The Year of DeFi | 2020 Recap

This is part one of my series reviewing 2020. The other entries can be found here: Part 1, Part 2, Part 3, Part 4, Part 5.

The explosive growth of DeFi was undoubtedly a highlight of 2020, with the fast-moving sector commanding attention in the second half of the year. If you want a reminder of what happened, or you if simply want to know why Ethereans are obsessed with farming, vampires, and rugs, I’ll be unpacking it all right here.

Before The Storm

DeFi came into 2020 with a lot of momentum behind it, having more than doubled in size the year before. In a relatively short space of time, it had grown from a handful of projects dominated by Maker to a relatively broad and ever-growing ecosystem that was, admittedly, still dominated by Maker.

The early months of 2020 saw some significant moments, such as the introduction of flash loans with the launch of Aave Protocol, the launch of stablecoin-swapping platform Curve, and celebrations as the Total Value Locked (TVL) in DeFi protocols crossed the $1bn mark for the first time in February. TVL is something of a vanity metric, and its sudden increase at the start of 2020 was partially driven by Ether increasing in price, but it was still a milestone moment that proved DeFi was worth paying attention to.

Celebrations were short-lived, as another event brought everyone back down to earth and reminded them that DeFi was still young and dangerous. In the space of just four days, lending protocol bZx was exploited twice, losing more than 3600 ETH in total, worth over $900k at the time.

These exploits were certainly concerning, raising some serious questions about the security of DeFi protocols in the age of flash loans, but they also served as a perverse form of advertisement for the power of these systems. Both attacks leveraged flash loans and DeFi’s legendary composability, with the first attack involving 5 different protocols and costing a total of $8.21, and the second using 4 protocols and costing $110 — and both attacks were conducted in a single block lasting around 15 seconds. It’s undeniably impressive.

The bZx attacks would go down as the first major hacks in DeFi history, and the first in a series of around 15 sophisticated attacks on DeFi protocols that took place throughout 2020.

Once the dust had settled and the attacks had been dissected and learned from, the DeFi community went back to building out their burgeoning financial system. They had no way of knowing that just a few weeks later, it would face its greatest challenge to date.

Black Thursday

Early March saw virtually every market worldwide struggling to cope as the realities of coronavirus finally set in, and crypto markets were no exception. In the space of 24 hours around March 12th, Bitcoin fell from $7300 to lows around $3900, while ETH’s price was cut almost in half, briefly dipping under $100. This spelt trouble for any DeFi protocol using ETH as collateral and Maker found itself hit hardest.

When a user wants to generate DAI, a dollar-pegged stablecoin, they open what Maker calls a Vault, a smart contract that holds the user’s collateral and allows them to access a loan of up to 2/3 of the US dollar value of that collateral in DAI. If the price of the collateral falls too much and the loan stops being sufficiently overcollateralised, and the user fails to either repay some DAI or provide additional collateral, then the vault will be liquidated. At this point, the collateral is auctioned off to entities called Keepers, who place bids in DAI to win some portion of that underlying collateral. In normal circumstances, these auctions remove the debt but leave users with most of their original collateral, minus a 13% liquidation fee.

However, Black Thursday was far from normal circumstances. As prices fell and DeFi users rushed to take advantage of the opportunity or to protect their loans, the network became congested and gas fees rocketed upwards, making it more expensive to transact. Maker’s auctions broke down as many Keepers weren’t smart enough to increase gas fees when their transactions got stuck, while others simply ran out of DAI and stopped bidding. Realising there would be no competition to bid them up to a fair price, one smart Keeper started placing 0 DAI bids for all the collateral in a vault. By the time things calmed down, more than one one-third of liquidations came at practically no cost to Keepers, and one lucky Keeper won 30,000 ETH with $0 bids.

