---
tags: production
---

# Saffron V2

Rx3a97¹
[e-mail]_

psykeeper¹
[e-mail]_

¹ Saffron Finance

> Note: This is a working version of the Saffron V2 paper printed on April 25th 2022.
## Abstract

The advent of smart contracts in the last few years has led to the emergence of a rich, interconnected decentralized finance (DeFi) system. DeFi protocols have democratized access to financial products by eliminating the need for intermediates. These protocols, like centralized systems preceding them, are not immune to failures. Here we introduce Saffron V2, a decentralized system that offers a set of tranches with different levels of covered yield protecting against failure of the underlying platform. In this whitepaper we will give a brief background on DeFi and show how Saffron V2 complements the ecosystem by offering lower risk opportunities. We will define Saffron V2 tranches and their properties, what determines a claimable event, and detail how the protocol calculates the yield for each tranche. Finally, we will show how Saffron V2 can work as a risk management tool. Our work follows the former Saffron V1 and brings simple yet effective asset protection to DeFi.

## Introduction

Decentralized Finance (DeFi) has been recently loosely defined in the past by several authors (Blockchain and Project, n.d.; Chohan, 2021; Leonhard, 2019; Schär, 2021; Zetzsche et al., 2020) and, in our opinion, more precisely by (Gudgeon et al., 2020) as a peer-to-peer financial paradigm which leverages blockchain-based smart contracts to ensure its integrity and security. In fact, DeFi relies on the smart contract ecosystem for its own execution as well. The utopian objective of DeFi is to democratize access to finance by eliminating intermediates. In the past few years, the DeFi development community has been actively releasing new on-chain tools including borrowing and lending platforms, yield farms, automated market makers, and other liquidity providing opportunities. All of which was accomplished without the need for intermediates and mostly through open source code.

However, as pointed out by others (Chohan, 2021; Zetzsche et al., 2020), there are considerable risks in the DeFi ecosystem. By October 2021 approximately US$ 1.2 billion of funds have been lost in various attacks and bugs (“Comprehensive List of DeFi Hacks & Exploits,” n.d.). This is despite the very nature of open source auditable contracts.

Indeed, one of the most relevant hallmarks of the DeFi ecosystem is the open source code which defines and executes the financial tools used by the investor. This radical transparency contrasts with the secrecy needed by intermediaries of traditional finance. In DeFi anyone can audit even the largest financial instruments. Investors with enough knowledge or resources can analyze these instruments and decide for themselves whether or not to use them.

Notwithstanding, we are not advocating that all investors in DeFi should know how to read code. Instead, we are highlighting that it is in the best interest of the DeFi ecosystem to provide a means to validate DeFi’s financial instruments. More importantly, it is our objective to make these financial instruments safe to attract investors with no programming knowledge. Furthermore, the infrastructure behind DeFi is flexible enough to provide opportunities for on-chain risk aversion, such as insurance providers and decentralized ratings for risk assessment.

Here, we describe the second version of the Saffron protocol, a tranche scheme for risk management. Through Saffron V2, users have risk-adjusted access to underlying financial products. They can choose the less risky path by buying coverage against failures in the underlying platform or they can earn enhanced yield by taking on more risk. For the risk-averse users, the total amount insured includes both the principal and the interest earned during the coverage period.

## Theory and Implementation

The most general version of Saffron V2 is composed of two tranches and an insurance fund. Below, we will describe the tranches, the insurance fund, and their possible configurations that can be combined to build over ten financial products.

### The Senior and Junior tranches

There are two tranches in Saffron V2, a senior tranche with lower yield and high coverage, and a junior tranche with high yield and limited coverage. The fundamental aspect of the tranche system in Saffron V2 is that the junior tranche provides coverage to the senior tranche while the senior tranche increases the yield of the junior tranche.

Furthermore, the capital in these tranches can be deployed with aggressive or conservative strategies. In the aggressive strategy, the protocol deploys capital in both tranches to the same platform. In case of failure, the insurance fund as well as any recovered funds from the underlying platform will pay the senior tranche back first, and the remainder will cover any outstanding junior tranche losses. We call this the waterfall payback scheme. In the conservative strategy, the protocol only deploys capital to the senior tranche, while funds in the junior tranche remain undeployed. In case of failure, the protocol will use the insurance fund to pay back the senior tranche first and the junior tranche later. If there are not enough funds in the insurance fund, the junior tranche will pay back the senior tranche.

