How closely do you model your biotech names
Working in a small generalist team where I am also in charge of healthcare investments. I have in the past tried to model things very granularly but my team has no appreciation for this, which is fine. Overtime I have therefore reduced my process to something like this for pharma/biotech names:
- Review their pipeline if they have a dedicated deck (which is the case 95% of the time), otherwise see if anyone in sellside has a current overview
- Find peak sales for each asset and determine the most important ones (usually only looking at phase 2b and beyond). Sometimes the company will have calculated this. If not available, I will try to see whether there are comparable drugs or otherwise try to approximate patients in the US, guess the price and then 2x that value for global sales. It is all pretty rudimentary, I am not here trying to split hairs. General idea being that if being off by 15-20% kills the investment thesis (specifically the upside) then I was not interested anyway to begin with.
- At this point I will round up the assets that I am most interested in after having reviewed the clinical data. Time for the model which becomes even more rudimentarily. E.g. ok this asset has peak sales of $5bn. Maybe it is phase 2 going into phase 3. Perhaps chance of success is 30% and then FDA approval is 80%, with commercial execution at 75% (this is not binary, essentially me saying the remaining 25% is released as the drug starts selling and eventually reaches peak sales). So my risk adjusted value is $0.9bn. Looking around it seems the market only has this at $0.6bn, but anyway this is kinda irrelevant to me. The point here is to get past phase 3 and erase the associated 70% discount. So then my Target Sales would be X + $0.9bn/0.3 = X + $ 3bn. I expect this to trade at 4x EV/Sales. So then I can calculate target EV and from there I can get the target share price.
So in short I don't even model out the revenue curve. Sure I have some assumptions of my own more or less in terms of what figures would be good enough, I typically will have a number written down in preparation for quarterly results, but I rarely if ever model launches beyond the quarter in question, seems so futile. Really and truly I mostly read sellside reports and twitter to gauge whether the market is happy or not and then if it is not happy I will do some thinking of my own to decide whether it can ramp up to a place where the market will be happy and if not then consider selling it (case in point being UCB, where I expected a certain figure for the previous quarter). Crucially I work as a stockpicker so our strategy allows us to be thoughtful and we rarely try to dump a position immediately.
FYI I do check the basic stuff like warrants and thus dilution, revenue split with partners, etc.
Anyway I never received any formal training on the buyside regarding biotech, though I have studied this stuff. And sure I can do pharmacokinetic and pharmacodynamic modelling (though again I keep this very rudimentary) to get an idea of whether the clinical data is decent enough. But it just feels like I am just stumbling my way through biotech as my work is extremely undetailed and primarily focused on the big picture. How bad is my process?
My PM, who used to work with a more detailed biotech guy maybe a decade ago when he was an analyst, jokes about how futile his models were. That the analyst would change a simple input and the stock would crater by 40% etc. Point being that he does not put much weight into the model and is more focused on me telling him what the upside is and the likelihood of success. That's it really.
Not to show off but for the most part my process has worked. Obviously I don't get them all right. What surprises me is that it does work which makes me wonder whether I am just getting lucky, because my process for non pharma/biotech names is completely different, with models that are easily over 1k lines. whilst you would be lucky to get more than a 100 lines from any of my biotech models, let alone having more than 20 columns.
It just makes me feel like a fraud. And yet here I am in charge of our healthcare investments. Life is so weird...
It’s a generalist fund. Your efforts in biotech seem as like “better than nothing”. Also the PM seems to have more of a betting approach for these names like what is upside and chance of getting there. If upside is $100 a share and it’s 50% of getting there, then fair value of that bet is $50. If stock is at $40 he would buy in theory.
Yes compared to dedicated biotech analysts you are probably much weaker and less in depth than modeling but seems like good enough for your current seat. Generally the better modeling you can do the better but there’s a trade off and figuring out the right balance and cost/benefit for you.
GL
Thanks, very helpful!
I think a lot of dedicated biotech analysts have valuation processes just as simple as yours. Modeling is not generally going to be a source of alpha in biotech and that discourages a lot of people from spending time on it.
However, the benefit of getting more granular is it forces you as as the investor to really think about what clinical data scenarios support which commercial outcomes. For many biotech catalysts, simply succeeding in clinical trials is not enough. If you're haphazardly crediting achievement of peak sales regardless of clinical outcome (as long as drug doesn't fail), you're understanding of the risk/reward of the bet is flawed. This is a good way to lose a lot of money even if the drug works.
Can second that I have experience seeing a great stock in the clinic fall and ultimately delay and put on hold it’s strong clinical expertise because it couldn’t figure out a good commercial model for the drug (specifically producing a particular treatment at scale and fast enough) to justify investing in pushing the drug through to FDA approval and having a patent cliff and exclusivity period wait out
Yes indeed, that is why I usually do not remain invested for the commercial portion unless it is big pharma as they usually tend to figure out how to sell their drugs, assuming they are not peddling a dud. I have lost enough money thinking a drug would sell smoothly only to eventually understand how thorny this path often is and that it is a different skillset. I personally found it to be a nightmare trying to figure out payor coverage, physician education etc. so I usually stay far away...
Thanks for the insight!
just model whatever management spoons feeds us.
Curious how confident you feel in your assigned probabilities as a generalist? Has never been part of my coverage, but biotech seems v hard to predict as a generalist (beyond just doing strawpolls on calls)
Just to clarify, I do have a life-science background but no PhD. My graduate studies covered everything that you can expect between thinking of a lab experiment and eventual FDA approval. So prior to starting I was well acquainted with this branch of science. Anyway I have been on the job for 5 years. When you do this long enough, and follow the clinical outcomes, you start to develop a sense for it. E.g. this team is a 6/10, the dataset is 8/10, prior assets (whether failed or successful) say this field or MOA is 4/10, etc. So, assuming you understand the science, I think over time this becomes more about investing in the situations where you think the probabilities are stacked in your favour. In terms of how I feel, I am always crossing my fingers in preparation for a catalyst haha. That said, I typically only invest in situations where I feel >80% chance of things going my way. I think that, if your stock is not being hotly debated on Twitter, generally speaking, then you are in a good position already as far as clinical catalysts go.
Nesciunt nostrum quia eaque unde et voluptates aut natus. Quis fugit aut aut eaque doloribus nam qui. Cumque magni est esse omnis odio dolor.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...