>the price is reflected in the first transfer of the good, not in any final consumer sale of the unit. Right, so this is trivially exploitable, as I described.

>the price is reflected in the first transfer of the good, not in any final consumer sale of the unit. Right, so this is trivially exploitable, as I described. Set up two entities. One of them makes the capacitor and "sells" them at cost, or lower, to the second one. The license fee will now be based on that deflated price. Arguably it's fraud, but proving that - and the extent to which it's out of the ordinary, since manufacturers do sell each other things at discounted bulk rates all the time - would be a nightmare. Even if you close that loophole (e.g. with perfect fraud enforcement (ha)), you're incentivizing inefficiencies by punishing vertical integration. What if it's cheaper (less labor, less waste, less machinery, easier logistics, ...) to produce a whole circuit as one unit rather than building the capacitors first and then assembling? If the license price depends on the first transfer point, then the circuit manufacturer is better off wasting money and resources to isolate the patent-dependent components so that they can be transferred independently, with the licence fee coming out of that small price. Otherwise, they'd have to pay a fraction of the entire final product, which is presumably much greater.