Why can DeFi master brokers allow DeFi to erode the centralized exchange market? What are the risks themselves?

In traditional finance, when a user clicks “buy” on a screen, they get a seamless experience and access to a financial product seconds later. Behind the scenes, a lot of financial engineering and infrastructure provide this seamless experience. A distributed system problem then arises between algorithmic traders, brokers, investors, rich regulations, and dozens of other intermediaries.

Decentralized Finance (DeFi) reverses this by allowing near-instant settlement and direct custody. That being said, as open financial primitives continue to develop, aggregated prime brokers such as protocols will emerge, squeezing inefficiencies out of the market and greatly increasing liquidity, thereby improving the utility of the entire market. The core question is what exactly these protocols look like?

The community has begun to discuss this, and the value narrative has tended to expand existing DeFi protocols by providing more and more financial products and protocol levels, leveraging rehypothication and margining. Dan Elitzer, one of the founding team members of Yam Finance , on aquaponic yield farming research and Trent Elmore, one of the founders of YAM, this wonderful thread well reflects this vision, and Synthetix is ​​the most obvious example of action in this direction [1]. However, my goal in this article is to design an alternative approach that seeks to separate the issuance of assets and derivatives from services such as prime brokers.

Financial markets become highly liquid under certain conditions. The first is real demand and trade flow, a trend that is increasingly being seen in the DeFi space as record DEX volumes further highlight[2]. But the second condition is derivatives markets and the associated risk infrastructure in the form of clearing and settlement. The simplest example of liquidation is an arbitrage trader who is long BTC with 10x leverage on exchange A and short BTC with 10x leverage on exchange B, and who does not need to provide full collateral (or even any collateral at all) due to the offsetting nature of the positions. For traders trading on centralized crypto derivatives exchanges like BitMEX , the cross margin provided by these exchanges is a simple example of this type of risk infrastructure. The limiting factor of this mechanism is that it only applies to that single exchange. Once clearing is introduced, hedge funds, quants, and traders are able to hedge and arbitrage with incredible capital efficiency.

Example of efficiency gains from liquidation

Centralized crypto markets have quickly stacked derivatives markets on top of this emerging industry, most notably futures and perpetual futures trading. However, clearing still lags far behind the standards of traditional finance. There are a few reasons why this infrastructure is not present in centralized crypto exchanges, all of which require a long explanation, but the long story short is that it does not exist yet, which presents an opportunity for DeFi to overtake the dominant centralized exchange CEX market share.

First consider the vertically integrated DeFi prime broker or “ liquidity black hole ”. Assuming that the protocol issues derivatives and synthetic assets, they can be traded efficiently using liquidation technology, which can calculate the required net value of positions through offsets, which is very important from the user's perspective because they can keep all their funds in one system. However, by combining liquidation and issuance, the protocol layered on two separate risks.

The first is protocol risk , including maintaining the core competitiveness of financial primitives. How closely are synthetic assets such as DAI or sBTC anchored to the collateral assets, and what mechanisms are needed to realize their value? For option contracts, what is the guarantee of exercise ability? Cash settlement or physical settlement? For prediction markets, who determines the outcome? Does the oracle require streaming data or one-time settlement? Is the contract fully collateralized? All of this is built on actual smart contracts and base layer security.

Liquidation risk, on the other hand, is an abstraction of protocol risk , an engine that generates the assumption that assets are similar or offsetting and willing to honor user positions. Each protocol represents a unique set of tail risks for the clearing agent. Any client of a brokerage firm wants ruthless and fair pricing of protocol risk to avoid bankruptcy when the protocol deviates. Equally important to risk elimination is the actual modeling of related but distinct products. Consider the liquidity provider LP share in the ETH/DAI liquidity pool and the 10x long BTC/USDC dydx perpetual contract. Calculating low risk and effective compensation here is non-trivial and requires skilled modeling of the derivatives and their underlying.

The point here is that risk comes in different forms. The entities trying to mitigate these two types of risk have different goals. The goal of protocols is to provide economic security, a strong peg, settlement guarantees, etc. Clearing brokers, on the other hand, offer a probabilistic approach to solvency, e.g., as long as the fees they collect are greater than the risk they take on in liquidating positions, they are solvent and profitable and can continue to exist. Many protocols do take on probabilistic risk in processes such as default actions, but these risks are concentrated risks that must be addressed by the architecture of the protocol! The vertically integrated DeFi protocols we are discussing layer on these unique risks, exacerbating their insolvency risk. This is not to say that in-protocol clearing shouldn’t happen: for example, Opyn is working to reduce margin on options packages like spreads to improve capital efficiency[3], and they should be doing this! Any options issuer needs to do in-protocol clearing to maximize efficiency. The point is that the Opyn system should not continue to offset positions in sETH or ETH/USDC LP shares, as this introduces a whole set of risks that are not relevant to the core business of the protocol.

Back to the DeFi prime broker issue. Brokers have no allegiance to any protocol, only to risk models that assess the interchangeability of various assets and the return profile of the portfolio. Brokers will allow margin reductions on long sBTC positions and short BTC-PERP positions. The result is tighter spreads and increased liquidity on all assets/markets that the broker clears. What will this look like in practice? At a high level, brokers accept any collateral they understand, and as a trader, you might just deposit a cash asset, whether it's USDC, ETH, or synthetic BTC. Once funds are deposited, the broker will match trades and credit/debit your account when you trade. In order to actually provide settlement on exit, brokers will need to hedge positions through interactions with the corresponding protocols through the main chain. Those who wish to continue using the core protocol now find that the market is more liquid because brokers feed funds back into the base protocol. Ideally there would be multiple DeFi clearing brokers, each with different products and risk settings, providing services for every level of the user's risk spectrum.

DeFi Prime Broker Process

The obvious problem is privacy. The broker’s hedging process can be preempted by sophisticated players who see the big trades inside the broker and preempt the broker’s input hedges. It’s not clear what real mitigation there is, whether it’s dark pools, private broker positions, OTC trading via aztec assets, etc… in fact, it may be a combination of all of the above. In order for the core merchants of DeFi to remain unscathed, the new financial system must simultaneously remain open and inspectable while also protecting the ability for traders to enter and exit positions privately, which is a unique and high bar.

Architecturally, the system will need to sacrifice some degree of decentralization to achieve this capital efficiency, and the ledger maintained by the broker will need to have some custom rules around things like position sizing and withdrawal wait times. Ideally, the actual trading would be done by a high-speed layer 2 trading system like ZK Rollup , or whatever the latest and greatest L2 design is. As Ethereum transitions to ETH 2.0 and sharding, L2 structures like this become more attractive. Aave could be deployed on one shard, Synthetix on another, and Augur on another. Since brokers already bear the latency risk, they are better able to handle the complexity and asynchronicity of cross-shard contract calls and communications.

If the open financial system wants to thrive, it needs to avoid the mistakes of its predecessors. Protocols can fail, aggressive brokers can fail, and unscrupulous profiteering funds can rush in and drive your favorite tokens down to zero value. Most importantly, make sure there are no sleeping giants that will collapse the entire system when such risks are released. How to avoid a global financial crisis in DeFi? There must not be a Lehman Brothers or Bear Stearns that can bring the entire system to the brink of collapse.