Things could have been even worse, though. Oracles, who provide price feeds for the system’s collateral, couldn’t update fast enough to track the declining prices. Eventually, they stopped working altogether, just as ETH slipped under $100. Some believe this was an intentional move by Oracle-operators, who realised that a vast wave of liquidations would have been triggered at that point, which would have caused a dramatic downward spiral that could have crushed ETH and the Maker system.

When prices finally stabilised, Maker was left with $4m missing from the system and a lawsuit from disgruntled users who’d found their entire Vaults liquidated. While an emergency shutdown was briefly considered, the obvious path forward was Debt Auction, where Maker’s governance token MKR is minted and sold until the missing funds are replaced.

But before it could hold that auction, Maker needed to fix another problem. In all the chaos, demand for DAI had pushed it far from its peg, with the supposedly stable currency trading as high as $1.14 at one point. To push DAI back towards zero, MKR holders voted to lower the stability fee to 0.5% (making it cheaper for users to mint new DAI), and the DAI Savings Rate to 0% (making it less appealing to hold DAI). However, even with these measures in place, DAI continued to trade above its peg. At that point, MKR holders made the controversial decision to onboard USDC as a new type of collateral. Some were unhappy with the decision to add a centralised form of collateral to the system as those funds could theoretically be frozen and become worthless, causing more issues down the line. But, the more pressing and present concerns about DAI’s inability to hold its peg won out. With USDC added, users could mint DAI with a token worth $1, sell the DAI above $1, and walk away with a profit.

With DAI working its way back towards its target price, Maker could get on with its auction. In the end, more than 17,000 MKR were created and sold, completely wiping out more than two year’s worth MKR token burns.

What was important, though, was that Maker — and DeFi as a whole — had survived a catastrophic black swan event. It was a little shaken, and some had lost a lot, but it had survived.

The DeFi Summer Begins

The months after Black Thursday were comparatively quiet, and mainly consisted of building DeFi back to full strength. TVL had taken a substantial hit in March and didn’t re-passed $1 billion until the end of May. Just a couple of weeks later, lending protocol Compound kickstarted the DeFi summer by launching their liquidity mining program.

Compound was developed and maintained by a company called Compound Labs, who had special administrative rights over the protocol that allowed them to make any changes they needed. This was useful for getting the project off the ground and keeping it safe in the early days, but ‘Decentralised Finance’ wouldn’t be decentralised if these privileged players all stuck around for the long term. So, Compound Labs came up with a plan to transfer ownership and control of their protocol to the masses: they created a governance token called COMP that allowed holders to vote on changes to the protocol, and then they distributed COMP to anyone that used Compound in proportion to how significant their usage was. It didn’t matter if they were lending or borrowing, everyone who used the service received COMP.

This made a lot of people very excited. For a long time, people have eagerly moved funds to whichever DeFi protocol provided the highest yield, and now there was an opportunity to earn not only interest but a funky new governance token with a skyrocketing price. What’s more, because COMP was being given to lenders and borrowers, you could earn money by taking out a loan. Those looking to maximise returns — or farm yield — would repeatedly lend to Compound, borrow from it, and then lend what they borrowed. Instadapp even created a recipe to do that as easily and efficiently as possible.

The result was a wave of capital flooding into Compound. Just 5 days after COMP distribution began, Compound overtook Maker to become the largest DeFi protocol by Total Value Locked, while COMP briefly traded at a market cap greater than every other DeFi token combined. In short, it was a major success, and others were paying attention.

In the weeks that followed, all sorts of projects launched their own liquidity mining schemes, desperate for users and attention. Money rushed into DeFi to take advantage of these yield farming opportunities, pushing the value locked in DeFi above $3 billion by the end of July, $8 billion in August, and $10 billion in September. Gas fees also jumped to all-time highs, pricing out many smaller players and making an urgent case for second-layer solutions. And, rather predictably, the dark side of the industry started to show through at scale for the first time since ICOs went out of favour.

