Interest-ing Times (Part II)
Exploring Crypto Lending: How Can it Generate Positive Cash Flows?
- In this note I take an in-depth look at crypto lending – a sector that witnessed explosive growth during the 2020-21 rally.
- Last year’s bear market slump provided answers to the question "where is the yield coming from?" Crypto lending is primarily driven by speculation making it highly procyclical.
- While DeFi lending protocols survived the crypto roller coaster, many CeFi lenders did not. Their multiple failures have harmed the sector’s reputation and given the doomsters a very valid stick by which to beat crypto.
- However, the sector is unlikely to vanish. Instead, regulators will require CeFi lenders to adopt business practices more in keeping with tradfi to continue operating.
- This is a step in the right direction, but for the crypto lending sector to really become useful as a reliable source of positive cash flow, demand for crypto loans will have to become more diverse and less driven by speculative motives.
Crypto Got Cash-flow
The first research note in this series outlined why the changing macro landscape – specifically higher fiat money interest rates – in my judgement does not constitute a crypto “extinction event” that many doomsters hope.
So as not to detract from the focus of my analysis, and to avoid the research note becoming overly lengthy, there was one deliberate and glaring (at least to those familiar with the sector) omission. As mentioned, natively - by which I mean Layer 1 use - Bitcoin does not generate cash flows, but its holders are, in fact, able earn income on their Bitcoins because of the way in which the crypto ecosystem has evolved. This is not just applicable to Bitcoin, but applies to many different cryptocurrencies. In this research note I will begin looking at how crypto users are able to generate positive cash flows and examine how each has been, and will be, influenced by the changing macro backdrop.
Currently, there are three ways cryptocurrency owners can generate positive cash flows: lending, liquidity mining and, for cryptocurrencies that use the Proof-of-Stake (PoS) consensus protocol (i.e. not Bitcoin), staking.
Of the three methods crypto lending, where users lend their crypto to others in return for a fee, is the easiest to understand and also the least innovative because it is very much a copy-and-paste of the tradfi system.
Crypto lending can be done either via centralized providers (CeFi) or decentralized providers (DeFi). In the case of the former, crypto holders deposit their tokens with the centralized entity who then lends them out on their behalf. Conversely in DeFi, crypto holders deposit their crypto into a protocol and these deposits are lent out subject to pre-defined conditions detailed in immutable smart contracts that both borrowers and lenders are obliged to sign at inception.
Obviously, there is no such thing as a free lunch in finance. The quid pro quo for being able to generate a rate of interest on an asset that does not natively have one (such as Bitcoin) is that there is an increase in risk. In CeFi, crypto holders transfer custody of their assets to a centralized entity, meaning the return of their crypto depends not only on the credit-worthiness of the end-borrower but also on the trustworthiness of the lender to deploy robust risk management. By contrast, in DeFi where crypto holders maintain custody, the primary risks are bugs in the coding of the smart contracts and the security of the underlying blockchain.
Aided by a benign macro backdrop, characterized by very low nominal interest rates in the fiat world, crypto prices rose strongly for reasons outlined in the previous research note. This rally triggered an explosion in the popularity of crypto lending programmes. I was unable to find a reliable data series for CeFi crypto lending, but if we look at DeFi crypto lending (transparency is much greater in DeFi) the amount of TVL (total value locked) in their protocols went from zero to $130bn in just over a year.
DeFi Lending TVL
At the time, many crypto fans viewed the surge in crypto lending and the creation of platforms and protocols – infrastructure necessary to support such activity – as a sign of the ecosystem evolving and maturing, thereby validating their belief that crypto was, indeed, the future of finance. It’s a compelling narrative. However, as the chart also shows, concomitant with the rise in fiat money interest rates last year and the crypto bull morphing into a bear, the amount of crypto locked in DeFi lending protocols slumped and currently stands around $30bn.
What the last couple of years aptly demonstrates is that the crypto lending business is extremely procyclical and, caught up in the euphoria of the bull market, many crypto users forgot to ask the most basic and fundamental question – where is the yield coming from?
