Interoperability and Market Segmentation

The starting point for my class on the cryptoeconomy is the concept that ownership data is currently housed in silos. Ledgers of currency ownership are siloed within banks, ownership records of stocks and bonds are siloed within brokerages, and land ownership records live within thousands of independent governmental recorders offices.

In order to reconcile and send information between the plethora of siloed ledgers, we’ve developed labyrinthine systems of intermediation like the SWIFT network and the Depository Trust & Clearing Corporation (DTCC).

I ask the students: wouldn’t it be easier to have a single system of record?

Years ago, I wrote: “The thesis underpinning the idea that everything will be tokenized is grounded in the aspiration that everything will be interoperable.” Blockchains promise global, market-based standards to create a system of record upon which everyone can build, but web3 is currently witnessing a new form of siloing.

Data is no longer siloed within corporations, but is increasingly siloed within protocol ecosystems. Ethereum, Solana, Terra, Avalanche, Polkadot, Tezos, and many others. The list of ecosystems with meaningful activity continues to grow longer, and there are numerous reasons why a project may choose to build on a particular chain versus another. We believe in a multi-chain world, where different ecosystems optimize for different features, but we simultaneously believe that the connection points between them need to be further developed. Collectively, we refer to these cross-chain connections as enabling interoperability.

Vitalik recently commented on the challenges inherent in cross-chain applications, specifically related to 51% attacks. However, we need not look any further than the circulating supply of WBTC, which is just a single cross-chain asset (and a centralized one at that), to see that the demand for cross-chain activity shows no signs of slowing down. The risks are real, but our focus should be on minimizing the number of attack surfaces because cross-chain activity is an inevitability.

Oxford defines interoperability as “the ability of computer systems or software to exchange and make use of information.” Lack of interoperability is widely recognized as a current limitation within web3 and there are numerous efforts underway to solve this issue. In this article I’ll focus on one approach, Axelar, which is part of the Collab+Currency portfolio.

The overarching theme is that we’ll all be better off once universal interoperability is solved and the data silos fade into the background from the user perspective.

There are a variety of reasons that data silos make us worse off, not the least of which is that they make the user experience more complicated and less approachable for new entrants. To onboard the masses, we need to address this friction. There is also a financial burden. Lack of interoperability increases costs, and one area where we can illustrate these costs is through the extant financial research on market segmentation.

Market segmentation is a term used in several different contexts, but within finance it refers to frictions that limit the ability (or desire) of buyers and sellers in one market from participating in another similar market. Market segmentation inhibits the free flow of capital. In crypto, one example might be the inability to use ETH as collateral within a lending platform on Terra. Or, the same trading pair appearing on Ethereum, Polygon, and Avalanche, each with a separate liquidity pool.

In traditional finance, a common setting for studying market segmentation is international equity markets, where investors in one jurisdiction have limited access to equity investments in a different jurisdiction. In many cases, the friction takes the form of restrictions on foreign equity ownership, meaning the local government imposes limits on the percentage of a firm’s equity that foreign investors can hold. The rationale behind these limits is grounded in national security concerns: governments may prefer that domestic corporations are owned (and controlled) by investors inside their borders.

There were numerous academic studies around international market segmentation in the mid-1990’s, such as: Bailey and Jagtiani (1994), Bekaert and Harvey (1995), and Domowitz et al. (1997). In a sentence, the main findings are that market segmentation induced by investment restrictions can have a meaningful impact on prices, and hence increases the cost of capital. This is a strictly suboptimal outcome for the issuers. One of the causes is the difference in liquidity.

Subsequently, several studies focused on the effects of liberalization, or in other words, efforts to integrate markets that were formerly segmented. One such liberalization was the introduction of American Depositary Receipts (ADRs), which allows foreign equities to be traded on US equity markets. Errunza and Miller (2000) study ADRs and report a reduction of over 42% in the cost of capital following the introduction of an ADR. Further, other studies of financial market integration such as Henry (2000), and Bekaert et al. (2005), have documented that they spur investment booms and real economic growth.

Taken together, the studies listed above suggest that market integration is something we should strive towards. That said, protocols aren’t companies and one might argue that these results don’t map cleanly to crypto. To address this question let’s consider a simple model based on trading platforms in segmented markets.

Various DeFi protocols are now operational on multiple chains, but the underlying liquidity pools remain segmented. For example, consider SushiSwap running the same trading pair, say WETH/USDC, on both Ethereum and Avalanche.

Here’s a screenshot of the quotes for each market, taken simultaneously (Avalanche on the left, Ethereum on the right):

WETH<>USDC pools

The immediate observation is that traders on the Avalanche market are worse off, since the price is better on the Ethereum market. But the more important insight is that market participants in both markets are worse off relative to markets with pooled liquidity.

Realized prices on decentralized exchanges (DEXs) using constant product Automated Market Makers (AMMs), like SushiSwap, are directly proportional to the size of the liquidity pool. Deeper pools = less price impact.

Here is a simple numerical example (h/t Daily DeFi). Say you have two markets for the same pair, but one market has a liquidity pool that is twice the size of the other.

