Hi all,
The following snippet by JR had me thinking for a few days about the blocksize debate and how it relates to sidechains, while reading the extensive back and forth between sidechain supporters and opponents here. I believe I have an answer, but it overturns some of the conventional wisdom.
Bitcoin could never survive as a high-fee settlement currency because high fees would arise due to block size rationing, not because transactions should naturally cost that much for technical reasons. Bitcoin would lose out to a competing currency with less/no rationing that would be less expensive to use for settlements.
High transaction rates on the main chain are the only way for Bitcoin to survive. Yes, getting there is a difficult technical problem to solve. Deal with it.
First of all, in a future where trillions of dollars' worth are transacted on the network, a 51% attack has to cost on the order of billions at least for good security. When the block reward ceases to fund miners enough to make a 51% attack cost billions, total revenue from transaction fees will have to be around that high to maintain security. But, as JR implies, tx fees will tend to fall to the actual marginal cost of including them in a block, which makes it seem near impossible for the revenue to be enough. After all, if Bitcoin devs choose to artificially limit block sizes as a handout to miners in lieu of the block reward, another coin that charges lower fees will eventually take its place...
...but no, this isn't quite right. Because somehow or other, a PoW network where trillions of dollars are securely transacted has to have plenty of mining to thwart 51% attacks, and in order to do that there must be a hefty incentive for miners to keep mining. That means a competing coin can only steal Bitcoin's thunder with lower tx fees
if it retains high block rewards - and in the long run it must be like Dogecoin, retaining constant, non-decreasing block rewards, in order to continue funding miners enough to secure the trillion-dollar network.
In other words, besides the dynamics of who pays (just the transactors, or every holder), the block rewards and the tx fees (and hence the blocksize limits) are two sides of the same coin: they're both ways of paying the miners, with the aim of ensuring the miners get enough to keep the network secure.
Since it cannot be known in advance how popular Bitcoin will ultimately be, nor how big the world economy may grow over the next hundred years under Bitcoin, it doesn't makes sense to go the Dogecoin route of perpetual inflation. The block reward is deemed sacred by investors so should not be altered. That ultimately leaves only transaction fees to be adjusted in line with the volume of transactions vs. the cost of securing the network. Who will adjust the fees? Of course the market, finally, but who will put a floor on the fees to ensure miners are getting paid enough? Rationing, as it were, but by the market. Perhaps there is a way to tie it to market conditions directly, or perhaps devs will simply release different versions with different max_blocksizes and see what the economic majority goes for. Either way, it is being left to the market. It is not centrally planned rationing, but market rationing, whether directly or indirectly. It is not like government rationing food or apartments, any more than Bitcoin's 21M coin limit and block reward schedule is.
Therefore I believe blocksize limits will increase, but not without bound. Any coin that tries to increase the blocksize limits without bound, in an attempt to undercut Bitcoin's transaction fees, will have to do that by having - at that point in time - a larger block reward, punishing savers a little more and spenders a little less. There is not free lunch. In the end the same tax is being paid from the users of the network to the miners. Any chain that tries to undercut on both aspects will be undersecured, not to mention the network effect making this whole undertaking difficult in the first place. The network effect makes is so the market or economic majority can mess up its guestimates quite a bit and still retain Bitcoin's dominance.
Nevertheless, this means that if most of the transactions are happening on sidechains the Bitcoin network could become grossly undersecured after a few more halvenings. Which suggests that ultimately paying miners with transaction fees for a ledger network that is designed, above all else, to store the value of its ledger is a bad idea, since the transactions can be moved off the network. The transactions are not something that must happen on the network, as sidechains make clear. Paying miners with tx fees thus fails to tether the service miners perform to the payment they receive.
As uncomfortable a prospect as this is, perhaps endless inflation like Dogecoin really is the answer, maybe with hardforks every few decades deciding on new inflation schedules that the economic majority prefers. The 21M coin limit is sacred to me, yet I have to admit it is meaningless if you have to pay a hefty fee to do anything with the bitcoins you own. What difference does it really make, as a hodler, if you lose 1% to inflation over 21M coins vs. losing 1% to transaction fees when you spend? In the end what matters is that there is no central authority dictating the inflation rate for political whim, and hence it will be very low, just as tx fees will be very low. Ultimately that rate, and hence the "tax" going to the miners (whether by inflation or tx fees), is determined by the market.
TL;DR: Sidechains (if they succeed) force Bitcoin's hand, although decades in the future, in setting an inflation rate determined by the market, in order to tie miner's subsidies more exactly to the primary service they perform: that of securing Bitcoin's store of value. This is because successful sidechains would allow Bitcoin transactions to be disconnected from Bitcoin miners. Bitcoin therefore could not rely on transaction fees for its network security incentives. Sidechains would have made it clear that transaction functionality is not intrinsic to or inseparable from the Bitcoin network; only the store of value function is, and therefore miners must be continually rewarded for maintaining this function alone, for the lifetime the network.