That said, perhaps that is a little harsh because even after last year’s events – more on this later - typing the phrase “crypto lending” into Google’s search engine generates a lot of results explaining how users can deposit their crypto and use them as collateral to borrow either fiat money or other crypto and what the various interest rates on offer are. There is very little information about who is actually borrowing the crypto and for what purpose.
One of the biggest, and rather obvious, sources of demand for crypto loans is to increase leverage (or to fund short positions) by borrowing direct from crypto lenders in the case of crypto funds or via margin trading accounts at exchanges for individuals.
Another significant – albeit less obvious - driver of demand, for Bitcoin loans specifically, arises from arbitrage opportunities related to Bitcoin futures. These cash-and-carry trades, where investors buy Bitcoin spot and sell the equivalent number of futures, are not speculative per se because the long and the short net off. Nevertheless, their popularity is pro-cyclical. To understand why we need to consider the mechanics of the trade – simplified below in the following tweet.
During the bull run of 2020/21, cash-and-carry trades generated double-digit returns and due to their relatively low volatility (certainly relative to spot) their Sharpe ratios were also often in double-digits – something rarely seen in traditional asset markets. Such returns were only possible because Bitcoin futures were typically in contango, meaning the futures price was persistently higher than the spot price. Unlike commodities markets, where contango can occur because of the cost of physical storage and delivery, for Bitcoin it reflected strong demand for crypto exposure by investors who wanted to gain from Bitcoin’s price rally but did not necessarily want the hassle of putting in place crypto-specific infrastructure such as wallets/custodians etc. Also, in the case of futures listed on exchanges like the CME, they are better suited for institutional players used to operating in regulated environments, which is not the case for crypto spot markets.
We can clearly see this from the following chart, which shows Bitcoin’s basis returns and funding rates for perpetual swaps (essentially a Bitcoin futures contract without an expiry date invented by BitMex in 2016) versus Bitcoin’s price. In early 2021 when Bitcoin’s price first rose above $60,000 both basis returns and funding rates for perpetual swaps were extremely high – upwards of 20% on an annualized basis. The subsequent pull-back in Bitcoin’s price saw these rates fall sharply and they remained low until Bitcoin resumed its upward march that culminated in a new all-time high on November 8, 2021. Since then, as everyone is fully aware, Bitcoin’s price has consistently declined. In tandem with this price move, the basis return/funding rate for perpetual swaps has also fallen, at times into negative territory, indicative of less demand for crypto exposure.
Bitcoin: Annualized Perpetual Funding Rates vs. 3m Rolling Basis
Source: glassnode (via cryptoslate)
These basis returns/funding rates do not constitute a Bitcoin interest rate as such, but they are closely related. The reason is simple: buying Bitcoin outright to put on these trades extracts the most alpha, but it is quite capital intensive. Borrowing Bitcoin (as noted in the aforementioned tweet) to do the same trade is less capital intensive. Hence, the price of Bitcoin, the profitability of these trades, the demand for Bitcoin loans and the interest rates charged on these loans, are all positively correlated.
In short, whether to gain leverage, fund a short position or finance cash-and-carry trades demand for Bitcoin/crypto loans is almost exclusively driven by speculation. Given how volatile crypto prices have been historically, this is perhaps not that surprising. Indeed, about the only example I can think of for taking a crypto loan other than for speculative purposes is Bitcoin mining companies borrowing Bitcoin to fund their business with the intention of paying off the loan with Bitcoin they subsequently mine.
For many people, speculation is a dirty word, synonymous with greed and an unproductive use of resources. However, speculation has its uses. It boosts market liquidity, aids price discovery (for example, there is nothing to stop doomsters backing up their views with action and borrowing Bitcoin to short it if they so inclined) and is a powerful driver of innovation. Moreover, having mature lending markets helps contribute to the well-functioning of the underling asset market. That is how it works in tradfi markets and crypto is no different. Where problems arise – as last year demonstrated – is how the lending process is managed.
When The Tide Turns
DeFi lending fell sharply last year as prices turned down, but the protocols functioned without a hiccup, i.e., the automated liquidation mechanisms coded into the smart contracts performed as intended.
