Having reviewed the sourcecode, I want to clarify a crossed wire that seems to be running through this discussion.

Igor's PR modifies transfer_payments() — the function that sends HBD from treasury to proposal receivers. It takes the hbd_print_rate signal and applies it to the payout amount itself:

period_pay.amount = period_pay.amount × hbd_print_rate / 100%

It does not touch the inflation funding side at all. Inflation still creates HBD and allocates it to the treasury at whatever rate the existing hbd_print_rate logic dictates. The change is purely on the payout side.

This matters because of how the protocol treats treasury HBD. Since HF24, the code explicitly subtracts treasury HBD from the debt ratio calculation (line 2855 of database.cpp: hbd_supply -= get_treasury().get_hbd_balance()). Treasury HBD is quarantined - it doesn't affect the debt ratio, doesn't create sell pressure and doesn't impact market cap. The payout is the event where inert HBD becomes liquid, sellable and debt-ratio-relevant.

So Igor's position, as I read it, is: the hbd_print_rate already functions as an emergency stress signal — "the HBD situation is abnormal." If that signal is firing, discretionary treasury outflows should not continue at full speed while the rest of the economy is under restriction. It's about ensuring the stress signal applies consistently to all HBD outflow paths, rather than carving out an exemption for the one path controlled by stake.

Without it, there's no mechanism preventing a feedback loop: payouts increase circulating HBD → debt ratio rises → hbd_print_rate suppresses ordinary HBD creation → but DHF payouts continue unabated → more circulating HBD → repeat.

Note: For almost all of my time on Steem/Hive I was under the impression that the DHF fund was part of the market cap that is calculated and visible publicly on sites like CMC - turns out that was wrong - so all my historic DHF voting was on the premise that I was 'unlocking' some pre-existing wealth that was already visible in the market cap for use in development, but in reality I was diluting the market cap.

I welcome all feedback and corrections!