### The Insurance Fund

In every pool, the protocol starts an insurance fund (IF). As we previously described, this fund will first cover the losses using the payback waterfall scheme paying out senior first and junior later. These funds are held in a smart contract that receives yield from the other tranches. There are two types of insurance funds: open and closed.

The Open Insurance Fund (OIF) is open for investors, and its size dynamically oscillates as more or fewer investors join it. These funds receive yield from the capital deployed according to the tranche deployment strategy of the pool. The Closed Insurance Fund (CIF) is closed for investors, controlled by the protocol, and accumulates yield indefinitely.

The insurance fund, in contrast to the junior tranche, offers a different of opportunity with its own set of unique risks. It offers coverage against devaluation of the underlying asset while still exposing yield from that asset. This yield is generally high when compared to the risk and return profiles of other stablecoin yields.

### Saffron V2 pools

Saffron V2 combines both flavors of tranche's capital deployment strategy and the types of insurance funds as described in the previous section (Table 1).

Table 1: Possible strategies of capital deployment in the Saffron V2 pools.
| | Open Insurance Fund | Closed Insurance Fund |
| ----------------------: | :-----------------: | :-------------------: |
| Aggressive deployment | AO | AC |
| Conservative deployment | CO | CC |

The possible configuration parameters outlined above allow us to deploy a total of ten different investment opportunities to a given asset (Table 2).

Table 2: Investments opportunities ranked with arbitrary qualitative units (6 high, 0 low)

| Product | Coverage | Yield | Risk Offset |
| --------: | :------: | :---: | :--------------: |
| AO-Senior | 5 | 2 | Platform failure |
| AC-Senior | 4 | 2 | Platform failure |
| CO-Senior | 6 | 1 | Platform failure |
| CC-Senior | 4 | 1 | Platform failure |
| AO-Junior | 3 | 6 | Platform failure |
| AC-Junior | 1 | 6 | Platform failure |
| CO-Junior | 2 | 5 | Platform failure |
| CC-Junior | 1 | 5 | Platform failure |
| AO-IF | 0 | 4 | Price volatility |
| CO-IF | 0 | 3 | Price volatility |

Together, the Saffron V2 protocol provides a range of financial products that allows the user to offset different risks at different levels with different yields.

### Junior tranche as optional

In some of the Saffron V2 pools, the junior tranche might be closed. First, the lack of the junior tranche simplifies the tranche system for less experienced users who can now choose a simplified high or low risk option. They can either seek coverage or earn more by providing cover using stablecoins in the insurance fund. In addition, the elimination of the junior tranche implies a larger incentive for insurance providers as well as a larger returns for the senior tranche.

### The SAFF-LP token

When a deposit is executed in either junior or senior tranches or the insurance fund, the Saffron protocol issues a position token called SAFF_j, SAFF_s or SAFF_if, respectively. Below, we will detail these tokens and how the algorithm calculates their price based on an exchange rate.

$$
\mathtt{SAFF}_{\{s,j,if\}} = \frac{\{S,J,IF\}}{X_{\{s,j,if\}_ｔ}} \qquad(1)
$$

where $S$ and $J$ are the amounts deposited, in units of the underlying asset, in the senior and junior tranche respectively, $IF$ is the amount deposited in the insurance fund in the stablecoin defined by the contract and $X_{\{s,j,if\}_t}$ is the rate of exchange between the position tokens and the underlying asset at the time of deposit $t$. For example,

### Evaluation of the SAFF_{s,j,if} exchange rates

The exchange rates are updated at the beginning of every deposit or withdrawal routine. The protocol calculates the exchange rates at time $t_1 > t_0$ according to the following:

First, it calculates the earnings’ increment, $\Delta e$, by comparing the difference of the total amount locked in the underlying platform at time $t_1$ since time $t_0$.

$$
\Delta e = T_{t_1} - T_{t_0} \qquad(2)
$$

where $\Delta e$ is the total earnings' increment and $T$ is the total number of tokens deployed. In Saffron V2, depending on the configuration of the pool, sometimes only the senior tranche is enabled.

