- a decentralized lending platform launching a Cosmos-based zone in September 2019.
- multi-collateral CDPs from day one; governance decides supported assets
- validators make governance decisions about critical system parameters
- KAVA token is staked, acts as the lender-of-last-resort, and used for governance
- “cross-chain first” principles — value prop grows w the number and liquidity of supported assets
- working w Binance Chain to create CDP functionality for BNB and BTC-B
- nodes may also participate in the oracle process
- badges earned via third-party validator audits to qualify for delegations from Kava Labs Inc.
First off, hello! And thank you to Figment for hosting this. I’m Kevin Davis, lead engineer at Kava [Labs Inc]. My background is as a researcher and data scientist focused on agriculture and invasive species up until 2017. Shortly before completing grad school I started researching and discussing blockchain with Scott and Ruaridh. A few months later we, along with our COO Brian Kerr, formed the company that became Kava.
Kava Labs Inc has focused on cross-chain settlement for the past 18 months, and considered many solutions circa 2017/2018 before discovering Interledger. The Kava team launched Switch, an Interledger-based product, in early Q2 2019.
Switch is an open-source, decentralized exchange app focused on BTC, ETH, and XRP. It works a little differently than most DEXes. To be clear, Switch is a separate product from the upcoming Cosmos zone, Kava.
While working on ILP (Interledger Protocol), the Kava team grew close with the Cosmos team. Despite being skeptical, the Kava team was consistently impressed by the Cosmos team’s practical design choices and ability to ship. They shared a common vision of blockchain interoperability, via architecture proposed by the Cosmos whitepaper and IBC (inter-blockchain protocol). Fast forward to the beginning of this year, where Kava Labs began researching and prototyping a CDP (collateralized debt position) system for Cosmos — what is now the present-day Kava zone slated to launch in September 2019.
Kava will be a decentralized lending platform, beginning with over-collateralized CDPs that are similar to MakerDAO. However, Kava is designed with “cross-chain first” principles, meaning that Kava’s value proposition grows with the number and liquidity of supported assets.
When asked by Hyung (B-Harvest) about plans to support connection and auto-exchange functionality with “big-corp crypto” like Libra, Kevin Davis felt that Libra is and will remain a moving target for some time to come.
Initially, Kava Labs are targeting chains that don’t natively support DeFi (decentralized finance), as well as synthetic and native assets created by zones within the Cosmos ecosystem. More specifically, they will be working with the decentralized exchange, Binance Chain, to create CDP functionality for BNB and BTC-B.
Multi-collateral CDPs will be supported from day one, and asset inclusion will be decided via governance, Kevin informed Henrik (Validator Network; e-Money). However, Kava Labs is initially targeting staked ATOM, XRP, and BNB via Binance Chain.
Henrik also wondered how Kava will avoid the large stability fees that Maker currently has. Kevin explained that the CDP design for Kava is inspired by Maker’s multi-collateral DAI (distinct from single-collateral DAI on Ethereum’s mainnet today).
Kava’s vs. Maker’s CDP design
- Governance participation rates:
CDPs will support any type of collateral that is decided upon by governance votes (similar to those on the Cosmos Hub). Kava Labs expects governance participation rates similar to Cosmos, and much higher than Maker’s.
- ‘CDP stability fee’ to be paired with a USDX savings rate:
Kava Labs believes that this will create a better demand equilibrium for the stablecoin than exists with DAI today. Combined with appropriate risk management (ie. slowly increasing the debt ceiling of USDX, not launching with an implausibly low stability rate), they think the stability fee will be both lower and more stable.
- KAVA tokens may be inflated in the case of accumulated bad debt:
Collateral auctions for seized collateral involve a) selling seized collateral for USDX and b) selling KAVA for USDX. This means that KAVA can be inflated in the case of accumulated bad debt, unlike Maker as it exists today.
USDX: the CPD-backed Stablecoin
USDX may be bonded and earn interest, similar to a staking asset, whereas unbonded USDX will forego this interest. Kevin found it hard to say what the USDX unbonding period will be, but it will be a governance parameter. The Kava team will recommend the same unbonding time as that of staking at mainnet launch — 21 days.
And bonding doesn’t have to constrain liquidity. cDAI, for example, provides liquidity for bonded DAI, and is gaining popularity is the ETH DiFi stack. We may expect to see something similar with USDX.
According to Scott, KAVA token sales have only involved validating organizations to date. He wouldn’t dive into specifics about the token allocations that equity-investors will be entitled to, however, he assured us that “their dividend will be consistent with the standard ‘token awareness’ clause found in most SAFEs in this industry.”
The KAVA token’s value proposition is its associated use in the system. When users create CDPs and accumulate interest, they repay that interest using KAVA, which is burned. Unlike MKR (the Maker DAO token) as it exists today, the KAVA token
1) is directly tied to the function of the platform as its native staking token, and
2) the value proposition is two-sided, because KAVA acts as the lender-of-last-resort in the case of bad debt accumulation.
Simply put, KAVA rewards its holders for good governance decisions that increase the usage of the platform, and presents downside risks if governance decisions lead to failures from failed collateral auctions or security breaches.
