Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Help needed to improve the "betting against beta" strategy!

According to the following paper, this strategy sgould yield positive returns during the day. It performs quiter well in early years, but then makes a lot of losses in the following years. Any help to improve the strategy is appreciated.
https://www.rhsmith.umd.edu/files/Documents/Departments/Finance/2017/bogousslavsky.pdf

Description of the strategy:
- Market beta for each stock is estimated using daily returns over the past year.
- The market return is the value-weighted return of all stocks in the sample excluding stocks with a price below $5 and is rebalanced once a month.
- The strategy shorts stocks with high beta.

1 response

My hunch is that betting against beta works when the market isn't highly correlated. When stocks are all moving in lock step (higher correlations) AND the market is going up then betting a stock will go down is probably a poor bet.

It turns out that between 2004-2008 stocks weren't very correlated with one another. However, starting in 2009 and then all the way to 2016 the correlations went up. The returns seem to track this pretty well. 2004-2008 had good returns while after that not so good. Looks like starting in 2016 the algorithm performance improved about the same time market correlations began to drop.

One way to maybe improve the algorithm is to check overall market correlation. When low then bet against beta. However, when high, then flip the strategy and bet with beta. Just a thought.

See this post for a good analysis of market correlation https://www.quantopian.com/posts/correlation-between-the-top-us-equities-over-time .

Good luck.