Hi QC Community

In the 3rd session of this series, we'll be using Smart Insider's Corporate Buybacks dataset to get information on buyback events, both the intention to do so and actual transactions. Although this demonstration is using it in a daily context, the dataset has resolution up to second that could come in handy for arbitrager, dated back to 2015.

In this algorithm, our idea is that either the company is putting faith in its own future or trying to boost its security prices, buyback is usually observed with an increase in longer-term stock price. We tested the correlation between buyback intention/transaction and forward return to see if we could trade on this insight. We found that log price return and buyback transaction's ratio have some weak to moderate correlation. Thus, by fitting the buyback information into an exponential function and using the correlation coefficient on that, we could come up with an expected return with odds for holding a portfolio. Please find and follow the attached notebook and then its backtest for a more detailed explanation. Enjoy!

Louis Szeto