Compilation of the original text: Nanfeng
Compilation of the original text: Nanfeng
June 20, 2007.
That day, Bear Stearns bailed out two of its hedge funds. It was the first sign that the subprime mortgage crisis had spread beyond the housing market, past lenders and securitizers, all the way to one of Wall Street's big investment banks.
Nine months later, Bear Stearns faced bankruptcy, and a year later, on March 6, 2009, the Dow hit its lowest point. No matter how depressed financial markets looked in mid-2007, things continued to get worse before they got better.
first level title
safety"safety"
The global financial crisis marked the end of the Great Moderation period. In the 30 years preceding that, there was an unrivaled bull market in which house prices increased fourfold. In general, people are irrationally overzealous and believe that housing will only go up in value. So banks started offering subprime mortgages to people with bad credit, assuming that in the worst case scenario, you could take your house as collateral. Since the house will be worth more, the risk seems minimal.
At the same time, there is a savings glut around the world as investors look for ways to invest. To meet this demand, banks decided to create more safe assets for investors to buy, packaging mortgages into mortgage-backed securities (MBS). These financial instruments provide income based on the underlying mortgage. Investors may be reluctant to take out a personal mortgage, but an MBS seems like a more attractive thing to do. They can buy many different mortgages without worrying about the risk of any one type of mortgage. One homeowner might default, but when in the world will everyone default at the same time? It's "diversified" -- in other words, it's safe.
What about the rare occasion when the quality of the mortgage is so poor that even the MBS looks unsafe? Banks simply repackage them into a type of collateralized debt obligation (CDO), bundling enough of them together until the pool of distressed MBS looks sufficiently "diversified". However, not just mortgages, but other loans are packaged into asset-backed securities. For example, auto companies make loans to help people buy cars, and then package those loans into asset-backed securities (ABS), which in turn are packaged into CDOs.
Once these ABS and CDOs are considered safe assets, they will be widely used, not just on the balance sheets of investment banks. To raise funds for operations, regular companies typically issue asset-backed commercial paper (ABCP), which is essentially a short-term loan. These ABCPs are supported by ABS and CDO. Companies such as General Electric, for example, would use CDOs derived from their auto loans to collateralize these ABCPs that were used to cover their day-to-day costs.
Money market mutual funds (MMMFs), whose goal is to make every dollar invested, buy up these ABCPs in droves because they're backed by seemingly risk-free assets. Even insurers like American International Group (AIG) are getting in on the action, offering insurance against ABS and CDOs to investment banks looking to reduce risk.
first level title
housing theory of everything
Suddenly, the housing bubble burst.
The immediate effect was a drop in the value of subprime mortgages. This means real estate investors are being scammed. For example, New Century is the second largest issuer of subprime mortgages in the world. With their assets dwindling, they filed for Chapter 11 bankruptcy on April 2, 2007.
Soon, the balance sheets of investment banks with ABS and CDOs were in the red. Bear Stearns will liquidate the hedge funds in July because of their exposure to these assets. By August 2007, BNP Paribas deemed it impossible to even value these financial instruments, preventing any investors from withdrawing money from their hedge funds. In September, Northern Rock, the UK bank, faced a run on depositors because it had acquired a large number of these assets. By early 2008, Northern Rock was nationalized by the UK government and Bear Stearns was bought by JPMorgan Chase along with the Federal Reserve.
Meanwhile, government-backed Fannie Mae and Freddie Mac began buying subprime mortgages from banks in an attempt to take some of the risk off their books. None of this was enough to stop the bleeding, and by September 7th Fannie and Freddie had to be taken over by the Fed. On September 15, Lehman Brothers went bankrupt and Merrill Lynch was bought by Bank of America. Just like that, three of the five largest investment banks (Bear Stearns, Lehman Brothers, and Merrill Lynch) disappeared.
The next day, huge insurance payouts from ABS and CDOs resulted in AIG being taken over by the Fed as well, and the oldest MMMF (Money Market Mutual Fund) Reserve Primary Fund also lost what every MMMF was expected to maintain1: 1 parity. In other words, it has "below $1" because its exposure to ABCPs is becoming as worthless as their underlying assets. As expected, investors across the country pulled $144 billion from US MMMFs the next day, causing companies to face problems covering day-to-day costs as they try to rollover their ABCPs.
With the interbank, asset-backed commercial paper (ABCP) and MMMF lending markets under pressure, there is only one source of short-term funding left: the secured repo lending market. Unlike the other three, borrowing in the repo market requires surrendering collateral. But as investors sold other assets to compensate for losses in ABS and CDOs, even those without exposure to those assets saw their balance sheets shrink sharply and the value of their collateral shrink, putting pressure on the repo market.
If it is clear and public whose balance sheets are worthless and insolvent, who is only temporarily illiquid, then perhaps they can be isolated and contained. But as the fog of war prevents us from figuring that out, borrowing costs have soared across the board. Banks stopped lending and credit dried up, bringing the real economy to a standstill.
So the housing crisis turned into a global financial crisis.
Fiscal Theory at the Price Level
That is the story of this global financial crisis.
Now I'm going to tell a different story. At least, it was a different story at first. Will the ending be the same? Let's find out.
