It had been below $1 for most of the last week. It has just been above $1 for an hour or so now. For me it's a bigger concern when the price goes below $1 than above. With our current low liquidity, a large order can often push the price below $1 anyway. The most important social contract with SBD is that you can get $1 worth of Steem with it, if you end up getting more, that's just gravy. It has to go a lot further above $1 for an upwards peg to be a real problem. It's fair to move more slowly on fixing the peg broken to the upside.
As long as it doesn't require a rebuild or some co-ordinated effort, I do think top witnesses should be on the ball on maintaining the peg. I think that expectation is fair given how much money is paid for the position.
The bias in particular should not need much to update every 2-3 days, it should be nothing more than a configured value. Your argument re: APR only applies on moving downwards. Nobody is going to complain when APR rises faster than expected.
I agree that liquidity/conversions is the first step in maintaining the peg. Like I mentioned in another comment I put several thousand dollars through conversion in the last 24 hours.
But in the past they did, loudly and often. There were months of very real concerns about the amount of debt in the system as SBD were far too numerous and the price of STEEM was tanking months ago. It became a systemic risk for the entire platform and SBD payouts had to suspended with liquid STEEM given out instead. Increasing SBD has downsides also which should be considered. Adding a feed bias impacts the conversion, but if the conversion takes 3.5 days anyway, then changing it day to day won't do much towards the goal of a stable SBD. Maybe a 2-3 day target makes sense, but I still think these systems take a while to self-regulate, and if we tweak the nobs too much or too often, we end up creating oscillations which can further increase instability.
Just my opinion, though. Maybe I'm relying too much on the market forces to do their thing. The $1 peg promise is an important one, but at what timescale? An hour? A day? A week? A month? I think that's the important question. I think waiting a week or two for things to settle out isn't unheard of. Maybe one fix would be a nice wallet notification along the lines of "We noticed SBD is currently trading under $1. Did you know you can convert your SBD for $1 worth of STEEM in 3.5 days? Would you like to do that now?" or something like that.
How much of that was due to APR rising faster than expected? As far as I'm aware, it was never higher than 10%. 10% APR can only possibly increase the debt burden by a maximum of 10% in a year. That's 7 years to double and 24 years to increase by an order of magnitude. Compared to other factors, it's a non-issue in regards to the debt burden.
My answer to that would be an ever decreasing timeframe. As the network gets larger, with more liquidity, it will be easier to maintain a stable peg. You will be able to do it on shorter and shorter time frames. You're not aiming to be the federal reserve, where you're responsible for the whole economy. You're only responsible for block production, adopting network rules updates and maintaining a price peg. I think on the contrary, faster responses will improve our ability to maintain a peg.
I also think that relying on market forces is the wrong philosophy here. A peg is by definition a method of manipulating the market to keep a stable price. The market can only maintain a peg by responding to the incentives presented to them (an unmanipulated market would have the token float all over the place like Steem does).
Thanks for the discussion.
It sounds like an automated price bias would be what you're in favor of. If the price ever drops below $1, then a bias gets implemented right away to always keep it at $1. That might be interesting. Ultimately, I have doubts about the current structure of the SBD ever working because unlike bitUSD, there isn't a conversion mechanism both directions which enforces the price peg. The 3.5 day lag time may also be too long to quickly adjust the market (it used to be 7 days). Maybe more liquidity would make this a non-issue as more people would use the conversion option, even without a price feed bias.
It doesn't have to be perfect, but it can get progressively better. Automation will be necessary eventually, but for now a human making a quick review and judgment every 2-3 days and applying the configuration change seems reasonable to me.
I argued in posts over the last month when the peg was well broken to the upside, that an upwards broken peg isn't a huge concern. The network as a whole still gets large benefits from that, as long as people are aware that there's no promise to keep the price that high. It undermines the ability to use the token for ordinary commerce, but we gain other benefits instead. I am in favor of allowing SBD to diminish in importance (such as by allowing rewards to be paid optionally in Steem), as long as the SBD contract itself still remains available.
I was unclear what price you get in conversion. Isn't the feed price the price you would receive. I think the stabilization factor that was added last fall is messing up things now. Back then, steem's survival was in question so it was trading with expected recovery value(50 percent chance of 0). Today it's not distressed debt anymore so it should trade with 3 days risk free interest discount to the expected steem value upon conversion.
What stabilization factor do you mean?
I was using your words. Is there something that does not use the price from the exchanges? Something that adds a discount or premium? I really don't know how it works but I remember people discussing a change to the price feeds when it traded at a large discount last fall.
I got confused by the "added last fall" part. The price feed bias is determined by the witnesses, so it's not something that is part of the blockchain itself. You can see the biases here under the "bias" column: https://steemd.com/witnesses
If most of the those are 0, then the price of Steem as far as the blockchain is concerned will be the same as the price on exchanges. I don't know the specific mechanics off hand as to how many are need to adjust the price feed, but right now it doesn't matter since few if any are using a bias.