Is DeFi Worth the Risk? Live w/ Ledn · Inversion · Rysk Finance

The discussion centers on whether current yields in Decentralized Finance (DeFi) adequately compensate for the inherent risks, especially in light of recent hacks and the increasing capabilities of malicious actors. April was noted as a particularly bad month for DeFi hacks, with the KelpDAO attack on Aave, a prominent lending protocol, causing significant concern. Several perspectives were presented. Santi and Mau argued that DeFi yields are not currently compensating for risk. Santi, referencing his early involvement in DeFi with yields exceeding 100% APY, believes that yields have been too low for the past two and a half years, coinciding with rising treasury rates. He points to a demand problem, with too much supply and a lack of borrowers, leading to low rates. While acknowledging smart contract innovation, Santi asserts that DeFi rates should not categorically fall below treasury yields, citing protocols like Morfo and Aave. He suggests that a high teens yield, at least, would be necessary to incentivize on-chain participation, comparing it to off-chain private credit yields of 12-18% without smart contract risk. The risk of losing all funds due to a catastrophic event, with a probability greater than 1%, warrants a higher yield across the board. Mau echoed this sentiment, focusing on credit markets. He highlighted that typical DeFi yields refer to stablecoin lending rates. Currently, USDT on Aave yields 3.92%, which is only 30 basis points higher than the Federal Reserve's rate, a risk-free option. He contrasted this with Leden's funding model, which issued the first S&P-rated Bitcoin ABS bond with a triple B-minus rating, pricing the senior tranche at 6.84%. This product, after extensive due diligence, still offered a significantly higher yield than Aave's stablecoin offering, despite Aave’s unresolved hack. Mau also pointed to Michael Saylor paying 11.5% for SDRC, a digital credit instrument, further illustrating the disconnect in pricing. He concluded that users are not adequately compensated for the risks taken in DeFi. Dan, on the other hand, presented a pro-DeFi perspective, arguing that yield is a use case, not the core product of DeFi. He views DeFi as creating a new economy where users want to earn on their on-chain assets. Yield is seen as a market-driven number influenced by supply and demand. Dan suggested that the current low yields are skewed and that the error lies in DeFi builders focusing too much on yield, leading to the creation of suboptimal products. He believes it's hard to accurately price DeFi risk, citing the difficulty in quantifying the probability of losing all funds. Dan argued that DeFi builders should not focus solely on yield, and that while some lending products might be inferior to traditional finance (tradFi), there are valid reasons for on-chain lending related to infrastructure and decentralization. He acknowledged that the market might not be pricing risk correctly but emphasized that DeFi's transparency and openness are superior to centralized systems. Santi questioned the products that offer good risk-reward in DeFi. He praised perpetuals and flash loans as interesting financial primitives but expressed concern about vaults, where trusting hedge fund managers on-chain is risky. He reiterated that tradFi players like Apollo and Blackstone offer higher yields for sophisticated investors, highlighting a lack of "A-players" in DeFi underwriting compared to their tradFi counterparts. Santi believes that bringing low-quality credit on-chain, combined with incorrect pricing, is a recipe for disaster, primarily affecting unsuspecting retail investors. He attributes the low yields to a captive pool of liquidity – individuals who made money in DeFi but cannot easily cash out due to tax implications or other reasons, thus accepting whatever yield is offered rather than earning zero. Mau expanded on this "trapped pool" concept. He noted that for Americans, earning yield on stablecoins in DeFi is more accessible than through regulated channels, which require significant legal and financial resources. DeFi offers a way to park digital tokens between trades. The second aspect is the trapped capital that cannot exit the ecosystem easily, leading to an oversupply of capital chasing yield. This capital cannot be easily arbitraged against other on-chain or off-chain credit products, creating a significant disconnect. He likened it to the "kimchi premium," now a "DeFi discount." Mau also argued that it's fundamentally impossible to assess risk in DeFi due to the lack of attribution guardrails. Unlike traditional finance, where clawbacks and legal recourse exist, in DeFi, early movers can escape loss entirely, making it difficult for serious allocators to underwrite risk properly. Dan countered that DeFi offers transparency, real-time awareness, and is significantly younger than tradFi. He argued that while CeFi might offer protections like broker calls before liquidation, DeFi provides a level of openness that tradFi lacks. He cited the example of banks collapsing without individuals going to jail, contrasting it with the potential for accountability in DeFi. Dan also mentioned that USDC can be frozen, and he personally experienced liquidation on Dai in 2020. He