What Can PoS And DeFi Learn From Mortgage-Backed Securities?
Editor's Note: This article comes fromEthereum enthusiasts (ID: ethfans)Editor's Note: This article comes from
Ethereum enthusiasts (ID: ethfans)
, Author: Tarun&Alex, translation & proofreading: Min Min, Zeng Mi & A Jian, reprinted with authorization by Odaily.
"I believe that the function of the interest rate is not to maintain the balance between the demand and supply of capital goods, but to maintain the balance between the demand for liquidity and the means of satisfying this demand." - Montesquieu
Since the stock market crash in 1929, the financial crisis in 2007 has made Americans fully understand the true meaning of "liquidity is king". Almost every market has been devastated, with the real estate market, which should be the safest under government protection, being the hardest hit. It was during this crisis that a cryptocurrency emerged, aiming to be a currency that was censorship-resistant and whose supply was not manipulated by governments.
Today, the cryptocurrency industry is saturated with a wealth of leverage provided by centralized and decentralized platforms. The problem with leverage is that once the market fluctuates, exchanges, custodians, and smart contracts will require people to call for margins in large numbers, resulting in a shortage of liquidity. Why do people need so much leverage? What caused the liquidity crunch?
In particular, one risk to be aware of is that certain market participants have grown to the point of being "too big to fail". But who is going to be too big to fail in the cryptocurrency industry? Imagine a world where Binance lent so much of its on-chain assets under management that it was overlending. Once these assets start to be liquidated on a large scale, just like Maker's Black Thursday, Binance may follow the example of "The DAO incident" and implement a hard fork to solve the liquidation crisis. This hard fork is similar to the emergency bailout measures taken by governments after the 2008 financial crisis.
We have attempted to generalize these and other risks in our recent article on staking derivatives. We discuss not only PoS networks, but also DeFi protocols (Aave, Compound, and Synthetix) and draw an analogy between their characteristics and those of traditional securitization structures.
text
What are pledged derivatives?
First of all, let's understand why pledged derivatives appear.
The PoS mechanism aims to provide network security by locking cryptographic assets instead of the PoW mechanism that consumes a lot of energy. PoW requires users to "lock" energy and hardware, while PoS requires users to lock funds (Staking). So what's the real difference between the two?
A key difference is liquidity constraints. In order to ensure network liveness (ability to continuously process transactions), PoS networks force users to lock collateral for a long time. With the collateral, the PoS network can take punitive measures (slashing) when validators behave badly. For example, if the network requires validators to be online 99.999% of the time to ensure liveness, a validator that goes offline during a block can be slashed.
In contrast, PoW miners “stake” by purchasing mining machines, and then receive block rewards and transaction fees in return. Due to the need to pay mining machine fees in advance, miners also face liquidity problems, because it takes a long time to return to profitability. If miners’ costs increase, for example, electricity bills go up, they may need to borrow money to pay for electricity. There are miners currently taking out loans of the tokens they earn. In theory, miners can use their own mining equipment as collateral for loans. However, Leo Zhang of Anicca Research pointed out in the article that the secondary market for mining machines is too small, and borrowers are not comfortable accepting mining machines as collateral. Miners turned to staking tokens for loans, which is the main reason for the surge in leveraged transactions on lending platforms such as BlockFi and Nexo.
Taking Bitcoin as an example, on average, miners can get an income every time they dig out 100 blocks (about 16.6 hours), so it is easy to obtain Bitcoin. This is not the case with the PoS mechanism. Many agreements specify that tokens are locked for weeks or even months. In order to allow PoS verifiers to obtain liquidity like PoW miners, some developers have proposed the concept of pledged derivatives. The method is to allow users to issue a virtual asset with their own locked pledge, and then take this virtual Assets to lend, so users can get some liquidity. The price of the synthetic asset will never exceed the value of the collateral, and once the borrower “defaults,” the value of the synthetic asset will drop. The simplest "default" scenario is when the validator is slashed, causing the value of the collateral to fall below the minimum collateral requirement for the lending transaction. When slashed, the network forfeits the validator's collateral and redistributes it appropriately.
So how does the network decide at what price validators can borrow? The key parameter is the derivative pricing function. Let's imagine how the borrower and the lender complete a loan transaction step by step. The borrower finds the lender to borrow money, and the lender checks the balance of the borrower on the PoS chain. If the loan amount and balance of the borrower are satisfactory to the lender, the lender can lend tokens to the borrower, record this loan in the loan book, and recalculate the price. Tracking prices is important as this allows lenders to accurately price subsequent loans. The lender then proceeds to monitor the PoS chain. If the borrower is forfeited, the lender updates the loan book and price. After the lock-up period expires, the lender can ask the borrower to repay the loan and/or calculate the price the borrower needs to pay. However, the above steps all assume that the lender can trust that the borrower will repay the loan after the lock-up period ends.
