Let fly over a scenario inside of the Haircut:
Here are numbers, Hive = 27c$
100.000.000,00$ Hive MC
30.000.000,00$ HBD aka 30m HBD
This is not inside the haircut. This is on the edge. Right in the doorway. That infers you might be one of those still thinking the haircut kicks in when someone says: "Good job guys, we have printed the last HBD, take some rest now" in the printing room. In practice, haircut kicks in when the witness price feed goes down and it does not stop in the door. The haircut that came to your house is not hitting the wall and stopping there. It is rushing through inside fully intending to go upstairs right in your bed.
Now the
HBD APY andgeneral Inflation will be paid in HP and Hive from that point on, and the HBD APY in liquid Hive.
Assuming I FYPed correctly to what you wanted to say, the former is true (note that HBD printing was already limited there since 20% debt ratio) but the latter is not. Noone wants to pay your HBD Savings interest in Hive. The whole point of debt ratio is that we stop pretending we are allowed to print dollars at one point. You want HBD coin? You get interest in HBD coin! HBD is not worth $1? Your problem!
So people can reinvest 30% of that into HBD, due to the Haircut and 70% will have to stay liquid.
How? Pretending all inflation is paid in HIVE and everyone wants HBD, there is 2.5M new HIVE coins and zero new HBD coins so HBD/USD goes up (I am not saying over 1.00 as it surely was well below 1.00 before the haircut). Unfortunately, the spike can only be followed on inside-your-head exchanges ("all HBD locked" is not very useful premise, is it?) and there is no way you can reinvest even 10 HIVE into HBD. And you claim it is 30% (of 2.5M) and not 29% or 31% based on something you do not explain (just because debt ratio is the same figure).
Even if you allow some liquid HBD for trades to actually happen, what does "people will arbitrage" mean? They can buy existing HBD driving the price up but no HBD is created to keep the peg so it overpegs (that sounds silly, and it really is - and that's because in practice all three "complexity-check points" are the exact opposite of what you can expect). I see no way to finish the mind experiment from that point on.