Jose Maria Macedo

@zemacedo

In defense of Ethereum and its fatness: why I’m still bullish on ETH

One of the primary drivers behind Ethereum‘s 500x returns over the last few years has been the so-called “Fat Protocols” thesis, initially put forth by Joel Monegro in his 2016 article. In it, Monegro argued that while the previous internet stack resulted in most of the wealth being captured at the application level (Facebook, Amazon, etc), the blockchain stack will see most of the wealth captured on the protocol level (Ethereum, Bitcoin, etc).

However, this thesis has recently come under attack from a series of commentators who argue the “fatness” of protocols, and consequently Ethereum, may have been overstated and Ethereum is actually thin. These fears and the so-called “Thin Protocols” narrative were reflected in a 30 page report issued by Tetras Capital on why they were short Ether where they directly address the fat protocol thesis. Since then, ETH shorts have hit record highs and the price of Ether has dropped over 50%, with some including TechCrunch declaring that “the collapse of ETH is inevitable”. The criticism is especially serious because it’s not the typical misinformed “No one is using it” argument peddled by bears. In fact, most of the critics are leading blockchain thinkers who are bullish on both the space and Ethereum as a network but do not believe Ether itself will capture the value created by the Ethereum network.. As they’ve put it: Long Ethereum, short ETH.

While many of these critics raise interesting points and force us to think more carefully about the way Ethereum captures value, on the whole the argument is unconvincing. In this article I’ll present some of the arguments used for a thin Ethereum and argue most of them do not apply to the current Ethereum network and none of them apply to an Ethereum with Casper and PoS.

What are fat protocols?

(If you’re familiar with fat protocols, feel free to skip this part and go straight to the arguments)

One of the primary drivers behind Ethereum‘s 500x returns over the last few years has been the so-called “Fat Protocols” thesis, initially put forth by Joel Monegro in his 2016 article.

In this article, Joel draws a comparison between the “value capture” characteristics of the internet technology stack of the 1990s and the blockchain one of the 2010s. While the early internet protocols such as HTTP , TCP/IP and SMTP created massive value as they form the underlying basis of the internet we use today, they captured none of this value. Instead, the application layer built on top of these protocols (Facebook, Amazon, NetFlix, Google) captured all the value and generated billions of dollars. The internet stack, in terms of value capture, can therefore be thought of as a “thin” protocol layer and a “fat” application layer.

However, blockchain, Joel argues, reverses this by: a) allowing protocols to be monetized through the use of cryptographic asset tokens (i.e. Ether in the case of Ethereum) and b) creating a shared data layer which reduces the monopolistic data silo advantages enjoyed by current internet applications. Thus, in terms of blockchain value capture, Joel argues that in contrast to the internet there will be a“fat” protocol layer and a “thin” application layer.

Source: James Kilroe

This view was widely accepted for several years, with the theses of many prestigious hedge funds such as Polychain centering around “investing at the protocol layer of web 3.0”.

The appeal of the thesis is easy to see: had TCP/IP been investable, it would undoubtedly have been one of the all-time great investments. Moreover, while investing in applications carries with it the 95% startup failure rate, investing in a protocol token theoretically allows one to diversify across all applications built on that protocol since protocols capture the value of everything built on top of them. As Monegro pointed out, this thesis is also backed by empirical observation as Bitcoin and Ethereum, the two largest protocol networks, are worth many times more than the most valuable application companies built on it such as Coinbase and Poloniex.

However, this thesis has recently come under attack from a series of commentators who argue Ethereum’s “fatness” may have been overstated or at least misunderstood. This is a crucial issue as the degree of fatness of Ethereum is very closely related to its market cap and thus its investment value.

Thank you to Thomas Ngai for the fat ETH photoshoppage

There are 3 separate arguments that are traditionally made: (1) Economic abstraction argument (2) Race to zero argument and (3) Velocity argument. I’ll now present these three arguments and seek to show why they do not apply to the current Ethereum implementation or to Casper.

