Raffael Huber: Let’s talk about the basics first – what is it that makes DeFi so powerful?
Rune Christensen: What really drove the recent explosion in popularity was the composability of DeFi. Composing protocols together stopped being a curiosity, and instead became a really powerful way to create new products. The best example is yearn.finance, which illustrates how to powerful products that rely entirely on other protocols can be built.
RH: So you were excited about the onboarding of YFI to Maker?
RC: What was exciting is the fact that one decentralized community came to another decentralized community and made a governance proposal – that enables a new type of interconnection. But most of yearn’s most popular products are built in completely permissionless fashion on Curve, Compound and Aave and so on, which is cool. There are more examples of this, such as InstaDApp – which was actually built by two Indian teenagers with no money and no connections! – that combined the Maker Protocol and Compound, or DeFiSaver that’s also built on top of Maker mainly. It showcases the potential of DeFi, blockchain, and permissionless innovation: You can create these core, low-level protocols that do some particular activity, and then several layers are built on top. Yearn is building on top of Curve which is building on top of DAI. That complexity is only going to increase over time.
RH: What else do you see emerge in the future? What other kinds of protocols will be built on top of the Maker protocol?
RC: Right now, there’s a mini-boom in apps that create two-sided markets between fixed and variable interest rates, so people can hedge their rates. This has been debated in the Maker community for years, but the conclusion always was that we’d wait until somebody else builds it. This is happening now, and others can build a service without any need to coordinate with the Maker community – but it will still significantly increase the value of the Maker protocol. I’ve formulated this basic thesis years ago in a semi-famous reddit comment which was actually more well-known than the Maker project. I explained that Ethereum was much further ahead than all other blockchains already at this early stage because it all comes down to the synergies, Metcalfe’s law: The number of connections and the value of the network scales exponentially with the number of nodes. So every time a new, interesting application is added to Ethereum, you’re not just expanding the value of Ethereum directly through the application, but also that of every other decentralized application (dApp) by a little bit because they can now access the synergy with this application. Through these powerful synergies, you get a flywheel effect – the more value you have, the more value you attract. That’s also why DeFi cannot move off Ethereum, and also underlies the big debate on scalability. It’s hard to coordinate actually moving elsewhere, everyone just wants to stay where the value is.
RH: Lots of protocols are now governed by governance tokens similar to Maker. We’ve seen some first, maybe not openly hostile, but surely profit-seeking attempts to influence governance, for example between Curve and Compound – how can this aspect of hostile takeovers of governance be controlled?
RC: That’s absolutely a reality already and a very real risk in the wild west of permissionless, pseudonymous blockchains. The Maker community emerged from an earlier community called BitShares, which had already been dealing with such long-term issues for years. Maker solves the issue of governance takeovers through the emergency shutdown, which represents a minority stakeholder protection. That is like a “mutually assured destruction” game theoretic approach to dealing with corrupt governance. Opposed to other DeFi protocols, Maker is focused on extreme scale, it is meant to be a neutral, unbiased world currency – so systemically important that it has to immediately and fully cover all issues, even of the long-tail kind. If you have any theoretical vulnerability, you can’t responsibly scale such a system. That’s why we deal with such issues upfront – the game theory of emergency shutdowns or strongly protected and resilient oracles with the tradeoff of being less efficient.
RH: So you don’t think Maker would switch to Chainlink-based oracles?
RC: It could incorporate them to benefit its own infrastructure in various ways. But oracles typically focus on efficiency and speed, on providing super-precise data as quickly as possible. In the Maker community, it’s more about being 100% sure that no oracle attacks are possible. A system designed for systemic distribution and importance in the global economy cannot have any unsolved theoretical issues. Maker has a bunch of solutions, which come with costly tradeoffs. But there have been lots of examples of what happens when you don’t take it seriously enough, such as EOS being completely taken over by a cartel of miners. Another crazy story was Binance doing a hostile takeover of Steem, and also recently the whole SushiSwap drama with a behind-the-scenes takeover. This shows that it’s easy to quickly bootstrap and go for full decentralization and power to the community, but these systems might end up in a suboptimal equilibrium. Sometimes decentralization actually works against decentralization, in a sense. Maker has always followed a careful approach of gradually handing over control to the community, so that the community is strong enough to handle unexpected situations or conflicts with some newly available aspect of governance.
RH: Very interesting. Speaking of reaching this global scale, surely you have also talked about scalability and layer 2 (L2) solutions. Do you think it is possible that Maker would move part of its system to second layer solutions? Creating a Maker Vault is quite costly, at the moment.
