Question: You are an alpha hunter! Your job is to identify assets with non-zero alphas, and to recommend appropriate trading strategies. You are aware of CAPM limitations, and to account for them, you consider Fama-French three-factor model the "better" framework. In the spread sheet posted on Blackboard ("Group assignment 3.xlsx") you will find monthly returns on ten "well-diversified" portfolios. These portfolios are identified by uppercase letters, i.e., A, B, etc. Note that these are returns, not risk premia, and that both CAPM and APT assess risk premia (excess returns). On "Sheet 2" of the same document, you can find the risk-free rate (rf), market risk premium (rm rf), SMB, and HML. In your empirical analysis, you stick with the traditional 5% significance level (type I error).

(a) Estimate CAPM for each portfolio, and determine whether or not, you find any non-zero alpha that is significantly different from zero (you need to estimate ten regressions).

(b) Are the significant and non-zero alpha(s), significantly positive/negative? [Hint: critical values used for one-sided and two-sided hypothesis tests differ.]

(c) Repeat the same steps using the Fama-French three-factor model. Which findings of yours are different?

(d) Do your findings possibly highlight empirical shortcomings of CAPM?

(e) Comment on implications of your findings for market efficiency.

(f) what trading strategy will you recommend based your analysis?

Note: Do not include data in the report you turn in. Please summarize your findings for CAPM and APT (coefficients/t-statistics/p-values) in separate tables, and refer to tables when explaining results.