MIP20: Target Price Adjustment Module (Vox)

MIP20: Target Price Adjustment Module (Vox)


MIP#: 20
Title: Target Price Adjustment Module (`Vox`)
Author(s): Lev Livnev (@equivrel), 🌧️ McRainface
Contributors: n/a
Type: Technical
Status: Accepted
Date Proposed: 2020-07-08
Date Ratified: 2020-10-27
Dependencies: n/a
Replaces: n/a
License: AGPL3+


Sentence summary

This proposal provides a smart contract implementation of Vox, a module which adjusts the target price of DAI according to a governance defined rate, allowing the institution of negative effective interest rates.

Paragraph summary

The Dai Stablecoin System is intended to reliably maintain dai’s exchange rate with respect to a reference asset (USD). Certain parameters of the System, such as the Stability Fee, are administered by Maker governance on an ongoing basis in response to evolving market conditions. This MIP implements the Vox module, which allows Maker governance to institute negative effective interest rates. In contrast to the Target Rate Feedback Mechanism (TRFM), the mechanism in this proposal does not adjust rates algorithmically using a DAI price oracle. Instead, the target rate is set directly by governance, similary to how the Stability Fees and Dai Savings Rate are set today.

Component summary

MIP20c1: Definitions: defines Target Price, Target Rate, Target Price Adjustment Module, and Target Price Cap.

MIP20c2: Desired properties: lists important properties that the Vox implementation must satisfy.

MIP20c3: Proposed code: contains snippet of proposed implementation.

MIP20c4: Test cases: lists existing test cases, including integration test

MIP20c5: Security considerations: comments on the limited nature of the security implications of adding the Vox.

MIP20c6: Other considerations: describes the changes to governance of monetary policy, and contrasts the proposal in this MIP with the TRFM.

MIP20c7: Formal verification/audit information: comments on the amenability of the proposed code to formal verification, even though formal specification, audit, or code review have yet to be conducted.

MIP20c8: Licensing: states the license under which the proposal and code are distributed.


In an environment where the supply of DAI is too low to meet demand, monetary policy might find itself at the “zero lower bound” where stimulus can no longer be effected through lowering interest rates. In this case, it may necessary to purseu a policy of negative interest rates, in which the direction of the cashflow is reversed, with savers possibly paying borrowers. In the system as it exists today, the combination of the CDP Stability Fee and Dai Savings Rate serves as a cash flow from borrowers (CDP users) to savers (DAI holders), or in other words, an interest rate. However, that interest rate is effectively constrained to be positive: though it is technically possible to accrue a negative interest rate to CDPs, depositing DAI into the Savings Contract is optional so savers would be able to avoid negative rates simply by not using it, leaving a deficit on MakerDAO’s balance sheet.

A crucial requirement in the implementation of this policy is for all DAI holders to be exposed to the negative rate. For technical reasons, it is not feasible to continuously manipulate on-chain user token balances, since this behaviour would undermine the implicit assumptions of well-behaved token semantics and could break integrations into other smart contract systems. A more compliant way of implementing negative interest rates is to manipulate the target price of DAI itself: this MIP implements such a system.

Adjusting the target price of DAI up or down causes an implicit value transfer from DAI borrowers to DAI holders, or from DAI holders to DAI borrowers, respectively, meaning that Target Price adjustment at a fixed rate of change acts like an effective interest rate, with a negative Target Rate corresponding to a negative effective interest rate.


MIP20c1: Definitions

  • Target Price: the system accounting price of 1 DAI in USD. This is currently represented as spot.par(), and is set to $1.00. Multi-collateral dai uses the Target Price in two places: when measuring the collateral ratio of a CDP, and when calculating the redemption value of DAI after Emergency Shutdown. The governance community may also use the Target Price as the price target for DAI/USD when setting interest rates and other risk parameters.
  • Target Rate: the annualised compounding rate of change of the Target Price (short for Target Price Adjustment Rate).
  • Target Price Adjustment Module: the smart contract which periodically adjusts the Target Price by the Target Rate.
  • Target Price Cap: the maximum Target Price that the Target Price Adjustment Module is able to set.