These transactions can all be executed through smart contracts. In fact, the smart contract can directly interact with the consensus algorithm of the PoS protocol to monitor the slashing situation and calculate the derivative pricing function. In addition, we can also use a constant function market maker (CFMM), such as Uniswap, to let the market determine the price of derivatives. From the interaction between the consensus algorithm and the DeFi protocol, we can see that when the consensus protocol can interact with the smart contract application layer, capital efficiency will increase.
Currently, many PoS networks do not natively support this functionality, so it would seem odd for users to lend directly to the network. However, the consensus algorithms of newer networks such as Celo, Terra, and Libra are all implemented through smart contracts executed on distributed virtual machines. In these systems, consensus is a first class citizen that can interact with other smart contracts. Additionally, DeFi contracts can interact with the consensus and read the current state of the system to determine if there is a breach (e.g., slashing). For example, the consensus protocol of the Celo blockchain adopts the CFMM algorithm to achieve the stability of its native algorithm stablecoin. These new network protocols all have built-in staking derivatives tools!
But are all these staking derivatives safe? After all, the security of a PoS network is directly proportional to the amount of pledge, because a double-spending attack requires at least 33% or even 50% of the pledge weight[0]. If the user takes the value of the pledged tokens as collateral for a loan (a form of re-hypothecation), it is equivalent to adding leverage, using X tokens to leverage X+Y tokens, where Y refers to the loan amount (determined by X is derived from the derivative pricing function). However, if the borrower defaults, the lender's collateral becomes XY tokens, making the network less secure. Therefore, if p is used to represent the probability that the borrower defaults on Y tokens, the amount of tokens pledged in the network is (1-p)X + pY, which may reduce the risk of security.
The value of p is extremely difficult to estimate because it varies with borrower credit quality, leading to serious incentive problems. For example, if the price of derivatives skyrockets, the borrower’s forfeiture cost becomes zero, and the borrower may forego the collateral. In addition, if p is very small, only one part per billion, then the security risk of the network is also very small, only one part per billion (XY). Then, the verifier can get the same income as the PoW miner.
A more interesting question is, what if collateralized derivatives loans are aggregated? In order to improve the liquidity and price discovery of pledged derivative assets, we hope that the PoS derivatives in the hands of verifiers can achieve homogenization. Just like the Dai generated by Maker is a homogeneous asset, but each CDP/Vault is non-homogeneous.
In this case, the price of the staking derivative is tied to the collective responsibility of the validator base. Issuing bonds backed by pools of secured loans is exactly what Mortgage-Backed Securities and related securitization structures do. But how do networks that offer collateralized loans achieve leverage the way mortgage-backed securities do? Before that, let's analyze why individuals and banks need leverage?
Let’s say you just bought a house for $640,000, with a down payment of $100,000, a monthly payment of $3,000, and 15 years to pay off. 10 years from now, let's say your home equity is $460,000[1]. Let's say you want to buy a Tesla at this point, and you don't have that much cash in your bank account -- but you have $460,000 in home equity! What if you could take out a $100,000 loan against that home equity and pay it off over 5 years? This is a home equity loan—a second loan that uses your home as collateral for an amount no greater than the equity in your home. Note that a home equity loan is a way to gain leverage, where you can get instant liquidity by pledging an illiquid asset (property) and only pay interest. If the borrower wrecks the Tesla, loses his job, and can't repay the $100,000 loan, the lender can take the property as collateral.
Let's assume again that you are a bank that provides home loans[2]. You take out a home loan worth $500,000, and you suddenly have $500,000 less cash available to lend. To make matters worse, the house used as collateral is illiquid—you can't directly mortgage the house to borrow more cash to lend to someone else. In this case, securitization can come into play. So-called securitization involves placing an asset in the custody of a company that does not hold other assets, and then selling shares in that company to investors. If the asset (eg, real estate, mortgage, mortgage package) is worth $100,000, and the issuer of the security issues 1,000 transferable shares, this is equivalent to dividing the claim on the asset into 1,000 shares.
Institutions offering securitization, such as banks, will charge a security issuance fee. Note that the lender here gains leverage by exchanging future cash flows (interest) from illiquid assets for liquidity (cash). If the owner of the property cannot pay the interest, the bank seizes the property and auctions it off, distributing the proceeds to the holders of 500 shares. In reality, a lender would bundle many properties together and sell their share of stock—this is a mortgage-backed security (MBS).