The team thinks that the Cosmos-based Kava zone will present new validator opportunities: a complex, multi-faceted governance system that enables differentiation.
Governance & the Roles of the Validator
Beyond “defining what to do at the boundaries of risk within the system,” governance systems like Kava may allow for significant differentiation in staking services for validators, according to Scott.
Though the mechanisms are roughly similar to that of Cosmos, governance of a DeFi lending system can be somewhat nuanced. Entrusting a validator to make good decisions about parameters in the system creates an opportunity for validators to differentiate beyond the current factors of liveness, security, and other technical metrics.
Kevin imagines a scenario where delegation decisions are dominated by governance considerations, balancing efficiency with risk control, with attention to security and commission rates being secondary. We should expect risk parameters and expectations to be explicit as Kava goes live, with the aim of starting slow and being risk-averse while the system is in its infancy.
Hyung (B-Harvest) pointed out that, beyond the Kava protocol, there should be a well-written governance procedure for situations involving expected risks, so that ecosystem participants can more calmly adopt prepared recovery processes. “I think we can provide some contribution on something like a constitution documentation for contingency situation governance procedure.”
Hyung wanted to know Kava’s risk-management strategy:
“Do you have any doc how you handle macro risk, stress testing scenarios, and contingency plans?”
Kava Labs Inc will produce more content around the mechanics of the system, but Scott thinks that it is likely a great time for validators interested in Kava to begin publishing their own materials, to cement their brand as ‘knowledgeable about the governance of the platform.’
Regarding the low participation rate for MakerDAO governance, Kevin stated that there’s pretty good evidence that native staking tokens increase governance participation, at least to date. In general, he thinks that tying governance to the validator role is useful (if not sufficient) for overcoming low participation rates.
Kava will begin with 100 validator slots, and since validator slots are now a governance parameter, that should be easy to adjust. The team’s goal is to onboard participants and convert them to long-term network participants — both in terms of staking and governance. Besides governance and validating, nodes will also be able to participate in the oracle process as a task for the Kava network.
We expect that liveness, security, and other parameters will mirror those of the Cosmos Hub: 21-day unbonding period, 5% bond slash for double-signing a block, 0.01% slash for failing to sign 5% of 10,000 blocks.
Third-Party Validator Audits
In general, Kava Labs will focus on development of the platform and plans to delegate, rather than to stake directly. How can validators win delegations from Kava Labs? Scott says that the team has plans to run third-party audits on validator setups, qualifying validators for a badge and enabling them to be delegated to by Kava Labs.
The Cosmos-based Kava zone is slated to launch in September 2019. The team published launch procedures for Kava’s second testnet the day after our AMA, on July 26. Keep an eye on Kava Labs’ Medium and join the validator Riot chat for more updates
Special thanks to Kevin and Scott for spending an hour with Staking Hub to answer our many questions. Thank you to Andrew Cronk for inviting the Kava Labs team and for organizing this AMA. Thanks to our Staking Hub community for the thoughtful and impactful questions that inspired high-quality answers. Since you’ve read this far, you might as well join us over in the Staking Hub Telegram channel :)
Hopefully you found this useful. Feedback is always welcome! I’m on Twitter.
Post-AMA Discussion (paraphrased)
Has it been determined what the USDX placeholder name of the stablecoin will be?
When staking with Kava, do you pay back in XRP or fiat? So if i borrow $50,000 and it costs 17,000 XRP, is that what I owe? If the price of XRP increases to $100, do I owe 500 XRP?
I think staking is different than taking out an over collateralized loans (CDPs). When you take out a CDP, you put in $150 worth of XRP and get $100 worth of USDX stablecoin. If the underlying XRP goes down in value, say to $125 worth, the loan is liquidated. So you’d lose your XRP but you still have the USDX.
It’s like how margin trading works in exchanges. Same logic. Can you deposit additional coins to the loan in order to keep your collateral? In case its value is decreasing?
In Maker, yes. Not sure about Kava.
Late to the party but here’s my take. Very cool idea, but I remain skeptical of validators’ ability to collectively manage the debt. While the participation rate will likely be exponentially higher than MKR’s, due to the coupling of governance duties with traditional PoS validation (and thus the inflationary block rewards), that isn’t inherently good. Maker’s governance is an example of even a concentrated group of supposed experts having trouble managing the parameters of CDPs for a single collateral type.
Inflationary block rewards create an incentive to stake, even if the validator has no risk management knowledge. And they’re now required to vote on parameters like collateralization ratios… on multiple different collateral types in tandem. Will definitely, at the very least, experience many hiccups. Governing slashing penalty and validator set size parameters are a lot different than this.
Scott, curious your thoughts. Nonetheless a very exciting project.
Do you think there’s a better governance incentive alignment? Agree that that doesn’t mean more knowledgeable constituents per se, but a better reason to try to “get it right”?
I think isolating the governance rights (with the inflationary auction in the same underlying token) *probably* represents better incentive alignment, but that’s not to say it isn’t worth experimenting with.
Yep, agree. Will be exciting to see the Cosmos governance software system evolve. Already good ideas and movement in first 12 Hub proposals.
Jeez I’m so excited to see this now. And this is the first time I’m seeing this. Switch looks like a solution to a business case that has been running in my head…