Who earns more money from generic drug sales: PBMs or drug manufacturers?

From a manufacturer perspective, is this surprising? Generics are the biopharma products most like "normal" ones, and for "normal" products I tend to think that if a manufacturer captures 25% of the retail price they're actually doing pretty well. For example, as a rule of thumb, I think retailers of "normal" goods often aim for a 50% gross margin when setting the retail price. (Net margin is vastly narrower, but that's be b/c retailing is actually a very costly activity!) Amazon's in hot water right now with the FTC for taking ~30% on average (for "fulfilled by margin") which to me sounds like a great deal compared to a traditional bricks-and-mortar supply chain. So compared to "normal" products, 30% for a generics maker is actually pretty good IMHO! What am I missing? No defense for the PBMs though. In fairness, it's undoubtedly a vastly more complicated payment system than "normal" products—bizarrely complex in fact. But surely that cannot explain why this category of activity—which doesn't exist for "normal" products—takes so much of the pie. Seems fishy.

Thanks for reviewing Jason! It’s important for everyone to read in the context that this was our best estimate with publicly available data, but individual benefit designs could vary substantially. PBM spread pricing or differential pricing or risk mitigation pricing (yes it has many names) may actually benefit some employers.

The paper focuses on gross profit. I don't have any intuition on the cost structures and inefficiencies. The paper only mentions lack of PBM competition at the end. Is there a reasonable explanation that yields similar net profits? Perhaps differential frictions across the various industries? Edit: perhaps more succinctly, the possibility of accounting driving economics?