The fact that it's a tiny part of what pharma companies do is why there is a disconnect. There is something like 70 billion in grants alone (so not even including salaries and stipends a lot of the time) going towards biomedical fundamental research, and it's disconnected from the price signals because pharma companies will never be the ones doing it. It's still a necessary part of the process.

So for the market it looks like 99% of the cost comes after academic research but in reality that's only 40-60% of it.

At the end of the day the price signal doesn't even reach most of the ressources and it is itself a poor proxy for the metric we actually want to optimize, which would be something like QALY.