RC: In Maker, we are a lot more focused on economic scalability rather than blockchain scalability. What’s great is that because of composability, DAI is already available on every single production L2, like Loopring or xDAI or zk-rollups. As an ERC-20, it’s incredibly easy to access DAI. However, you’re right that Vaults should be more accessible to smaller users – that’s important, but not so much a priority. The logic is that it’s more important that DAI is very solid and scalable, and that it’s okay if the vaults are more focused on large holders and advanced users. If DAI is ubiquitous and really works well, L2 lending platforms will take over that role. Compound and Aave will have some scalable solution on Optimism maybe very early. I think that Optimism has the highest probability, people might converge to it in the short run, and it might be the first time where we see a mass migration. DAI will just be available there from day 1. At some point, it’s also important that that Maker can run and generate DAI on more than one blockchain at the same time – other blockchains and scalability networks, maybe even private, corporate, federated centralized chains. But that’s not a priority either – because the real problem that Maker is facing right now is simply scaling the supply of DAI. DAI is worth more than $1, even though there’s $900M in circulation, there’s demand for even more. It’s not so easy to have that extra DAI generated because there isn’t enough quality collateral available on Ethereum. When people in the Maker community think about scalability and growth in the long run, it’s about real world assets – opening up the floodgates for the traditional financial system and tokenized versions of real world business activity to interoperate with DeFi and Maker. DAI could be based not just on BTC and ETH, but also for example gold, stocks, bonds, other commodities, real estate – real estate drives most of the real world banking sector and is sort of what underpins money in general. We’ve been thinking of this for years, and this month of October, for the first time, there’s a concrete Maker governance proposal that would onboard the first real world assets onto Maker to power DAI generation. Governance is considering to extend a loan of $15M for the construction of an auto parts shop in the U.S. – the holy grail of real world assets condensed into the mundane “we want to build an auto parts shop”. But once it proves to be possible, there’s nothing preventing it from scaling up massively.
RH: Speaking of real world assets: How will the oracle problem be solved?
RC: It’s not really a problem because you just rely on traditional methodology. In the previous example, the approach to oracles – which really means when do you proceed with liquidation and not use the assets as collateral anymore – is that the community will do it manually through a governance decision. Unlike the wildly volatile ETH where you need to react in real time, something like real estate is not going to lose 50% of its value overnight, given proper risk assessment: If someone defrauded you and no assets are there, you’re already stuck with worthless assets. What has been holding up real world assets for years is the basic question of legal recourse – how do you legally enforce a claim by a decentralized organization (DAO)? There are now some credible proposals how to replicate the traditional legal structure and get the legal recourse similar to how a bank extending a loan for real estate does it. A DAO would instruct some real world proxy to enforce the claim, because the DAO itself doesn’t have a legal personality. It turns out you can set this up with the trust structure in the U.S., and there are similar models around the world that can be used.
RH: So that would also help with the economic scalability of DAI and it would get back to $1. Now the idea is to go to another peg at some point, to another fiat currency or even something like the Consumer Price Index – where do you see that part of Maker going?
RC: That has always been the vision. In the short run, it’s enough trouble scaling a simple USD peg and get DAI back to $1 while having a nice DAI savings rate. It’s really the ability of Maker to provide a low risk savings rate that creates a new paradigm for business models built on top of the DAI stablecoin and DeFi. For example, we were involved in a dozen or so initiatives in Africa, all using DAI as a store of value for financial inclusion projects where the whole point was to provide banking services and basic savings accounts to rural, disenfranchised people completely cut off from the financial system, maybe even the Internet – and give them access to assets that are protected from inflation and even get a decent return with low financial and counterparty risk. The startup that provides the service can even monetize because they can also take a little cut of the DAI savings rate. This is possible because the Maker protocol is so efficient, itself it needs to take only a very small cut of what is paid out in DSR and what is earned in stability rates. It has been very frustrating that all of this was crushed because demand for DAI was just too much and Maker was not able to scale the supply in time. Once that gets back to equilibrium, DAI can realize its full potential and you might see the balance flip to the other side – when the ball really starts rolling on e.g. funding real estate projects through Maker, there may not be enough people that want to hold DAI versus how many people want to finance new real estate projects etc. And when that happens, it will finally be the time to return to the original plan of expanding the DAI stablecoin. If you need to grow supply, you onboard more collateral – if you need to grow demand, you onboard more synthetic assets, like a Euro, Yen, Pound and so on version of DAI, pretty much anything. Once we reach that point, you’ll really see the crazy scale of the system – that’s what it’s really meant to do, facilitate this incredible financial scale in an open, transparent system without opaque fee structures. This will ultimately benefit the end user. In the very long run, circular economies entirely based on DAI might emerge, where the Maker governance and monetary policy of Maker is trying to stabilize those economies rather than achieving a peg. That’s also when you could consider creating a version of DAI that is free-floating, a currency that focuses on its own economy basically. But these are just some long term ideas so far in the future that it’s impossible to imagine what Maker governance would look like then.
