Some Thoughts About DeFi Risk and Complexity at Scale
More sophisticated modeling is definitely needed.
The DeFi ecosystem is experiencing great momentum, with some of the highest Total Value Locked (TVL) figures since 2022. New ecosystems are launching, and others are achieving unprecedented levels of scale. An incredible pace of innovation has been the catalyst behind the recent rally. From a macro perspective, DeFi has benefited from favorable conditions, such as the recent rate cuts signaled by several central banks, which have made DeFi yields very attractive. Obviously, the recent price rally in crypto has also been a significant factor.
One of the hallmarks of this new phase in DeFi is the emergence of protocols that have no parallel with the previous generation. Ethena, Pendle, and the new generation of EigenLayer LRTs are some of the most notable examples in this category. With many of these protocols reaching billion-dollar TVLs in a matter of weeks, there have been rising concerns about whether DeFi is accumulating too much risk again.
- Can Ethena work if funding rates go negative for a long time?
- Who pays for all these point systems?
- Are LRT derivatives creating layers of risk in EigenLayer?
- What are the ramifications if one of these large protocols gets hacked?
These are all valid concerns, and they become more pronounced in an ecosystem that is moving incredibly fast. However, I believe some of these risk alarms are quite trivialized. Unlike the previous generation, these new groups of protocols are taking risk management extremely seriously, and they are run by highly sophisticated teams. I know this because IntoTheBlock works with most of these projects on both risk management and capital allocation, and we are regularly impressed by the robustness of their approach.
The question of risk in the new generation of DeFi should not be about recklessness, vulnerable mechanism designs, or speed but rather about a different factor that is typically ignored: scale.
DeFi Risk and Scale
The main concern about risk in the new generation of DeFi protocols should be related to its scale and interconnectivity. Simply put, after a certain scale, risk manifests itself in very unpredictable and nearly impossible-to-model ways. If you want a recent example, let’s use the 2008 financial crisis, in which layers of well-known derivative products introduced very complex layers of risk into the financial system. Was this due to the complexities of products such as Collateralized Debt Obligations (CDOs)? Only partially. The biggest reason was that risk in CDOs at scale was very different from the modeling you can do in constrained environments.
Going back to DeFi, modeling a large depeg or the ramifications of an exploit in a protocol with several billions in TVL and interdependencies with other protocols is very different from doing the same exercise with a few hundred million and a few protocols.
From a mathematical standpoint, modeling risk at that scale starts to enter the territory of disciplines such as complexity theory. This is a relatively new theory that tries to understand the emergent properties of complex and interconnected systems such as economies, large supply chains, or micro-organisms. DeFi at scale is a large and interconnected ecosystem and should be modeled as such.
Imagine a depeg in a major EigenLayer LRT leading to massive withdrawals in one protocol and major swaps in others. Token holders might have positions in other LRTs, and the pressure can lead to subsequent depegs, etc. Sounds complex? Well, there’s a theory for that 😉
Modeling risk in DeFi at those levels is not an easy endeavor, but it is, without a doubt, a required one. I will share more ideas about this area in future posts.
