Index | Recent Threads | Who's Online | Search

New Thread
This topic has been viewed 234 times and has 8 replies
Jaume
R&D question

I'm trying to implement a ranking that ranks the companies based in R&D and "price to R&D", but I'm not sure what is the best way.

I have these three options:

A- A composite with two formulas inside (50% weight each):

-RAndDA/MktCap (Higher better)

-RAndDTTM/MktCap (Higher better)

B- A composite with 4 factors inside (equal weighted):

PEInclRDQ
PEInclRDTTM
Pr2CFInclRDQ
Pr2CFInclRDTTM

C- A composite with one formula inside:

Price/IsNA( RandDTTM, RandDA ) - Lower is better. -


Which one do you think it's best? (if not all of them together)

Is there any other way?

Thanks.

Mar 30, 2021 11:34:43 AM       
SpacemanJones
Re: R&D question

fwiw, I use the "yield" version of the calcs - just makes more sense to me that way - just have to be sure to control for possible negatives in the denominator depending on what you may put in the denominator. My RD factor has about 13 variation inside. Might be overkill, but overkill seems to work for me on many factors. (can vary over time as well as vs universe, sector, industry sorts).

Mar 30, 2021 7:53:22 PM       
yuvaltaylor
Re: R&D question

I agree with Michael. IsNA(RandDTTM,RandDA)/MktCap is, I've found, a pretty good way to go.

Yuval Taylor
Product Manager, Portfolio123
invest(igations)
Any opinions or recommendations in this message are not opinions or recommendations of Portfolio123 Securities LLC.

Mar 30, 2021 8:47:02 PM       
Jaume
Re: R&D question

Great!! Thanks Yuval and Michael. :)

Apr 1, 2021 7:02:36 AM       
Jaume
Re: R&D question

Yuval:

I tested all those factors, or factors combination in the Ranking's Performance tab, and I set all of them as a one single factor's ranking system.

I can see the one that performs worst is the one you recommend me. I do not understand why.

Perhaps my approach, test each factor alone, not as a bunch of factors, it's not a proper way to test a ranking?

Attachment Captura de pantalla 2021-04-03 a las 16.01.07.png (222976 bytes) (Download count: 48)


Attachment Captura de pantalla 2021-04-03 a las 15.57.12.png (172452 bytes) (Download count: 48)


Attachment Captura de pantalla 2021-04-03 a las 16.07.40.png (153684 bytes) (Download count: 46)


Apr 3, 2021 9:19:26 AM       
yuvaltaylor
Re: R&D question

You have three very different things going on here. The first one uses price to earnings and price to free cash flow but includes R&D expenses in the earnings and free cash flow. The second one compares R&D to market cap to all stocks, so it's going to heavily favor companies in certain sectors (healthcare and technology). The third one compares R&D to market cap to other stocks in the same sector, so it's not going to favor stocks in certain sectors.

Also, the first one has comparatively few N/As, the second and third have a lot of N/As and treat them entirely differently. In the second, all the N/As will go at the bottom; in the third, N/As for sectors which have no R&D expenses at all (e.g. Financial) will get high ranks (since if 99% of stocks get N/A, they'll all be ranked 98) and N/As for sectors with some R&D will go at the bottom. So the performance chart is going to be rather difficult to make any sense of.

I'm not sure why you used universe ranking for two of these and sector ranking for the third. If you're going to use sector ranking, I suggest you exclude a bunch of sectors and subsectors from your universe before testing. This factor is mostly useful for healthcare and tech.

Yuval Taylor
Product Manager, Portfolio123
invest(igations)
Any opinions or recommendations in this message are not opinions or recommendations of Portfolio123 Securities LLC.

Apr 3, 2021 4:58:09 PM       
Edit 2 times, last edit by yuvaltaylor at Apr 3, 2021 5:12:08 PM
Jaume
Re: R&D question

I used sector ranking for the third factor just by mistake. :D

Thank you for clarifying this to me. :)

Apr 5, 2021 5:14:36 AM       
ahha
Re: R&D question

R&D should work only for high-tech industries, try to define your own universe of high-tech industries - this may be your secret sauce.
Then benchmark the strategy to your own universe.

Norbert

Apr 6, 2021 2:53:37 AM       
Edit 1 times, last edit by ahha at Apr 6, 2021 3:28:55 AM
Jaume
Re: R&D question

Yes. In fact just I use this factors in healthcare and software sectors.

Apr 6, 2021 7:07:33 AM