#DeFi Summit — San Francisco
Last Thursday, #DeFi organized our first summit, a one-day community event comprised of collaborative strategic discussions on how to best build the open financial system. We hosted the event on the eve of ETHSanFrancisco, bringing together the enthusiasts, builders, and subject matter experts necessary for creating the open financial system we all strive toward. The end result was nothing short of spectacular: with topics ranging from wallet UX design to on-chain financial instruments, the caliber of the conversations blew us away. This blog post is meant to serve as an overview of the day’s meetings, providing you with summaries of the talks as well as more information behind each team.
Challenges to Trading on DEXs
0x enables peer-to-peer cryptocurrency trading by providing a protocol for decentralized exchange. Any organization can build a relayer on top of the protocol to coordinate market makers with takers, eliminating the concerns over custodial and counterparty risks associated with centralized exchanges. The majority of existing liquidity on the protocol is comprised of ERC20 tokens, but the release of v2 has led to ERC721-enabled relayers.
While tightening spreads are indicative of the protocol’s success, the 0x team understands that future growth will be predicated on support for a greater variety of asset categorizations. Security tokens represent the most compelling opportunity for expansion, due to the sheer volume represented by institutional investors. KYC compliance may be the elephant in the room when it comes to security tokens and DEXs, but the advances being made by Harbor and Wyre with on-chain solutions suggest that such models will be feasible within a greater decentralized finance stack. Greater options for stable coin pairings will also contribute to increased volume and liquidity.
Blockchain vs. Traditional Lending
Within traditional finance, money markets provide banks, companies, and mutual funds with access to short term lending. Compound has created a crypto-native equivalent that allows for lenders to put up ETH, REP, BAT, and ZRX tokens for block-length durations. Borrowers primarily use these loans for trading, either for incremental liquidity to catch dips or for assets which to then short sell. While the volatility of crypto assets is far greater than the collateral used for traditional money markets (e.g. treasury bonds, commercial bills), Compound addresses this issue with forced over-collateralization.
The Compound team faced another design obstacle in Ethereum’s on-chain computation limitations. Because constantly calculating interest rates for individual loans requires significant overhead (i.e. smart contracts can’t call themselves and block-lengths are constantly varying), the protocol serves as the counterparty for all loans and sets global interest rates on an asset-by-asset basis. To incentive the supply side of the protocol, interest rates on borrowing far outpace those on lending.
By keeping the interest rate model on-chain, Compound keeps the system transparent to both the supply and demand sides of the market. When asked about the team’s concern regarding someone forking their smart contracts, the consensus was that the building of transaction logging and server-accessible API layers on top of them makes that far less likely.
Wallets as the Gateway
Wallets are more than just a tool to hold your crypto assets — we should consider them the gateways to the decentralized web. Coinbase’s customer research has confirmed just that, with surveyed individuals ranking interest-bearing financial instruments, decentralized exchange, and easy remittances at the top of their wish lists. The team’s workshop at the DeFi summit demonstrated that whether the value proposition behind designing a wallet is investing, trading, or borrowing, identity remains most the significant UX hurdle. It’s unlikely that wallet experiences that require reading hashed address and signing digital signatures will usher in true mainstream crypto adoption. Considering this, it’ll be the projects that understand the importance of common identity and human-legible abstraction that design popular gateways to crypto.
Origin Protocol is a platform for creating sharing economy marketplaces within decentralized networks. The team at Origin knew early on that they would have to find a solution to the issue of trust at the core of markets like AirBnb and Uber that did not depend on centralized review or dispute resolution systems. They opted for the ERC725 token standard, first proposed in October 2017 by Fabian Vogelsteller, which allows for the representation of identities via smart contracts. As a result, users can transact across the P2P markets built upon the Origin Protocol with their common identities and foster their reputation through iterated transactions.
The benefits of integrating with ERC725 are clear: with a universal login, user experiences become streamlined and platforms gain immediate access to existing networks. While users will want to provide different platforms with varying levels of permissioned access to their identity, recent pushes for KYC processes provide a compelling near-term use case for the standard.
Challenges to Unsecured Lending
Dharma is building a protocol for borrowing and lending — crucial infrastructure for capital markets. The protocol issues and collects debt in the form of ERC721 non-fungible tokens and, like 0x, allows for anyone to build applications on top of its Ethereum-based smart contracts. In Dharma Protocol, underwriters perform the crucial task of evaluating the credit risk of a borrower. There aren’t many underwriters in the ecosystem today, so borrowers are forced to over-collateralize their loans. Eventually, underwriters will develop the standardized risk rating schemes necessary for unsecured lending.
