This is a two-part series which outlines the problems that exist amongst crypto-backed lending solutions today. The second post in our series focuses on the improvements that decentralized lending solutions like MakerDAO and ETHLend bring, but also their own unique set of challenges and problems. You can read Part 1 about centralized lending solutions here.

Previously on The Issues with Crypto-Backed Loans Today

Two days late for rent, two weeks away from my next paycheck, and still two thousand dollars short of paying off my next four months of rent. But despite all that, I wasn’t worried. With the seemingly countless number of crypto-backed lending solutions out there, what was there to fuss about? There must be something out there that fit my needs.

Now with less than a week before getting evicted, I was in a real cash crunch. Problem is, all the centralized lending solutions I was seeing out there were plagued by the very same four constraints and issues. Using centralized lending platforms means:

  • Putting my funds at risk of being stolen or confiscated throughout the duration of the loan. Their centralized architecture means they are highly susceptible to hacks, downtime, and government intervention.
  • Disclosing my personal information during KYC/AML and risking the theft of my identity and personal information.
  • Being told what rate to borrow at, how long to borrow for, and how much collateral to borrow at — sounds a little bit like borrowing from a bank, doesn’t it?
  • Often having to go to an exchange, and convert my BTC/ETH for some token that can be used for just one thing on that one lending platform (eg. SALT, Nexo etc.)

So when it became clear that the likes of SALT, Nexo, and BlockFi just wouldn’t work for me, decentralized lending solutions were clearly the way to go, right? Well, not so fast…

Decentralized lending solutions — One step forward, two steps back?

Just like how decentralized exchanges marked the first time in which counterparties can exchange value without needing to deposit their cryptocurrency on an exchange, decentralized lending solutions lock collateral on the blockchain through smart contracts, alleviating the need to store collateral or funds with a centralized custodian. Instead, users like me can have full control of my own funds, allowing me the ability to interact with the service from the safety of my own wallet. Additionally, the lack of jurisdiction requirements and KYC/AML processes means that decentralized lending is truly cross-border, as well as more private and secure than centralized counterparts, eliminating the risk of identity theft. Lastly, unlike centralized providers who dictate the terms of each loan, many decentralized lending solutions offer greater flexibility in the form of a peer-to-peer marketplace.

So why am I complaining?

Limited asset choice

MakerDAO, ETHLend, Dharma, and Compound. All four of these aforementioned platforms are lending solutions that leverage their decentralized architecture to serve users in differing ways. However, all four of these decentralized lending solutions share a common issue — everything must be done on the Ethereum blockchain. Whether it’s the collateral, principal, or the interest payments, everything must be denominated in ether or ERC20 tokens, which, at the time of writing, makes up just 15.8% of the total market cap. This means an astounding 84.2% of the entire cryptocurrency market is currently untapped by decentralized lending solutions. Be it miners, traders, or holders, that’s a USD $98B marketplace of crypto capital being left out in the cold by decentralized lending. An $98B segment that I, being a bitcoin holder, was a part of.

Market cap overview chart from CoinPaprika.

While centralized lending platforms can conceivably support any crypto-asset so long as they can be held within a wallet or some kind of centralized custodian, the decentralized lending solutions of today are limited to the crypto-assets found on smart contract enabled blockchains. The reasoning behind this is simple. In order to know whether a loan is in progress, repaid, or defaulted on, decentralized debt agreements require a way to record and update state. Current decentralized lending solutions solve this problem by recording state as a variable within a smart contract, but an immediate consequence is that these solutions are then limited to the assets of the blockchain in which their underlying smart contract is written.

Case study: Wrapped Bitcoin (WBTC)
Recently, one attempt to resolve the lack of Bitcoin support on decentralized Ethereum applications and protocols has arisen in the form of Wrapped Bitcoin (WBTC), which is a bitcoin-backed ERC20 token. Through the use of centralized custodians, a new WBTC is minted for every bitcoin that is held in custody, ensuring that the WBTC in circulation is fully backed 1:1 by BTC stored with the custodian. On the surface, while WBTC seems to finally allow bitcoin holders to leverage Ethereum’s smart contract capabilities and take advantage of Ethereum dApps and DEX’s, it comes at a cost. To generate WBTC for example, I would first have to go through the necessary KYC/AML processes with a “merchant” to verify my identity, thus exposing me to hackers and identity theft. Subsequently, the merchant would deposit my BTC with a centralized custodian before minting the corresponding WBTC to be sent back to me.