What do these two situations have in common? First, asset owners and debtors need to stick to long-term commitments—paying/collecting mortgage interest over a period of up to 15 to 30 years. However, asset owners will have many short-term liabilities and/or situations that require short-term liquidity. In order to obtain liquidity, they need collateral assets to lend against. This type of secured loan provides both liquidity for the borrower and downside protection for the lender - the lender can liquidate the collateral. This ability to convert long-term assets into short-term liabilities is the cornerstone of finance.
Translator's note: The author's explanation of secured loans and MBS (mortgage-backed securities) is a bit mixed here. In a secured loan (such as a mortgage to buy a house), the borrower (lender) obtains the house and bears the debt; while the lender (lender, bank) obtains the creditor's rights and mortgage property rights; the lender has the right to require the borrower to pay interest on time (creditor's rights), and if the borrower defaults, the property can be acquired and sold (mortgage title). MBS is not the liquidity obtained by the bank by mortgaging something, but to obtain cash by directly selling creditor's rights and mortgage property rights. If a company only holds these creditor's rights and collateral rights, then the company's only income is the interest on the mortgage, and the company's stock becomes MBS, because the value of these stocks is purely supported by the mortgage. The purpose of the bank's operation is not only to obtain cash, but to resell the risks related to the mortgage. After the sale, the income has nothing to do with the bank, but the risk has nothing to do with the bank.
Similarly, in a PoS network, verifiers can implement pledge operations with the help of pledge derivatives. If there are many validators simultaneously taking out loans backed by pledged derivatives, the network will hold each validator's corresponding debt according to its probability of default. The network must price the security loss, estimated by the following formula to aggregate the total security capital of the network.
in
Adjusted Security Budget for PoS Networks
in
S represents the amount of money spent on securing the network
p is a vector representing the probability of default (for example: pi represents the probability of default of the i-th validator)
X is a vector representing the pledge (for example: Xi represents the pledged lockup amount of the i-th validator)
Y is a vector representing borrowing (eg: Yi represents the amount of funds borrowed by the i-th validator via staking derivatives)
Just like MBS represents the total value of the many loans behind it, a pool of pledged derivatives can aggregate the value of pledged assets into a package of assets and tokenize them. And the price of that aggregated token is the most important metric — it represents the price of network security.
But wouldn't pledged derivatives be as bad as mortgage-backed securities? Will we trigger a new round of financial crises (just not in the world of state-backed capital)?
Maybe, and there is a key difference between "crypto securitization" and MBS. First, let's take a step back and look at the characteristics of securitization. The simplest function of securitization is to pool assets and issue securities backed by cash flows. In our (obviously ironic) cartoon example of MBS below, a mortgage originator makes home loans to consumers. The originator then sold the loans to a special-targeted entity in exchange for cash. The entity finances the purchase of the loans by issuing securities, known as MBS, to investors. These securities represent the cash flow of potentially thousands of individual loans behind them, yet are far more liquid than the loans that make up them. This allows larger buyers to purchase a single liquid asset that matches their risk appetite without having to wade through thousands of individual loans.
If we look at DeFi lending protocols like Aave and Compound, their structure is very similar. A series of different types of loans are pooled into one smart contract. The protocol funnels these loans back into the financial market by issuing bonds (such as cDai) backed by interest income from the loans. In this sense, Compound and Aave are not only lending agreements, but also securitization agreements.
The key difference is that lending protocols can both originate loans and perform securitization automatically. In conventional securitization, by contrast, a separate entity within the issuer usually handles the relationship with the borrower. This difference is critical for PoS, as PoS protocols require automatic forced liquidation of overleveraged borrowers. DeFi protocols like Liquity plan to do the same.
Securitization transactions are often thought of as tiered transactions: issuing different grades of securities backed by the same pool of loans to cater to different risk appetites. At first glance, it seems that there is no grading in the case of the above-mentioned Compound, which issues bonds in the unified form of cDAI. However, we are seeing lower risk (such as cDAI combined with insurance through Opyn or Nexus Mutual) and higher risk (such as PoolTogether) options starting to emerge to cater to lenders with different risk appetites.
One of the problems that led to the catastrophic collapse of mortgage-backed securities in 2007-2008 was that they were difficult to price transparently. Part of the reason comes from the principal-agent problem, but the more important reason is that the time delay between the occurrence of a default and the impact on the value of MBS is too long. For example, if a borrower defaults on a home loan, the unpaid payments will affect the lender first. The lender then reports the outstanding payment to the mortgage securitization issuer, who then tells all shareholders that their lien has been defaulted. Due to the unique nature of the US housing market, this process can take months to propagate defaults through the complex financial system [4].