RH: So you would much rather see DAI going back to its current peg through an increase in supply rather than something like negative interest rates, something that changes the whole dynamic of how vaults operate? There is a proposal that would effectively allow negative interest rates.
RC: There is a community-created mechanism that would allow to break the peg and decrease the target price of DAI over time. In my opinion, it’s really obvious for anyone that actually understands what Maker’s user base looks like that this would be total suicide – it would pretty much kill the project. The reason why DAI is successful is that it is actually serving a real user base of real people that use it as money, e.g. in South America or Argentina, not crypto nerds or super tech savvy people, but just regular people that use it with their smart phones to escape inflation in their own currency. They trust DAI because they can create that mental image where DAI on the phone is like cash in the pocket. Nobody can inflate it, it’s reliable, hard money that even if not precisely at peg is incredibly stable. Money is so much about trust and brand – what money you use is a psychological decision based on things like fear of inflation and trust and so on. If you break the 1 DAI = $1 idea, you’re betraying your user base. DAI at $0.99 isn’t so bad in principle, but the promise, the guarantee that I thought I have – turns out I didn’t have that. The people proposing this mechanism have a different relationship with the currency, coming from crazy DeFi whale-farming, but that’s not where the longer term future of Maker lies. It’s supposed to get out of the crypto bubble and serve the real world. So it would be a fatal mistake to remove that possibility up front by destroying the trust in DAI. Even though it’s above the peg, that’s currently good enough and backing it with centralized stablecoins works until real world asset scalability allows for more organic growth. Maker will have really succeeded once it’s seen as a boring project – at the moment, DAI can be boring because it’s trusted. It’s important to keep it that way.
RH: Going back to the topic of composability – how do you think Ethereum 2 (ETH2) will interact this? There are some different approaches to ETH2 now like “Phase 1.5 and done”. Will there be DeFi specific shards?
RC: Composability as we see now happening on the current Ethereum chain will not be possible to do at an ultra-large scale in the long run. There will need to be some bolt-on solutions that help maintain it across a sharded ecosystem. But even before ETH2, we’ll be dealing with this issue when DeFi tries rollups at scale. So I completely agree with Vitalik’s rollup-centric view of ETH2 – either rollups are going to work, or they’re not going to work – and if they don’t work, it’s like a dead end for the entire space. Maybe then, DeFi would be able to migrate to another blockchain with better native scalability, but you still have to solve the physical limitations that underlie the fundamental characteristics of the technology.
RH: One thing that I’m wondering is: Can you somehow do undercollateralized loans through DeFi or through Maker? Is there a way, perhaps with an identity system or so?
RC: The real problem or misconception are the terms “secured” and “unsecured” lending from traditional finance. The concept of an undercollateralized loan is non-sensical. It cannot happen – because it’s equivalent to saying you would lend $1M to someone who you know can only possibly have recourse for $500k – a transaction with an expected loss of $500k for you. So you always have overcollateralization, not matter what. The question is what form this collateralization takes. For secured lending, some physical asset is deposited, like putting up ETH as collateral in Maker. For unsecured lending or microlending, the legal claim against that person becomes the collateral. The question then becomes how to properly tokenize and maybe bundle or manage legal claims against individuals to use them as collateral in some DeFi protocol. In Maker, as a first way to implement this, perhaps a consumer credit company would tokenize their portfolio of consumer loans into an asset-backed security. The bundled legal claims of, let’s say 10’000 people, are represented as tokens, which could then be used as collateral in the Maker protocol. The first step is to solve the fundamental problem of how to onboard real-world assets, how to enforce legal recourse against those assets using some sort of proxy in the real world. It needs to be a reliable mechanism, which would then allow you to also do proper risk management.
RH: Thank you very much, Rune, this has been very insightful!