The Bitcoin blockchain was launched in 2009, providing for the first time a secure, decentralized, and electronic platform for monetary exchange. It was not purely anonymous – public account information was available – but users did not have to know or trust each other in order to participate in “money” transfers. Over the next decade, a flurry of applications and advances were built upon this foundation, including: the tracking of goods along supply chains, registration of ownership, and even certification of marriage. But despite these advances, blockchain intermediated finance has yet to develop a method of decentralized, pseudonymous, and electronic lending.
In this post, we find that (1) pseudonymous borrowing and lending on the blockchain pose additional incentive problems beyond those in money exchange; (2) in order to make progress on financial applications, blockchain will have to interface with traditional, trusted, and identity-laden institutions; and (3) new technology like Oracle has the potential to make progress on this front. Far from vitiating the distinctiveness of trustless electronic exchange, we believe that interfacing with traditional institutions has the potential to bring together the advantages of both trustless exchange and off-chain finance.
Why Finance is Harder than Exchange
Successful financial intermediation requires stricter conditions for incentive compatibility than for monetary exchange. An institution is incentive-compatible when an opportunity exists to make everyone better off, and given everyone else’s strategies, no one can do better for himself by making other participants worse off.
For example, imagine an anonymous and uncollateralized loan. The lender sends you a sum of money, expecting you to pay it back in some amount of time. But if you are anonymous, the lender has no way to make sure you pay it back. Your best option is to abscond with the money. The lender, knowing this possibility, refuses to make the loan in the first place. Anonymous and uncollateralized lending, therefore, is notincentive-compatible.
Monetary exchange, on the other hand, is incentive-compatible. As Satoshi Nakamoto’s white paper lays out, a decentralized network for transferring money makes it too costly for users to manipulate the ledger for their own advantage. In other words, no one can do better by cheating. The core of our argument is that the use of blockchain protocol ensures incentive-compatibility in organizing a transaction. However, if we think about blockchain-based anonymous uncollateralized lending, it is not incentive-compatibility. Thus, what adjustments can be made to ensure incentive compatibility in blockchain-based lending?
Collateral is one way of ensuring incentive compatibility. By requiring the borrower to provide collateral that reverts to the lender if the borrower defaults, the lender can ensure that the borrower does not have an incentive to defect. In our paper, we analyze that collateral worth more than 100% of the value of the principal is necessary for repayment to be incentive-compatible. Consistent with this prediction, the existing avenues for direct borrowing and lending on the blockchain, like ETHLend and Ripio require collateral in the range of 125%. While better than nothing, a requirement that loans be more-than-fully-collateralized severely limits the usefulness and applications of blockchain finance.
Reputation is another way of ensuring incentive compatibility. Most cryptocurrencies are pseudonymous rather than anonymous, and lenders – even if they have no identifying information about borrowers – could in principle see borrowers’ loan history. New prospects get small loans, borrowers with a strong track record of repayment get larger loans, and borrowers who default get blacklisted. Therefore, the benefits of defaulting have to be weighed against the drawbacks of not getting loans in the future.
Although reputation helps compared to anonymous lending, we argue that the value of having a good reputation cannot be counted on to exceed the value of defaulting if the borrower can exit the market at will. In the long-run, the value of a reputation is proportional to the value of the largest loan one can acquire. Provision of a larger loan by a lender raises the value of defaulting for the borrower. Something other than reputation will be necessary in order to ensure the incentive to repay, and therefore to make lending possible.
The difficulties compound when we move from simple direct lending to more advanced forms. For instance, in the case of tradeable securities, buyers face the same dilemma with regard to the originating pseudonymous lender. If originators know that they can exit the market after originating a security, they will have no incentive to invest in gathering enough information to ensure quality loans. Buyers, therefore, will not be willing to buy without some assurance of due diligence, which is costly to verify in a pseudonymous environment. Here, as before, reputation somewhat diminishes the incentive to defect, but not enough to ensure incentive compatibility ex ante.
Conditions for Maintaining Pseudonymity: The Promise of Oracles
We consider another question–if reputation by itself cannot support blockchain finance, and quality financial instruments cannot be originated under conditions of pure anonymity, is uncollateralized lending on the blockchain possible at all? Traditionally, these problems have been solved by tracking the history of a non-alienable identity across different domains; for example, a credit score linked to a social security number. It follows a user and summarizes past credit behavior in a way that makes it possible for a lender to assess risk.
It appears, however, that maintenance of anonymity is incompatible with a credit score linked to a persistent identity. Oracle, however, is a technology with the potential to fill the gap.
An oracle provides a means of securely importing information external to a blockchain. Oracles may import data from weather sensors or radio-frequency identification (RFID). They may even access information from a private account using account ID and password without revealing this information to any users.
Regarding the maintenance of pseudonymity, it is possible that users maintain pseudonymity as long as they abide by the terms of the smart contract that defines the lending agreement. This makes pseudonymity contingent, treating revelation of identity as collateral. A smart contract may hold the identity of the borrower and originator as a private attribute, only to be made public in the case of a default. Borrowers and originators can also provide identifying information to an oracle and evidence their ability to repay or repurchase a loan. The oracle can also confirm that a trusted third party,such as a bank, vouches for the ownership of a minimum value of a borrower’s assets, credit score, or similar criteria while maintaining pseudonymity, thereby increasing theconfidence of lenders in the ability of, and incentive for, the borrower to repay. In the case of default, identities of the borrower or originator are revealed, and the affected persons can pursue legal recourse.With an automated and transparent smart contract, each party can ensure that pseudonymity is compromised only to the minimum extent to support financial applications.
Though entrepreneurs have made impressive progress in the blockchain space, blockchain lending is in a nascent stage. Both direct lending and the creation and transfer of saleable financial instruments are already supported by dapps on the Ethereum blockchain, though in highly limited forms. Because finance poses different kinds of problems than monetary exchange, further development will require more than the purely technical innovations that have been the mainstay of blockchain tech so far. In particular, blockchains will need to be able to interface with the institutions that can provide sufficient identifying information to ensure loan quality and legal redress against a loan default.
Oracles are a way forward on exactly this, allowing users to confirm personally identifying information without revealing or otherwise compromising it except when necessary. Complementary technologies are also being developed and proliferating, though “killer” applications remain elusive. Nevertheless, the shape and pace of developments in blockchain applications, and especially in their interface with off-chain identity and data, bode well for the future of a decentralized finance.