It is 3.5% now. What if it get's to 9% again?
Since we have been in that situation we have a pretty good idea.
The SBD price will likely drop due to the distressed nature and asymmetric payoff unfavorable to the SBD holder, even during a conversion window (though the size of the price discount would likely be less since the conversion window has changed from 7 days to 3.5 days). To maintain the peg witnesses will need to increase the interest rate, and should also introduce some conversion discounts. Last time something like 15% APR was sufficient (it might have briefly been higher). Given less price risk for conversion we could probably make do with less now, but to be safe we'll assume 15%. At 10% debt and 15% APR that's a cost of 1.5% of STEEM market cap if maintained for an entire year, though that is quite unlikely. Alternately we could state the cost as 0.125% per month.
At that point (>5% debt ratio) production of new SBD would be stopped. Any conversions that occur would necessarily reduce outstanding SBD supply, and this includes large conversions that would probably (though we can't be 100% sure) occur by important stakeholders (for example Steemit itself owns about 300K SBD) specifically for the purpose of reducing debt. While this doesn't entirely guarantee that the debt ratio would reduce, it likely would (and it literally can't go above 10% given the current code). Some additional costs would incur due to the conversion discounts. If all 10% were converted (a very unlikely prospect) at a 3% discount (previously discounts were higher, but that was using a 7-day conversion process instead of the current 3.5), this would add an additional one-time cost of 0.3% of STEEM market cap.
So again, while there would be increased costs to this situation we are still looking at costs measured in low single digit percentages.
No one denies that SBD adds some costs and risks, but the actual numerical magnitude of these of often greatly exaggerated. It isn't anything close to the huge boogeyman that it is often made out to be. Even the leverage component, which isn't a net cost since it applies equally during up markets as down markets, is not all that large. At a maximum the leverage ratio is 1.1x (currently about 1.04x). This is not even easily measurable, much less a huge factor in practice.
Thanks for explaining.
Like I said I'm not an economist but I'm trying to expand my knowledge of such things. Having spoken to other people too - it seems the likely impact is smaller than I imagined.
Further as suggested by @tombstone the impact could be reduced further by reducing or eliminating the interest.
I had also underestimated the PR effect and attention that SD brings in terms of outside interest from non-crypto people. I think it is sometimes easy to get stuck in your own perspective and I am used to fluctuating values with BTC etc.
Merry Christmas!