Cracks in the Fat Protocol Theory, and its effect on crypto portfolio construction
Cracks in the Fat Protocol Theory, introducing the Reffal Curve, and why & how crypto portfolio construction are affected by your position on the Fat Protocol Theory
In 2016, Joel Monegro of Union Square Ventures published a seminal thesis on how crypto protocols will be able to capture more value than the dapps on top of them. Put another way, the Fat Protocol Theory stated that for every dollar of value capture that the dapp layer had, the protocol layer would be able to capture at least that same dollar, or more, given the need of the dapp layer to utilize the native protocol token to power all interaction.
This thesis was critical in framing where value will be captured in crypto, where investors should invest, and how to think about the application vs the protocol layer. Unfortunately, due to a range of recent developments, we have seen that this thesis is under significant pressure, and explore if the death kneel for the Fat Protocol Theory can be sounded. The importance of understanding this thesis is where it affects our view on where we should allocate investment focus across protocols and dapps, and the stage of first entry. Ashwath of Delphi Research noted that the “legitimacy of the fat protocol thesis is key to Ethereum’s continued dominance.”
The main drivers of this theory’s decline include the following:
Reduced monopolism: Multi-chain Dapps/ Dapp branding
Overstated network effects: Long-term protocol competition leading to lower fees
Reffal Curve: The role of Roll-Ups reducing total demand
New framework of thinking of L1 value capture: Currency v Nations
Why the Fat Protocol Theory made sense
The Fat Protocol Theory was based on 3 main descriptions of the interaction between protocols and dapps:
Any dapp building on top of a protocol will require the protocol’s native token to power any transactions within the dapp.
It is far easier to clone dapps and attract users than it is to clone protocols and attract users, allowing protocols to take on a more monopolistic competitive position relative to dapps
Protocols will thus capture more value than dapps, since the protocol’s value capture will be reflected in any transactions that the dapp layer processes (value creation), let alone the value that the dapp itself will capture.
Measuring the cracks in the Fat Protocol Theory
Joel John of LedgerPrime noted that the first cracks of the Fat Protocol Theory emerged when capital-intensive use-cases like stablecoins emerged. With stablecoins like Tether driving adoption of ERC-20 tokens, the market cap of ETH did not rise as much in proportion to the rise in market cap of stablecoins, indicating a fundamental break in the relationship between protocol/dapp value capture (at least for stablecoins).
The main way that the Fat Protocol Theory is proved is that we should expect the market capitalization of dapps on Ethereum to be a fraction of the market capitalization of Ethereum itself. The alternative situation would be described as a ‘flippening’.
The thinking is that temporary speculative bubbles in dapp valuations can exist, but must fundamentally revert to being under ETH’s market cap. However, since the thesis has been published, we have seen multiple long-term instances where the ERC-20 dapp layer market cap has exceeded ETH’s market cap since 2020, 2021, and today in 2022 (320b v 311b).
(Snapshot: Calculated by summing the 308 largest ERC-20 token market caps on CoinMarketCap on 20 Feb 2022)
On the one hand, it can be argued that the dapp market cap ‘flippening’ is a result of temporary speculative bubbles in dapp valuations, and we should expect a long-term reversion of ETH’s market capitalization to be above that of its underlying dapps. On the other hand, it may be reflective of the increasing dapp-product-market-fit in a maturing crypto ecosystem that will look to stress the Fat Protocol Theory even further. Furthermore, it can be argued that the current divergence is understated. Centralized organizations that interact with protocol may contribute to the protocol development, but not necessarily have their value-capture reflected in tokens. A good example of this is Opensea or Coinbase, where value capture by these organizations is reflected in the equity value of their company, rather than the tokens that were not issued.
Overstated network effects: Long-term protocol competition leading to lower fees
Joel’s thesis was written in a time when there was an overwhelming sentiment toward single-chain maximalists. It seemed that protocols provided the ultimate network effect that would be hard to replicate. We see today that due to ETH scalability issues, and general capital/ financial incentives, there was a strong motivation and wedge for competing, increasingly specialized, protocols to emerge.
One of the main ways that value capture by the protocol level would take place would be through the transaction/gas fees that the protocol captures in a transaction it facilitates.
The emergence of cross-chain infrastructure helped to promote an increasingly sustainable multi-chain world. Protocol competition then sparked a drive for lower protocol fees, even if the mechanism for value capture existed (see Solana’s waived rent fees).
Multi-chain Dapps & Limitation of Forking
A dapp building on a single ecosystem restricts their user base, which is increasingly non-sustainable given the popularity of multiple competing protocols. Dapps will increasingly be multi-homing, the act of ensuring native interoperability across multiple protocols, to prevent being ‘stuck’ on one protocol. For example, increasing ETH gas fees saw a migration of Dapps and crypto holders into alternative L1s like AVAX, SOL, etc. Dapps, seeing the trend, are able to diversify their product offerings/user base by providing multiple integrations to serve communities across multiple protocol ecosystems. AAVE building on Polygon as well as AVAX is a good example of this phenomenon. As Dragonfly Capital noted, “Binance Smart Chain taught everyone: if you don’t launch here, we’ll just launch a fork of you and take the revenue you would’ve gotten”.
