The Changing Face of Risk in DeFi

Decentralized finance (DeFi) is experiencing renewed momentum. The activity in new ecosystems, and the high yields, resemble the famous 2021 DeFi Summer. The variety of innovative protocols makes it incredibly hard for investors to keep up, while at the same time, the impressive growth raises concerns about risks accumulating in the DeFi ecosystem.

You might have heard the doomsday analyses comparing the most successful protocols of this wave, like Ethena or Eigen Layer LRTs, with risk management disasters like Terra, without really providing any credible evidence of the parallels. Fact is, this new generation of fast-growing DeFi protocols is much more mature and a lot of thought has gone into risk management. However, there is still plenty of risk.

Jesus Rodriguez, the CEO of IntoTheBlock, is a speaker on the AI Stage at Consensus 2024, May 29-31.

The biggest risk in the current DeFi market is not based on mechanistic failures such as those that caused the collapse of Terra, but rather on three key factors: scale, complexity, and interconnectivity.

Protocols in this DeFi wave have grown quite large in a matter of months, they enable more complex financial primitives, and they are incredibly interconnected. That combination of complexity, size and interconnectivity has drastically outpaced the capabilities of the risk models in the current DeFi market. In simple terms, there are plenty of risk conditions in the current DeFi markets for which we don’t have credible risk models. And that gap seems to be increasing, not shrinking.

The Four Biggest Risks in Modern DeFi

Risk has been part of the DeFi narrative since the beginning, and it’s very easy to discuss it in broad, generic terms. This new era of DeFi brings novel innovations and has grown significantly fast. As a result, risk is taking on a different connotation than before. Taking a first-principles approach to analyze risk in this era of DeFi highlights four fundamental factors: scale, speed, complexity and interconnectivity.

To illustrate these factors, consider the differences in quantifying risk for a basic AMM with a few hundred million in TVL versus an AMM that uses restaked assets with their corresponding point systems and introduces its own tokens and points. The former risk model can be solved with basic statistical or machine learning methods. The latter enters the domain of much more advanced branches of mathematics and economics such as complexity or chaos theory, which are nowhere near being applied in DeFi.

Let’s look at the different factors in more detail.

1) Scale

The principle of the relationship between risk and scale in DeFi is incredibly simple. In financial markets, modeling risk at a smaller scale, say a few hundred million, is very different than at a few hundred billion. At a larger scale, there are always risk conditions that surface that were not present at smaller scales. This principle certainly applies to DeFi as a parallel financial system with many interconnected primitives.

Ethena is one of the most innovative projects of the current wave of DeFi and has attracted billions in TVL in just a few months. The biggest challenge for Ethena in the current market is to adapt its risk and insurance models to that scale in the event of funding rates going negative for a long time.

2) Speed

The relationship between risk and speed is the traditional friction between growing too big too fast. As a risk condition, speed acts as an accelerator to scale. A protocol that goes from a few million to a few billion in TVL in just a few months might not have the time to adjust its risk models to the new scale before unforeseen risk conditions appear.

The rapid rise of EigenLayer triggered an entire movement of LRTs, several of which grew to several billions in TVL in just a few months while still lacking basic functionalities such as withdrawals. The combination of speed and scale can exacerbate simple depegging conditions into really impactful risk factors in some of these protocols.

3) Complexity

The entire field of complexity theory was born to study systems that escape the laws of predictive models. Economic risk has been at the center of complexity theory almost since its early days as world economies rapidly outgrew risk models post-World War II. Modeling risk in a simple economic system is, well, just simple.

In the new wave of DeFi, we have protocols such as Pendle or Gearbox, which abstract quite sophisticated primitives such as yield derivatives and leverage. The risk models for these protocols are fundamentally more difficult than those from the previous generation of DeFi protocols.

4) Interconnectivity

Widely interconnected economic systems can be a nightmare from the risk perspective as any condition can have numerous cascading effects. However, interconnectivity is a natural step in the evolution of economic systems.

The current DeFi ecosystem is much more interconnected than its predecessors. We have restaking derivatives in EigenLayer being tokenized and trading in pools in Pendle or being used with leverage in Gearbox. The result is that risk conditions in one protocol can rapidly permeate through different key building blocks of the DeFi ecosystem, which makes risk models incredibly challenging to build.

Transitioning from Technical to Economic Risk

Hacks and exploits have been the dominant risk theme in DeFi for the last few years, but that might be starting to change. The new generation of DeFi protocols is not only more innovative but also much more robust from the technical security standpoint. Auditing firms have gotten smarter, and protocols are taking security much more seriously.

As an evolving financial system, the risk in DeFi seems to be transitioning from technical to economic. The large scale, fast growth speed, increasing complexity, and deep interconnectivity are moving DeFi into unforeseen territories from the risk perspective. With only a handful of companies working on risk in DeFi, the challenge is now to catch up.