CRYPTO YIELD: MASKING CREdit RISKS and leverage ALL ALONG?
June was a month for the record books! Bitcoin fell 34% on the month and 58% on the quarter. ETH fell 41% and 69% on the month and quarter respectively. These are the biggest declines since 2011 as the crypto markets dove into full liquidation mode. Luna's spectacular unraveling in May 2022 resulted in escalating contagion in June. A $18bn crypto fund blew up and defaulted on its obligations to crypto lending services. Participants withdrew capital from yield providers en masse taking a fearful precaution against the risk of their capital being gated. I have been in deep reflection these past few weeks, returning to my past articles on yield. As you might expect, my conviction on bitcoin is unshaken, but this is certainly a time for learning. TLDR: Desire for stability combined with centralisation and rehypothecation in crypto yield products masked deep credit risks and leverage under the hood. Crypto is violently solving these challenges, humbling participants with another bear market. We will arise from the liquidations stronger than before but there are major lessons to digest on that path.
Key Insights
Crypto yields are not like traditional yields, they are pro-cyclical
Stable yields and centralisation masked credit risk and leverage
CEFI lenders are banks without a consumer protection regulations and central bank back-stop
Fractional reserve banks lead to boom bust cycles
Cleansing is healthy and is now reflected in higher bitcoin dominance
Sound Money’s foundational principles have stood us in good stead in the face of extreme volatility
Backstory to this gripping crypto tale
Before walking into the reflections on yield, leverage and centralisation, let's create the scene of this months gripping crypto tale! One of the largest crypto funds, 3 Arrows Capital (3AC) with $18bn AUM was heavily invested in Luna (Luna’s unraveling was the subject of last months article). Despite the losses, 3AC maintained an aggressive outlook and were heavily leveraged. The market downturn pressured them to default on their debts to centralised yield providers like Celsius and Blockfi. Celsius gated withdrawals and market participants are now fearful that Blockfi will walk down the same path so Blockfi is desperately trying to secure funding in order to stave off a deeper crisis in their business.
Side note: Organisations like Celsius, Nexo and Blockfi are often referred to as CEFI. "C" for centralised finance vs. "D" for decentralised finance. CEFI lenders take client funds and lend them out, providing users with yield in return for putting their capital at risk.
There is a lot in there so let’s take a step back for a second… Tightening global liquidity is violently exposing the cracks within the unhealthy elements of financial markets and the economy, including crypto. The trouble ran deeper than just a poorly created coin, Luna, taking us into the world of crypto yields and CEFI lenders. In August 2021 I wrote a favourable article on crypto yields called “Crypto yields are sucking in trad-fi” where I laid out a positive case for yield on the back of the long-term opportunity offered by crypto and the utility of global stablecoin networks. I clearly argued that counter-party of protocol risks were paramount. Nevertheless, on reflection I could have done more to outline the risk of these products. I looked at the opportunities with rose-tinted glasses in the past so I need to dive deeper into the risks here.
Sources of returns: Crypto ‘yields’ are pro-cyclical
I still think that reasonable yield opportunities exist in crypto but let me be very clear, crypto yields are not the steady, stable income generators seen in traditional finance. Traditionally, yield serves as regular/periodic income distribution to investors who have put capital at risk. I have said it before and I will say it again, nothing in crypto produces a stable rate of return! Crypto yields are volatile and pro-cyclical. To understand, let’s dive into the potential sources of crypto return, aside from merely holding protocols/tokens.
Demand for leverage
Traders speculate on price increases, which can result in large positive yields that can be harvested as a risk premium.
Given that I am bullish on crypto, I suspect that demand for leverage will be positive on average over time so this is a potential source of return in the future. However, large fluctuations in leverage demand are expected, mirroring crypto price volatility so this is a pro-cyclical source of return.
2. Arbitrage
Due to crypto’s volatility, market segmentation and constant innovation where new products are popping up regularly, price discrepancies often exist between various crypto markets (different exchanges, or geographies, for example). Arbitragers can harvest these discrepancies as a risk premium. Arbitrage returns are less likely to be pro-cyclical but are generally whittled away because they present risk free investment opportunities. However, I have witnessed first hand that these arbitrage returns can last far longer than expected. Regulatory barriers to entry and price volatility are two factors that can cause arbitrage returns to remain intact longer than expected.
3. Protocol revenue
Proof stake blockchains pay stakers who allocate their capital to the network a proportion of transaction volumes. Earning protocol revenue through a non-dilutive staking protocol probably has the closest resemblance to traditional yield. However, almost all crypto transaction volumes are dependent on market conditions, waxing and waning between bull and bear markets. Additionally, proof of stake blockchains tend to be higher risk crypto assets vs. bitcoin, which can imply that the incomes are lower value during weaker market conditions. So while protocol revenue is promising, it is also pro-cyclical.
4. Protocol issuance
New blockchains often distribute issuance through tokens to users to incentivize activity and holding. It is difficult to argue that this provides holders with a return because new issuance is dilutive. Holders who decide to stake just choose to avoid dilution vs. those who do not.
During bear markets new protocol issuance is particularly useless because very few users want to allocate time or attention to an inflationary and untested protocol. Even if they did, tokens outside of BTC and ETH tend to lose substantial value during bear markets, calling into question the efficacy of the yields.
I would be very wary to call protocol issuance a source of return, and even if it is, it is strongly pro-cyclical in nature.
Unlike trad-fi where yields tend to be counter-cyclical, rising during times market stress, crypto yields are currently pro-cyclical, rising during periods of market exuberance and declining during stress. Arbitrage can buck that trend but generally, this pro-cyclicality calls into question the efficacy of yields during current conditions and they certainly are not stable yields.
Is the credit risk worth the yield?
