The scripts in this repository aim to generate a positive alpha by proposing an active portfolio strategy. For a thorough introduction see the Wiki tab.
- All codes and analyses are subject to error.
- All investments are subject to risk.
- Past performance may not reflective of future gain/losses.
- Provision of the codes and portfolios is not a financial advice.
For enquiries and commercial use, please get in touch via [email protected]
The codes are built on a class structure to simplify backtesting. In order to replicate the backtest results you need to run backtest_ik.py. You can fine-tune the specification in the script.
– The weekly stock data and indexes are from Refinitiv Eikon. You need an APP KEY to access Eikon's API.
– The annual fundamentals are from merged Compustat/CRSP file on WRDS.
– The risk-free rate and Fama-French factors are also from WRDS.
I) The contents in the R script "knockoffs_matlab" are from @msesia 's GitHub repository providing MATLAB and R codes for the paper entitled "Controlling the False Discovery Rate via Knockoffs” by Barber and Candès in Annals of Statistics (2015). Warning: Running the scripts may ask for admin priviliges. You do not have to provide that access for running the codes.
To access the data you must have an active subscription with WRDS see (https://wrds-web.wharton.upenn.edu/wrds/). Specifically, you need active subscriptions to CRSP and Compustat schemas on WRDS. To acquire the data you need a functioning JAR driver for MATLAB to access WRDS through Matlab. See https://wrds-www.wharton.upenn.edu/pages/support/programming-wrds/programming-matlab/matlab-from-your-computer/