BCOs & Fair Launches

It’s a positive that ICOs didn’t make a real comeback this year; it proves they’re properly dead. Instead, we saw a few different forms of token distribution for new projects in 2020: the ICO-like Bonding Curve Offerings (BCOs), the return of the coveted ‘Fair Launch’, and the Uniswap-style Airdrop.

Bonding Curve Offerings, sometimes called Initial DEX Offerings, became the distribution method of choice in early 2020, with projects like UMA, bZx, and Perpetual Protocol releasing tokens using them.

A bonding curve sets a mathematical relationship between a token’s price and its supply, and they form the backbone of decentralised exchanges like Uniswap. When a team wants to sell a brand-new token using a BCO, they create a pool containing their token on a decentralised exchange. As soon as this is done, users can buy and sell into the curve to determine the fair market price for the token.

This approach has several potential benefits over other forms of token distribution: theoretically, they prioritise distribution and price discovery over teams making money, they create instant markets for tokens, and they’re a completely permissionless, so anyone in the world can participate. While certainly a step in the right direction compared to ICOs and IEOs, BDOs still have a few kinks to work out, such as rampant front-running issues that saw the price of some tokens skyrocket in seconds last year. Assuming BDOs stick around for some time and aren’t replaced by other innovations, I’m sure solutions will eventually be found.

Later in the year, after the explosion in Yield Farming, we saw some projects choosing to forgo any kind of fundraising and simply distributing tokens to anyone that performed certain tasks. This was the reinvention of the Fair Launch, and it was kickstarted by Yearn Finance in July.

In one of his pieces for Deribit Insights, Hasu defined a fair launch as having four attributes: “there is no free lunch — everyone has to work for acquiring the tokens”, “the work performed is provably costly to prevent Sybil attacks”, “participation in the launch is permissionless and transparent”, and “most importantly, there is no premine or founder allocation, and no early access. The founders are playing along with everyone else”. Basically, everyone and anyone that wants to own a fairly launched token must perform some costly action to earn that token or buy it from someone that has. In DeFi, the ‘costly action’ is often providing liquidity to something, but it can be as simple as incurring opportunity cost by locking tokens in a smart contract.

On paper, Fair Launches look like the best way to launch a cryptoasset, and you only have to look at the hyper-engaged community around Yearn to see the benefits, but there are some significant drawbacks to them. Most obviously, there’s no source of funding for the project, potentially causing significant problems down the road as we’re now seeing with Yearn, but more immediately forcing it to publish unaudited code that could cost users millions of dollars. An initiative called Fair Launch Capital is hoping to solve this issue by covering audit costs upfront with the hope of being paid back later, but they’ve been fairly quiet since launch. There are also big questions about just how fair fair launches really are. In some cases, founders could perform ’synthetic pre-mines’ by purchasing a handful of assets ahead of launch, then require others to stake those same assets to earn their token, giving the founders a chance to profit as people rush in to buy. Additionally, in an environment with extremely high gas fees, only those wealthy enough to risk tens of thousands of dollars can realistically participate in a fair launch — though that’s still a significant step forward from the ‘shitcoin waterfall’ model of 2017, where you needed to be both extremely wealthy and have the right connections to access the best deals.

Perhaps the toughest challenge for fair launches, though, is that they’re simply too difficult to replicate after the first iteration. Nothing has ever and will ever come close to replicating Bitcoin’s fair launch because it took place in the purest and least profit-driven environment possible. Similarly, no DeFi token can replicate the launch of YFI.

The effect is even more obvious with UNI-style airdrops. Before Uniswap released tokens to every address that had ever used their products, all the users had been genuine. Once the possibility of future airdrops became clear, people started using all sorts of protocols with the hope of one day receiving free money, not because they wanted or needed to. Essentially, once the fairest possible version of a launch has taken place and rewarded engaged, positive-sum participants, all future launches will be plagued with parasitic, negative-sum participants.

