R Factor Analysis with BAB and QMJ factors (Table 4) in the "Buffett's Alpha" paper

1. Edit frazinni.r, choose the csv file to use.
2. ffdata_brk.csv: BRK.A stock prices from 1980 to 2013
3. ffdata_brk13f.csv: Berkshire Hathaway 13F-HR portfolio's simulation return from 1999 to 2013
4. Choose which factor analysis to display - reg_ff1, reg_ff2, ref_ff3.
5. Run it by Rscript frazinni.r

BRK.A has alpha of 12% and beta of 63%

We attempt to reproduce the loadings shown in Table 4.

Here we see that the loadings are strong on rmrf, smb, hml, but not on umd.

Here we see that BAB has strong loading, and it weakens both hml and umd. Alpha is reduced from 12% to 8%, indicating BAB has explanatory power.

Here we see that QMJ has strong loading, and it weakens smb and umd. Alpha is reduced from 8% to 3%, indicating BAB and QMJ has explanatory power.

We don't have enough data (since 1980) to reproduce the 13F-HR analysis in the paper -

In the short history we have (since 1999), we can observe loading on QMJ, but not on BAB.

Buffett's Alpha

https://www.econ.yale.edu/~af227/pdf/Buffett%27s%20Alpha%20-%20Frazzini,%20Kabiller%20and%20Pedersen.pdf

Factor Analysis in R

https://www.calculatinginvestor.com/2011/04/19/fama-french-tutorial/