Pool 1:

• USDC = 200,000
• ETH = 100
• Constant Product = 20,000,000
• Market Price of ETH in terms of USDC= 2,000

Pool 2:

• USDC = 100,000
• ETH = 50
• Constant Product = 5,000,000
• Market Price of ETH in terms of USDC= 2,000

Trade 10,000 USDC for ETH on Pool 1:

After swap

• USDC = 210,000 (because we added 10,000 to the pool)
• Constant Product = 20,000,000 (stays the same)
• ETH = 95.24 (constant product / new usdc amount)

ETH received = 4.762 (old eth amount — new eth amount)

Price paid per ETH = 2100 USDC

Price impact = 5%

Trade 10,000 USDC for ETH on Pool 2:

After swap

• USDC = 110,000 (because we added 10,000 to the pool)
• Constant Product = 5,000,000 (stays the same)
• ETH = 45.45 (constant product / new usdc amount)

ETH received = 4.545 (old eth amount — new eth amount)

Price paid per ETH = 2200 USDC

Price impact = 10%

Now, if these pools are combined we have:

Pool 1 + Pool 2:

• USDC = 300,000
• ETH = 150
• Constant Product = 45,000,000
• Market Price of ETH in terms of USDC= 2,000

Trade 10,000 USDC for ETH on combined pool:

After swap

• USDC = 310,000 (because we added 10,000 to the pool)
• Constant Product = 45,000,000 (stays the same)
• ETH = 145.16 (constant product / new usdc amount)

ETH received = 4.839 (old eth amount — new eth amount)

Price paid per ETH = 2066.67 USDC

Price impact = 3.33%

In summary, traders on the combined pool are far better off than traders on the small pool, but even a bit better off than traders on the bigger pool. As pools get closer in size the improvement from combining them increases. The takeaway is that trading platforms on chains with smaller markets stand to gain a lot by tapping into pools on chains with larger markets.

It’s worth reviewing the current state of interoperability for a moment. Virtually all models lock assets on one database and mirror the state on another ledger.

One approach is to establish bridges between each ecosystem. The idea is to deposit the asset into the bridge, which locks up the asset on the entry chain while simultaneously issuing a replica of the asset on the exit chain. These bridges can be administered by a centralized third party, like WBTC, federated, like Wormhole, or decentralized, like tBTC. The challenge is that the bridges are pairwise, not any-to-any. As the number of ecosystems grows, the number of pairwise connections required grows exponentially.

Another approach is the interoperable ecosystem approach, like Cosmos or Polkadot, whereby the ecosystem itself is multi-chain, all of which work together. The challenge with this approach is that while the chains inside the ecosystem all work together, there are always going to be additional chains outside the ecosystem that are incompatible.

By connecting different ecosystems in a universally consistent way, Axelar is solving fragmentation across apps. Axelar meets growing demand for cross-chain services with minimum risk and maximum composability, replacing pairwise cross-chain bridges with a network architecture that provides a uniform code base and governance structure.

Today, you have to deploy a DEX on every chain, and fragment pools of liquidity. With Axelar, users will be able to interact with any asset and any application on any chain. It will be sufficient to have one copy of the DEX but make it talk with the axelar-API and make it cross-chain enabled (i.e. allow users to interact with it). Liquidity will be shared, traders will rejoice.

How might this work in practice? There are a couple of approaches…

First, consider the scenario above where there are two pools, say, one on Avalanche and one on Ethereum. Axelar will allow the assets to move to the deeper pool, in effect, allowing liquidity pools to be composable. Behind the scenes, this might include three separate trades:

1. Trade AVAX to UST on Avalanche on source chain
2. UST routed to Ethereum chain
3. Another contract call to swap UST to ETH, then return to the user’s key.

This approach solves for fragmented liquidity, but does so at the expense of multiple trades, each with slippage.

Alternatively, and perhaps optimally, we may not need pools on separate chains because once the markets are integrated, participants can trade natively against a single pool on any chain from wherever the assets are located. Say you want to swap AVAX for ETH and you hold the AVAX on the Avalanche chain. If there’s already a pool on Ethereum for ETH/WAVAX, then Axelar moves the AVAX as a wrapped asset, swaps, and then passes the ETH back to the user wrapped on Avalanche.

Since this approach involves only a single trade, slippage is minimized.

Ultimately, the question is: how do we abstract away the technical details to make it easy for users to transact across chains? The answer is to push the cross-chain operations to the background, so that the user doesn’t even have to think about it. Axelar has begun to talk about this as providing an overlay network for web3.

This model isn’t new, overlay networks provide much of the functionality on the internet that users expect today. Akamai is a good example. To understand Axelar as a universal overlay network for web3, let’s rewind for a moment.

The internet is effectively a network of networks. Historically it was fragmented, like web3 is today. It was not designed for the demands of the apps that would be developed much later in its evolution.

Enter Akamai, a content delivery network that sits on top of a multitude of underlay networks. In this brilliant paper, they outline the history of overlay networks and the deficiencies of the internet that overlays were designed to solve: (i) Outages, (ii) Congestion, (iii) Lack of scalability, (iv) slow adaptability, and (v) lack of security. Those deficiencies sound shockingly familiar.

Akamai realized that we need a more efficient network to deliver content from website to user. They built gateway servers in every underlay network across the globe, and developed standards to route and deliver content based on request of an application. Today 85% of the world’s internet users are within a single network hop of an Akamai server. You are likely using the Akamai infrastructure right now to read this article.

In Axelar’s version of an overlay network for web3, instead of gateway servers we have gateway smart contracts. There will also need to be a service layer on top of the network. It is anticipated that this service would comprise SDKs that will make it easier for developers to interact with the network. The key features of Axelar services will be (i) finding paths (routing), (ii) translating messages, and (iii) providing security.

The combination of the Axelar Network plus the services layer presents a viable solution to market segmentation in crypto.

It’s a big step towards the grand vision: stitching together the fabric of web3 into a unified, interoperable set of ledgers.

A special thanks to the Axelar team and Derek Edws for their feedback and insights during the construction of this piece.

Disclosure: Collab+Currency is an investor in Axelar.

You can find more thoughts from our team on Twitter: @Collab_Currency