In December, we saw DEX Aggregator 1inch add a few crucial changes to Uniswap’s model to reduce the number of negative-sum participants earning tokens. While this won’t solve the problem entirely, and may sadly exclude some legitimate users as well, this kind of amended airdrop/distribution event is likely the best path forward for any projects attempting something similar. Certainly, I expect to see many currently token-less projects following 1inch’s lead in 2021, modifying distribution parameters to ensure the project’s future is placed in the hands of those that actually care about it.

When Vampires Attack

The yield farming craze made tokens cool again for this first time since early 2018, and in a complete reversal of trends, it quickly became apparent that not having a token was a risky decision. Projects that failed to properly incentivise users to stick with them were at risk of being abandoned in favour of something else, and nothing made this more obvious than DeFi’s first ‘Vampire Attack’.

Apparently taking inspiration from a tweet by The Block’s Larry Cermak, an anonymous individual going by the name Chef Nomi decided to fork Uniswap’s code, rebrand it as SushiSwap, and add a governance token called SUSHI that would be used to incentivise usage and help ‘migrate liquidity’ from Uniswap.

SushiSwap would charge users the same 0.30% fee as Uniswap, but instead of giving the full reward to liquidity providers, it would convert 0.05% into SUSHI and distribute it to existing SUSHI holders. Additionally, to fund the project over the long-term, 10% of every SUSHI distribution would be placed in a ‘development fund’.

To earn SUSHI ahead of launch, users had to stake their Uniswap Liquidity Provider (LP) Tokens (earned by providing liquidity to Uniswap pools) in a SushiSwap contract. After a couple of weeks, SushiSwap would redeem those LP tokens for the underlying assets in Uniswap and redeposit them in SushiSwap pools — essentially transplanting liquidity from Uniswap to SushiSwap. It was genius, it was dastardly, and it was a joy to watch… until Chef Nomi threw the whole thing into jeopardy.

In early September, just before SushiSwap’s planned liquidity migration, DeFi tokens started selling off. Perhaps spooked by this, Chef Nomi withdrew SUSHI from the development fund and sold it into ETH. Despite his insistence that it wasn’t an exit scam and that him taking the ETH was somehow best for the SushiSwap community, SUSHI’s price fell 70%.

That could well have been the end of SushiSwap, preventing us from seeing how the liquidity migration, or vampire attack, would work out. But then Sam Bankman-Fried (SBF) of FTX and Alameda Research fame stepped in to save the day. He suggested Chef Nomi should hand control of the project over to someone else, someone the community could trust. Nomi decided that SBF was the man for the job and gave control of the project to him, who then transferred control to a multisig formed of individuals voted in by SUSHI holders. Members of the multisig included SBF himself, SushiSwap’s muse Larry Cermak, and Compound’s Robert Leschner.

With Nomi firmly out of the picture and trust restored, migration finally took place on September 9th. Incredibly, this drained about half of Uniswap’s liquidity and instantly took SushiSwap’s TVL to $1.2 billion.

While that sounds bad for Uniswap, it wasn’t actually a disaster. As The Defiant’s Cami Russo pointed out at the time, Uniswap actually had more liquidity after the migration than it had before the ‘Sushi Saga’ began. So, while undisputedly a parasite attack, SushiSwap had been net positive for Uniswap. Still, to stave off further losses to the new exchange, Uniswap were forced to respond and introduce their own token towards the end of September.

SushiSwap has gone on to prove many of its early doubters (myself included) wrong. Its murky start has given rise to a genuinely innovative, well-intentioned project that does significant volume — it accounted for 14% of all DEX volume in December. That wouldn’t have been possible, certainly not in such a short timeframe, without the accelerated bootstrapping of a vampire attack. It also proves, in the words of Messari’s Connor Dempsey, that “liquidity is not a moat when capital can move at the speed of the internet”, and that DeFi protocols have to be smart to catch and then maintain a significant user base.