Argument 1: Economic abstraction- ETH isn’t necessary to pay gas

The first argument I’ll address is the “Economic Abstraction” argument as put forth by Jeremy Rubin in his article “The Collapse of ETH Is Inevitable”. Rubin compares the Ethereum network, as a decentralized world computer, to a shared car:

“The Ethereum network is like a shared car. When a contract wants to be driven by the shared car, the car uses up fuel, which you have to pay the driver for. How much gas money you owe depends on how far you had to be driven, and how much trash you left in the car.”

In this metaphor, ETH (gas) is paid to a miner (the driver) to process computation (how far you want to be driven) and contract storage (how much trash you left in the car). In this case, the value of ETH could be derived from the demand for computation and storage or in other words, demand for gas.

However, Rubin argues, the metaphor doesn’t work because whereas gas is actually necessary to the operation of the internal combustion engine of a car, there is no hard (read: physical) requirement for Gas or ETH in an Ethereum contract. In fact, through a phenomenon referred to as “Economic Abstraction”, it is possible to pay for Ethereum fees in other currencies such as ERC-20 tokens. As Rubin tells us:

Suppose we’re building a new decentralized application, BuzzwordCoin. By default, following a standard ERC-20 Token template, every transaction on BuzzwordCoin will pay gas in $ETH. Requiring every BuzzwordCoin transaction to also depend on ETH for fees creates substantial risk, third party dependency, and artificial downwards pressure on the price of the underlying token (if one must sell BuzzwordCoin for ETH ahead of time to run a BuzzwordCoin transaction, then the sell-pressure will happen before the transaction requires it, and must be a larger sale than necessary to ensure sufficient funds to cover the transaction).

Instead of paying for Gas in ETH, we could make every BuzzwordCoin transaction deposit a small amount of BuzzwordCoin directly to the block’s miner’s address to pay for the contract’s execution. Paying for Gas in a non-ETH asset is sometimes referred to as economic abstraction in the Ethereum community.

Basically, Rubin’s argument can be summarised as follows:
(1) The value of ETH comes from its use in paying gas fees for decentralized computation.
(2) There is no reason for ETH to be used to pay gas fees, as through economic abstraction any other currencies can be used.
(3) Rational, independent and self-interested miners will choose to be paid in assets of their own choosing rather than in ETH.
(4) There is no reason for ETH to be valuable and “the collapse of ETH is inevitable.”

Counter argument

First of all, it’s worth nothing that this same argument can be applied to almost all other PoW protocols including, for instance, Bitcoin. However, in the case of Ethereum, the argument is not true today and it will become even less true in future once Casper is released.

There are two clear benefits to paying gas fees in ETH today: (1) ETH is the only medium of exchange on Ethereum where the gas cost of transactions is 21,000 GWEI rather than 40,000 GWEI, a 47.5% discount. (2) Paying for gas fees in ETH is built-in and has no gas cost of its own, so there is no “tax tax” (i.e. paying gas for the gas paying transaction). If miners are truly “uncoordinated, mutually disinterested, and rational” as Rubin assumes, then there’s no reason for them to harm their own profits by accepting any currency other than ETH.

If we consider the future roadmap of Ethereum, the argument becomes even weaker. In an Ethereum in which Proof-of-Stake (PoS) is implemented, owning/staking ETH becomes a requirement to become a validator, create blocks and receive gas fees. As such, even with full economic abstraction where gas fees can be paid in any currency, we can use a Discounted Cashflow Model to estimate ETH’s valuation as ETH acts as the requirement to receive gas fees, regardless of what currency they’re paid in. ETH thus acts similarly to taxi medallions which you must buy and own for the right to work/mine for the network, even if payouts happen in dollars/other currencies.