In the world of cryptocurrencies, this is no longer an issue. With the help of the blockchain, all participants are aware of whether the loan is in default - whether it is cryptographic or pledged collateral. This means that financial products that rely on these loans, such as collateralized derivatives, can price defaults and execute liquidations immediately. Although the risk of liquidation still exists—for example, market and liquidity risks may inhibit liquidators in agreements such as Compound and Maker from performing liquidation[3]—it is much more efficient than traditional markets. Relatively convenient liquidation The efficiency of blockchain means we can price, package and reuse more complex securities in ways that are difficult to do in the traditional financial world.
However, realizing this advantage comes at a price. Unlike creditors who lend directly to debtors, cryptocurrency protocols do not know the credit quality of their debtors. Additionally, collateral received in the form of cryptocurrencies tends to be of lower quality. DeFi protocols attempt to address these issues by setting higher collateral rates and more aggressive liquidation policies.
Additionally, collateralized derivatives introduce additional risk as liquidations reduce the money supply. Because when a validator defaults on the loan, the PoS protocol will destroy or remove its pledged share in the system. A reduction in the money supply will increase the future expected rewards of other validators, but will reduce the security of the system. Even if the protocol compensates for this by issuing additional currency, the value of these assets in the market may be discounted. Therefore, the security of the network is greatly reduced at each liquidation.
Finally, traditional securitization structures rely heavily on diverse pools of loans representing many different borrowers. The use of pseudonyms makes it difficult in DeFi and PoS protocols to assess the protocol's exposure to any single entity. Therefore, it is particularly important to study the wealth distribution among debtors in PoS. An uneven distribution of wealth among PoS participants would concentrate risk on a small number of debtors, thereby compromising the security of the system.
This is the focus of this paper, where we examine how staking derivatives affect inequality and returns in PoS networks. We find that, under certain conditions, staking derivatives reduces inequality. There are two intuitive reasons for this effect. First of all, staking derivatives create a level playing field by allowing validators to recycle funds, so that no matter how large the validator is, they have the opportunity to be selected into the set of block producers. Second, by destroying the stake of over-leveraged defaulting validators, it increases the earnings of other validators. This effect can be quite dramatic when large validators default. But we also think that specific "too big to fail" validators will change this result. In order to maintain the good performance of staking derivatives, validators must bear the default risk proportionally and collectively punish defaults, especially those of large entities.
What does this mean for protocol developers? If they decide to add collateralized derivatives to their protocol, they need to consider that they are essentially adding trustless mortgage-backed securities to the system. On the bright side, this allows validators to optimize their capital efficiency for higher overall returns, while the liquidation mechanism ensures that these liens are always priced correctly. On the other hand, the well-known MBS risks still exist in PoS. If the standards for lending in the network are too loose, then these derivatives will significantly reduce the security of the network. Protocol designers need to carefully design the pricing function of derivatives to tell verifiers their borrowing amount. The function should properly combine paid fees (interest), credit limits, and liquidation policies to avoid adverse situations.
We evaluated their theoretical model using Gauntlet's agent-based simulation platform and found that there is indeed a "sweet spot" where these derivatives can be added without compromising safety. Although the advantages often outweigh the disadvantages in these systems, it is still difficult to effectively maintain a balance when adding staking derivatives to PoS protocols. Failure to price default risk correctly, or force liquidation if necessary, can result in a significant loss of safety.
[0] 33% if the network guarantees determinism and uses a Byzantine fault-tolerant system, and 50% if the network uses the "longest chain" Satoshi mechanism
[1] Home equity is the difference between your mortgage and the current value of your home. Your home equity increases as your mortgage payments are paid off. If you owe $180,000 on your mortgage and your home is worth $640,000, then your home equity is $460,000.
[2] Technical notes from Haseeb Qureshi: Technically speaking, banks do not need to maintain a 1:1 deposit ratio. But in an abstract sense, you do need to withdraw deposits from the bank, so the bank needs to maintain a deposit reserve ratio. Assuming banks' RRRs are already close to the required RRR (or their actual RRR is "satisfactory"), if they make small loans, they will need to take in more deposits to maintain RRR balance . Thus, in equilibrium (unless banks slowly lower their reserve requirement ratios, which happens from time to time in banking regulation), banks need to take in deposits to make new loans.