Contrast this with CeFi lending, which suffered multiple bankruptcies and failures. Indeed, last year alone high profile CeFi crypto lenders such as Celsius, Vauld, Gemini Earn, BlockFi, Midas Investments, Inlock, MyConstant and Hodlnaut (it’s a long list) all suspended operations and blocked customers from withdrawing their funds from their platforms. More recently, Genesis, the crypto lender which is part of the Digital Currency Group, filed for Chapter 11 bankruptcy due to losses associated with crypto loans to Three Arrows Capital, a Singapore-based crypto hedge fund, and Alameda Research, a crypto trading company co-founded by former FTX CEO Sam Bankman-Fried.
The reason why CeFi was hit much harder than DeFi is because they “goosed” their returns offered to users seeking to earn interest on cryptocurrencies. One of the ways this did this was by engaging in rehypothecation.
Unlike DeFi protocols, where the crypto collateral of users is locked in a smart contract, CeFi lenders gain the rights to the crypto given to them and are able to re-use the collateral to back additional loans, this is what is meant by rehypothecation. By spreading the same collateral over a larger number of interest-paying loans in this way, it gives CeFi lenders the resources to pay higher rates of interest on crypto deposits than available in DeFi. In fact, many CeFi lenders relied on DeFi lending protocols to rehypothecate customers deposits. They were, in many instances, effectively offering their customers a leveraged DeFi product with a easy-to-use interface, something many of their customers were seemingly unaware of due to a lack of transparency of CeFi business models and investment strategies.
Rehypothecation was not invented by the crypto community - it has long been used in tradfi (albeit it less so after Lehman’s failure in 2008). Neither is it inherently bad because it does helps to lower borrowing costs/boost returns. What it is though is riskier.
There is nothing to stop individual DeFi users from rehypothecating on their own assets to leverage up the return. After all, the smart contracts that are used in DeFi execute the lending transaction as long as the conditions of the contract are met. Importantly though, when things go wrong, only their funds are at risk of being liquidated if they do not provide additional collateral to comply with the terms of the smart contract. Contagion is limited.
In CeFi, lending firms may have control of the crypto assets deposited by users but the assets ultimately belong to the users. Unnerved by the decline in cryptocurrency prices, user demand for withdrawals understandably surged. Unfortunately for the CeFi lenders, the drop in crypto prices also required them to post more collateral to avoid automatic liquidation mechanisms in the DeFi smart contracts they used to rehypothecate users’ crypto assets. CeFi had a liquidity crisis and unlike the tradfi world there was no centralized entity to act as a lender of last resort.
Those CeFi lenders most badly impacted by the squeeze responded by limiting, and eventually, blocking user withdrawals from their platforms. For a community whose founding ethos was “Don’t trust, verify”, this was akin to shouting fire in a crowded room. As is now coming to light, some CeFi lenders even went even further, using new customer deposits to fund customer withdrawals – the very definition of a Ponzi scheme.
The depressing effect higher nominal fiat interest rates had on crypto prices proved so toxic to many CeFi lenders for an age-old reason: humankind’s inability to overcome the temptation for easy money.
The Regulatory Response
Without doubt, the numerous failures of CeFi lenders last year harmed crypto’s reputation and it has given the doomsters a very valid stick to beat-up the sector. However, the sector is unlikely to vanish.
In the tradfi world, the risks associated with rehypothecation are (in theory at least) mitigated by rules and regulations imposed by government bodies, but crypto has much lighter regulation and weaker safeguards to protect users. This will now be changed. Regulation will be applied much more stringently to CeFi lenders. They will be required to adopt business practices more in keeping with the tradfi banking sector in order to operate. Indeed, it is quite plausible that tradfi institutions will, once regulations are put in place, become significant players in the CeFi lending space.
That said, even if governments act with unprecedented speed, there is no escaping the fact that the legislative process is, by its very nature, slow and it will take time for regulations to be put in place. This represents an opportunity for DeFi. The surge in self-custody of crypto witnessed late last year following the bankruptcy of FTX is a strong indication that the broader crypto community has become more cognoscente of the risks associated with depositing their crypto with CeFi lenders. Moreover, in DeFi, crypto users do not need to wait for regulatory protection, they can rely on the code underpinning the smart contracts to safeguard their assets which are visible on-chain in a fully transparent manner.