believes that DeFi's transparency is a fundamental advantage, even if it's still evolving. Santi pushed back on the idea that DeFi is inherently better due to transparency, pointing out that in the Aave hack, individuals on Twitter became aware of the exploit before others and were able to withdraw funds. He argued that this isn't true transparency if some have privileged information. He acknowledged that real-time pricing in DeFi, like stablecoin prices, is a positive aspect. However, he reiterated that yield should not be the primary driver for using DeFi, suggesting users joined for reasons like unstoppable money or a belief in a better financial system. Dan framed his perspective through his experience in Venezuela, where access to stable, interest-bearing dollar accounts was crucial for wealth preservation. He sees DeFi as providing this essential product, offering an alternative to traditional banks for individuals worldwide. He also highlighted the significance of providing loans to individuals who have never had access to credit before. Mau then discussed CeFi risks, acknowledging that many CeFi lenders collapsed post-FTX due to a lack of transparency and undisclosed risks. However, he contrasted this with DeFi, where early withdrawals during crises might be subject to clawbacks in a legal process, and perpetrators can face justice. He questioned the lack of accountability for those involved in the Torchain bankruptcy. Mau argued that while DeFi has transparency, it lacks guardrails and appropriate risk pricing, making it fundamentally different from CeFi, which, despite its flaws, has established mechanisms for recourse and accountability. Santi elaborated on the Aave hack, explaining how pooled assets in DeFi, unlike siloed positions in banks, amplify risk. He argued for siloed markets in DeFi to contain risk, even at the expense of efficiency. He criticized the automatic deposit and borrowing against newly minted assets in Aave, calling it "ludicrous" and emphasizing the need for guardrails and human oversight, even in a technology-driven system. He suggested that while smart contracts and transparency are valuable, they should not replace basic risk management principles seen in traditional finance. Dan defended DeFi's age, stating it's only six years old and still innovating. He argued that CeFi is also imperfect, citing the lack of jail time for Lehman Brothers executives and the Bank of America/Merrill Lynch acquisition. He believes DeFi's openness and transparency are preferable, even with its current flaws. Dan attributed the incentive problem in DeFi to the pressure to offer high yields to attract users, leading to increased risk-taking, as seen in Aave's decision to accept certain staking projects as collateral. He stressed that while mistakes are made, DeFi's public governance proposals offer a level of transparency absent in tradFi. Mau, while agreeing that DeFi risk is mispriced, suggested the on-chain economy's desire to exit DeFi contributes to an oversupply and low yields. He believes this will eventually harmonize with reality as more arbitrage opportunities emerge. He also pointed out that the lack of personal information in DeFi makes it difficult to perform traditional outreach, like calling a user before liquidation. Dan countered that on-chain notification systems, like DeFi Saver, can provide alerts, though they might require some personal information. He also brought up the issue of stablecoins like USDC being frozen, questioning the absolute transparency argument. Santi emphasized the need for more transparency regarding protocol dependencies, such as the use of a one-of-one DVN with LayerZero in the KelpDAO exploit, which many were unaware of. He stressed the importance of understanding "known unknowns" and "unknown unknowns." Mau suggested hiring more risk managers, focusing on siloed markets over pooled assets, and improving oracle design. He also noted that traditional financial institutions are increasingly interested in DeFi, but bringing real-world assets on-chain is complex, requiring lawyers and acknowledging that DeFi is not always good at handling these complexities. He highlighted AMMs as useful for price discovery on illiquid assets and suggested that DeFi can offer unique hedging opportunities. He concluded optimistically that these are solvable problems, but DeFi protocols need to invest heavily in security, operational security (OpSec), and endpoint detection due to the growing threat from AI and state actors. Dan agreed on the need for increased risk management and called out incorrect practices like excessive looping for yield. He emphasized that while DeFi has issues, its core principles of transparency and global access remain valuable. He also highlighted the challenge of protecting against smart contract risk, suggesting innovation in this area, potentially using AI for building better smart contracts. The panel concluded that DeFi yields are not currently compensating for risk, and builders need to significantly improve risk management. The threat from AI and state actors is substantial, and the future of DeFi hinges on its ability to become safer and more investable. While acknowledging the current shortcomings, there was an optimistic outlook that DeFi will emerge stronger and more robust.