MIP20c2: Desired properties

  • At every call to vox.prod(), spot.par() should change by the target rate accrued over the period of time since the last call to vox.prod().
  • The Target Rate should be initialised at 0% per year, which results in no change to the Target Price.
  • The Target Price should not be adjusted after Emergency Shutdown, regardless of the value of the Target Rate prior to Emergency Shutdown.
  • The Target Price should not be adjusted above the Target Price Cap. When the Target Price Cap is reached, the Target Price should be set to exactly the Target Price Cap and further upward adjustments should have no effect.

MIP20c3: Proposed code

see vox.sol. The core price adjustment functionality is simple:

function prod() external {
    uint256 age = sub(now, tau);
    tau = now;

    if (age == 0) return;          // optimised
    if (way == ONE) return;        // optimised
    if (spot.live() == 0) return;  // no adjustment after cage

    uint256 par = min(cap, rmul(rpow(way, age, ONE), spot.par()));
    spot.file("par", par);

MIP20c4: Test cases

see vox.t.sol

  • test_prod_noop
  • test_prod_basic
  • test_prod_cap
  • test_mainnet_target_adjustment (this one is an integration test using the mainnet deployment of dss)

MIP20c5: Security considerations

The proposed solution is simple and non-invasive, interacting with only one other component of the system (the Spotter) through an existing method for adjusting the par price. Even though the core system was designed with the possibility of a changing par in mind, peripheral components and integrations should be carefully inspected for reliance on a fixed par.

MIP20c6: Other considerations

Upon adoption of this MIP, the Target Rate parameter can be adjusted by governance as an additional monetary policy lever, similar to the current notion of “Base Rate”. Monetary policy processes may have to be amended in order to leverage this facility, and this MIP may be expanded in order to specify them.

This MIP can be compared with the folkloric Target Rate Feedback Mechanism (TRFM), an unused implementation of which was present in single-collateral dai. The TRFM relies on the same notion of adjusting the target price of DAI as a monetary policy tool. The crucial difference between the TRFM and the mechananism proposed in this MIP is that while the TRFM algorithm sets the Target Rate automatically and continuously depending on a DAI price oracle, this MIP does not propose for the Target Rate to be set algorithmically.

MIP20c7: Formal verification/audit information

The proposed contract is written in a way which is amenable to formal specification and verification, in accordance with the style and practices of the core multi-collateral DAI contracts, though it has not been formally specified. No audit or code review has taken place yet.

MIP20c8: Licensing


I like new ideas that can easily give us moremonetary levers/tools.

It is not entirely clear to me how the above would incentivize creation of DAI to provide liquidity to markets and help with the PEG. Could you give us an example of how the above would help manage the PEG either from a vault owner or new opener perspective or from a market participant perspective.

Thanks in advance.

@MakerMan there’s a discussion of the incentives induced by a changing par in this older post:

(see “market price > target price” section for the current scenario)

1 Like

Thanks for reminding me of that since I had already read it and there was good examples.

Everytime I read the market low side here we have more than ample tools to bring the PEG back up but when I read the market high price example in the above it really looks weak again to me.

This incremental change in CR I don’t see as eliciting a lot of extra DAI. It is an incremental change.

Frees up vault holders to mint some more DAI but even if we assume ALL vault holders take advantage of this or even an average amount using this model basically would only at best mint the TRFM rate difference. If this is 1% it would be 1% on say current 160M or 1.6M DAI. The question becomes what happens in this mechanism if there is a constant demand for DAI as some fixed price point (say 1.01 or 1.015). Eats up the 1% TRFM change and the PEG doesn’t change. From what I see this TRFM will continue to change the target price until enough DAI is minted to finally drive the PEG down but then we end up with a target that is quite low.

So here the 1:1 redemption value is kept because on ES the vaults will redeem their collateral based on the targetDAI price used in the vaults and the whole 1:1 is kept intact throughout this whole process simply because CR against LR valuation mechanics is kept intact.

What is unclear to me is the rate this kind of system acts at. Since it may simply not be able to respond quickly enough or provide enough liquidity to deal with something like this comp issue.

The other thing this model doesn’t consider is the nominal fee percentage as a change against the TRFM value change. If I have a 10K vault and get a 1% boost in CR (for a whopping $100) but tx fees cost me $2 to do anything with that vault am I going to do this or not? I already am dealing with this in all of my DeFI activities - considering that doing anything is going to cost me not $2 but more like $10 and so I better be moving something north of 5K at a minimum to make more than that $10 it cost me. This may drive an incremental change but what if this system due to comp would end up pushing the target off by more than 20%. In most normal market circumstance I think this is a great system to push gently that retains the 1:1 DAI collateral backing. But I think the TRFM or variant might itself create new stresses as well as take significant amount of time to react especially when there are really big constant drivers affecting the PEG.

