CAPM and Factor Models

Part I: CAPM

(Mandatory for all groups)

1. Request a password for WRDS (Wharton Research Data Service) database.

2. In WRDS, find Compustat database. There you can find information on the book value of equity. Also, find information on the market value of equity and market capitalization.

3. Based on the above information, select two companies that you will work with throughout this project. The first company should be a large cap stock with high book to market (BM) ratio. The second company should be a small cap stock with low BM ratio. You can check how Prof. Kenneth French constructs these ratios. Look at the explanations from the data library on his website: …show more content…
Note that the use of a stock portfolio as a proxy for M leaves out such risky assets as real estate, most foreign stocks, and risky debt securities. Note also, that the Value Line Index is not value weighted but equally weighted. It is thus not an appropriate proxy for M.

3. For some carefully chosen past time period – say Jan1, 2000 – Dec. 31, 2004 – we construct the following data set:

(assumes time period is one month)

Jan 1, 2000-Jan 31 2000, Dec 1, 2004-Dec 31, 2004

RM RM1/1/00-1/31/00 ,….., RM12/1/04-12/31/04

Rj Rj1/1/00-1/31/00 ,….., Rj12/1/04-12/31/04

where, for example Dividends, if any, paid in June 2000 Total dividends paid in June 2000 to all S&P500 stocks

Market Value of S&P500 portfolio

So we would have 512 = 60 “observations”. The choice of the historical period is very important. Not only should it be long enough so that our period estimates have statistical significance, but it should also be a time in which the economic outlook of the firm whose stock we are studying is similar to its current (at the time of the estimation) outlook (as best as we are able to judge).

4. The next step is to perform the following regression:

(R1): , where the are chosen to minimize deviations from observed values. From this equation we know that

measures systematic risk of stock j, and measures its unique or diversifiable risk relative to our choice of market proxy.

Slope