Risk Parameter Updates 2022-12-01

[jakeaujus]

Venus Parameter Recommendations 2022-12-01

Gauntlet makes the following recommendations to optimize risk and capital efficiency for Venus:

Summary Recommendations

CF:

• We recommend decreasing BNB collateral factor from 80% to 75%. This would reduce the Value at Risk (VAR) for Venus from $7.7m → $270k to minimize risk for “Black Thursday” events.

Supply Caps and Borrow Caps (denominated in TOKEN NOT USD):

Symbol	Current Supply Cap	Current Borrow Cap	VIP 77 supply caps	Recommended Supply Cap	Recommended Borrow Cap
USDT	-	-	-	736,300,000	245,500,000
XVS	-	0	-	1,311,000	0
XRP	-	-	1,000,000,000	35,400,000	3,029,000
USDC	-	-	-	258,000,000	124,700,000
TUSD	-	-	-	21,130,000	11,810,000
TRX	-	-	-	48,300,000	16,400,000
SXP	-	-	25,000,000	25,000,000	6,410,000
MATIC	-	-	262,000,000	6,718,000	1,470,000
LTC	-	-	2,000,000	254,100	25,410
LINK	-	-	15,000,000	2,388,000	238,800
FIL	-	-	9,000,000	908,500	122,200
ETH	-	-	-	222,300	28,740
DOT	-	-	33,000,000	2,209,000	776,400
DOGE	-	-	4,000,000,000	157,700,000	23,240,000
DAI	-	-	-	13,910,000	5,414,000
CAKE	-	2,500,000	7,000,000	7,000,000	3,749,000
BUSD	-	-	-	680,800,000	184,500,000
BTCB	-	-	-	22,770	3,531
BNB	-	-	-	2,672,000	2,008,000
BETH	-	-	-	21,890	16,450
BCH	-	-	500,000	26,820	4,331
ADA	-	-	1,000,000,000	37,510,000	14,430,000
AAVE	-	-	422,000	19,160	4,808

Methodology

This set of parameter updates seeks to maintain the overall risk tolerance of the protocol while making risk trade-offs between specific assets.

Gauntlet’s parameter recommendations are driven by an optimization function that balances 3 core metrics: insolvencies, liquidations, and borrow usage. Parameter recommendations seek to optimize for this objective function. Our agent-based simulations use a wide array of varied input data that changes on a daily basis (including but not limited to asset volatility, asset correlation, asset collateral usage, DEX / CEX liquidity, trading volume, expected market impact of trades, and liquidator behavior). Existing insolvencies are excluded from our simulations and VaR and LaR numbers. Gauntlet’s simulations tease out complex relationships between these inputs that cannot be simply expressed as heuristics. As such, the input metrics we show below can help understand why some of the param recs have been made but should not be taken as the only reason for recommendation. The individual collateral pages on the Gauntlet’s Venus Risk Dashboard cover other key statistics and outputs from our simulations that can help with understanding interesting inputs and results related to our simulations.

For more details, please see Gauntlet’s Parameter Recommendation Methodology 1 and Gauntlet’s Model Methodology 1.

Risk Off Liquidations:

Whenever we lower collateral factors, there’s a chance that some users may immediately become liquidatable as a result. There are a number of accounts that lend and borrow stablecoins and keep their balances just slightly above the liquidation threshold.

Reducing the collateral factor of BNB by 5% would push 90 accounts from
just above the collateral factor to just below the collateral factor. The accounts have a total borrow value of $165k. The top account has a borrow balance of $150k. Most of the accounts are extremely small.

Supply Caps and Borrow Caps:

In light of the recent abnormal liquidations and activity across many lending protocols, Gauntlet recommends conservative borrow and supply caps on all assets, particularly on those who have low liquidity and/or supply on BSC, in order to significantly lower the potential for price manipulation. Over time, as the Venus lending market grows for many of these assets, we will continue to monitor these parameters and recommend adjustments as needed.

Supporting Data

Top 10 borrowers, Aggregate Positions and Borrow usages

Top 10 borrowers, Breakdown of Borrower Supply

Top 10 Borrowers Breakdown of Borrows

Above shows the distribution in risk among the top borrowers - A large amount of BNB is concentrated in one borrower, which could lead to insolvencies. Because of this, we can observe why a large portion of the VaR is in BNB.