The returns generated in DeFi will be lower than crypto users have become accustomed to during the boom years because they will not be “goosed” as they were by CeFi lenders. This is a good thing. Removing opacity to the crypto lending sector and being able to independently verify ownership of their crypto assets will help rebuild confidence. It will also reduce the amount of leverage in the crypto system thereby reducing the propensity for price bubbles to form.
More generally though, for the crypto lending sector to really become useful as a reliable source of positive cash flow, demand for crypto loans will have to become more diverse and less driven by speculative motives. I anticipate this will occur, but it will only be in tandem with much wider crypto adoption (which regulation will also encourage).
Tokenization of so-called real world assets is a rather obvious candidate. Indeed, MakerDAO an Ethereum-based decentralized lending protocol launched a $220m fund with BlockTower Credit to invest in real world assets (RWA), the largest on-chain RWA allocation in DeFi, late last year. This follows an earlier $500m investment in US Treasuries and corporate bonds. Both are investments that generate cash flows by non-speculative means. After last year’s CeFi debacle, it is rather likely that other crypto lenders will follow MakerDAO’s lead.
The changing macro landscape also raises an interesting challenge to fiat-backed stablecoin issuers like USDC and USDT. Neither of these US dollar backed stablecoins pay their holders any interest on their coins. When interest rates in the fiat world were very low, this was understandable. Stablecoin issuers were not generating much revenue from their fiat collateral and the opportunity cost for crypto users to hold them versus fiat was insufficient to offset the benefits from avoiding the on-off ramp. However, we are no longer in such a world. The issuers of these coins are now generating substantial revenues on their fiat holdings, and with DAI – the crypto-backed stablecoin issued by MakerDAO – having recently increased its DAI Savings Rate (DSR) to 1% on the back of increased revenues for the protocol, perhaps it’s time fiat-backed stablecoin issuers followed suit. If implemented, it would provide crypto users with another source of non-speculative-based cash flow. Will that ever happen? Let’s see, but competitive forces suggest it might.
In the final research note of this series, I will delve into the other two methods by which crypto users are able to generate positive cash flow: liquidity mining and staking.
Until next time.
Ryan Shea, crypto economist
 See: https://blog.trakx.io/interest-ing-times-part-i/
 Bitcoin runs a Proof-of-Work consensus protocol so it’s owners are not able to earn staking rewards.
 These smart contracts are not built on the Bitcoin blockchain, but primarily Ethereum and to a lesser extent Solana. In order to deposit Bitcoin into these smart contracts, they first need to be converted into a form that can be represented on this non-native blockchain. This process is known as wrapping. Wrapped Bitcoin is a token that represents Bitcoin on Ethereum’s blockchain – see: https://decrypt.co/resources/what-is-wbtc-explained-bitcoin-ethereum-defi
 If you think you have such a free lunch opportunity please do not get in touch but instead contact the Royal Swedish Academy of Sciences because, if confirmed, you should – rightly – receive a Nobel Prize – see: https://www.nobelprize.org/about/the-economic-sciences-prize-committee/
 The vast majority of crypto lending - CeFi or DeFi - is against collateral (fiat or other cryptocurrencies subject to a suitable valuation haircut) put up by the borrower, namely lending is not unsecured.
 Some DeFi lenders like Aave have a Safety Module, which is an investor funded insurance pool that aims to cover losses from smart contract bugs etc.
 Because this is an aggregate of all crypto DeFi lending programmes, denominated in different cryptocurrencies, the numbers are converted into a $ amount. As a result, these estimates are influenced both by crypto-fiat prices and lending volumes.
 CeFi and DeFi lending volumes, as I shall explain later, are highly connected.
 Borrowing Bitcoin to sell spot in the hope of being able to buy it back at a lower price in the future.
