On Digital Collateral
Edit: An update from a recent Bankless post.
The Bankless website reads, "Instead of central bank money, we have cryptocurrencies. (2) Instead of commercial banks, we have money protocols. In a world where private keys act as money and code makes that money programmable, the entire money system changes. So Bankless is about banking less. And less banking. And less dependence on banks."
Importantly, the hosts do not claim that smart contracts can replace banking in their entirety, but rather leave us less dependent on banks. We cannot yet free ourselves entirely from the banking system. But why? I can answer this in short: Decentralized lending protocols have no way of effectively collateralizing loans. If you want a more in-depth answer, I’ll attempt to provide that below.
Calamaris, Haber, and Bitcoin
The peer-to-peer exchange system in Bitcoin allows for maintaining accounts (a tally of who has what amount of each asset) in the absence of banks. This decentralized database is the major difference between electronic and non-electronic peer-to-peer exchanges. Meaning: exchange between two parties without a trusted third party has always been possible, even easy. However, the blockchain provides a universal accounting system in the absence of a third party.
Most importantly, this allows for trust minimized transactions. Trust acts as a barrier to what would otherwise be many mutually useful exchanges. Any technology that allows for exchange without trust increases access to useful exchanges and improves the lives of those involved. Trust minimized transactions, paired with the improvements in cryptography that make wallets secure by requiring more resources to hack than any theft could be worth, create a major social improvement for society.
There is a common objection regarding security, which is to point out that Bitcoin, like fiat money, has been stolen. But this is true only for bitcoin kept on centralized exchanges—crypto banks. No one has successfully hacked cryptographically sealed wallets, which is a big difference. Crypto banks mirror the structure of traditional banks, but store cryptocurrencies instead of fiat money. So, it makes sense that hackers can steal from them, as they could a traditional bank. Cryptographically sealed wallets, however, are the innovation that allows for greater security. The failure of crypto banks does not equal the failure of cryptography.
When I began writing this essay, I had hoped to connect this development to the Game of Bank Bargains described by Calamris and Haber in Fragile by Design. My hope was that, in the same way, that Bitcoin allows for exchange in the absence of the government (or any trusted third party), decentralized finance could do something similar in banking. This untethering of the banking system from governance would, like trustless transactions, have immense positive downstream effects.
Unfortunately, this does not seem to be possible with smart contracts alone. Decentralized finance can only provide overcollateralized loans, which allows for severely limited use cases. It seems that one can be bankless when storing and transacting, but not when borrowing and lending.
Banking and Government
In this section, I hope to answer the following question: Why do banks require property rights that can only be provided by the government? I hope to answer this by recapitulating the discussion in Chapter 2 of Fragile by Design, titled The Game of Bank Bargains.
The business model of banking is exceptionally simple. Both the major inputs and outputs of banks are loans. When you deposit your money to the bank, you are making a loan to them in return for interest. The bank then takes that same money and lends it out to others at a higher rate, and makes a profit on the difference.
There are three major risks inherent in banking: credit risk, liquidity risk, and bank runs.
Credit Risk — banks cannot predict with certainty whether the borrowers will repay their loans
Liquidity risk — the impossibility for bankers to perfectly match the duration of their contracts between depositors and debtors
Bank Run — If the bankers manage to handle both credit and liquidity risks, depositors may still show up in masses to demand repayment
These risks can be protected against cushions, which can be either cash assets (borrowing more than they lend out), or equity capital (a block of money invested in the bank by stockholders). If the banks' cushion is equity capital, the risks are simply transferred from depositors to investors. In the case of default, the depositors are repaid first. The investors are repaid second or, more likely, not at all. Without equity capital, the banks must loan at an extremely small rate, as their only protection against insolvency is the difference between their loans and deposits.
Only governments, through limited liability, can provide outside investors incentives to take on the risks inherent in banking. Thus, limited liability allows banks to expand significantly beyond the rate they would be limited to without equity capital. This allows for a greater level of lending and borrowing, and, in turn, a rapidly growing economy.
Governments offer limited liability through bank charters, but this solution exacerbates three fundamental property-rights problems: expropriation risk, tunneling, and mass default. An effective banking system requires (1) incentives against bank investors and depositors tunneling the banks' resources, effectively stealing from depositors and minority shareholders. It must also (2) compensate those for bearing the risk of borrowers reneging on their contract. At the same time, it must (3) have mechanisms to prevent, or compensate for the risk of, government expropriation. Unfortunately, the government is the only institution that can provide insurance against tunneling & default, but doing so means increases the risk of expropriation.
