1. After you merged both datasets as described in the Instructions: Compute the market capitalization as [abs(PRC) * SHROUT] for each stock in each month. Why are some prices negative? (You might need to lookup the definition of PRC in WRDS for example with aninternetsearch to answer this question.)
Why can closing pricesat time t be negative? [select every answer that you think is correct]:
10 points
Question 2
1. For each stock compute the average excess return and the standard deviation over the period from January 2000 to December 2010 (to calculate such summary statistics the PivotTable is very useful). Which stock has the highest averageexcessreturn? (Reminder: excess return is the difference between the return and the risk-free rate)
10 points
Question 3
1. Now compute the Sharpe ratio for each stock over this period. Which stock looks most attractive to you in terms of the tradeoff between return and risk over this period?
10 points
Question 4
1. Do you observe any pattern regarding the Sharpe ratio of a stock and its average market cap (computed per PERMNO over the whole sample period)? Run a simple cross-sectional regression to check.
10 points
Question 5
1. Now run a formal CAPM (market model) time-series regression for each of the 100 firms (you can use the LINEST function in EXCEL to do so). Estimate the market model over the entire sample period. Which stock has the largest beta estimate? Insert the PERMNO of the stock with the largest Beta coefficient below.
10 points
Question 6
1. Make a scatter plot with on the vertical axis the historical averageexcess returnof all stocks and on the horizontal axis the stocks beta estimates over the full sample period. Also add a linear trend line which is the security market line (SML). Does this plot support the CAPM predictions?
Indicate every correct statement from the list below.
10 points
Question 7
1. Please indicate theCoefficient of determination (R2) from the regression in Question 6 (i.e. the estimated SML) that is closest to what you find:
10 points
Question 8
1. Now run a cross-sectional regression. Explain thewholesampleaverage excess returns of each stock by thewholesampleaverage MARKETCAPof each stockand the estimated beta (from Question 5). What do you find?Assume statistical significance is indicated by at-statistic below -2 or above 2.Select every answer that corresponds with your findings.
10 points
Question 9
1. Now run a so called Fama-MacBeth regression (2nd step) that is: each month regress excess returns on MarketCap only then compute the average coefficient on Market Cap as well as the Fama-MacBeth t-statistic which is avg(X)/[stddev(X)/sqrt(T)] where: X is the monthly estimated slope coefficient when explaining Returns by MarketCap and T is the number of observations (the number of months in the sample).
10 points
Question 10
1. As before run a so called Fama-MacBeth regression and compute the Fama-MacBeth t-statistic. But this time explain Returns byMarketCap as of January for each year i.e. in February to December you use the market cap estimated in January of each year for each stock.
10 points
Question 11
1. Of course smaller stocks are also associated with higher risk. Hence redo the Fama-MacBeth regressions use the MarketCap as of January for each year and the CAPM-beta (estimated over the whole sample) as a control variable to explain monthly returns (as before).
10 points
Question 12
1. Are your results so far consistent with the data you received from Ken Frenchs website (using data from Jan-2000 to Dec-2010)?
10 points
Question 13
1.
Report the Fama-MacBeth test statistic i.e.sqrt(N)*avg(X)/stddev(X)where N is the number of observations (the number of months) and X is the monthly estimated slope coefficient on MarketCap when explaining Returns by MarketCap and CAPM-Beta (i.e. the slope coefficients from the previous regression).
Round the value to two decimal digits and use the dot to separate decimal from non-decimal digits i.e. enter like:
12.23
Use all slope coefficients from 2005 (i.e. N=12).