Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Research and Development Factor Models

Hi All,

I was wondering if anyone here has any insights in this space? I started modyfying an older notebook on ranking universes by factors. I decided to investigate a classic factor on increases (decreases) in research and development as a percentage of sales. There are a few papers I've come across and am trying to implement. The paper below highlights using revenue growth and a firms past performance in relation to RnD percentage as an indicator of future performance.

http://www.people.hbs.edu/lcohen/pdffiles/dimalco.pdf

Please take a look at the notebook attached. I first filter a universe by technology equities using the Morningstar classification symbols. I then extract RnD expenses and total revenue and calculate a ratio. I also extract revenue growth and created a new factor that takes the average or RnD to Total Revenue and Revenue Growth.

Interestingly enough RnD as a percentage of sales does not have a positive correlation in 30 day equity returns, but when I include revenue growth as another factor, there is a positive correlation.

Interested in anyones thoughts and critiques.

Thank you,

Matt