DeFi Promise and Fabian Schar Pitfalls
Decentralized finance could support new financial infrastructure if challenges are overcome
Digital innovation has brought major improvements to the financial system. But the architecture of the system remains essentially the same. It is always centralized.
Decentralized finance (DeFi) offers an alternative. It uses public blockchain networks to conduct transactions without having to rely on centralized service providers such as custodians, central clearinghouses or escrow agents. Instead, these roles are taken over by so-called smart contracts.
Smart contracts are instructions in the form of computer code. The code is stored on public blockchains and executed under the system’s consensus rules. DeFi protocols can be designed in a way that prohibits intervention and manipulation. All participants can observe the rules before committing and check that everything is executed accordingly. Status changes (e.g. updates to account balances) are reflected on the blockchain and can be checked by anyone.
In the context of DeFi, smart contracts are mainly used to ensure the atomic transfer (simultaneous and inseparable) of two assets or to hold collateral in an escrow account. In both cases, the assets are subject to the rules of the smart contract and can only be released if the predefined conditions are met.
By using these properties, DeFi can mitigate counterparty risk and replicate many financial services without the need for intermediaries and centralized platform operators. This can reduce costs and the risk of errors. Lending markets, exchange protocols, financial derivatives and asset management protocols are just a few examples.
Smart contracts can refer to other smart contracts and use the services they provide. If, for example, an asset management protocol uses a decentralized exchange, incoming assets can be exchanged as part of the same transaction. This concept, of actions on multiple smart contracts that can take place within a single transaction, is called “intra-transaction composability” and can effectively mitigate counterparty risk (the likelihood that other parties will not fulfill their share of agreement).
Benefits of decentralization
Many of the benefits typically attributed to DeFi – or blockchains in general – can also be achieved through centralized infrastructure. Smart contracts are not limited to decentralized systems. In fact, the same standards and runtimes can be used on centralized registries. There are countless examples of the Ethereum virtual machine (a virtual machine that runs on all computers in the blockchain network and executes smart contracts) being used alongside highly centralized consensus protocols. Similarly, the same token standards and financial protocols can be used on centralized platforms. Even composability can work on such systems.
Moreover, well-managed centralized systems are much more efficient than public blockchains. This could lead to the conclusion that public blockchains and DeFi are inferior to centralized systems.
However, centralized systems are based on a very strong postulate: trust in largely opaque intermediaries and institutions. But such trust should not be taken for granted. History offers countless examples of corruption and errors within institutions. Yet when economists discuss financial infrastructure and compare the properties of public blockchains with those of centralized ledgers, they generally assume that centralized entities are benevolent, making it difficult to see the benefits of decentralization.
Public blockchains are transparent. Because they are not controlled by a single entity, they can provide a neutral, independent, and immutable infrastructure for financial transactions. The code is stored and executed on an open system. All data is available and verifiable. This allows researchers and policymakers to analyze transactions, conduct empirical studies, and calculate risk measures in real time.
More importantly, access is not restricted. This has two implications.
First, the absence of access restrictions provides a neutral foundation that cannot discriminate between use cases or stakeholders. This is in stark contrast to authorized registries, whose rules are defined by a centralized entity. Because it is so centralized, universally accepted standards can be difficult to achieve, and rights to access and use the infrastructure could easily be politicized. In anticipation of such problems, participants who feel it may be to their disadvantage will not use the centralized infrastructure in the first place. Decentralized systems can mitigate these delays, potentially avoiding the problem of no or minimal cooperation.
Second, DeFi is built on a layered infrastructure (see Schär 2021). A decentralized ledger does not mean that everything deployed on it has to be decentralized as well. There may be good reasons why access to certain tokens or financial protocols is restricted or subject to intervention. These restrictions can be implemented at the smart contract level without compromising the general neutrality of the core infrastructure. However, if the ledger itself (settlement layer) was already centralized, it would be impossible to credibly decentralize everything built on it.
It is very likely that we will see a move towards ledgers that combine payments, tokenized assets and financial protocols, such as exchanges and lending markets. DeFi is the first example of this development, but there will be similar developments in centralized infrastructure. The logic is that intra-transaction composability only works if assets and financial protocols are in the same ledger. There are strong network effects, and neither crypto assets nor central bank digital currencies would be particularly compelling if deployed on a ledger without other financial assets or protocols. It is possible to create a centralized composable infrastructure with additional financial assets and protocols, but this would be risky and difficult to manage given the challenges associated with authorized ledgers. This makes a strong case for decentralization.
Centralized systems rely on a very strong assumption: trust in intermediaries and institutions
Challenges and risks
DeFi has many advantages, but there are challenges and trade-offs to consider.
First, there is the risk of deception, or “theatre of decentralization”. What is generally referred to as DeFi is, in fact, often highly centralized. In many cases, DeFi protocols are subject to centralized data feeds and can be shaped or influenced by people with “admin keys” or a highly concentrated allocation of governance tokens (voting rights). While partial centralization isn’t necessarily a bad thing, it’s important to make a strict distinction between true decentralization and companies claiming to be DeFi when they actually provide centralized infrastructure.
Second, immutability can introduce new risks. Enforcing investor protection can be more difficult, and smart contract programming errors can have devastating consequences. Composability and complex token-wrapping schemes (Nadler and Schär, forthcoming) that resemble collateral rehypothecation contribute to the propagation of shocks through the system and can affect the real economy.
Third, the transparent nature of the blockchain and the creation of decentralized blocks can be problematic from a privacy perspective. Additionally, it enables the extraction of rents through widespread front-running – a phenomenon known as mining/maximum extractable value (MEV). Those who observe a transaction containing an order to exchange assets on a decentralized exchange may try to anticipate (or sandwich) this action by issuing their own transaction. The precursor thus benefits at the expense of the issuer of the initial transaction. There are potential solutions that can at least partially alleviate this problem, but they involve tradeoffs.
Finally, scaling public blockchains cannot be done easily without compromising some of their unique properties. Decentralized block creation inflicts significant costs. The hardware requirements to run a node cannot be arbitrarily high, as this would cost many stakeholders and undermine decentralization. This limits on-chain scalability, increasing transaction fees. This trade-off between security, decentralization, and scalability is commonly described as a trilemma. One potential solution is something called Layer 2s. These are designed to take some of the burden off the blockchain while allowing participants to assert their rights on the blockchain if something goes wrong. This is a promising approach, but in many cases it still requires trust and various forms of centralized infrastructure.
DeFi still faces many challenges. However, it can also create an independent infrastructure, mitigate some of the risks of traditional finance, and provide an alternative to over-centralization. The open source nature of DeFi encourages innovation, and many talented people, academics and practitioners, are working on these challenges. If they can find solutions without compromising the unique properties at the heart of DeFi, it could become an important part of the future of finance.
Opinions expressed in articles and other materials are those of the authors; they do not necessarily reflect IMF policy.