As suggested in Dan's post, discounting price feed is a way to improve SBD peg. After his suggestion at 2016/9/19, each witness started to discount its feed price by 7~14%, and the SBD market price went up from $0.86 to $0.96. Aside from this positive effect, I want to point out another effect and related market opportunity.
- Period: 2016/7/19 ~ 2016/10/9
- Price sources
- Poloniex: STEEM/BTC, SBD/BTC, BTC/USD
- Steem blockchain: FEED/USD
The meaning of price feed discount is simple: giving premiums on SBD by discounting STEEM. This has direct impact on total reward amounts since the market cap of STEEM determined by the feed price is also discounted. Meanwhile, premiums on SBD incentivize to buy SBD at higher price because there are less risks in SBD-STEEM conversion.
In addition to these direct effects, I found one more impact. Discouting the feed price can lead to facilitate STEEM price decrease. Let's look at a graph first.
The blue line(steemusd) represents market prices of STEEM in real dollar (STEEM/BTC * BTC/USD). Orange line means STEEM price in SBD (median price feed). Green line means STEEM price via conversion (median feed price * SBD market price at 7 days ago). So an implication of this graph is simple. If the green line is below the blue line, you can earn money via conversion. For instace, we have +28% difference at 8/8. We can (1) buy SBD at $0.81 at 8/1, (2) convert it to STEEM, (3) buy STEEM via conversion at the median feed price of $2.07, and (4) Sell them at the market price of $2.14. So 28% comes from
(2.14 / 2.07 * 0.87) - 1
However, if the market price fastly decreases, the lagged feed price can be higher than actual market price. Actually, without cheap SBD in market (orange line), SBD conversion is not profitable most of times, and we can even see that the green line is above the blue line after some price deeps (e.g. between 8/9~8/13)
But what happens if we discount the price feed? There is more possibility of earning through conversion since the orange and green line are shifted down. The next graph shows the differences between STEEM market price and STEEM price via conversion, if it aboves zero, it means you can earn via conversion.
Here, it is notable that 3.5 days after Dan's suggestion, the graph is always above zero.
Then, how much incentives does the discount provide? If we assume the price moves perfectly linearly, the discount gives x2 percent of a safety zone. That is, if discount is 10%, the conversion stil earns money unless STEEM decreases 20% (10*2) a week. Although the sample size is small, I plotted STEEM market price change (x-axis) and the difference between STEEM market price and STEEM conversion price (y-axis), and tried to regress before and after the discount, respectively. Since the discount level is 8~10% (witness's median), I expected the change between the before and the after is 16~20%.
According to the regression results, the conversion tends to lose money if STEEM price decrease a week is over 19% before the discount. However, after the discount, the conversion is more protected (36%) from the price decrease. The difference is +17%, which falls in my expectation of 16~20%.
This implies three things.
First, we can have more opportunity to earn money via SBD conversion
Second, thanks to the incentive, there will be more buying power on SBD
Thrid, however this can generate more selling pressure on STEEM
Explaining the third point, traders who convert SBD easily can sell STEEM because their breakeven point is lower(e.g. 36% decrease a week). So they can do the process (convert SBD, sell STEEM, buy SBD, convert SBD ...) more actively. In this sense, I think that we need to make a balance in the discount rate, because low or no discount rates can break the peg while high rates can push STEEM price down quickly. Therefore, witnesses should consider this mechanism and provide their feed prices wisely, and voters should prudently select their witnesses.
FYI: Currently my witness provides price feed discounted by 5% (and maximum 7% depending on the market condition)