I don't recall ever seeing anything about the effects of price feed bias, except maybe a small mention in the whitepaper. I'm not really a fan of experimenting with that. I have some family obligations, so can't spend time researching for a while, unfortunately. Can only comment from cell phone.

Of the three, I think a burn proposal would be most effective - as long as STEEM is below the haircut threshold. It might make sense to create multiple proposals (100, 500, 5000, 10k,...) so that stakeholders could dynamically adjust the size of the daily burn.

On the other side, paying SBD interest might encourage exchanges to offer SBDs... The objection was that exchanges benefit most, but maybe that's a feature...