Decentralized Lending and Borrowing
My original motivation for exploring this topic was to determine if developments in cryptography have made it possible to solve the property rights problems necessary for banking without the need for violence. By violence, I am referring to the force of the legal system (the government’s monopoly on legitimate violence). If so, I thought, we may be able to untether banking from politics and transcend the suboptimal nash equilibrium described in Fragile by Design. In short, I wanted to ask: If we can now exchange directly, peer-to-peer, can we lend and borrow in the same way? And, if so, what are the consequences of that development?
Enter Decentralized Finance (DeFi, for short). In the same way that Bitcoin uses the blockchain as a decentralized database for monetary accounts, Ethereum uses the blockchain to decentralize computer code. Each computer in the Ethereum network hosts the code necessary to run applications in the same way that each computer in the bitcoin network maintains its monetary tally. If Bitcoin is a massive decentralized bank, then Ethereum is a massive decentralized computer.
On this decentralized computer—the Ethereum network—two successful apps have emerged to orchestrate borrowing and lending: Aave and Compound. Users lend by pooling their assets into a smart contract that pays interest to the lender. Borrowers can then take from this same contract, and must eventually repay their loans with interest. Quite unlike traditional borrowing and lending practices, all loans are permissionless. There is no third-party organization to review the borrower’s financial history. Instead, the smart contracts execute the exchange automatically when certain conditions are met.
However, for a few reasons, the protocols are still unable to take on the responsibilities of the traditional banking system. Most importantly, borrowers often have to overcollateralize their loans. Meaning, that the borrower must supply tokens worth more than the loan as collateral. This is mainly to protect against price fluctuations. The issue here is that, as far as I can tell, the protocol locks up the collateral so that it cannot be used in another way until the loan is repaid. In traditional systems of collateral, the borrower can still make use of the collateral (his home, for example) while using the loan to finance an investment. Locking up collateral so that it cannot be used effectively defeats the purpose of borrowing, sparing a few unique situations.
Again, the question is: Can property rights traditionally granted by governments instead be supplied by smart contracts in the DeFi money market?
The answer, as long as over-collateralization is required, is no. If, however, the only reason for over-collateralization is price fluctuation, then a future with a less volatile Ethereum price is all that is required. The rest of the pieces are in place. But I’m skeptical that price fluctuation is the only reason for over-collateralization.
The smart contract method requires any collateral to be digital. There are few ways around this, as the absence of a police force means the loaner has no way to collect non-digital collateral in the event of a default. The use of force, on the other hand, would bring us right back to the same problem. So, what could act as digital collateral while still being of use? NFTs, maybe. Depending on whether or not they prove to be of value in the long term.
Another objection may be: What separates lending protocols such as Aave and Compound from being anything more than what Calamaris and Haber describe as a “medieval moneylender—an agent who could make loans and charge interest but not mobilize savings en masse”? Well, the protocols quite literally “pool” savings. The issue lies on the other end. The protocols actually have difficulty lending to people in ways that can be productive, which I guess can be considered difficulty “mobilizing” them. But, the savings are there if the opportunity for mobilization arises.
The Rewards of Escape
What would it mean to escape the game of bank bargains?
Calamaris and Haber write, “The implication is inescapable: the property-rights system that structures banking has not evolved in response to some efficiency criterion in an anonymous “market” for institutions. Rather it is the product of deals arranged and enforced within an existing set of political institutions and hammered out by coalitions of market participants and the group in control of the government. These deals are intended to improve the welfare of the members of those coalitions and of the group in control of the government, not of society at large.”
Depending on the form of government, ranging from democracies to autocracies, the results of who benefits widely differ. Even so, some countries are able to hold together much less fragile banking systems than others (Canada compared to the United States, for example). But for those with fragile banking systems, the common denominator is the same: the part benefits more than the whole. An escape from the game of bank bargains would mean we could reject the part for the whole, and benefit all instead of some.
Conclusion
In conclusion, the inability to provide working collateral to decentralized finance protocols has kept us trapped in the Game of Bank Bargains. This incapacity means that we can, at most, be bankless as “banking less” rather than without banks, and countries with fragile banking systems are still stuck with fragile banking systems. For this to change, there needs to be a way to effectively collateralize loans on DeFi.