We use tranche multipliers $m_s$, $m_j$, $m_{if}$ to calculate the earnings' increment fraction of each tranche:

$$
\begin{array}{l}
\Delta e_{s} = m_s * \Delta e \\
\Delta e_{j} = m_j * \Delta e \\
\Delta e_{if} = m_{if} * \Delta e
\end{array} \qquad(3)
$$

where $\Delta e_{\{s,j,if\}_{t_1}}$ is the earnings' increment of each tranche between time $t_0$ and $t_1$. If there is a significant amount of earnings during a status update, the protocol recalculates the exchange rate similarly to how Compound calculates cToken exchange rates (Leshner and Hayes, 2019).

$$
\begin{array}{l}
X_{s_{t_1}} = X_{s_{t_0}} + \dfrac{\Delta e_s} {T_{{\mathtt{SAFF}_s}_{t_0}}} \\
X_{j_{t_1}} = X_{j_{t_0}} + \dfrac{\Delta e_j} {T_{{\mathtt{SAFF}_j}_{t_0}}} \\
X_{if_{t_1}} = X_{if_{t_0}} + \dfrac{\Delta e_{if}} {T_{{\mathtt{SAFF}_{if}}_{t_0}}} \\
\end{array} \qquad(4)
$$

where $T_{{\mathtt{SAFF}_{\{s,j,if\}}}_{t_0}}$ are the total supply of each SAFF token issued at the time $t_0$. Importantly, when $T_{{\mathtt{SAFF}_{\{s,j,if\}}}_{t_0}} = 0$, the respective exchange rate is set to 1. Since the exchange rates are updated at every status change, we can calculate the exchange rates for any given time $t$ since time $t_0$ of deposit as:

$$
\begin{array}{l}
X_{s_{t}} = X_{s_{t_0}} + \sum_{i=1}^{t} \dfrac{\Delta {e_s}_{i}} {T_{{\mathtt{SAFF}_s}_{i-1}}} \\
X_{j_{t}} = X_{j_{t_0}} + \sum_{i=1}^{t} \dfrac{\Delta {e_j}_{i}} {T_{{\mathtt{SAFF}_j}_{i-1}}} \\
X_{if_{t}} = X_{if_{t_0}} + \sum_{i=1}^{t} \dfrac{\Delta {e_{if}}_{i}} {T_{{\mathtt{SAFF}_{if}}_{i-1}}} \\
\end{array} \qquad(5)
$$

where $\Delta e_{\{s,j,if\}_{i}}$ is the earnings' increment of each tranche between time $i$ and $i-1$ and $T_{{\mathtt{SAFF}_{\{s,j,if\}}}_{i}}$ are the total supply of each SAFF token at the time $i$.

### Performance Fee

There are no premiums, deposit fees, or withdrawal fees in the Saffron V2 protocol. Instead, the protocol collects a performance fee as a percentage of the earnings. The collected fees are converted and sent to the Treasury when called by Governance.

Fees emerge as a mathematical consequence of the multipliers described above.

$$
m_s + m_j + m_{if} < 1 \qquad(6)
$$

Thus, the fees are simply calculated as the remainder of the earnings:

$$
f_{t_1} = f_{t_0} + (\Delta e - \Delta e_s - \Delta e_j - \Delta e_{if})
\qquad(7)
$$

### Redeeming the SAFF_{s,j,if} tokens

Both tranches SAFF_{s,j,if} are locked by default. To redeem, the owner needs to request it and wait one week until the funds can be withdrawn. The price of the SAFF token is set at the time, $t$, of the request, and the total amount to be received is calculated as:

$$
\{S,J,IF\}_{t} = X_{\{s,j,if\}_{t}} \mathtt{SAFF}_{\{s,j,if\}} \qquad(8)
$$

### After a claimable state

In the event of a failure that leads to a claimable state the exchange rates are set according to the following protocol.

