1. The network capacity is half a million transactions per day no matter how high the fees are. Fees don't change capacity.
2. If people are forced onto layers, then the fees that go to securing the network will go to the layers also, depriving the miners of needed compensation when the blockreward gets halved. If the 1 MB limit is never raised, then the network will eventually be funded by those same half million people even if billions are using it in layers and side chains. That will present a security problem.
extrapolating out hashing difficulty, the mining network in a couple of years will be several gigawatts, all funded by fees because the block reward will be next to nothing. That means that fees will have to reflect a cost of hundreds of dollars per transaction if we keep the 1 MB limit!
1. Economics is about dealing with scarce resources, not free shit, especially of the self-defeating kind.
2. Therefore the upper-layers managers will do exactly what you ask: paying higher btc fees thanks to the higher volume that they manage -not work at a loss and compensating it with volume :-D
Economics is about measuring efficiency. You think having an artificial scarcity makes something more valuable, but you can't raise the value of CocaCola by reducing the amount you bottle. People will just buy Pepsi. Some very loyal customers may be willing to pay more, but there is a limit to how much more you can charge.
You can't make up a loss on volume, but you can make up a smaller profit. Bitcoin fees are low now because it is an introductory offer, intended to attract new customers. We need several orders of magnitude more customers before we can raise fees.