Not only this, the Ethereum community is currently considering two proposals, both of which enshrine the need to pay gas fees in ETH at the protocol level. The first proposal, as described in this paper, is a self-adjusting minimum transaction fee charged to the block proposer and payable in ETH. This means that even under Economic Abstraction where users can pay gas fees in Buzzword Coin or any other currency, the block proposer still has to pay this fee in ETH. The second, as described by Vitalik is “a storage maintenance fee (aka “rent”)” where users “pay N wei per byte per block to keep data in storage”. Both of these would have the effect of making ETH’s use mandatory at the protocol level. Additionally, since both minfee and storage fee will be burned, it will cause little increased velocity (more on this later).

Argument 2: Race to zero — The protocol layer may be fat, but market forces will ensure individual protocols are commoditised and thin

This is the argument made by many prominent thinkers in the crypto space, including James Kilroe in his piece about applications being the better investment, by Teemu Paivinen in his blog about thin protocols and mentioned by prominent names such as Travis Kling and Rocco in their podcast with Crypto Bobby.

The argument states that while the protocol layer itself may be fat, due to various competitive market forces (including scaling, forking, competition and interoperability) it is unlikely to be dominated by one large protocol (i.e. Ethereum) but it will instead be thin, split up amongst multiple smaller, specialized players.

Source

In order to break down this argument, it is important to look at each of these market forces in turn.

Scaling

While Ethereum’s value is derived from the gas fees paid to miners which should grow in number as usage increases, Ethereum is also implementing various scaling solutions, meaning the fee required to process each transaction will drop. Thus, it is argued, even as the number of transactions processed may increase, the amount of gas fees paid may decrease as long as price falls faster than the number of transactions processed.

However, this argument is assuming that the price will drop faster than adoption will increase. In other words, it is assuming that the demand curve for gas is convex (price inelastic) rather than concave (price elastic). This doesn’t make sense as demand for technology/computing power has always been elastic as there’s no upper bound on demand for computation. In order to understand this, we can look at the following two graphs representing supply and demand equilibria for the gas market, with the Y-Axis denoting the price per unit of gas and the X-Axis denoting quantity supplied/demanded of gas.

Inelastic (convex) demand curve for gas

What we can immediately see is that in the case of a convex demand curve, any scalability gains, symbolized by a shifting down of the supply curve, will lead to a smaller increase in quantity of gas demanded. This will result in a decrease in total gas fees collected, as represented by the area under the graph where S3 meets the demand curve. Once again, it is immediately apparent that the area under (S3;D) is ~34 units, ~28 units smaller than the area under (S1;D) which is ~62 units. This shows that, if the demand curve is convex, increases in scalability will lead to decreases in total gas fees collected and therefore in demand for ETH.

Elastic (concave) demand curve for gas

On the other hand, in the case of the concave demand curve, any scalability gains, symbolized once again by a shifting down of the supply curve, will lead to an exponential increase in quantity of gas demanded. This will result in an increase in total gas fees collected, as represented by the area under the graph where S3 meets the demand curve. As is immediately apparent, the area under S3;D is 36 units, 11 units larger than the area under S1;D which is 25 units. This shows that, as long as the demand curve is concave, increases in scalability will lead to increases in total gas fees collected and therefore in demand for ETH.

As such, the validity of the scalability argument hinges on the shape of the demand curve for gas. If the curve is convex, critics are right and scalability will drive the price of ETH down. If the curve is concave, critics are wrong and scalability will drive the price of ETH up.

In this author’s opinion, it makes no sense for the demand for computation to be convex as demand for computation and technology more generally has always followed a concave demand curve and this would require a fundamental shift in the characteristics of demand for computation. As “Getting the Most out of Information Systems” tells us:

“When technology gets cheap, price elasticity kicks in. Tech products are highly price elastic, meaning consumers buy more products as they become cheaper.A s opposed to goods and services that are price inelastic (like health care and housing), which consumers will try their best to buy even if prices go up. And it’s not just that existing customers load up on more tech; entire new markets open up as firms find new uses for these new chips.”