Let $A$ be the recovered amount of the underlying asset plus the capital that has not been deployed. Let $X_{\{s,j,if\}_f}$ be the last exchange rates prior to the failure event. These exchange rates are states of the Saffron V2 pool and do not need to be calculated again. Instead, they are set variables for the contract to calculate the new exchange rates using the waterfall protocol according to the following algorithm:

// X_s_f, X_j_f and X_if_f are given by the pool at the time of failure

let rest = A; // amount recovered
let X_s_c = 0;
let X_j_c = 0;
let X_if_c = 0;

// calculates tentative exchange rate for senior
X_s_c = A / SAFF_s;
// if it is equal or smaller than the rate pre failure return,
// if not, calculate the rest
if (X_s_c <= X_s_f) return;
// sets the exchange rate post failure as pre failure for senior
X_s_c = X_s_f;
// calculates the amount of recovered funds left
rest -= X_s_c * SAFF_s;
// calculates the junior exchange rate
X_j_c = rest / SAFF_j;
// if it is equal or smaller than the rate pre failure return,
// if not, calculate the rest
if (X_j_c <= X_j_f) return;
// sets the exchange rate post failure as pre failure for senior
X_j_c = X_j_f;
// calculates the amount of recovered funds left
rest -= X_j_c * SAFF_j;
// calculates the exchange rate for the insurance fund
X_if_c = rest / SAFF_if;

The pool is open for withdrawal after setting the exchange rates as demonstrated above. The amount to be received for each user can be calculated based on the amount of SAFF_{s,j,if} they held using equation (8). It is worth notice that in the Saffron V2 protocol, the insurance policy includes the profits acquired during the lock period.

### Cost of Insurance

The coverage offered by Saffron V2 pools has a cost that can be conceptually conceived as the difference between the expected returns of deploying the capital to the underlying asset alone and deploying it to the insured tranches.

Let Alice deposit her capital $C_A$ into the senior tranche at time $t=0$ when the exchange rate of the senior tranche is ${X_{S}}_0$. She wonders how much the coverage costs because Saffron V2 pools do not have premiums. The cost of insurance is calculated from the yield that the tranche has in comparison to the underlying platform it is insuring. It can be seen as a performance fee.

To calculate the cost of insurance we can calculate the return of the senior tranche using equations (1), (5), and (8):

$$
{C^\prime_A}_S = \mathtt{SAFF}_{S}{X_{S}}_t = C_A \frac{{X_{S}}_t}{{X_{S}}_0} = C_A \left (1 + \frac{m_S}{{X_S}_0} \sum_{i=1}^{t} \frac{\Delta e_i}{{T_S}_i}\right ) \quad(9)
$$

Thus, to calculate Alice's earnings:

$$
\Delta {C^\prime_A}_S = {C^\prime_A}_S - C_A = C_A \frac{m_S}{{X_S}_0} \sum_{i=1}^{t} \frac{\Delta e_i}{{T_S}_i} \quad(10)
$$

and relative earnings in relation to her initial deposit:

$$
\frac{\Delta {C^\prime_A}_S}{{C_A}_S} = \frac{m_S}{{X_S}_0} \sum_{i=1}^{t} \frac{\Delta e_i}{{T_S}_i} \quad(11)
$$

similarly, we can calculate what her return would have been had she deposited directly into the underlying platform rather than through Saffron:

$$
\frac{\Delta {C^\prime_A}}{C_A} = \frac{1}{X_0}\sum_{i=1}^{t} \frac{\Delta e_i}{T_i} \quad(12)
$$

Thus the cost of insurance is the difference from what she could have received from the underlying platform directly versus what she will receive from the senior tranche given the same performance of the underlying platform.

$$
\mathtt{cost Of Insurance} = {\Delta C^\prime_A} - \Delta {C^\prime_A}_S = C_A \left (\frac{1}{X_0}\sum_{i=1}^{t} \frac{\Delta e_i}{T_i} \right ) - C_A \left (\frac{m_S}{{X_S}_0} \sum_{i=1}^{t} \frac{\Delta e_i}{{T_S}_i} \right ) \quad(13)
$$