Risk Dashboard

The community should use Gauntlet’s Venus Risk Dashboard to better understand the updated parameter suggestions and general market risk in Venus.

Value at Risk represents the 95th percentile insolvency value that occurs from simulations we run over a range of volatilities to approximate a tail event.

Liquidations at Risk represents the 95th percentile liquidation volume that occurs from simulations we run over a range of volatilities to approximate a tail event.

Next Steps

• Gauntlet to put up a snapshot vote for the community to vote on

[wonderomg]

I’m glad that Gauntlet is working with us but this will destroy our revenue. For example if we apply lower borrow cap for BNB and we will reach it soo we will lose $millions during every Binance launchpad. Our interest rate model is built to let us earn maximum when there is no liquidity and force borrowers to repay loans because of high %. If we apply those parameters it can drastically reduce our income and growth.

Every time someone borrows the maximum amount of an asset for a limited time, as sometimes happens with BETH, where there is up to -257% borrow apy, then within a few hours or 1 day it produces a profit greater than at normal rates for the whole year. And with these “caps” we will absolutely cut ourselves off from this income. We would have to modify our ‘‘jump rate model’’ so that the higher percentage rate starts earlier, in other words, change the interest rate curve for borrowing, which in turn would make borrowing more expensive on average, which in turn may discourage users who they use it (It means we cannot use this solution).

To be honest i expected some changes but this is like suspension of 50% of the production force in the factory so we will need probably to discuss positive vs negative for each change in the parameter of individual assets.

Of course, this problem will concern us only at the moment when the supply exceeds the maximum borrow cap. Because at that moment, our maximum asset APY will be reduced many times over at maximum utilization.

[davd.mbt]

Why is the borrow cap for BUSD limited to 35% of the supply cap?

Also, why do other ‘stable’ assets like Bitcoin and Ether have a borrow cap nearer 10% of the Supply Cap?

[davd.mbt]

How will the Supply and Borrow Caps effect the interest rate model for each asset?

[jody]

i
I very much agree with some of the views in it, but the lending factor of some coins may affect the overall income of the project. The Black Thursday you mentioned may have , but Venus already has a complete security audit model, as well as Binance’s audit. A similar situation can be prevented from happening.

[Emrarss]

-The measure of minimizing the value-at-risk for Venus is very effective. but BNB is very important to Protocol revenues. This measure, along with the risk, can also reduce revenues.

-The parameter recommendation methodology of the Gauntlet for exchange risks in the Protocol provides important statistics. This is valuable.

-Risk boards and reports are important for keeping the community informed.

Thanks for the work.

[SK-Mousavi]

1- Existing bad debts are one of the great risk on protocol. Do we know why it’s excluded from the assessments by Gauntlet?
We should know exactly why such huge bad debts are accumulated and how they can be avoided in future?

The lending business can not be without bad debts , but it should be always much lower than the revenue .

2- The proposed parameters should not wave out the main revenue we are collecting during the launchpads.

[JR-ARG]

I agree that we need to be cautious and setting a max cap to supply and borrow is a reasonable risk management measure.

Though I’m surprised to see that some markets, like LTC or TRX for example, have a borrow cap as little as 10% of the supply cap. Also, a very important market in term of volume like BUSD has less than 20%.

I think that before moving this proposal forward, we need to fully understand and eventually reconsider some of this suggestions, as they will probably have a direct impact on the protocol revenues.

[Mariana.Red.Star]

I appreciate Gauntlet’s proposal very much but this leads me to some questions: how this would affect the protocol’s revenue and its users?

Since Venus’ users are the number one priority how the proposal is going to be managed for the ones who are going to get liquidated?

[NoOneVII]

I like the community’s critiques for Gauntlet. I wrote it 2 days ago when I first browsed it, even the BNB change is completely wrong. Looks like Gauntlet is underestimating us

If this Snapshot is implemented:
The CEO’s goals;
For higher TVL,
Usage by more people,
Will be number 1 in the industry will be destroyed before the market recovers

I think they couldn’t analyze Venus well due to dealing with AAVE issues and they lost confidence in their first offer.

With the numbers here, it completely reduced
revenues, TVL does not increase and we become an unpreferred platform.