Putting cyclicality risks to the side, centralised crypto yield providers like Celsius, Nexo and Blockfi take these highly variable yields and attempt to provide a stable yield to clients through rehypothecation. That is a complex undertaking that requires extensive risk management skills. Over and above the business undertaking, large centralised entities are prone to human error and potentially even fraud by senior executives.
CEFI lenders have a strong incentive to lend out more capital than they have on deposit and make longer term loans to juice profitability (Austrian economists call this fractional reserve banking). These actions create credit risk, leverage and expose the lenders to bank runs, which is exactly what CEFI lenders are experiencing. Traditional banks also conduct fractional reserve lending, but CEFI lenders are not regulated like banks. There is no consumer protection and there certainly is not a lender of last resort that is going to bailout CEFI lenders if they get overextended (like traditional bank bailouts in 2009). Users of these centralised yield products become unsecured lenders - users are fully liable for the risks taken by organisation with no legal recourse.
Users need to conduct specialised credit risk assessment of each organisation in order to mitigate against these risks. This is far easier said than done and it begs the question of whether the yield is worthwhile in the first place. Users cannot get a full look-through into metrics like loans outstanding, collateral ratios and whether there are any uncollateralised loans. Most of these organisations are a blackbox. Hopefully we will get to a place in the future where the industry forces disclosure of a few key metrics to confirm organisational credit health. Until then, users are making a decision based on judgement.
I remember in 2020 when I was investigating these organisations my judgement call was that Celsius was riskier than Blockfi because they had a token which they issued in the 2017 ICO craze. Given the complexity of their undertaking, I prefer less experimental centralised yield providers. I also thought that Celsius' yields were a little too good to be true. This judgement proved correct but it would have helped to have more data.
Masked Leverage impacted collateral assets too
I naively underappreciated the hidden leverage below the surface in crypto. I monitor open interest leverage, which is very low as a percentage of market cap. I was under the impression the over-collateralised nature of the loans would limit leverage, however there is definitely a weakness to this argument. First, leverage is still possible with over-collateralised loans. I can give 1 BTC to a CEFI lender, take out a USD loan, and purchase more bitcoin with the proceeds. Over and above this, there is evidence that some centralised parties were partaking in uncollateralised lending. Outside participants would never know this until liquidation shocks cause the cockroaches to come out of the woodwork.
Bitcoin and ethereum remain the key collateral assets in crypto. As investors liquidated in their crypto exposure, they clamoured for the most liquid holdings in their portfolios, BTC and ETH to extinguish debts. However bills are not paid in BTC or ETH and investors do not seek redemption’s in BTC or ETH. During a liquidity crunch, everyone wants their hands on USDs so the collateral assets were also sold indiscriminately
I have scratched my head to determine how I should have tracked crypto leverage better than I did. I continue to return to bitcoin dominance. I have always included dominance as part of our investment process but seeing the liquidation engines working overtime over the past few months has really elevated its importance in my eyes. As bitcoin dominance falls, broader and riskier financial media are effectively being leveraged on top of bitcoin's pristine collateral. Recently, we have actually started to see bitcoin dominance rise again. Bitcoin accounts for >50% of the non-stablecoin crypto market cap, which is an encouraging sign that health is returning to the market.
masking the risk and hiding the true value
Returning to the article “Crypto yields are sucking in trad-fi” from August 2021. Let me clarify. The conclusion in my previous article was hyperbolic. Crypto yields are interesting but are certainly NOT crypto's number one use-case! Demand for leverage, arbitrage opportunities and potentially protocol revenue offer interesting ways to access crypto. Direct leverage financing is a pretty low risk but it is pro-cyclical, arbitrage is a non-cyclical opportunity but it will likely be whittled away and protocol revenue is almost certainly pro-cyclical. I do think some of these opportunities will suck institutional investors into market neutral strategies, but investors will still be exposed to crypto cyclicality.
The nature of CEFI lenders implies that they are likely to amplify leverage and credit risk during bull-markets. Users of these platforms need to be acutely aware of the risks and should not follow passive strategies. The only proven passive crypto strategy is buying and holding bitcoin for periods longer than 4 years. It is highly unlikely that large institutions would have much interest in allocating their capital to CEFI lenders, other than if they were making VC investments. On that note, it is insightful that a $420 Canadian pension fund Caisse invested approximately $150mn into Celsius. Caisse were wary of making a direct investment into an underlying protocol and thought Celsius was a lower risk bet. This proved wildly inaccurate and makes an important point.
I remain convinced that the biggest and lowest risk gains in crypto will come from holding the large underlying protocols.
Yes, bitcoin is volatile but that volatility does not mask the risk! The risk of a volatile asset is abundantly clear for everyone to see and if you have a long enough time horizon that does not matter. On the other hand, a stable yield offering masks centralisation, rehypothecation, leverage and credit risks. I am not dismissing the possibility that the yield model can work but I think we need to be very clear about the risks. Without a lender of last resort we should expect boom and bust cycles in these lenders as human nature implies that companies will get greedy and overextend themselves during good times. Liquidations during a bust are painful but they clean up the bad businesses, reallocate capital towards good businesses and avoids the negative consequences of central banking.
The centralisation and credit risks exposed by CEFI lenders are exactly what bitcoin and crypto is trying to solve. Bitcoin has no centralised individual or organisation, there is no counterparty risk and there is no rehypothecation of your assets. This is not to say that everything must be decentralised and anything centralised is bad. Centralisation will always exist when humans are trying to coordinate. But users need to be very clear about the desired degree of centralisation and the implications.
At Sound Money we could certainly have navigated the recent volatility better than we did, but our foundational principles have stood us in good stead, adding value for clients and missing the landmines.