To see this in action, we can look at the five waves of computing, as per Michael Copeland’s “How to Ride the Fifth Wave”. What we see is that since Moore’s Law (the doubling of computational capacity every 6 months) kicked in in the 1970’s, it has been accompanied by a far greater than double increase in demand for computation:

“In the first wave in the 1960s, computing was limited to large, room-sized mainframe computers that only governments and big corporations could afford. Moore’s Law kicked in during the 1970s for the second wave, and minicomputers were a hit. These were refrigerator-sized computers that were as speedy as or speedier than the prior generation of mainframes, yet were affordable by work groups, factories, and smaller organizations. The 1980s brought wave three in the form of PCs, and by the end of the decade nearly every white-collar worker in America had a fast and cheap computer on their desk. In the 1990s wave four came in the form of Internet computing — cheap servers and networks made it possible to scatter data around the world, and with more power, personal computers displayed graphical interfaces that replaced complex commands with easy-to-understand menus accessible by a mouse click. At the close of the last century, the majority of the population in many developed countries had home PCs, as did most libraries and schools.
 
Now we’re in wave five, where computers are so fast and so inexpensive that they have become ubiquitous — woven into products in ways few imagined years before. Silicon is everywhere! It’s in the throwaway radio frequency identification (RFID) tags that track your luggage at the airport. It provides the smarts in the world’s billion-plus mobile phones. It’s the brains inside robot vacuum cleaners, next generation Legos, and the table lamps that change color when the stock market moves up or down.”

If we think of Moore’s law as being similar to scaling measures which shift the supply curve down, we can thus see that this increase in efficiency , far from reducing demand for computation as would seem to occur with a convex demand curve, took us from mainframes → supercomputers →personal computers → smartphones →(eventually) IOT devices, with each increasing total computation demanded. It makes sense that a similar phenomenon would occur with blockchain scalability.

As Vitalik puts it in his 2018 deconomy presentation:

Source: https://www.youtube.com/watch?v=7WL9hr445uo

It is assumed that to use blockchain technology, you must value its benefits (security, decentralization) more highly than its costs (scalability and efficiency losses). Thus, due to the current scalability limits of Ethereum (high price per gas), only the applications that benefit from Ethereum the most can afford the efficiency losses and the number of beneficiaries (demand for gas) is small. As scalability improves and efficiency losses are smaller (lower price per gas), applications that benefit less from blockchain will be able to use it, and thus the number of beneficiaries (demand for gas) will increase.

Crucially, each increase in scalability, which will lower the price per gas, will cause an increase in the number of beneficiaries which previously could not use the network, increasing amount of gas demanded. Initially, these beneficiaries will come from current users of centralized computation networks. However, as the efficiency of blockchain overtakes that of centralized networks, the beneficiaries will begin to come from use cases which are not possible given current scalability limits. These include things like climate modeling, life sciences and generally machine learning. Indeed, according to McKinsey, in 2016, the world produced 16 zettabytes of data, and yet only analysed 1% of it. By 2025, the world’s data generation is expected to surpass 160 zettabytes. The demand for computation to analyze all this data is not likely to level off anytime soon.

As such, it would seem to me that if we make the very reasonable assumption that the demand curve for gas, similar to the demand curve for computation, is elastic/concave, then scalability increases should lead to an increase in total gas demanded (and thus in the price of ETH), rather than a decrease.

Forking and competition

The argument here is that ‘general’ base level protocols such as Ethereum will eventually be commoditized due to: a) Competition from EOS, NEO, Dffinity, Tezos and all the other specialized smart contract protocols and 
b) The open-source nature of these protocols which allows for easy forking and the creation of bespoke designs for niche use-cases.

There are actually two related but separate arguments here.

