Research report analysis: Should investors buy ETH or L2 tokens?

Author: Michael Nadeau, The DeFi Report; Compiler: Tao Zhu, Golden Finance

Ethereum is executing on its roadmap. The network is expanding with Layer 2. Sometime this summer, we should have ETFs trading. Larry Fink won't stop talking about tokenization. We are on the brink of a Fed rate cut.

Now is a good time to be bullish on Ethereum.

Should you own ETH? Or a basket of L2s? Or both?

In this week's report, we provide a data-driven framework for portfolio construction.

L2 Data vs. Ethereum

We'll start our analysis with some high-level data, comparing the top-ranked L2s vs. ETH across multiple KPIs.

We also looked at the top L2s in each metric.

Total value locked:

Data: Token Terminal as of June 23, 2024

The top 12 L2s together account for 15% of Ethereum’s daily locked value.

Here is the TVL ranking in L2:

Data: Token Terminal as of June 23, 2024

Arbitrum ranks first, and its top 5 drivers are as follows:

1. Aave - US$790 million;

2. Pendle - US$659 million; $612 million;

3. GMX - $612 million;

4. Renzo - $371 million;

5. Uniswap - $306 million.

Notably, Blast made it to the second spot by offering yield in Ethereum (via staking) and stablecoins (via MakerDAO, Treasury Bills). We’ll be keeping an eye on this to see how sticky it is after the token drop.

Daily active addresses:

Data: Token Terminal as of June 23, 2024

Currently, the number of daily active addresses of the top L2 is 4.7 times that of Ethereum. Here is the L2 ranking list:

Data: Token Terminal as of June 23, 2024

Arbitrum once again took the first place and is the first L2 with more daily sustained users than Ethereum L1.

Data: Token Terminal as of June 23, 2024

The fees of the top 11 L2s account for 15% of the fees collected by Ethereum in the past 90 days.

Let’s take a quick look at the L2 leaderboard:

Data: Token Terminal as of June 23, 2024

In the past 90 days, Base’s fees are more than twice that of the second-ranked Scroll, accounting for 6% of Ethereum’s total fees.

90-day transaction volume:

Data: Token Terminal as of June 23, 2024

On average,the top L2 now processes 6.7 times as many transactions per day as Ethereum L1.

Here is the breakdown of L2:

Data: Token Terminal as of June 23, 2024

Base's average daily user volume is about 66% of Arbitrum. However, Base users trade 6.9 times more on Arbitrum every day, while only 2.5 times on Arbitrum.

Cost/Transaction:

Data: Token Terminal as of June 23, 2024

Average cost/transaction reveals why Arbitrum, despite being second in terms of transactions, is in the middle of the pack in terms of fees: transactions on Arbitrum cost $0.02, while transactions on Base cost $0.15. For reference, Ethereum’s average transaction fee for the same period was $5.30.

Developers:

Data: Token Terminal as of June 23, 2024

The number of active core developers on top L2s is currently about double what it is today. *Token Terminal defines core developers as the number of distinct Github users who have made 1+ commits to a project’s public repository in the past 30 days. These numbers do not include ecosystem/application developers.

Let’s see how L2 compares:

Data: Token Terminal as of June 23, 2024

Here, Optimism appears in the first place for the first time, with more than twice as many active developers on Arbitrum today. Notably, Base is in the middle - this suggests that they are earning more from developers than their competitors, taking into account applications, users, and on-chain fees.

But what do market participants hold? ETH or L2 tokens?

Data: Token Terminal as of June 23, 2024

The L2 portfolio accounts for only 2% of Ethereum’s current token holders (note that 4 top projects do not yet have tokens: Base, Blast, Linea and Scroll).

Here are the details of the 8 leaders who own tokens in the market:

Data: Token Terminal as of June 23, 2024

In terms of L2 token holder/investor mind share, Optimism and Arbitrum have 80% of the market share, but only 1% of Ethereum L1.

Finally, let’s summarize the valuations:

Data: Token Terminal as of June 23, 2024

As mentioned above, the top L2 has:

• 15% of Ethereum TVL;

• 15% of Ethereum fees;

• 4.7 of the number of active addresses times; 6.7 times the number of daily transactions; 2 times the number of developers.

Just…

Here are the fully diluted market caps of the top L2s:

Data: Token Terminal

On a fully diluted basis, Arbitrum and Optimism account for 40% of the top 10 and 3.7% of Ethereum’s market cap (1.06% of circulating market cap).

L2 Token Utility/Demand vs. ETH

Now that we have some data, let’s examine the utility and demand of ETH tokens vs. L2 tokens.

ETH

Today, ETH is used for the following:

• Paying gas fees on the Ethereum network (including L2!). If you want to transfer stablecoins, trade on a DEX, mint NFTs, play games on-chain, get loans, etc., you need some ETH.

• Collateral for yield. Want to earn yield on the Ethereum network? You need some ETH to do that.

