The GSK-Pfizer merger did cut costs. A rival put prices up 8.55 percent anyway.

Start with the pass-through, because it is the one quantity here that can be computed rather than asserted. Pfizer's estimated cost of supply in the category fell 9.43 percent, and the prices of those same products fell 6.57 percent ^[3][4]. The ratio of the two is about 0.70, so roughly 70 percent of the measured saving reached the till and about 30 percent stayed inside the firm ^[14].

The awkward part is that the products getting cheaper and the products getting dearer sat inside the same combined business. GSK's prices rose an estimated 3.25 percent ^[6] while its new stablemate's fell ^[4]. Sanofi, which was not a party to the transaction, went up 8.55 percent ^[5], about 2.6 times GSK's move ^[16] and roughly two percentage points larger in magnitude than Pfizer's decline ^[15]. Of the four suppliers the study names, exactly one lowered prices ^[17]. Unilab, the cheaper local manufacturer, stayed broadly where it was ^[7].

That pattern is where the coordination finding comes from. The authors report evidence consistent with greater coordination between GSK/Pfizer and Sanofi after the deal, and Bokhari is explicit that this does not mean the firms agreed on prices: with fewer independent competitors, coordinating gets easier without any agreement, and prices settle above what competition would produce ^[9]. The finding names the branded suppliers and not Unilab ^[9][7], which is the shape you would expect if the cheap local line was never in the same pricing game to begin with.

Two notes on evidentiary weight. Ennis points out that confirming efficiencies for one of the merging firms is not widely studied ^[11], and that cuts both ways: the paper strengthens the case that efficiency claims can be literally true, then shows that being true is not the same as being sufficient. And the Philippine Competition Commission is one of the co-author institutions ^[13], so the recommendation to weigh coordination alongside efficiency ^[10] is not arriving from outside the agency world.

The £500 million a year the companies forecast ^[8] is a deal-level figure, while what the study measures is one product category in one country ^[2]. Nothing here tests that forecast. What it tests is whether a saving of that kind reaches shoppers, and even then the reported numbers are firm-level percentage changes with no volume weights attached, so the net effect on the average Philippine cough-and-cold buyer cannot be recovered from them ^[18]. The narrower claim is the one that is hard to argue with: the efficiency was real, and the unmerged rival's prices rose by more than the merged one's fell.