The first is to do with monopolistic power. If a protocol like Ethereum becomes fat, it is argued that this must mean it is using its monopoly power to charge economic rent and capture a disproportionate amount of value (in economic terms, generating “Supernormal Profits”). These profits will attract new entrants which, given the low costs of forking and/or creating a competitor, will enter and charge lower fees, competing away the Supernormal Profits until only normal profits remain and the fees reflect costs of computation.

The second is to do with specialization and goes something like this: If a general base level protocol exists, it will not suit every use case perfectly. Since protocols like Ethereum are open-source, users can simply fork it or create a competitor that suits their specialized use case, providing some portion of the functionality in a more efficient way.

There are several problems with both of these arguments. The first and biggest flaw is the idea that forking/creating a competitor is a zero cost activity and therefore Ethereum has zero pricing power. In reality, I would argue forking has extremely high coordination costs and Ethereum has significant pricing power, equivalent to the size of its network effects. In accordance with Metcalfe’s law, the value of a network is proportional to the square of the number of connected users. As such, network effect (number of users) is a quadratic factor in deriving network value. Given this, the total value of a split community is necessarily lower than that of one unified community. This emphasizes that forking/creating a competitor has costs greater than zero. More importantly, the cost of forking goes up exponentially as the size of the community increases, also known as network effects.

Not only this, a fork/competition to implement lower prices isn’t as simple as creating a competitor and charging lower prices in traditional markets, because whereas in traditional markets companies can use equity/debt to finance undercutting their competitors’ prices, this is not feasible in blockchain. In fact, in blockchain you’ll have to convince a sufficient number of rational, self-interested and largely uncoordinated miners to join in with you and, presuming you’re undercutting Ethereum, undercut their own profits to bootstrap this new chain. This is perhaps why, as Savant Specter tells us, looking at the major open source communities surrounding programming languages/OS’s, there aren’t infinite forks of every project. Instead, there are a few forks, but most people just use the major release of the biggest projects.

Crucially, forking/creating a competitor to Ethereum is likely to become even more difficult once Casper and PoS come around. Whereas miners in Ethereum under PoW are currently not necessarily invested in ETH as they can mine ETH but instantly sell block rewards, making it theoretically easier to convince them to fork/join a competitor, in the case of Ethereum under PoS miners will be forced to hold and therefore be invested into ETH in order to mine and receive block rewards. As such, their economic incentives will align with maximising the value of ETH and minimising competition and forks.

Addressing specifically the second argument regarding specialization, while it is true that specialization is important and there will likely be many different blockchains making different trade-offs to serve specific use cases, I would argue the primary competitive advantages of blockchain that differentiate it from the extremely efficient centralized databases like AWS are security and decentralization.

Given this, any scalability gains made by sacrificing these two properties need to be considered extremely carefully and in my opinion make much more sense to be implemented on a Layer 2 solution like Plasma Chains which can allow for greater scalability through off-chain processing while still benefitting from the security and censorship resistance provided by the main chain. Why do we need to rely on the 21 delegates in EOS not acting maliciously if it could simply be implemented as a Plasma Chain on Ethereum, achieving similar scalability while being grounded in the security of the main chain? In my opinion, Ethereum is making the best trade-offs in seeking to negotiate the famous blockchain trilemma, scaling up while maintaining the security and decentralization that provide blockchain’s USP and make it preferable to other technologies.

Indeed, this seems to be being borne out in practice as while EOS is trying to brand itself as the “gaming” blockchain, LOOM is already scaling an Ethereum NFT card game called ZBCardGame using layer 2 DPOS sidechains called DAppChains that are bridged to Ethereum. Similar to EOS, Loom enables fast transactions at zero cost to users, while keeping the security of a decentralized blockchain (unlike EOS).