Don’t get me wrong. The more tools the better here. I think we do need to classify these tools in terms of real practical applicability towards various real situations. I think this comp issue is probably one of the most powerful and difficult issues to deal with since the amount of liquidity to bring the PEG back isn’t probably 1% but more like 25-50% or more of the current outstanding supply which is going to take any normally operating mechanism right out of whatever normal operating ranges are…


I agree that the incentives lack empirical validation “in the wild”. Many things work differently in practice. One thing we know–under current conditions, the system is recognizing a price of 1 USD / 1 DAI, yet the market stubbornly sits above that.

I think there’s an argument to be made that this module is something we’d rather have than not have–if the peg gets so bad that the only choices are 1) try negative rates or 2) ES, having this ready to go will be something we’re glad for.

It can be another tool in the toolbox–which there’s no obligation to ever use.


I am for MakerDao preparing and testing an implementation of this so that if it was ever necessary, it would be ready to deploy swiftly.


@MakerMan @Kurt_Barry Kurt’s TRFM Introduction covers the essentials on Target Price adjustment very well. However it is missing one important angle: the use of Target Price as the monetary policy target for future parameter adjustments (i.e. replacing our current target of $1.00). The effect of this is actually crucial, since it aims to create the credible expectation that the negative interest rate “on paper” is eventually realised into an effective negative interest rate.

It’s important to understand what this would mean in practice: suppose that because of a supply/demand imbalance that persists near the zero lower bound, with DAI/USD trading at $1.02 as a result, a target rate of -4% is set by governance. Suppose that some time later, the Target Price has come down to $0.99 (from the initial level of $1.00), and that DAI is now trading at $1.00. With the Target Price used as the monetary policy target, it is of utmost importance that the price is not considered to be “at the peg” (the new “peg” is $0.99!): instead, doveish monetary policy (like low SFs and negative TR) must continue until the price of DAI has reached the new (lower) price target. Indeed, with the TP at $0.99, $1.00 is the new $1.01, $0.99 is the new $1.00, etc.

Once we have credibly commited to pursuing the new Target Price as the monetary policy price target, only then will Target Rate Adjustment serve as an effective negative real interest rate.


With the Target Price used as the monetary policy target, it is of utmost importance that the price is not considered to be “at the peg” (the new “peg” is $0.99!) … doveish monetary policy … must continue.

I’m not sure I follow. You mean that the borrower should know that they will be able to repurchase DAI at a lower price that the one in effect when governance set -4%? Wouldn’t that already be the case with the price going from $1.02 to $1? “-4%” doesn’t have to be the expectation-setting negative rate.

Another kind of credible commitment would be to a $1 market price. Having 1DAI!=$1 internally would be a means to that end by adjusting expectations in case of ES and increasing borrowing ability. If we acknowledge that the internal and the market price can legitimately deviate, I don’t see why the target market price couldn’t be $1 forever and the target internal price be variable.


I’m not sure I follow. You mean that the borrower should know that they will be able to repurchase DAI at a lower price that the on in effect when governance set -4%? Wouldn’t that already be the case with the price going from $1.02 to $1? “-4%” doesn’t have to be the expectation-setting negative rate.

In a world where the Target Price is $0.99, i.e.

  • the system’s accounting value for collateral ratios
  • settlement value for global settlement
  • current monetary policy target

are all $0.99, a DAI price of $1.00 represents a 1% premium in the same way that $1.01 is a 1% premium in today’s world. The three factors above are the only places in the system where “1 DAI = $1.00” is present as an assumption: changing this is what gives Target Price adjustment teeth as a monetary policy tool.

If we credibly commit to using the Target Price as the monetary policy target, then a DAI saver or borrower should anticipate that in the long run, the price of DAI should follow the Target Price, for exactly the same reason that their expectations currently are anchored to $1.00, and therefore should perceive the continuous adjustment of the Target Price as an effective interest rate.


