Actually a few points here.
Lets look at the mechanics of what HAS been happening typically (from SCD/SAI history)
ETH price rising. CDP owners are ABLE to mint and sell more DAI (more DAI liquidity, typically lower DAI price than if the markets are stable or decreasing). IF we factor in SF in this a higher SF typically means less minting, and more likely for loans to be paid back (less DAI liquidity). SF rate has some effect here in terms of controlling the amount of DAI borrowers would be willing to mint.
ETH price dropping. CDP owners scrambling for collateral or better DAI to pay back loans. Here one needs DAI to short markets and so what we see is a PEG that goes > 1 because of lack of liquidity - particularly by CDP owners. This happens quite often. The SF rate effect here is basically NIL because the overwhelming concern of CDP holders is to not get liquidated vs. paying fees. In this case imo the SF rate is mostly irrellevant.
ETH price stable. This is the steady state collateral case where we will get minting if the PEG > 1 and sopping up to pay loans if < 1, provided the PEG fluctuation is significantly outside trade fee costs. The SF is in pretty good control here but based on the SCD data from last year it appears borrowers only react to changes of more than 5% up or down. It took the SF rate going from 1% to 18% before the PEG started to reverse from being as low as .98 at times to as high as 1.02 with fluctuations to .96 and 1.04.
Now if we add in the DSR effect. Think about what depositors think if the PEG goes below $1. This basically means they have lost all their DSR gains. Literally if in 1 month $1 DAI drops to .995 this is like a 6%/yr interest change in price and if they bought $1000DAI at the beginning of the month and withdrew and sold at the end of this month they would have gotten ~$3.33DAI worth of DAI in interest but then sold the $1003.33 for .995 to get 998.3 USD losing ~2%/yr effectively in interest. None of this includes fees on these transactions which are starting to be substantial for the ‘underserved’ (to the tune of .2-.5% or higher to exchange DAI to USDC on many exchanges now). Now lets reverse this. If the PEG is > 1 then DAI can come out of the DSR to sell and capture any extra return. Even a price of 1.005 provided fees don’t kill you can provide extra return to DSR holders. Certainly if the PEG is 1.01 or higher this can lead to return rates by selling DAI into markets from the DSR significantly higher than the DSR rate itself.
I think this is one point underlooked and underdiscussed about the DSR. It is a great place for liquidity to appear if the PEG is high, but won’t do dink except be very upset and lose money if the PEG is below $1. Again none of this factors in fees to convert back to cash. The fees to convert to cash I consider to be the biggest reason why the PEGs are going to fluctuate about $1. It is why I think the absolute BEST PEG price target is .995-1.005 to perhaps 3 sigma. But when you consider this floating price against the return earned and the prices obtained when people need to convert their DAI to cash you see how easily their DSR return could be washed out immediately.
This is one of the many reasons the PEG being at $1 or higher is actually desireable and hence a higher than what markets might want SF, and a lower DAI growth trajectory for MKR holders, to facilitate the stated Maker objective. At least if Maker really wants to serve the ‘underserved’ - i.e. peoples who probably will never have more than $1000 DAI in their lives at any time. What Maker really should be focused on is enabling key markets where people can exchange DAI directly for goods these underserved might need/want to buy and so offloading the conversion fees to vendors who are more likely to have lower fees and more easily be able to absorb these fee costs. Maker should also be considering that the PEG needs to be maintained at or above $1 if possible. A better PEG target for these underserved markets is something like .997-1.007 with the mean at or around 1.002.