The problem of developing a decentralized credit scoring mechanism remains Dharma’s most pressing concern. With little infrastructure providing identity or reputation in the crypto space, contemporary solutions like Bloom are forced to go off-chain through traditional channels to determine initial baselines. Even with an on-chain identity system, vulnerability to Sybil attacks could remain a weakness. The goal for now is to start with a few trusted entities who will attest to identity, and eventually switch to a multi-tiered architecture with more at stake for underwriters.
Paradigm removes the need for centralized order relayers by providing an exchange protocol with on-chain logic capable of supporting spot rates, loans, derivatives, and more. The exchanges built on Paradigm simply serve as a front-ends with order matching engines. With this model, developers can build their own settlement logic that any exchange can utilize, and then receive fees proportionate to their contract’s usage. This model is a departure from 0x’s, in which centralized relayers build on the protocol, and closer to Numerai, where developers design strategies and compete for network share.
By allowing for liquidity to flow on the basis of standards rather than relationships, the protocol provides for more efficient markets with lower spreads and fewer arbitrage opportunities. The drawbacks to this design, however, lie in countering order front-running and incentivizing market makers. As Paradigm attracts more users, it’ll be interesting to note whether the gains in efficiency outweigh the demand for centralized relayers.
Better Mobile Wallet UX
Despite being the primary interface for crypto holders, wallets often relegate user experience to an afterthought. Between gas payments, key management, and the need for web3 extensions, wallets often lack the intuitive design required to onboard new users. This friction stems from the fact that web3 design has shifted the burden of signing and writing transactions from a server to the user. Combine that with confusion over Metamask usernames (why do I have to sign transactions if I’m already logged in?) and unclear transaction costs, and now you’ve raised barriers to entry for the vast majority of people.
A possible solution for wallet interactions lies between two Ethereum standards: ERC1077 and ERC725. ERC1077 allows for “meta-transactions” in which the user signs their intent for a certain transaction, but delegates the actual execution to another address. With ERC725, users can assign their identity to a specific smart contract address. Combine the two, and now users can log-in to their wallets with traditional username + password access, and simply sign transactions to be executed by their respective ERC725 smart contracts. This abstracts away the need for users to interact with dApps at the layer one level.
Wyre considers on-chain identity to be the next major driver of crypto usage. This is especially true in the decentralized finance ecosystem, where reputation scoring will be key in issuing loans and other financial instruments. Accordingly, their platform provides major exchanges and OTC desks with an API for KYC compliance. This same functionality will soon allow dApps to deploy seamless fiat on- and off-ramps, making it far easier to onboard users without the need for web3 extensions or crypto wallets.
Protocol vs. Product
dYdX is a decentralized protocol for financial derivatives and margin trading. Unlike other projects in the #DeFi space, however, it has taken a product-first approach. Whereas a project like 0x has focused on providing developer tooling and open source libraries for relayers, the dYdX team has put all its effort toward an end user product. This mindset comes from the belief that the early user experience is crucial in bootstrapping a network; if you don’t find the market fit you’re expecting, there’s always room to pivot. The current dYdX user experience centers on Expo, their recently-launched platform that lets users purchase short-Ether.
The team also opted to raised funding through equity, rather than through token sales, to preserve this flexibility. Because protocols require either tokens or fees to capture the value they create, it becomes ill-advised to implement a token before product-market fit is even determined. A lack of flexibility can prove deadly, especially when dealing with open source protocols that derive 100% of their value from network effects and customer moats.
Crypto in Emerging Markets
Celo is a collateralized stable coin project that allows users to map their phone numbers to wallet addresses. While their protocol is still being built out, they’ve conducted a case study in Kenya to both confirm demand for mobile payments and identify the key user experience requirements.
For this case study, the Celo team conducted a survey gauging sentiment towards M-PESA, the top mobile payments solution in Kenya. The survey results were clear: people love M-PESA. Celo researchers learned that this success was primarily a function of understanding and trust: M-PESA scaled in part because it represented financial applications with simple icons and employed local community ambassadors. While Celo offers the functionality of M-PESA but with cheaper fees and easier fiat on-ramps, it’ll require nailing these aspects of the UX to gain traction in emerging markets where word of mouth is so vital to customer acquisition.
Open Finance began its security token journey as a decentralized exchange, shifting compliance risk onto the projects issuing their securities on the network. While DEXs are often lauded in traditional crypto circles, the institutional investors security tokens are meant to lure into the space are put off by them. The majority of companies listing on the network prefer dealing with a standardized white-list solution, which Open Finance has provided through the investor passport. Most investors want to hand their custodianship over to someone else because no one wants the liability of owning the private keys for a large corporation or institution.
Custodianship solutions in crypto are often a combination of both hot and cold storage. Cold storage, meaning that the server has no connection to the outside world or might even just be off, is effective at retaining security but makes it difficult to trade on your crypto assets. Hot storage, which has many levels to it depending on cloud usage, exchange trust, and insurance needs, is far more susceptible to hacking but provides the necessary flexibility for a 24/7 market. A lesser known option due to regulatory uncertainty is the broker-dealer approach, in which the assets are stored in a multisig wallet on your behalf.