(numbered your list for clarity) What I had in mind was decoupling 1&2 from 3. Take this hypothetical scenario: market price is $1.02. Internal price is $1. Governance sets a $0.99 Target Price. At this moment, the expectation is that if market price hasn’t reached $1 when the Target Price of $0.99 is reached, a lower Target Price ($0.98, $0.97…) will be set. As the market price reaches $1, governance immediately starts raising the Target Price and behaves like a human PID controller.

In effect, that would be accepting the premium that the market seems to be placing on DAI on top of its intrinsic properties and adjusting those properties until, combined with the premium, a market price of $1 is reached. I don’t see a need for targeting any market price other than $1. That is: it seems unnecessary to promise that the market price will ever reach the Target Price. It seems good enough to promise that the value of 1 and 2 (from the quoted list) will always decrease until market $1 is reached.

edit: I am of course for the proposal itself.

1 Like

Thank you so much for additional helpful info on this.

So lets use your example. In an ES how much collateral value will 1 DAI recieve if the target price is .99 in the system? If it is .99 USD worth of collateral then in effect we are using this to do exactly what I was suggesting in terms of running with a negative surplus. It is just a different approach to effectively diluting the value of DAI in terms of backing collateral received during an ES.

What I am trying to wrap my head around is the delay cycle time in this PID loop since we have to take a measurement of the DAI market price and then re-adjust the DAI target price in the system. How often does this module - take a measurement - and then adjust the target rate?

Also what would you expect to be a normal working range on such a feature? How far would you be willing to allow the target rate to change before you would consider the system in danger either for DAI holders or vault owners (either the low or high side of the DAI target price).

Thanks again Lev always appreciate your intelligent thoughtful inputs. I like that the code change for this looks pretty simple too btw. I just think given fees and incremental changes that the real working rate and range will be of limited use in circumstances like this comp farming.

So if you would please address first what the system rate of change is, what you consider nominal working ranges for the target price, and whether you think these would address the comp farming issue or not.

1 Like

What I had in mind was decoupling 1&2 from 3.

Yes, I understand your position, and I have thought about this option too. I think it could possibly work, since indeed there must be a discount for (1) and (2) at which the market price of DAI starts to be affected. My concern is that under this monetary policy regime, the Target Rate wouldn’t have enough “teeth”, and that decoupling (3) from (1) and (2) introduces too much dissonance to monetary policy signaling. In practical terms, it would probably require a steeper Target Price adjustment than under the Target Price targeting regime, so we could end up in a scenario where DAI is only partially “backed” w.r.t. the monetary policy target of $1.00, making this scheme similar to the proposals of minting unbacked DAI.

1 Like

$0.99’s worth. The effect of this would have some similarities to proposals to create a system deficit, but the difference is that there will be no deficit, since the new price target will be $0.99 too. I believe this should also make the price adjustment effect much more effective (i.e. have a greater impact on supply and demand per $0.01 adjustment), since it’s not only the vague threat of lower redemption in ES which “realises” the price adjustment, but also the full force of all future monetary policy.

What I am trying to wrap my head around is the delay cycle time in this PID loop since we have to take a measurement of the DAI market price and then re-adjust the DAI target price in the system. How often does this module - take a measurement - and then adjust the target rate?

Unlike the TRFM, this module does not take any measurements, it simply adjusts the Target Price continuously using the Target Rate. It is up to governance to set the Target Rate, just like we set the Base Rate today. For example, if the Target Rate is -2% per year, then the Target Price will be adjusted downwards by about $0.00005 every day (in practice the adjustment happens whenever the module is prodded, similarly to how DSR and Stability Fees are accumulated).

Also what would you expect to be a normal working range on such a feature? How far would you be willing to allow the target rate to change before you would consider the system in danger either for DAI holders or vault owners (either the low or high side of the DAI target price).

I don’t think there’s any obvious range here outside of which the mechanism doesn’t make sense, just as with positive Base Rates. Last year when there was a lot of SAI supply we had to go as +19.5% to try to keep things under control (though we did not have the DSR back then to push from the other side). If the price remains above the target, then we probably have to keep adjust the Target Rate down.


In this model balance is achieved in a demand heavy market for Dai by:

  • Dai supply side - Vault owners supplying more Dai to the market as the target price falls in the system (they do so for various reasons).

  • Dai demand side - Demand for Dai falling to a balanced level with supply over time as the value of Dai decreases in the market as a result of increased Dai being supplied, a higher risk to holding Dai as a result of the potential global settlement being under $1.00 and over time getting lower and lower until balance is achieved.