• Collateral for loans. Currently, over 1% (2.3M) of all ETH in circulation is locked in MakerDAO smart contracts as collateral for on-chain loans.

• Medium of Exchange. Want to buy your favorite NFT on OpenSea? You need some ETH.

• Real-World Assets. Want to tokenize an asset so it can be freely traded around the world? You need some ETH.

• Deploy Smart Contracts. Want to build something on Ethereum? You need some ETH.

• Re-collateralization. ETH can be used as collateral to earn fees directly from Ethereum, or from applications and protocols within the Ethereum ecosystem. This is done by “re-collateralizing” ETH. To earn this yield, you need some ETH.

As we have seen, ETH has a wealth of utility in the Ethereum ecosystem. This utility drives demand for assets in a similar way to how oil’s utility drives demand for assets/commodities.

For these reasons, we believe ETH has more utility than any other asset in crypto today.

Layer 2 Tokens

Today, Layer 2 tokens are used for… governance.

That’s true. Some networks allow users to pay for gas fees with L2 tokens, but most of the time, those fees are paid in ETH.

In addition, L2s must pay for block space on Ethereum (where transactions settle) in ETH.

Key Point: L2s have no utility or incentive structure to drive demand for tokens today. Ethereum does.

L2 Token Value Accrual vs. ETH

How does ETH accrue value? Through real yields generated by staking (and re-staking).

How do L2 tokens increase in value? Today, no such value accrual mechanism exists, as user fees are paid to L2 sorters, not a distributed network of validators.

Execution (L2) vs. Settlement (ETH)

What’s more valuable? Execution? Or Settlement?

First, let’s horizontally set the difference between the two:

• Execution: A useful analogy might be to think of execution as ordering food at a restaurant. You choose what you want. You talk to the waiter. The waiter then confirms your order.

• Settlement: Settlement can be thought of as paying the bill, receiving the receipt, and having the transaction accounted for.

Today, the relationship between L1 and L2 is this: users execute transactions on L2, and L2 settles transactions on L1 (pays a fee for doing so).

So, what’s more valuable?

Some thoughts from the “L2/Execution” camp:

• In traditional finance, the vast majority of value accrues to the execution layer of the tech stack — brokers, market makers, and high-frequency traders. Less value is generated by settlement — think of it as clearing/accounting (DTCC).

• It’s possible that transaction fees on all blockchains will eventually drop to zero (or close to zero). If that’s the case, then MEV is how these networks will be monetized in the future. For now. If all execution happens on L2, then that’s where we’d expect MEV to come from — since execution includes batching and ordering of transactions.

Some thoughts from the Ethereum L1/Settlement camp:

• As far as the Ethereum ecosystem is concerned, there is only 1 settlement layer. Everything on top eventually lands on Ethereum, and pays ETH fees for it.

• L2 inherits Ethereum’s security and decentralization.

• The utility of the token (and subsequent network effects) may ultimately be the only thing that matters. For example, Ethereum may lose most of its fees at the L1 level, but still get cash flow from other applications and protocols that "rent" Ethereum security (including L2 without tokens such as Base).

L2 Security vs ETH

Ethereum’s security is basically 100% and will not change in the future.

Here are the current L2 profit rates, which have increased significantly since EIP4844:

Data: Token Terminal

• Base has lower profit rates due to paying 15% of its total revenue to Optimism.

• Blast has a lower balance due to delays in upgrading the network to support upgrades related to EIP4844. Since the May 27 upgrade, Blast profit margins have increased to 91%.

L2 Correlation with ETH

ARB & OP price to ETH

Data: Token Terminal

The historical price correlation coefficient between ARB and ETH is 0.71 (highly correlated).

Similarly, optimism has a low correlation with ETH, with a historical correlation of 0.61 (moderately correlated with periods of complete uncorrelation with ETH).

ARB Price to Users

Data: Token Terminal

Arbitrum has a historical correlation of -0.009 (negative correlation). This is not a typo.

How does the negative correlation between user activity and price compare?

Token unlocking. This is largely due to a lack of regulation from the SEC and Congress.

Stay tuned, we’ll cover unlocking in the next section.

ARB Price vs. Fees

Data: Token Terminal

Not surprisingly, price is also negatively correlated with fees (historical correlation is -0.004). You can’t make it up.

L2 Catalysts vs ETH

Ethereum Catalysts

The biggest catalyst for ETH next year is an ETF - which should be trading sometime this summer. Other catalysts include re-hypothecation (additional yield for ETH holders), tokenization of real-world assets, enterprise adoption, and maturation of DeFi and NFT use cases.

With the approval of an ETF, we expect Wall Street to start doing some real due diligence on the network and the asset. We ultimately believe Ethereum has a larger addressable market than Bitcoin, so it will be interesting to see if large institutions ultimately come to the same conclusion.

L2 Catalysts

We see two potential catalysts for L2 next year.