Furthermore, there was increasing recognition of the value of the brand equity for dapps. Uniswap expanding into another chain like AVAX would in all likelihood be more successful than a 3rd party Uniswap fork on Polygon, due to Uniswap’s ability to leverage existing liquidity relationships, talent, capital, distribution, etc. Consumers would also be more likely to trust Uniswap given their prior brand equity built up (even with no experience on Polygon itself).
The premise of monopolistic value capture by the protocol layer has been weakened by these developments.
The role of Roll-Ups
“But you mentioned L2s like Polygon that are on ETH!”
L2s are helpful in helping to solve some of the key constraints of ETH, and retain transaction volume on ETH. However, as L2s roll-up and consolidate a larger number of transactions into a single L1 transaction, it will reduce the amount of ETH that is needed for the same number of transactions.
One argument is that, similar to tax revenues, lower transaction costs (tax rates) can still produce a higher aggregate demand for ETH (tax revenues), as it would help increase the rate of dapp development/transaction.
Creating the above, I call this theory the Reffal Curve of Protocol Value Capture, in respect of my economist teacher, Mr. Reffal (/s). Although we talk about this theory in relation to ETH L2s and their impact on demand on L1 ETH, it is also broadly applicable to any alternative L1s and their internal gas policy. It is important to note that the Reffal Curve, particularly the areas to the right of B, could not have existed in the view of the Fat Protocol theory since it assumed a monopolistic view of protocols. Here, the curve to the right of B indicates users and dapps forking or switching to alternative L1s to conduct transactions.
Accordingly to the Reffal Curve, alongside increased protocol competition and reduced gas fees from L2s, I believe we are far more likely to see a situation closer to point A, rather than point B. Ethereum maximalists will disagree and believe that we are headed to a scenario closer to point B from point C, while proponents of a multi-chain world may find themselves on the side of point A, given the speed at which alternative L1s have been able to take market share away from ETH, or alternatively believe in point B if they believe that the crypto macro-trend will drive transaction volume/demand far more than existing capacity is able to serve.
One pushback could be that, given what we know about the price elasticity of ETH gas costs (last calculated by Vitalik in 2018) being above 1 (i.e. is price elastic, where a change in demand is greater than a change in price). Therefore, a drop in gas price will be more likely to move us into the side of B, rather than the side of C.
My personal belief is to A as we are far more bottlenecked in terms of good crypto projects than in transaction throughput, especially taking into account essentially infinitely scalable alt-L1 throughput.
Currency v. Nations
One of the most useful criticisms to have arisen out of the Fat Protocol Theory debate has been the use of currency rather than GDP as a means to evaluate protocols. As noted by Nick Hotz of Arca, in the US for example, money supply is worth about $20T while equity capitalization is $50T, with the ratio of money supply: equity capitalization varying based on how much value that government (layer 1) is providing to its companies (native apps).
The reframing of a fundamental view of protocol value has been helpful in getting a more intuitive understanding of protocols, but also served as a further weakening of the Fat Protocol Theory.
This is given that if we view protocols as currency rather than GDP, the total market capitalization of the underlying blockchain/currency money supply would not necessarily need to exceed the market capitalization of the dapp layer.
Nick Hotz further notes that while the Fat Protocol thesis has been very profitable and ahead of its time, the main driver for its success has been due to a lack of alternative places to invest, and a lack of traction from the applications themselves, not necessarily value capture dynamics from the protocol level. Given the huge amount of relative dapp product-market fit that we have seen since 2016 when Joel wrote the Fat Protocol Thesis, the Fat Protocol Thesis is showing its limitations.
Why it matters for portfolio construction
On the one hand, what the Fat Protocol Thesis fundamentally gets correct is that there is a larger value capture by the protocol level, but we have seen value capture compression on the protocol level. However, it is indisputable that even today, the best-performing investors have seen their returns come from protocols rather than dapps. Solana for example reached 100B market cap in token value after 4 years.
At the very least, we can at least conclude that the growth of TVL/utilization beyond a certain point does not correspondingly increase protocol value. This has interesting implications for whether growth investments in mature protocols would necessarily make financial sense (in the Venture Capital sense), given the repudiation of a monopolistic single-protocol world. Its repudiation also increases the value capture potential of the dapp layer, and increased exposure to dapps within an investment portfolio.
The Fat Protocol Theory has left us with a few principles that have so far stood the test of time, and while we can expect investing into emerging protocols to continue to be lucrative in the short term, we should expect heavier portfolio allocation into dapps as a way to hedge against increased protocol competition. Cracks and limitations of the theory have developed since 2018, and is under pressure from a wide range of developments, such as increasing dapp product-market-fit, multi-chain competition, and overstated blockchain network effects, as evidenced by the dapp:protocol market cap ratio. Yet, protocol returns have so far spoken for themselves, indicating some soundness in the investment strategy in the short/medium term, though we no longer hold conviction in its long-term viability.
Thanks to Jose L Sampedro Mazon, Chiyoung Kim, Mike Giampapa for their thoughts/comments (thoughts does not imply endorsement)