We ensure that there will be no new bad debts by ensuring that there is no borrowing anyway lol, and this requires expensive analysis??

[ChristoWarriorCrypto]

I really appreciate your work but I know that some terms have to be fixed to avoid risks but I have a doubt.

How would these adjustments affect the income of the protocol and likewise the users that can be liquidated ?

Because something that characterizes us is that our priority is our users.

Because as we saw a few days ago, there was a great demand of users who interacted with the protocol to participate in the Binance launchpad.

[jakeaujus]

Hey wonderomg, just wanted to clarify:

Our cap table is denominated in tokens (not USD) as this is how borrow caps are denominated in the smart contract. Our recommended Borrow Cap of 2,008,000 BNB, at a current price of $288, would enable up to $578M in borrows, which is significantly above the last peak of $406M. Therefore there is no expected impact on protocol revenue or borrower UX. I added a note into the post to call this out.

[wonderomg]

I know that it is denominated in tokens the only thing i wanted to say is that for example in last Binance launchpad we earned about 2.500 BNB cuz supply APY was in range between 50-60% mostly with zero or minimum liquidity. If we will use same example with your proposed caps (they are almost 2x higher so our launchpad profit would be approx 5.000 BNB) but your caps means that we can borrow only 75% of supplied BNB and with our jump rate model if BNB supply increases to the maximum cap during the next launchpad and there will be a borrow cap of 75% APY on BNB market will be max. 5.53% instead of 50-60% so instead of 5.000 BNB our profit will be around 500 BNB and that is of course a huge difference, especially if we consider some future Bull market phase. If you understand me i know that with our actual supply and borrow nothing changes, our profit would be the same but if supply will growth fast to maximum during Launchpads we will lost millions of USD (in BNB tokens) in a few weeks. So I’m not sure if this is the best solution for us. And that’s not counting the reduction of CF at BNB from 80% to 75%, which is another significant loss of revenue.

[Frayst]

Hi,

If gauntlet calculates the recommended supply & borrow caps for each asset by taking into account the peak usage + 10%, I do not see why we will loose revenues. Especially if these parameters will be updated frequently. Could you please confirm the frequency of this analysis and associated VIP? Is it every two weeks, months?

In addition, If a cap is full (ex: bull run, bnb launchpad) we can still increase it with a fast vip (new governance feature as part of Venus v4).

It could be great to implement Dynamic supply cap to avoid vip in case of bull run though. Feasibility to be checked with Venus developers.

[Criptoeli]

If the goal is to balance capital efficiency and risk in the protocol, and they try among other things to introduce conservative lending and bid limits on all assets, particularly those that are illiquid. I have 2 questions left: 1) It is in relation to the Methodology in the network congestion model, which uses ETH as a reference. Was this the case for Venus? Not only do we have an Ethereum in another scenario thanks to its fork, but we also have a network like BNBChain with another level of capacity. and 2) and would lowering the BNB collateral factor and additional percentage adjustments really allow for worst-case solvency?

[mopgawka]

@wonderomg

I’m glad that Gauntlet is working with us but this will destroy our revenue. For example if we apply lower borrow cap for BNB and we will reach it soo we will lose $millions during every Binance launchpad. Our interest rate model is built to let us earn maximum when there is no liquidity and force borrowers to repay loans because of high %. If we apply those parameters it can drastically reduce our income and growth.

Every time someone borrows the maximum amount of an asset for a limited time, as sometimes happens with BETH, where there is up to -257% borrow apy, then within a few hours or 1 day it produces a profit greater than at normal rates for the whole year. And with these “caps” we will absolutely cut ourselves off from this income. We would have to modify our ‘‘jump rate model’’ so that the higher percentage rate starts earlier, in other words, change the interest rate curve for borrowing, which in turn would make borrowing more expensive on average, which in turn may discourage users who they use it (It means we cannot use this solution).

To be honest i expected some changes but this is like suspension of 50% of the production force in the factory so we will need probably to discuss positive vs negative for each change in the parameter of individual assets.

Of course, this problem will concern us only at the moment when the supply exceeds the maximum borrow cap. Because at that moment, our maximum asset APY will be reduced many times over at maximum utilization.