 See: https://www.andrew.cmu.edu/user/azj/files/CarryTrade.v1.0.pdf
 Some Bitcoin were borrowed to create shares in the Grayscale Bitcoin Trust as a means to gain extra additional alpha arising from the fact that the Trust had a premium over its NAV due to investors seeking crypto exposure in a regulated environment. From January 2021 onwards the premium became a discount (see: https://ycharts.com/companies/GBTC/discount_or_premium_to_nav) and when crypto prices turned south this has quite serious P&L implications for the participants in this trade and, eventually Grayscale. For a deeper (speculative – non pun intended) dive into the mechanics of this – see: https://datafinnovation.medium.com/3ac-dcg-amazing-coincidences-c14eec941c06
 They also offer leverage.
 See: https://www.crypto-reporter.com/press-releases/bitmex-launches-the-first-ever-fx-perpetual-swap-contracts-34931/
 See: https://www.bitmex.com/app/perpetualContractsGuide
 If the annual cost of borrowing Bitcoin is 5% then users can borrow 20 Bitcoin for the equivalent cost of buying one Bitcoin.
 Some companies, such as Milo Finance, offer mortgage loans backed by crypto assets, such as Bitcoin and Ethereum. It provides a means by which crypto users can utilize their crypto assets without having to off-ramp. Over time it may well develop into a fully crypto-based lending universe, but it should be noted at this stage the loan and any interest payments use fiat currency as the unit of account – see: https://www.milo.io/
 Speculation is also an inherent human characteristic and impossible to ban.
 The trader Avraham Eisenberg, who was behind last October’s hack of Mango Markets, attempted to destabilize the DeFi lending platform Aave, but his attempt failed – see: https://decrypt.co/115390/mango-market-hacker-loses-millions-in-failed-aave-scheme
 Another crypto lender Nexo, was sued by two of its customers claiming they were blocked from withdrawing their funds from the platform. In response, Nexo said its customers do not have the “unfettered” right to withdraw their cryptocurrencies only the right to “request” withdrawals – see: https://www.cityam.com/crypto-exchange-nexo-users-do-not-have-unfettered-right-to-withdraw/
 The Digital Currency Group owns (owned?) the Grayscale Bitcoin Trust mentioned in footnote 10 above.
 It is also possible that CeFi lenders utilized capital raised from their investors to subsidize their interest rates in order to gain market share. Critics make consider such financial operations as Ponzi-like, but using capital in this way is not unknown in the start-up world. After all, many startup companies lose money in their first few years in order to acquire customers.
 “Not your keys, Not your coins” – sage crypto advice.
 See: https://www.imf.org/external/pubs/ft/wp/2009/wp0942.pdf
 See: https://www.theblock.co/post/207007/celsius-lost-800-million-risky-bets?utm_source=twitter&utm_medium=social
 For reasons I outlined in my previous research note – see: https://blog.trakx.io/interest-ing-times-part-i/
 This applies to both users who deposited their crypto but also the management of the bankrupt CeFi lenders.
 See: https://blog.trakx.io/winter-extended/
 Moreover, as the regulators are finding out, applying rules to decentralized entities is nigh impossible due to the fact that have no obvious leadership/management and because the are geographically fluid and no respecter of nation-state boundaries – see: https://blog.trakx.io/tornado-crash-or-has-it/
 Return of capital always trumps return on capital.
 I am not suggesting it will remove it entirely, human nature being what it is.
 See: https://blog.trakx.io/roadmap-to-utopia/
 For more information on MakerDAO and the recent challenges facing the protocol - see: https://blog.trakx.io/maker-dao-tao/
 See: https://medium.com/centrifuge/blocktower-credit-and-makerdao-to-fund-220-million-of-real-world-assets-through-centrifuge-b52d0fab0fee
 See: https://cointelegraph.com/news/makerdao-goes-ahead-with-500m-investment-in-treasuries-and-bonds
 Meaning the cost of converting between fiat and crypto measured both in terms of monetary and time costs.
 Tether – the issuer of USDT – has circulating supply of $67bn. According to their reserves attestation (not audit to be clear), 82% of its reserves are in Cash and cash equivalents. Given rates along the entire US yield curve are above 4%, this share of Tether’s reserve alone are generating $2.7bn per year!!!!! - see: https://tether.to/en/transparency/#reports
 See: https://www.finyear.com/MakerDAO-to-Activate-Fixed-Yield-for-DAI-Stablecoin-Holders_a48632.html