Here, $X_0 = {X_S}_0$ and $T_i = {T_S}_i$ as it emulates an alternative single tranche with the same characteristics as the senior tranche but with a multiplier of 1.

$$
\mathtt{cost Of Insurance} = C_A \left (\frac{1}{{X}_0} \sum_{i=1}^{t} \frac{\Delta e_i}{{T}_i} \right ) ( 1 - m_S )\quad(14)
$$

Thus, the relative cost of insurance of a Saffron V2 pool is proportional to $1-m_S$. We can futher calculate the relative cost of insurance per insured amount. The coverage provided by the junior tranche is relative to the size of the tranches which varies according to market conditions. In a situation with balanced tranches, the covered percentage is the ratio cap between the two tranches, $\kappa = \frac{J}{S}$. Thus the relative cost of insurance per unit of capital covered is proportional to $\kappa(1-m_S)$.

## Study case

### Saffron V2 CO pool and failure of the underlying platform

Let AcmeLending be a lending platform with a pool using their token ACME. AcmeLending is new and its pool has high yield. Let the Saffron-AcmeLending-CO pool be a Saffron V2 CO pool using the AcmeLending protocol as the underlying platform. The Saffron-AcmeLending-CO pool has a cap on its junior tranche set to $1/5$ of the senior tranche. As previously described, Saffron V2 CO pools have an insurance fund open to deposits. This pool only deploys the capital deposited in the senior tranche to the underlying protocol, and yield earned is distributed according to the multipliers found in Table 3.

Table 3: Example of yield multipliers in the V2 Saffron protocol
| Multiplier | Value |
| :--------: | :---: |
| $m_s$ | 0.150 |
| $m_j$ | 0.600 |
| $m_{if}$ | 0.125 |

Now, let Alice, Bob, and Carol be users with different risk profiles. Alice is a conservative user, she is impressed by the yields in the AcmeLending platform, but she is skeptical given how new the protocol is. Bob is an aggressive user and is interested in the highest yield possible. He believes that AcmeLending has a low chance to fail but thinks that the yield of the platform is not high enough. Carol is also an aggressive user that completely trusts that the AcmeLending protocol will not fail. However, Carol believes that the ACME coin will lose its value against stablecoins, so she does not want to invest in the underlying asset of the AcmeLending pool, although she appreciates the relatively high yield on its platform.

Considering this scenario, let’s suppose that the APR of the AcmeLending ACME coin pool is 12%. For simplicity, let’s assume that Alice, Bob, and Carol all had their initial deposits at the same time. Fitting her profile, Alice deposits 1000 ACME to the senior tranche of the Saffron-AcmeLending-CO pool and receives 1000 ${\mathtt{SAFF}_{s}}$ tokens when the exchange rate was 1. Bob can only deposit 200 ACME to the junior tranche because of the cap and got back 200 ${\mathtt{SAFF}_{j}}$ tokens. Finally, Carol adds 300 stablecoins to the insurance fund to get 300 ${\mathtt{SAFF}_{if}}$ tokens. Table 4 summarizes the initial condition of the pool.

Table 4: Initial position and deposited asset for each user
| User | Position | Deposited asset |
|:-:|:-:|:-:|
|Alice | 1000 ${\mathtt{SAFF}_{s}}$ | ACME |
| Bob | 200 ${\mathtt{SAFF}_{j}}$ | ACME |
| Carol | 300 ${\mathtt{SAFF}_{if}}$ | Stablecoin |

The CO pools only deploy the senior tranche assets to the underlying platform. Thus, after a month, there are 10 ACME coins earned in the Saffron-AcmeLending-CO pool. To update the new exchange rate, let’s use the set of equations (3) and (4) assuming a 1:1 exchange rate between ACME and stablecoins (Table 5).

Table 5: Earnings and Exchange rate calculated for each tranche and IF.
| Tranche | Earnings | Exchange rate |
|:-:|:-:|:-|
| Senior | 1.5 | $1 + \dfrac{1.5}{1000} = 1.0015$ |
| Junior | 6.0 | $1 + \dfrac{6.0}{200} = 1.03$|
| Insurance Fund | 1.25 | $1 + \dfrac{1.25}{300} = 1.004$|

In the scenario where each user was to withdraw their funds, their SAFF token value can be calculated using equation (7). In addition, the effective APR for each position can be calculated by taking the earnings in the percentage of the initial deposit and multiplying it by 12, Table 6.

