I was referring to start-ups that are not acquired. For example a VC may fund a range of biotech companies, but only recoup on those that are acquired. Equally there are many examples of drug-based biotech that simply fail for a range of reasons, losing all the money invested in them.

The point being that the general concept that acquired research may be more efficient compared to in-house research would have to account not only for the failed in-house research, but also for the failed research within companies that are not acquired, or which fail for other reasons.