I agree with Dominik. Of course, we have to consider options for reducing the loan interest for some assets and this is very important, but in the case of BNB, this does not seem to be entirely logical. Not so long ago, there was an offer from one of the employees of the BNB chain, where an offer was received to cover loans in the BNB, with the help of Binance. Based on this, I do not fully understand the practical meaning of reducing the threshold value. In my humble opinion, the proposal to reduce the threshold needs to be combined with the profit alternative of the protocol. As Dominik rightly pointed out, the protocol profits from proportionate rate hikes and liquidations. If we change this, then we need to rework the model, but this is not noted in the proposal. Perhaps there are some other, more weighty arguments for changing the percentage from 80 to 75?

[jakeaujus]

One purpose of the borrow and supply cap is to help prevent 100% utilization of the market. 100% utilization is not only poor user experience (as it prevents withdrawals), but it also poses risk to liquidations (can hinder atomic liquidations because liquidators are not able to obtain the collateral asset directly). If liquidations are hindered, it can lead to insolvency. Interest rate curves don’t protect against 100% utilization in all scenarios, which is why borrow caps can be set lower than supply caps to help protect against 100% utilization.

[jakeaujus]

As clarified above, the initial borrow caps are set with the objective of mitigating outsized risk while not impeding regular borrowers and suppliers (defined as historical borrower / supplier behavior as distinct from an attack like one off behavior).

Borrow Caps in particular seek to protect against sudden large short positions caused by someone borrowing an asset and switching it into stables. Allowing short positions that are of significant size relative to the circulating supply of the asset can subject the protocol to outsized insolvency risk, as there might not be enough liquidity left in the open market to effectively liquidate insolvent positions.

In addition, borrow caps can protect against Mango squeeze type events. This is because attackers can borrow assets to avoid price slippage. Thus, having borrow caps increases the capital cost (thus reducing the likelihood) of price manipulation attacks.

Supply Caps seek to protect against the opposite - someone suddenly depositing a large amount of a given token and using it as collateral. In the event of a successful price manipulation (be it through recursive borrows, perps, or even an oracle exploit), an uncapped market leaves the protocol vulnerable to being drained completely. Our methodology seeks to limit exposure to these extreme usage patterns while minimizing impact to users. If set appropriately, user experience should not be materially impacted. Borrow and Supply positions change frequently, therefore we do have continuous alerting set up internally to identify tokens that are nearing their caps, so that we can proactively raise them as needed.

We would also like to speak to our role as risk managers of market risk. We seek to help Venus increase their TVL while mitigating existential risks. Recent market events have shown that leaving any lending market uncapped leaves that market vulnerable. Venus is well positioned by having functionality for both Borrow and Supply caps (which is not yet standard in the industry) and we want to help you use them without impacting revenues.

The one recommendation that does have immediate impact on borrowers (and revenue) is the Collateral Factor (CF) change for BNB. When making CF recommendations, we generate a “liquidation impact table” identifying any accounts that would be immediately liquidatable. As mentioned in our post, a CF decrease of 5% would make 90 accounts have a health factor less than 1 (be eligible for liquidation). “The accounts have a total borrow value of $165k. The top account has a borrow balance of $150k. Most of the accounts are extremely small.” This is what we call a “risk off” recommendation which reduces capital efficiency but with a greater expected benefit to the protocol. As you’ll see in the expected VaR and LaR numbers, the majority of the risk on the protocol currently stems from loans collateralized with BNB. The liquidity, which is required for successful liquidations, is significantly lower for BNB than other “stable” (USDC, DAI) and “blue chip” assets (ETH). Realistically, this CF should be decreased even further. Our sensitivities to User Experience, and avoiding large liquidations, is what capped our recommendation to a 5% decrease.

[Akegpak]

This is the good development for Venus this year, although many users didn’t realize, but it will bring sustainability to the protocol.

[jakeaujus]

Hi Venus Community, Gauntlet has updated our recommendations, please see below:

TLDR:

• Gauntlet’s recommendation remains to reduce the CF of BNB to 75%, and we have updated our recommendations for borrow and supply caps in the table below.

Collateral Factor Recommendations

• BNB: Lower CF from 80% to 75%.
  • To clarify, Gauntlet did a fresh analysis of the protocol and this adjustment is still being recommended.