The Dark Side of DeFi

Among the fun and money-making frenzy of the DeFi summer, there were a number of strange and even dark moments. The Sushi Saga serves as a good microcosm of the whole DeFi summer: it showcases the dramatic, quirky, and creative side of the sector, but also the hazards and sometimes-scaminess. For every genius idea and 10x-er, there were multiple semi-scams, rug pulls, hacks, and more.

One project that was emblematic of many of the negative DeFi trends was Yam, a relatively early entry in the ‘food token’ niche that consisted almost entirely of zero-sum memecoins clearly designed for gambling but claiming to be something much more significant.

Yam was a kind of Frankenstein-token that had been thrown together in 10 days to capitalise on a number of emerging DeFi trends like fair launches and rebasing mechanisms, and that approach seemed to pay off. In just 36 hours, $600m worth of cryptoassets were staked to earn YAM, its price ploughed through its ‘intended’ target of $1 to peak a little under $170, and then the entire thing collapsed as a bug was discovered in its unaudited code. Luckily, only $750,000 worth of funds were actually lost because of the bug, with Yam Farmers primarily losing opportunity cost and gas fees, but it’s easy to imagine similar scenarios turning out far worse.

After the incident, Ethereum OG Lefteris Karapetsas published a furious blog post calling on DeFi builders and users to do better: to audit code, to act responsibly, to stop shilling, and to think about building a safe, sustainable, professional financial ecosystem.

The problem with Yam, and other efforts like it, isn’t that they exist or that there are people that want to use them. Gambling apps and Ponzi-games are fine, as long as everyone knows what they’re getting into. Profit-driven degenerates that ape into unaudited projects are fine, as long as they’re in the minority and their mindset does infect the entire sector. But that’s what seemed to be happening. Previously respectable, influential Ethereans were shilling Ponzi-games and fetishising reckless and irresponsible behaviour, largely in an effort to enrich themselves. This was made clear in one of the lowest moments of the year: the $Few Scandal.

As the DeFi summer cooled off and prices and yields collapsed, some prominent Ethereans made one last attempt to bag some easy money. The general idea was to gather 50 ethereum ‘influencers’ in a telegram chat, airdrop tokens to them, and have them promote those tokens to their followers in the hope of pumping the price. The token had no purpose — its creators hoped somebody would think of one later — and came to be named $Few after the ‘few understand’ meme. Before the plan could become reality, but not before the $Few shilling began, someone leaked the Telegram channel and exposed the plan, preventing any Uniswap listing of the token — though fake versions were created and sold by opportunistic scammers.

Once the chat was leaked, many of the individuals involved apologised for their actions and said they’d believed $Few to be a joke or that they hadn’t actively participated. In the best-case scenario, those claims are true and those ’influencers’ simply lack any sense of responsibility or awareness of the damage they can cause. At worst, they’re deliberately trying to leverage their trusted positions and exploit the greed of others to benefit themselves.

DeFi, and cryptoassets in general, will attract scalpers and profit-driven individuals, and that’s fine. But the people who act as the heart or the face of the industry cannot be like that, not if they truly want it to be better than the systems that exist today.

DeFi will likely have an even bigger year in 2021 than it did in 2020, which means more newcomers and more bad actors trying to make easy money off them. It’s on all of us that want to see DeFi succeed to call them out and push back against them, but last year’s evidence suggests that sadly may not happen.

Disclaimer

Anything expressed here is my own opinion stated for informational and educational purposes; nothing I say should be taken as investment or financial advice. Any projects mentioned are not recommendations and may be highly experimental and therefore risky. Please evaluate your own risk tolerance before trying them out.

I may own some of the cryptoassets mentioned.

At the time of upload (February 2021), I own:

• Long Term Holdings: Bitcoin (BTC), Ether (ETH)
• Short-Medium Term Holdings: Aave (AAVE), Alpha Finance (ALPHA), Terra (LUNA), THORChain (RUNE), SushiSwap (SUSHI), Uniswap (UNI), Nexus Mutual (NXM), Yearn.Finance (YFI)
• Stablecoins: USDC