The tax CAN NOT be reduced by halving, because the PoW network NEEDS to maintain 10% expense in order to not be laughably easy to 51% attack. Therefore it will reach a new equilibrium that rests at 10%, and continue to suck wealth out of the Bitcoin eco-system at that rate.
You lost me here. The halving will happen regardless of conditions. Over time, fees are supposed to replace the block subsidy.
Hash-power follows price. If Bitcoin is laughably easy to attack now: that is due to poor adoption. That implies that Bitcoin is not important enough to attack by well-funded actors. The hope is that by the time Bitcoin becomes a clear threat to well-funded actors, it will not be so cheap to attack.
Of course, in that case, a well-funded adversary can use the "hack every node" strategy I outlined for PoS coins. In a PoW system, it would be easier to recover from such a scenario: since the block-chain would still be largely trusted due to the proof-of-work it contains.
As I said in
this post:
Bitcoin is a revolutionary experiment. It is the "first secure networked application ever created in the history of computers." (
Out-of-context Bruce Schneier quote) Governments and banks will attack it through legal and technical means. It may only have a small niche over the next 50 years. If that happens, it will be OK: because it is just an experiment.
Converting Bitcoin to PoS would invalidate the experimental results. I really am curious if Bitcoin will turn out to be the first secure networked application in the history of man-kind. The main thing that makes me doubtful is that today's computer systems are inherently insecure. Formal correctness proofs are generally limited to military applications where you want to be able to deploy special weapon modes in war-time without prior testing. Even safety-critical systems like automotives still use Ad-hoc debugging. This allows bugs to creep in.