Furthermore, with PoS it is the price of ETH, rather than hash power, that determines the security of the network. Indeed, while in PoW the security of the blockchain is derived from the amount of hash power connected to it as an attacker must seek control at least 51% of the hash power to execute an attack, in PoS the security of the blockchain is derived from the price of ETH (as well as the percentage of ETH staked) as an attacker must buy up and control 51% of staked ETH in order to execute an attack. As such, under PoS, the “fatness” of ETH in the form of its price/market capitalization is directly related to its security. Given security is one of blockchain and Ethereum’s main competitive advantages, it is now also in the interest of every Ethereum stakeholder (including protocol developers — perhaps Vitalik will be more careful saying stuff like this) to maximise the value of ETH. This could also provide additional network effects to Ethereum since the market cap of ETH is directly proportional to its security; one of blockchain’s primary competitive advantages.

Argument 3: ETH is high velocity, “money, not equity”

This is the argument made by John Pfeffer, Trent Eady and others. While he mentions it elsewhere, this argument is laid out most clearly in Pfeffer’s “Doubts About the Long Term Viability of Utility Cryptoassets”:

“…if a cryptoasset isn’t a dominant non-sovereign monetary store of value (“SoV”), it’s somebody’s working capital. Economic agents seek to minimise working capital (because of the opportunity cost of capital) and the level they hold is a function of the friction, latency and uncertainty of replenishment. Protocol-land will be frictionless, interoperable, forkable and open-source, so users won’t need to tie up capital in stocks of utility protocols, which will push their velocity to very high levels… High velocity will mean that the network value of a cryptoasset (as measured in some external measure of value) will be low compared to the economic activity denominated in that cryptoasset (measured in the same external value measure). This circumstance means that it will not be possible to secure the blockchain in question without reliance on transaction and/or other kinds of usage fees paid in non-native currencies (could be fiat or a dominant non-sovereign monetary SoV cryptoasset). At that point, there’s no reason to have a native currency for that protocol anymore, the native currency collapses and the protocol switches to a transaction/usage fee-only model paid in a non-native currency.”

Basically, his argument is that if protocol/utility tokens such as ETH aren’t a store of value, then they must be money or what he calls “working capital”. Given the opportunity costs of money (i.e. the return you’d make from investing it elsewhere), people seek to minimise the amount they hold, and the amount they hold will be dependent on the friction, latency and uncertainty of replenishment (i.e. how difficult it is and how long it takes to acquire more of it). Since protocols will be frictionless, open-source and forkable, users will not need to hold ETH as they can easily acquire it when they need to use it and sell it straight afterwards, leading to extremely high velocity. This high velocity will lead to bad value capture mechanics (for more on this, see my previous blog post) which will mean that the market cap of Ether will be low compared to the economic activity or transaction volume denominated in it. Given this, it will be impossible to secure the Ethereum blockchain using ETH (since both amount of hash power in PoW and value of ETH staked in PoS are dependent on market cap) which will be rendered useless and collapse as we instead move to a non-native token which Pfeffer suggests will either be FIAT or a non-sovereign SoV (i.e. Bitcoin).

Crucially, Pfeffer directly addresses PoS and argues that this is not a solution to the velocity problem, as although some tokens will be locked up as stakes which will reduce overall velocity, the very high velocity of non-staked will still result in very high average overall velocity:

“…let’s assume for simplicity the absolute floor on 1/V is the block time of the chain in question. Let’s then take ETH with a 2.5 minute block time as an example (highly theoretical, just to make a simple maths point). This implies each token could be used (assuming fixed block times, which in fact will likely shorten) 210,240 times a year. Buterin, Choi, etc. talk about, say, 10% of ETH being staked (let’s assume staked tokens never move at all). That would bring V down to 189,216 per year. Assume 50%, then V=105,120. Multiply this last number by $50b of network value (i.e., ETH just maintains its current value, and you’d need $5.25 quadrillion of economic activity denominated in ETH (i.e., excluding any ERC20/ ERC721-denominated economic activity), that is to say, 65x the current global GDP of $80 trillion. These numbers are all just varying shades of silly. That’s the point. As long as some of your tokens are circulating at a high V, your overall V is high.”

