Real vs Fake TVL

This is your last chance. After this, there is no turning back. You take the blue pill — the story ends, you wake up in your bed and believe whatever TVL numbers you want to believe. You take the red pill — you stay in DeFi land, and I show you what real fundamental analysis looks like.

Remember: all I’m offering is the truth. Nothing more.

Utility of TVL as a Metric

In the early days of DeFi (2018–2019), one of the most common debates on cryptotwitter was about which metric to use to describe the success of various DeFi protocols. DEXs used trading volume as their primary metric of success, lending platforms used value of loans originated, and others used number of unique users. Over time, the DeFi industry converged on a single north star, TVL (Total Value Locked). This metric loosely captures how much value a protocol “owns”, which could be thought of as a proxy of how valuable the protocol is because people only deposit assets into a protocol if they can get some value out of it. This metric was popularized by the team behind DeFi Pulse.

For a long time, this metric was difficult to corrupt. A protocol needed to recruit real assets to grow its TVL. Although there were some issues double counting assets across different protocols, it still served as a fairly good gauge of what people found valuable.

Yield Farming — the death blow for TVL

What started out as a mechanism for incentivizing liquidity, yield farming quickly became the dominant way for most new projects to distribute tokens— and people loved it. Users could deposit their assets into a protocol and get that protocol’s native token for free. In many cases, those tokens would also subsequently 2-10x in value, leading to astronomical yields.

Because of yield farming, TVL’s of many different protocols were starting to rise at a shocking rate. YAM, one of the first projects to use yield farming, went from zero to $400m in TVL in less than 24 hours. More recently, another project called Big Data Protocol recruited over $6 billion in TVL in a few days.

Yield farming rendered the TVL metric useless by most measures. In the case of BDP, reasonable people would agree that this random protocol did not overtake everything else in terms of utility overnight, although its TVL did. It is clear that most of the assets that were being put to farm BDP was “fake TVL” which would be withdrawn after a week. Although we can easily dismiss the TVL of BDP as “fake TVL”, the line between what real and fake TVL is for other protocols is not as clear.

Measuring Real TVL

The way we should think about “real TVL” is how much value is being locked up that actively contributes to the application’s use case. For example, if the application we are measuring is an AMM, assets locked up are providing more liquidity for traders, so this is real TVL.

A recent example of misleading TVL numbers is Alchemix, a synthetic debt protocol. Much like MakerDAO, a useful metric for evaluating its success is the amount of collateral in the system that is used to mint debt against. Defi Llama reports their TVL to be at $920m at time of writing, so you may expect $900m of assets in the protocol being used as collateral to mint debt against, but this could not be further from the truth.

In fact, there is only $75m of Dai in the protocol that is being used as collateral — less than 10% of the reported TVL number. $500m of the TVL comes from Yield Farming incentives for providing liquidity for their native asset and the synthetic debt asset on Sushiswap and Curve. One could claim that this $500m of TVL is making the native and debt assets more liquid, but this is quite tangential to the core value proposition of the protocol (deposit collater -> mint debt). Furthermore, the remaining $140m of TVL comes from “staking pools”, which is simply locking up the native and debt asset for free rewards. This serves zero utility to the protocol, yet accounts for 2x the “real” TVL. At best, the protocol’s TVL is 60% inflated. At worst, the TVL is 12x inflated. *

Confusion about TVL attribution

Although I cherry-picked Alchemix, there are many other protocols who account for TVL the same way. The most common tricks to inflate TVL are:

  • Native asset staking for extra rewards
  • Staking LP shares for extra rewards

Conclusion

When you are trying to evaluate the next DeFi blue chip project and immediately look at the TVL, be wary of the tricks that teams use to inflate TVL numbers. Always question whether the TVL in a project meaningfully describes the core financial value proposition of a project.

If you look close and think hard enough, you will often be able to uncover some hidden truth about a project through how their TVL is described.

*these numbers were obtained from https://vfat.tools/alcx/ on 21st March 2021, and may not currently reflect the state of the project.