This would scale infinitely because as long as there was any collateral in the system whatsoever a target price of approaching zero would be able to balance demand to supply. The idea is that just the knowledge of this system actually being able to balance the system properly is fear enough for Dai holders that they keep the price close to $1.00 in the markets. Anytime they choose to pay more than the target price for Dai they should expect to take a loss.

As far as I understand the incentives still work in this model because Dai holders still get the expectation of stability of Dai moving towards $1.00 in the market at all times. On the vault side the vault collateral should in theory increase as the target price gets lower because it is now more profitable and less risk to produce Dai.

The current issue is Vault owners are not willing to risk producing Dai knowing that the price can keep rising and producing Dai at $1.015 or $1.02 is not worth the risk of having to buy Dai back higher than they sold it for or risk liquidation costing them potentially $1.13. This would alleviate this fear over the long term. However, there is still short term liquidity risk of the collateral falling sharply and Dai price rising in the short term. Once this final issue is addressed and corrected in the system I expect MakerDao will be able to finally achieve proper stability in the system and price of Dai in the markets.

One last note: This solution is a good one, but MakerDao should also look for ways to increase the collateral in the system in other ways because even with balanced supply and demand for Dai the maximum value of all Dai in the market (viewed from the perspective of a global settlement) is still limited by the collateral backing Dai rather than the actual market demand for Dai at a $1.00 price. Ideally the target price of Dai is simply $1.00 and demand can be met at this target price. Anything less is stunting the growth and demand for Dai. If demand is too high and there is no other way though, I fully support this as a way to balance market forces because it balances them nicely.


Any DAI is partially backed once the settlement price becomes <$1. As a side note, if for some reason the market persistently puts a premium on DAI such that the internal price is <$1, the market price is $1 and Target Rate is 0, it will be simple to have only (2) (lower settlement value), not (1) (riskier collateral ratios): adjust those CRs upward so that they are effectively calculated against a price of $1. If the market is long-term happy with partially backed DAI this would be an interesting policy to follow.

(For clarity, by (3) I will mean: a governance commitment to try to actually reach its set Target Price in the market and not use it as a purely nominal adjustment knob).

I am confused by your model of (3) being able to bring down DAI price faster / with a higher Target Price. Why would the commitment to keep going after the market price has crossed $1 make that very event happen at a higher Target Price than in the world where the commitment is only to keep un-backing DAI until $1 has been reached?

Supposing that it is easier to to bring the price down with (3), is that worth it? Leaving out (3) means the $1=1DAI meme can be kept alive in market terms. In your model the cost to that seems to be a more unbacked DAI, relative to using (3). How bad is that compared to the good of keeping $1=1DAI?

Thanks for clarifying. I am sure you and a lot of other people know this system way better than I do. I am just looking at this from a monetary policy perspective and what it means to stake holders. I agree there are two factors that play on this. The psychological idea of ‘threatening’ a move, then actually doing it. Both should have some effect.

So the Target Rate is a governance thing so it would be a rather slow change unless we set a target rate pretty high here. This is a pretty slow response tool then and will only make incremental changes. One could say the same of the SF but in terms of making a change that is significant for a vault holder to take the tx cost hit to mint more DAI or not I think is a valid issue with this approach simply not having enough meat to really move the PEG. Even so I like it as a general monetary policy tool. It would be nice to have governance decide a working range and to code this up as something that changes periodically when someone takes out a DAI loan or repays DAI to update the TR based on what the PEG is telling us. This way your Vox idea becomes an actual PEG management mechanism that governance simply would have to monitor if it hits the end of actual working range (say 1.05 and .95 or whatever governance set on it).

Well I think as a DAI holder I would start getting concerned if the target price was < .9 above I only care about as a vault holder borrowing. But I as a DAI holder and vault borrower DAI PEG being off doesn’t really bother me much. I think I get more concerned if the PEG is like .98 than 1.02 but that is just me.

One thing I guess you are still mulling over is a response to whether you think this mechanism has enough teeth to manage something like this comp problem. Each time I am turning a response to PEG management when I think comp is driving this so far off I wonder no matter what tools we come up with. How far is governance going to have to go in terms of either ‘growing the system’ or ‘taking on risk’ to provide the kind of liquidity growth that could still manage the PEG. I really don’t think anyone other than myself and perhaps yourself has even tried to engage the community on this question.