• Ethereum Price. If ETH outperforms the ETF following ETF approval, we may see some investors jump straight to L2 - seeking allocation to smaller market cap projects.

• L2 Re-rating Relative to ETH. Currently, the top L2s represent only 2.7% of Ethereum’s market cap (9% fully diluted). If the market believes that it should be closer to 10%, then L2s could outperform.

Token Unlocks

This is probably the most important section of this report.

There is no "token unlocking" for Ethereum as the supply is fully circulated. In addition, Ethereum whales who received large allocations via the initial ICO have now experienced two bull runs and can sell/recycle their tokens back into the market/new hands.

That being said, Ethereum supply may increase during periods of slow on-chain activity and "fee burn" is not enough to offset the new issuance/consensus rewards paid to incentivize validators for their services.

For reference, Ethereum was deflationary in 2023 (-.28%) and has been deflationary so far in 2024 (-.02%). Here is Ethereum’s inflation rate since its inception:

Data: Etherscan

It is worth noting that with the decline in L2 fees due to EIP4844 (network upgrade), we observe that the rate of “fee burn” is slowing down.

In fact, Ethereum’s ETH issuance has increased by 75,951 since March 13, or about 767 new ETH per day (EIP 4844 implementation). This year, the network’s supply has still decreased by 23.3k ETH. But if we don’t see increased on-chain activity in the second half of this year, the network could see mild inflation this year.

L2 is different in the following factors:

L2 distributes tokens to investors, advisors, and contributors, typically over a vesting period of 4+ years.

Today, the vast majority of L2 tokens (especially major ones) have no “buyback” or “burn” mechanism.

L2 tokens have almost zero utility, so there is no structural purchase support.

As mentioned earlier, we predict that Ethereum will see mild deflation this year.

We can contrast this with what is known about L2 token unlocks. For example, Arbitrum will unlock 1.15 billion tokens next year (36% inflation on existing supply).

Arbitrum will unlock 3.2 billion tokens over the next three years. Most of these tokens are distributed to investors and contributors with large unrealized gains.

We should expect them to be sold at the end of the vesting period - increasing the supply of tokens on the market.

We can see the impact of token unlocks in the chart below. For example, Arbitrum’s unlocks started in March, and we can see the impact on price below. This offset the incredible growth in Arbitrum’s fundamentals (pink represents active user growth).

Data: Token Terminal

Given that Arbitrum will unlock 1.15 billion tokens over the next year, nearly $1 billion in new money must enter the asset for its price to remain at $0.82 - all else being equal (ARB price at time of writing).

Predicting Prices

Remember that no one can predict prices, no matter how much data, foresight, or luck we have. In this section, we are just sharing some high-level predictions so that we can get back to reasonable price targets.

Data: Token Terminal

*The above predictions assume that ETH's circulating token supply remains constant. BTC's supply is known. Therefore, use actual values ​​and add one year of inflation. Additionally, “BTC dominance” and “ETH as % of BTC MC” are taken from cycle peaks.

We use our forecast for ETH to regress potential valuations on the top 2 Layer 2s: Arbitrum and Optimism.

Data: Token Terminal

The above data table assumes the following:

In its base case and bear case scenarios, Arbitrum accounts for 0.62% of ETH's projected market cap (as it is today).

In the bull case, Arbitrum accounts for 1.24% of ETH's market cap (2x its current value).

In the base and bear cases, Optimism accounts for 0.46% of ETH's projected market cap (as it is today).

In the bull case, Optimism accounts for 0.92% of ETH's projected market cap (2x its current value).

One year of token unlocks is factored into the price/token of each asset.

A bet on Ethereum L2 is essentially a bet on ETH. That's why our forecast starts with a high-level forecast for ETH.

Conclusion

If you are bullish on the Ethereum ecosystem, we think you need to emphasize having ETH as part of your portfolio. Why? Because investing too much in any one L2 can produce negative returns - even if the thesis on Ethereum is correct. For this reason, we believe that at least 50% of the Ethereum portion of a liquid portfolio should be in ETH. Note that this number depends on risk tolerance, goals, investment timeline, etc.

Token unlocks are real. They obviously affect forecast returns. For example, Arbitrum's base case predicts a 315% increase in market cap next year, but only a 210% increase in price per token. It is for this reasonthat ETH outperforms L2 in both the base and bear scenarios.

For the top L2s to outperform ETH, the market needs to re-rate based on their % of Ethereum market cap. We saw this happen with Solana last year - in December 2022, it had a market cap of just 2.5% of Ethereum's market cap. This was clearly wrong. The market has since re-rated Solana at 15% of Ethereum.

On a fully diluted basis, ARB and OP represent 3.7% of Ethereum's market cap. If you believe the top L2s should represent anything close to 10% of Ethereum's market cap, this may change your portfolio allocation. Given the lack of utility, unlocks, and competition for the token, we don't think this will happen next year, but a re-rating is likely in the long run.