BU.232.620 Empirical Project 1 (Linear Econometrics for Finance)
In this series of (two) projects, you will analyze portfolios that are building blocks of most quantitative investment strategies in U.S. common stocks.
The following is what you are expected to do in the project
• Read the following seminal article (you can download it from our Canvas course site).
Fama, Eugene F. French, Kenneth R., 1993. “Common risk factors in the returns on stocks and bonds"Journal of Financial Economics, Elsevier, vol. 33(1), pages 3-56.
You are expected to understand and explain how the 25 stock portfolios are formed on size and book-to-market equity.
• Download the monthly return series of the 25 stock portfolios, as well as the market excess return factor, from Ken French’s Data Library
https://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html
Note: you are expected to be able to find the right data; this is one of the most important skills in practice!
• Before diving into the regressions, you want to first summarize returns of the 25 stock portfolios
– For the market excess return, refer to the “Explanatory returns” panel of Table 2 in Fama and French (1993) (page 14)
– For the 25 stock portfolios, refer to the last panel “Book-to-market equity (BE/ME) quintiles of Table 2 in Fama and French (1993) (page 15)
– Some questions you want to answer: Are the means significantly different from zero?
Do these returns follow a normal distribution? Hint: a normal random variable is symmetric (skewness=0), and its kurtosis is equal to 3.
• Now you want to run a simple linear regression of each of the 25 return series on the market excess return to investigate how each portfolio depends on market portfolio. In Fama and French (1993), this regression is written as
R (t) − RF(t) = a + b[RM(t) − RF(t)] + e(t)
– Refer to the panel “Book-to-market equity (BE/ME) quintiles” of Table 4 in Fama and French (1993) (page 20)
– Some questions you want to answer: Are the loadings of the 25 portfolios on the market portfolio significantly different from zero, e.g., at the 5% significance level? If so, are the loadings positive or negative? How much of the time series variation in the return of each portfolio is accounted for by the time series variation in the market excess return?
Now that you have done the reading and analyses, write a project report, which consists of the following three sections in general (basically, a short version of Fama and French (1993)):
1. The first section is Introduction, which should contain a high-level summary of the re- search question, model, and findings.
2. The second section should describe the data, including the data source and summary (the first table can be put here).
3. The third section should explain the model used and the findings (the second table can be put here).
4. The last section is Conclusion; you can add your own thoughts on how to improve the model, e.g., educated guess on adding explanatory variables/factors on top of the market excess return in explaining the 25 portfolios.
What You Need to Submit
• The project report
• The code you use in generating all the empirical results in the report
– The TA Regression Demonstration Session will teach you the key Python procedures in conducting the data analyses, but you are expected to write your own code!
– Some notes that discuss the main Python procedures are be provided; see Canvas announcements.
– I will check the code to make sure no one simply copies the code from other people (the code will look somewhat different if one really writes his/her own code)!