We definitely want to explore monetary policy tools, implications for PEG stability, system risk, and growth but then we really have to come to grips with what comp has thrown at us. Do we even want to use all of these tools to drive the system to attempt to compensate for this massive driver, or do we allow other forces to drive Maker here via PEG, DAI demand etc. I think it is interesting @cmooney pointing out we have a good problem - high DAI demand - which everyone appreciated. Left out of this comment is the possible fact the only reason there is high DAI demand is because of compound and comp farming is most efficient with DAI - and perhaps no other. Without this driver - today we might have had a PEG close to 1 and then we would all feel less pressure to come up with new monetary policy tools and deal with the additional issues of risk, demand growth etc.

So I think this whole comp thing may end up being a blessing in disguise since it has caused the community to identify new tools, and implications to system as well as system participants causing a lot of work to be done on this issue. Where I think the discussion has not really matured on this is an understanding of what, in the future, if we experience more of these crazy secondary drivers messing with DAI demand what is governance guidance on how to proceed to deal with these things.

Do we consider devaluing DAI off the 1:1 ES guarantee. Do we have one or more PSMs (I am working on a post regarding my general concept of an ideal PSM system operation just to have as a example to work off of while constructing the first and later versions of the PSM tool should governance decide to push forward on this tool as well as others). Do we have a secondary reserve fund that governance uses to buy and sell assets to inject or remove DAI liquidity directly from markets. Do we allow MKR in governance to have access to MKR vaults with very conservative operating values (DC, LR, SF) to allow MKR in governance to be used not just to vote, but also to help manage monetary policy directly? etc.

I really want to hear from others on whether we would have to drive this system to extreme points to even have any hope of managing the PEG in the face of this comp problem. As a note. I watch carefully the uniswap comp/ETH paid and this ratio has consistently been coming down at something like a few percent a day. My rough estimate is that if this decay continues at this rate comp mining should ease off in about 2 months. This is mostly due to the fact people are using pretty high cash leverage here to access this COMP and from what I can see the total deposits are growing as well further diluting the return against real cash invested. WIth BAT and ZRX there was at least pricing risk by holding any of those coins long or short depending on the position. With DAI the pricing risk is significantly less.

Anyway. Really appreciate you putting MIP20 forward in detail Lev and I would support using this as an additional monetary policy tool generally. I think governance may need stronger tools that can bring DAI liquidity to bear directly for PEG management but we need to get a real direction from governance and the entire community as to how far we should go with these tools, and/or how far off of 1 the PEG can be for how long before it affects various players adversly in the DAI ecosystem.

1 Like

I support the proposal. The current tools do not work.

1 Like

@rune, @cyrus, @g_dip, @monet-supply

I would greatly appreciate if you could weigh in on this. I feel like this is the proper solution to the demand imbalance that is causing the peg to drift. MakerDao could implement this to solve the root problem and still have a PSM for the added liquidity it can provide. This is the first time in several months I have felt at peace over the peg drifting. This solution is comforting because it would actually work (as far as I understand). Thank you in advance for your time to weigh in on this everyone!

Sure, it breaks the 1 Dai is $1.00 promise (temporarily), but not in the markets which is what the average Dai holder is likely most concerned with. I feel like most people would be OK with this solution because it will keep Dai stable in the long term and as a result in the short term as well as the forward looking expectations play out.

I feel like for things to truly be comfortable in the ecosystem there needs to be a massive short term liquidity of Dai available for emergency type situations when collateral is falling fast in price and Dai rising do to deleveraging and panic. After the “Target Price Adjustment Module” and also a short term liquidity solution has been put in place I feel like MakerDao’s Dai will be in a really good place.

@BrianMcMakerDAO Thanks for the explanation of the MIP on the community call!

One follow up question… If the Vox functionally adjusts the collateral ratio to allow people to mint more dai, do you think people who choose “safe” ratios (eg 200-300%) would be enticed enough to mint more dai and cause a meaningful change to the PEG?


That depends on each individual vault owner and how they perceive and interact with the system. I will say that it shouldn’t too much matter because anyone can open up a vault and create Dai, so it doesn’t have to be the safe holders it can be instead a new owner who is looking to take advantage of the profit opportunity.