Table 6: Earnings for each user after one month
| Investor| Initial position | Withdraw amount | Withdrawn asset | Effective APR |
|:-:|:-:|:-:|:-:|:-:|
|Alice | 1000 ${\mathtt{SAFF}_{s}}$ | 1001.5 | ACME | 1.8% |
| Bob | 200 ${\mathtt{SAFF}_{j}}$ | 206 | ACME | 36% |
| Carol | 300 ${\mathtt{SAFF}_{if}}$ | 301.25 |stablecoin | 5% |

The sum of these positions adds up to 1508.78 ACME coins, while the total holding of the pool is 1510 ACME coins. The difference, 1.25 ACME, is collected as Governance fees. We can see that this amount equals the fees calculated using equation (7), explicitly:

$$
f = 10 - 10(0.15 + 0.6 + 0.125) = 10(1 - 0.875) = 10 * 0.125 = 1.25
$$

### Claimable state

Let’s assume that a catastrophic failure of the AcmeLending protocol triggers the claimable state. In this case, all positions remain locked because governance sets the exchange rates to all tranches and IF to zero. Also, all the exchange rates are set according to the calculations in the waterfall protocol.

Let’s assume that 50% of the deployed capital was recovered, $1010 / 2 = 505$. This means that the total holdings of the pool are now the sum of the recovered capital, the undeployed capital from the junior tranche, and the IF: $505 + 200 + 300 = 1005$. This is set as the total recovered assets. The protocol follows the algorithm previously described to calculate the claimable exchange rates for tranches and IF.

First it calculates the tentative ${X_{s}}_c = 1005 / 1000 = 1.005$ which is larger than ${X_{s}}_f = 1.0015$. In this case, the protocol sets ${X_{s}}_c = {X_{s}}_f$ and calculates the remainder of the funds available as $R = 1005 - 1000 * 1.0015 = 1005 - 1001.5 = 3.5$.Next, it proceeds to calculate the tentative exchange rate for the junior tranche as ${X_{s}}_c = 3.5 / 200 = 0.0175$ which is less than ${X_{s}}_f = 1.03$ which makes the algorithm to return without changing the IF exchange rate. At this point, Alice has all her principal and gains covered by the junior tranche and IF. Bob has lost 98.25% of his principal and all of his gains, and Carol lost all her funds, Table 7.

Table 7: Final state after a failure in the underlying asset
| User | Initial portfolio | Final portfolio | Asset | Principal recovered (%) | Gains recovered (%) |
|:-:|:-:|:-:|:-:|:-:|:-:|
|Alice | 1000 | 1001.5 | ACME | 100% | 100% |
| Bob | 200 | 3.5 | ACME | 1.75% | 0% |
| Carol | 300 | 0 | stablecoin | 0% | 0% |

From this state, the pool will be permanently closed for new deposits, and the exchange rates will not be updated. All $\mathtt{SAFF}$ tokens with a larger than zero exchange rate can be redeemed.

### Cost of Insurance

Interestingly, using equation (13) Alice theoretically paid $10 (1 - 0.15) = 7.5$ ACME for the insurance. Even if none of the deployed capital were recovered, the junior tranche and IF would have provided insurance for half of Alice’s capital. In other words, in this example Alice recovered 500 ACME - close to half of her capital - paying only 7.5 ACME, which is equivalent to 0.7% of the total capital deployed and only 1.5% of the capital recovered.

## Price protection

In the previous examples, we considered that the exchange rate of the pair ACME/stablecoin was invariant over the course of the analysis. Let's now consider, instead a failure of the underlying platform, a sudden rise of 25% in the ACME/stablecoin pair, which means a devaluation of 25% of the ACME coin against the stablecoin. For that, let's consider that the initial exchange rate of the pair was ${(A/S)}_0 = 1$ and that at the time of withdraw, it was ${(A/S)}_f = 0.75$. In other words, 1 ACME is equivalent to 0.75 Stable at the time of widthdraw.

