently want the return for all of AMEX. Instead of making your own, see if you can download an AMEX index. A quick web search for "AME composite index" suggests the "ticker" XAX might do. If you have that, and confirm that it serves your needs, then you can use its return for the third factor above.
By the way, any CRSP variable whose name begins with EW usually means "equal weighted". That may not be what you want. And of course, you have to find out what are the constituents of this Equal-Weighted measure. And another thing - usually CRSP vars that end in RETD usually mean daily returns, not monthly.
In any case, you will have to match the market return and the risk-free return with all of your stocks, by date before you can run your model. And of course, the proc reg would have to be run BY PERMNO to establish the alpha and beta for each stock covering the period of interest. I'd suggest just doing 4 or 5 stocks to make sure everything is working properly before you produce results for your entire data sample. And you should also make sure you are using the same time range for all your PERMNO's to make useful comparisons.
Join us for SAS Innovate 2025, our biggest and most exciting global event of the year, in Orlando, FL, from May 6-9. Sign up by March 14 for just $795.
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.
Ready to level-up your skills? Choose your own adventure.