Methodology

Our methodology remains the same:

This set of parameter updates seeks to maintain the overall risk tolerance of the protocol while making risk trade-offs between specific assets.

Gauntlet’s parameter recommendations are driven by an optimization function that balances 3 core metrics: insolvencies, liquidations, and borrow usage. Parameter recommendations seek to optimize for this objective function. Our agent-based simulations use a wide array of varied input data that changes on a daily basis (including but not limited to asset volatility, asset correlation, asset collateral usage, DEX / CEX liquidity, trading volume, expected market impact of trades, and liquidator behavior). Existing insolvencies are excluded from our simulations and VaR and LaR numbers, since we can only mitigate risk of future insolvencies. Gauntlet’s simulations tease out complex relationships between these inputs that cannot be simply expressed as heuristics. As such, the input metrics we show below can help understand why some of the param recs have been made but should not be taken as the only reason for recommendation. The individual collateral pages on the Gauntlet’s Venus Risk Dashboard cover other key statistics and outputs from our simulations that can help with understanding interesting inputs and results related to our simulations.

For more details, please see Gauntlet’s Parameter Recommendation Methodology 1 and Gauntlet’s Model Methodology 1.

Borrow / Supply Cap Recommendations

In light of the recent abnormal liquidations and activity across many lending protocols, Gauntlet recommends conservative borrow and supply caps on all assets, particularly on those who have low liquidity and/or supply on BSC, in order to significantly lower the potential for price manipulation. Over time, as the Venus lending market grows for many of these assets, we will continue to monitor these parameters and recommend adjustments as needed. We are amending our proposed supply and borrow caps for six assets (TRX, SXP, DOT, DAI, AAVE, and XRP) pulled to the top of the following table. We recommend proceeding with the remaining recommendations as previously presented.

Token	Current supply caps	Current Borrow Cap	Recommended Supply Cap	Recommended Borrow Cap
TRX	-	-	81,207,000	18,853,000
SXP	25,000,000	-	55,052,000	8,100,000
DOT	33,000,000	-	2,209,000	925,000
DAI	-	-	13,910,000	7,041,000
AAVE	422,000	-	30,000	10,000
XRP	1,000,000,000	-	35,938,000	4,248,000
————	————-	———-	———-	———-
USDT	-	-	736,300,000	245,500,000
XVS	25,000,000	0	1,311,000	0
USDC	-	-	258,000,000	124,700,000
TUSD	-	-	21,130,000	11,810,000
MATIC	262,000,000		6,718,000	1,470,000
LTC	2,000,000		254,100	25,410
LINK	15,000,000		2,388,000	238,800
FIL	-	-	908,500	122,200
ETH	-	-	222,300	28,740
DOGE	4,000,000,000		157,700,000	23,240,000
CAKE	7,000,000	2,500,000	7,000,000	3,749,000
BUSD	-	-	680,800,000	184,500,000
BTCB	-	-	22,770	3,531
BNB	-	-	2,672,000	2,008,000
BETH	-	-	21,890	16,450
BCH	500,000	-	26,820	4,331
ADA	1,000,000,000	-	37,510,000	14,430,000

Risk Off Liquidations:

Whenever we lower collateral factors, there’s a chance that some users may immediately become liquidatable as a result. There are a number of accounts that lend and borrow stablecoins and keep their balances just slightly above the liquidation threshold.

Reducing the collateral factor of BNB by 5% would push 101 accounts from
just above the collateral factor to just below the collateral factor. The accounts have a total borrow value of $35k. The top account has a borrow balance of $28k. Most of the accounts are extremely small with recursive holdings.

Supporting Data

The figures above show the distribution in risk among the top borrowers - A large amount of BNB is concentrated in one borrower, which could lead to insolvencies. Because of this, we can observe why a large portion of the VaR is in BNB.

Risk Dashboard

The community should use Gauntlet’s Venus Risk Dashboard to better understand the updated parameter suggestions and general market risk in Venus.

Value at Risk represents the 95th percentile insolvency value that occurs from simulations we run over a range of volatilities to approximate a tail event.

Liquidations at Risk represents the 95th percentile liquidation volume that occurs from simulations we run over a range of volatilities to approximate a tail event.

Next Steps

• Gauntlet will be putting up a snapshot vote for the community to vote on the implementation of these recommendations