While the logic of this argument is sound and applied when it was written in April, it doesn’t address the latest targeted implementation of Casper and PoS which includes two proposals designed to counterbalance the velocity issue by creating a proportional deflationary force: (1) A self-adjusting minimum transaction fee charged to the block proposer and payable in ETH which is burned and (2) A storage maintenance fee (“rent”) where uses pay N wei per byte per block to keep data in storage, with N being burned. Vitalik has previously said that he estimates “well over 2/3rds of transaction fees paid could end up being burned through these mechanisms”.

The effect of the fee burns is to counter the downward price pressure caused by increased velocity with an upwards price pressure from a decrease in total supply, since each increase in velocity will cause additional transaction fees to be burned. To see how this works, let us examine the Equation of Exchange as applied to crypto by Burniske and Vitalik.

Vitalik takes MV=PT and in order to simplify the analysis of cryptocurrencies recasts it as MC=TH, where:

M= total money supply (or total number of coins) 
C= price of the cryptocurrency (or 1/P, with P being price level) 
T= transaction volume (the economic value of transactions per time) 
H= 1/V (the average time that a user holds a coin before using it to make a transaction)

Therefore, the left part of the equation (MC) is simply the market cap (total supply*price) whereas the right side is the economic value transacted per time period (T) multiplied by the average time a user holds a coin (H).

To solve for the token price, one must therefore solve for C:

C=TH/M

What we can now see that although higher velocity (or reduced holding time H) leads to lower token price, a decrease in total supply (M) through the fee burns will have the opposite effect, leading to an increased token price.

It’s also worth noting that unlike in PoW where miners receive their block reward payouts and can instantly sell them the end of each block (~every 2.5 minutes), resulting in extremely high velocity, in PoS validators will only be paid 3x per year (~every 4 months), greatly reducing token velocity as compared to PoW.

I’d be very interested in seeing someone model out the effect of the fee burn at different velocities on the overall transaction volume required to maintain ETH’s $50B market cap. This would also help the Ethereum team be able to use the min fee to target a deflation rate based on different token velocities. However, I’d suspect the transaction volume required will be far lower than the 5 quadrillion suggested by Pfeffer.

Conclusion

While “Thin Protocols” is an interesting and contrarian thesis that forces us to think more clearly about what makes Ethereum fat, (i.e.what are the mechanisms by which it actually captures value), I believe that upon closer inspection most of the arguments do not actually apply to Ethereum in its current state and particularly do not apply to the targeted implementation with Casper and PoS, which is what we should be judging Ethereum on given it has been made abundantly clear we will not be sticking with PoW.

In fact, even in its current state with PoW, a recent study by Sebastian Wurst examined Ethereum’s top 250 ERC20’s with a view at empirically examining the fat protocol thesis via value accrual in 4 layers of the stack. The conclusion was that, given the market cap of these projects was about $10B compared to the $20B of Ethereum, the initial premise of the fat protocol thesis seems to hold true.

Overall, I believe the popularity of the thin protocol thesis is a phenomenon similar in character to the “Blockchain, not Bitcoin” thesis that emerged during the Bitcoin bear market in 2015. Indeed, as prices fall for no apparent reason, causing panic and fear among investors, these theses emerge as a post-hoc attempt to explain these falls and seek to provide some sort of rational justification for ultimately irrational short-term price movements. In reality, in the short-term, markets are driven more by sentiment than fundamentals and I believe, once ETH inevitably recovers as Bitcoin eventually did in 2016, the “Thin Protocol” thesis will go the way of the “Blockchain, not Bitcoin” thesis which has largely been abandoned as most private blockchains fail to live up to expectation. In the meanwhile, I’ll be buying.

Thanks very much to Colm Buckley for his feedback in drafting this post.

Topics of interest

More Related Stories