Process of generating WBTC through a “merchant”. Note that the merchant must send the bitcoin they receive from you to a centralized “qualified custodian” who stores the bitcoin throughout the duration of the loan.

If this sounds familiar, it’s because it should. Using WBTC on a DEX or decentralized lending solution such as MakerDAO has many of the same downsides as using a centralized exchange or platform like SALT — a risk of stolen funds and lack of privacy. And to rub salt on the wound, Compound and MakerDAO have actively expressed interest in adding support for WBTC within their respective protocols. To build a decentralized lending solution only to support centralized assets as collateral would defeat the purpose of making it decentralized in the first place. If I was a lender, so what if the protocol I’m lending on is decentralized if the collateral backing the loan can be frozen at any time at the click of a button?

Do you agree with these issues? Are there other problems you see with centralized loan platforms? Comment below or add your email here.

Intermediated contracts

Case study: Bancor
In July of 2018, the decentralized exchange Bancor became the victims of a USD$23.5 million hack. Weren’t decentralized exchanges specifically invented so that these types of hacks wouldn’t happen?

When developing their token and protocol smart contracts, Bancor included certain backdoors allowing wallets/keys held by the team to have special privileges, such as the ability to arbitrarily issue, freeze, and destroy any BNT tokens whenever they wanted. But what they didn’t realize is that by having so much power concentrated centrally within these wallets, they created massive points of failure and undermined the principles of decentralization. Additionally, law enforcement could force the project to freeze or destroy tokens as a deterrent, or worse, malicious actors can leverage these wallets to do harm if the keys become compromised.

So when one of the team wallets became compromised, millions of dollars worth of Bancor’s own BNT tokens and ETH were stolen from their reserves and smart contracts that were accessible by the compromised wallet. Subsequently, by using another backdoor to freeze the stolen BNT and prevent attackers from liquidating them, Bancor proceeded to further accentuate the centralized nature of their platform.

Litecoin creator Charlie Lee’s take on the Bancor hack.

The backlash was swift and immediate. And as I saw the tweets roll in one by one, I started to wonder whether these types of backdoors and “intermediated” smart contracts were standard across other supposed decentralized exchanges and applications. Using decentralized protocols that are intermediated with centralized controls within the underlying smart contracts could, in fact, leave you exposed to some of the same vulnerabilities of centralized platforms. Fast forward to September, as I was my doing my research on the decentralized lending solution out there today, it was this very lesson I put into use.

A quick scan of the smart contracts of ETHLend, RCN, Dharma, and Compound reveal a consistent behavior of including various administrative privileges or backdoors within the contracts. Take Compound and Dharma for example. Both of these protocols designate “admin” addresses with the ability to entirely “pausethe functioning of the underlying smart contract by merely calling a function on their respective smart contracts, essentially intermediating themselves between the user and the protocol. For instance, once the Compound smart contract is paused, all interaction with the protocol is suspended, meaning everyone’s funds are stuck within the contract, and neither lenders nor borrowers can withdraw their funds or collateral. And again, even if the Compound team themselves wouldn’t go as far as acting maliciously towards their own protocol, these backdoors and points of failure make the protocol highly susceptible to government intervention and malicious actors.

The withdraw function within Compound’s MoneyMarket contract will fail and become inaccessible in the event an admin pauses the contract.
The transferFrom function within Dharma’s DebtToken contract will fail and users will be unable to transfer tokens to or from the debt agreement in the event an admin pauses the contract.