Security Tokens as Catalysts for Decentralized Finance
Harbor is the project behind R-token, an ERC20 token with additional restrictions encoded to allow for the compliant trading of private securities. These restrictions are enforced through references to an off-chain “trade controller,” which attributes wallet addresses with specific permissions. These permissions are derived from data provided by identity-check vendors and in-house broker dealers. Holders of any given R-token would only be able to send them to users with similarly permissioned wallets. An attempt to transfer a R-token to a non-permissioned address would result in a failed transaction.
If you were to sketch out the security token stack, Harbor would be square in the middle — building upon smart contracts and token standards — interoperable with exchange layers and accessed via dApps. Given the fragmented regulatory environment for cryptocurrencies worldwide, Harbor is approaching this “compliance layer” strategy on a jurisdictional basis, one-by-one. But as the market for security tokens expands, this integration with other financial primitives will allow for complex instruments to be issued almost immediately.
The Potentials of Bundling
Industries swing between trends of bundling and unbundling as they undergo cycles of innovation. Take music: CDs were replaced by individual songs on iTunes, which have now been spurned in favor of subscribed libraries like Spotify. Similarly, finance is undergoing an unbundling phase now, in which various bank functions (e.g. loans, derivatives, money markets) are being recreated with individual crypto analogs. Set anticipates that the future of decentralized finance will be user experiences that bundle these financial primitives together.
Set Protocol allows for baskets of fungible crypto assets to be represented by individual tokens. Users can deploy smart contracts with specific asset allocation parameters and then redeem set tokens upon collateralizing it. As long as it’s an ERC20, Set can bundle it — whether it’s a BAT token, a sETH dYdX derivative, or an Augur prediction market share, it can all be boiled down to a single token. This reduces gas overhead for transacting multiple ERC20s, and also enables for portfolios/indexes/funds to be tokenized themselves. Further rebalancing functionality allows for asset allocations to be altered post-deployment. Future development will focus on bundling ERC721s as well as real asset ownership.
Price Feeds and Oracles
Maker’s stablecoin, DAI, relies on collateralized Ether to peg its price at $1. As the price of Ether fluctuates, the platform takes certain measures to ensure that the circulating supply of DAI remains adequately backed — this requires a trustless price information source, i.e. an oracle. The MakerDAO oracle currently determines the median price of Ether by aggregating information from top exchanges across a series of time-delayed fail-safes (e.g. multiple broadcasters, separate smart contract caches).
The success of the Maker oracle comes at the cost of efficiency, however, as the numerous gas-intensive transactions required make it susceptible to periods of high network congestion. Its new design ditches on-chain price discovery for a P2P gossip network that relies on a wisdom of the crowds approach for price reporting. With this Secure Scuttlebutt implementation, Maker has shifted its on-chain computational overhead to a Sybil-resistant, cryptographic off-chain solution; all the “medianizer” will have to do now is verify transaction signatures.
This transition reflect the challenges in designing oracles. While their utility extends from reporting real world events to querying APIs for platform data, it is difficult to balance off-chain accuracy with on-chain performance. Maker’s hybrid oracle has achieved this on testnet, but future iterations could pull it off on-chain by verifying trustworthiness through token staking or KYC.
Compliance with the Delaware General Corporation Law
Abacus, like Harbor and Polymath, is building a securities trading platform on top of Ethereum. While the liquidity and cross-border capital flow advantages offered by blockchains are well documented, Abacus honed in on ease of settlement as being the key differentiator between traditional systems. They cite the Depository Trust and Clearing Corporation (DTCC), through which most of the financial industry’s transactions run through, as ripe for disruption: their asset tracking is often incorrect, and even then, they charge significant fees for the service.
Public securities trading on blockchain systems is still a ways off, however. The added complexities around cap table management and dividends, to name just a few, make it more likely that projects will focus on private securities.
When we first launched in August, it was to build on the overlap between our projects’ developer communities and ambitions for an open financial system. Our focus since then has been on open standards and interoperability, and this event was just that: member teams coming together to tackle the shared challenges facing our field. If the DeFi Summit and ETHSanFrancisco hackathon were any indication, it’s becoming clear that the potential for a decentralized finance ecosystem is enormous.
Special thanks to Max Bronstein, Felix Feng, Brendan Forster, Ian Macalinao, Dylan Macalinao, Sid Ramesh, Alex Soong, and Inje Yeo for leading DeFi Summit coordination, as well as Derek Hsue, Chris McCann, Matt Taylor, German Espitia and Brian Weickmann for volunteering their notes.