In this scenario, it is important to calculate the potifolio of Alice, Bob and Carol in units of the stablecoin, at the time of deeposit and after widthdraw, Table 8.

Table 8: Evolution of Alice, Bob, and Carol's portfolios (in units of stablecoins) in the scenario of a devaluation of ACME relative to stablecoins.
| User | Initial portfolio | Final portfolio | Change |
|:-:|:-:|:-:|:-:|
| Alice | 1000 | 751.12 | -24.89% |
| Bob | 200 | 154.5 | -22.75% |
| Carol | 300 | 300.9375 | +0.31% |

It is worth noticing that in this scenario, the gain acquaired by the IF is in ACME and therefore suffers a significant devaluation against the Stable. In this example, Carol's gain of equivalent of 1.25 ACME now is worth 0.9475 Stable which thus still leads to an effective APR of 3.75%. Also, that some of the price-related loss in the capital of the other tranches was mitigated by the earnings mechanism. Overall all tranches performed better than locked ACME with zero interest.

## Discussion

The Saffron V2 pools offer flexibility in covering risk management in DeFi. The pools encoded by the Saffron V2 protocol are financial tools to offset the risks of capital allocated to external DeFi protocols. More specifically, it tackles two types of risks: 1) platform failure and 2) price volatility. It works by separating the yield of the underlying assets into two tranches. The senior tranche forfeits part of their return to boost the return of the junior tranche, which in turn, provides coverage to the capital on the senior tranche. In addition to this mechanism, Saffron V2 pools are coupled to an insurance fund. This new element serves to lower the price of coverage for senior tranches and may offer coverage to junior tranches. The Closed Insurance Fund builds up from the yield of the tranches. The Open Insurance Funds builds up from users who provide this extra coverage for the tranches in exchange for a high yield in stablecoins. For the insurance fund, this is an opportunity to receive some of the high yield from the underlying platform while offsetting the price volatility risk of the underlying asset.

These financial tools provide a range of opportunities for different risk profiles. Although junior tranche and Insurance Fund are vehicles intended to provide high yield, Saffron’s main goal is to provide coverage to conservative users. In addition, Saffron's aggressive coverage policy includes the gains acquired during the period capital is deployed.

## Conclusion

In this whitepaper, we presented the Saffron V2 protocol. We demonstrated how the system of tranches paired with an Insurance fund provides a range of financial tools to offset different types of risks and high yield opportunities. Contrary to the Saffron V1 protocol, the second version introduced here allows for continuous capital deployment. Also, it can easily deployed against a variety of underlying protocols in the DeFi ecosystem.

## Acknowledgements
We thank the Saffron team for their insightful comments, in particular for the several rounds of discussion with Darrow, Dingo, Maze, Tupelo, and Turpintinz.

## References

* Blockchain, W., Project, D.A., 2021. DeFi Beyond the Hype 20.
* Chohan, U.W., 2021. Decentralized Finance (DeFi): An Emergent Alternative Financial Architecture (SSRN Scholarly Paper No. ID 3791921). Social Science Research Network, Rochester, NY. https://doi.org/10.2139/ssrn.3791921
* Comprehensive List of DeFi Hacks & Exploits, n.d. . CryptoSec. URL https://cryptosec.info/defi-hacks/ (accessed 10.6.21).
* Leonhard, R., 2019. Decentralized Finance on the Ethereum Blockchain (SSRN Scholarly Paper No. ID 3359732). Social Science Research Network, Rochester, NY. https://doi.org/10.2139/ssrn.3359732
* Leshner, R., Hayes, G., 2019. Compound: The Money Market Protocol.
* Schär, F., 2021. Decentralized Finance: On Blockchain- and Smart Contract-Based Financial Markets (SSRN Scholarly Paper No. ID 3843844). Social Science Research Network, Rochester, NY. https://doi.org/10.20955/r.103.153-74
* Zetzsche, D.A., Arner, D.W., Buckley, R.P., 2020. Decentralized Finance. J. Financ. Regul. 6, 172–203. https://doi.org/10.1093/jfr/fjaa010