You may have noticed that I made no mention of MakerDAO in the paragraph above. Widely considered to be one of the most architecturally decentralized projects in the space, Maker is not intermediated in the traditional sense by an administrator of the contract like the others. Rather, the Maker protocol is intermediated by an elaborate governance system centered around the MKR token which is used to vote on protocol-wide changes that affect every CDP holder and DAI borrower. This sounds democratic on the surface, until you realize that the top three MKR holders easily comprise 55.06% of the available MKR token supply (with one holder owning 27% of all MKR tokens). As such, if there ever comes a day where the top three token holders decide to collude, every user of the Maker protocol would be powerless to stop them.

As seen from EtherScan, the top 3 MKR token holders could easily collude and perform a 51% attack on the Maker protocol through its governance system.

Token friction

As discussed in Part 1, many centralized loan providers have introduced a variety of token models for the primary purpose of fundraising under the guise of providing utility to the tokenholders. By intermediating the loan process with their own token, dApps are making users go out of their way to purchase a highly volatile token that can be used for just one thing. To reuse my analogy, why would anyone store their wealth in Air Miles or loyalty points as opposed to cold, hard cash that can be used anywhere? So when ETHLend tries to justify their ICO by offering a 25% discount on all platform fees paid with LEND (instead of ETH) and Maker requires all CDP holders to pay a stability fee in MKR tokens when closing their the CDP’s, they’re just adding friction for the end user.

Decentralization ≠ Disintermediation

As you can see, many of the decentralized lending solutions today simply don’t achieve the benefits I would expect from decentralization. In my view, this has been due in large part to the ecosystem’s overemphasis on decentralization, and underemphasis on the related but distinct concept of disintermediation. Decentralization is the architecture choice to build a platform on a smart contract blockchain, rather than the removal of a middleman or intermediary from a system. On the other hand, disintermediation is the decision to remove yourself as the middleman, and enable a truly open, neutral, and censorship-resistant protocol that is not susceptible to the many hacks and vulnerabilities that centralized platforms also fall prey to. Perhaps there is no greater example of a decentralized but intermediated platform than Bancor, who implemented their smart contracts such that Bancor themselves would always serve as an intermediary between the user and the platform. In spite of being termed a decentralized exchange, the hack it fell prey to clearly demonstrates how disintermediation is a whole other beast on its own and that havoc can ensue when it isn’t given sufficient emphasis by developers.

How did I avoid eviction?

And so, after all that time and research on both centralized and decentralized lending solutions, it became apparent that there is simply no way to leverage my bitcoin without having to delegate trust to an intermediary. While centralized solutions give me the option of using my bitcoin as collateral for a loan, using one would’ve meant risking my funds getting stolen or confiscated at some point during my loan. Being the decentralization (and disintermediation) maximalist that I am, there was no way that I could force myself to use the likes of SALT, Nexo, or BlockFi and lock my collateral with them for any prolonged period of time. On the other hand, as promising as some of these decentralized loan providers are, they’re just limited to Ethereum. There’s simply no way to obtain a loan in a decentralized manner if you’re part of that $98B segment who is holding something other than ETH or ERC20 tokens.

A visual comparison of some existing crypto-lending solutions in the market today — both centralized and decentralized. The horizontal axis represents the degree of centralization. The vertical axis represents the extent to which assets on multiple blockchains can be used as collateral or borrowed on the platform/protocol.

As I began to run out of options, so began the adventure of using MakerDAO as a bitcoin holder. First, I converted my bitcoin to ether. Then, using the ether, I created a CDP to borrow the remaining $2000 worth of DAI I owed. As much as I wish I could say I had a crypto-savvy landlord who was happy to accept DAI, I didn’t quite get that lucky. So I then proceeded to exchange DAI for bitcoin, and offramp the bitcoin to fiat through LocalBitcoins. By the end of this ordeal, I was nearly evicted, almost the owner of a liquidated CDP (due to ETH’s high volatility), and felt as if I aged two additional years within the span of two weeks.

There has got to be a better way.

Do the problems we highlight strike a chord with you? Are there other problems you see with decentralized loan platforms? Comment below!
We are currently building a solution that uses atomic swap technology to make crypto-backed loans more accessible, flexible, and secure. Want to be first to know more? Add your email below.