A while back a friend asked what I think of Novo Nordisk. I said I didn’t know. Last edition I told you I was building a workflow to value pharma companies without becoming a pharma expert.

I figured I’d have to learn some of the industry anyway: the drugs, the trials, the patents. That’s true. But I’m also learning a lot about how the people who price these companies think.

Two lessons so far.

Lesson 1: This model likely won’t give me a huge edge

How do you know when a model is right? One way is to reference some reliable benchmark. Market price and analyst consensus come to mind.

However, if I’m tweaking my model to generally align with analyst consensus, then what is it really telling me?

Not much when it agrees. If the model lands near the market, all I’ve learned is that I built a slower way to get the same answer. The value is in the deviations. My expectation is that the model will agree with these benchmarks most of the time, and I’ll have to watch for the exceptions. Then I do a deep dive on why each one exists. Lilly is the best example so far.

Lesson 2: When the model disagrees with the market, ask what you’d have to believe

Two of my five companies landed near their prices. Lilly didn’t. The stock trades at about $1,273, and my model says roughly half that.

So I asked a simple question: what would I have to believe for Lilly to be worth its price?

Lilly spends about $15 billion a year on R&D. In my model, I assume each R&D dollar creates about $1.25 of value, which is in line with what the industry has historically done. For Lilly to be worth $1,273 a share, each R&D dollar would have to create about $5.

Novo is the opposite case. Its price requires about $1.19 of value per R&D dollar, almost exactly what I assumed.

I can’t judge whether Lilly’s next drug will work. But I can ask whether I believe Lilly’s R&D is about four times more productive than the industry’s. That’s a question I can reason about, and it tells me what I’d be signing up for if I bought the stock.

Two cautions. First, the market has a fair argument: today’s biggest drugmakers are the ones whose R&D worked, so they’ve earned some premium over the industry average. Second, the $5 also absorbs things my model misses, like a thin pipeline count and no credit for acquisitions. So treat it as a ceiling, not a verdict.

Back to Novo

On today’s products and pipeline, my model puts Novo roughly in line with its price, with a wide range around it. That’s not a call. Novo reports November 4, and my numbers are based on a June balance sheet, so I can give a more accurate update next week.

Next up: more model iterations, and running it on companies I didn’t build it around, like Pfizer and Bristol, to see whether it holds up or just reflects my assumptions. 😁