Let's start with Terra first. Terra is a cryptocurrency ecosystem with two key securities: Luna and TerraUSD (UST). Luna is a cryptocurrency used for transactions on the Terra protocol. You can think of it like equity with no fixed value: its price reflects the extent to which people are building useful things on the Terra blockchain.
Meanwhile, UST is a stablecoin designed to be pegged to the U.S. dollar. You can think of it like debt: a fixed-income asset that you expect to be able to sell for the price you paid for it while still earning interest. Unlike many other stablecoins that have fiat assets backing their value, UST is an algorithmic stablecoin. This means that its peg to the US dollar is secured by an algorithm that will ensure that 1 UST is always exchangeable for $1 worth of Luna. If the value of UST is too high, the Terra protocol will introduce inflation by selling UST to the market; if the value of UST is too low, it will introduce deflation by selling Luna to redeem UST.
Algorithmic stablecoins may seem strange, so let’s map it to a more obvious concept in traditional finance.
Take the US dollar as an example.
One way to look at the dollar is to think of it as some sort of algorithmic stablecoin. Much like UST, it's an asset that you want to be able to redeem for the value you paid plus interest, although here the interest rate is negative (because of inflation). Like UST, its purchasing power is destined to follow a steady path, although here it is destined to decline by 2% per year. Just like UST, the inflationary policy involves issuing more stablecoins through monetary expansion. Similar to UST, a deflationary policy involves exchanging the market's stablecoin for an asset that represents a share of the system's economic output, either a government surplus or deficit. The only difference is that, for the dollar, this can be done by raising taxes to absorb the dollar, rather than by relying on consensual market transactions.
first level title
Mundell-Fleming Trilemma
The next traditional finance analogy is how countries create and defend exchange rate pegs.
The Mundell-Fleming Trilemma tells us that a country can only choose two of the following three options: free movement of capital, a fixed exchange rate, and monetary policy independence. When places like Hong Kong decide to peg the Hong Kong dollar to the U.S. dollar, they do two things: First, they make sure they have huge foreign exchange reserves that they can use to defend the peg. Second, they accept that if they want free capital flows, they have outsourced monetary policy to the central bank of whatever currency they are pegged to, such as the Fed.
Terra does it a little differently. they do have one"foreign exchange reserves", the Luna Foundation Guard, which holds billions of dollars worth of other cryptocurrencies. However, they try to have their cake and eat it too, refusing to accept this trilemma. In order to attract people to hold UST, they set up the Anchor protocol. If people buy UST and lend (deposit) it to the Anchor protocol, they will get extremely high returns. In other words, they wanted UST to have a fixed exchange rate against the dollar, refused to accept any barriers to capital movement, and wanted to set their own interest rates.
What happened when they started reducing the unsustainably high interest rates paid by the Anchor protocol? Almost immediately, there was a massive outflow of UST deposits. There are two ways to sell UST: the first is to rely on the stablecoin's protocol, using UST for Luna; the second is to sell UST on a decentralized exchange (DEX) called Curve for other stablecoins , Curve provides liquidity pools between various cryptocurrency pairs.
So when people start selling UST, the first way above means that the supply of Luna increases substantially. This means that the price of Luna drops, incentivizing people to sell their Luna for something else. This in turn will lead to more people trying to flee UST and Luna, causingdeath spiral. After all, nothing can sustain Luna's value.
As for the second way, the massive sale means that the amount of UST on the Curve exchange exceeds that of other stablecoins. To rebalance this, the exchange started offering UST at a discounted price, but no one wanted to buy it. As this imbalance grew, Curve had to keep offering bigger and bigger discounts. Both of these therefore destabilize UST's peg to the dollar.
first level title
Yield farming has been on the decline
While LFG's defense failed to prevent UST from collapsing, it drove down the price of many of the cryptocurrencies it sold (bringing selling pressure). Of course, it wasn't just them: everyone else who was overexposed to the UST and Luna debacles started selling their other cryptocurrencies.
In particular, there is one other asset that has been pushed off its “peg,” and that is stETH. Ethereum is currently preparing to upgrade to Ethereum 2.0, and as part of that, people can stake their ETH. This means that ETH will be locked on the yet-to-be-announced Ethereum 2.0 network, and it cannot be used or sold until the new network goes live, at which point the pledged ETH can be withdrawn. Of course, they will also be rewarded for participating in staking.
For those looking to maintain liquidity, the Lido Finance protocol offers a liquid collateralized token called stETH. For every ETH given to Lido, you get one stETH. This is a token that can be used like any other cryptocurrency. A popular transaction is to deposit ETH into Lido, get stETH, then use stETH as collateral to lend ETH on a lending protocol such as Aave, and repeat the process.
This cycle of yield farming seems pretty safe. Most of the time, stETH is exchanged for ETH at a ratio of 1:1. However, this is not a real peg, and until Ethereum 2.0 arrives, there is no reason to keep it completely 1:1. In particular, once you hand over your ETH, Lido stakes the ETH on your behalf, so there is no way to get the ETH back. The only way to unwrap this deal is to sell stETH on the secondary market, and if the market demand for liquidity increases, or people think the new network will be delayed, it will likely be lower than the anchor deal.
domino effect
domino effect
The UST and stETH panic started to cause systemic contagion, hitting key financial institutions in the cryptocurrency space.
One of the biggest dominoes to fall was Celsius.
Celsius is one of the largest centralized cryptocurrency financial service providers. You can use Celsius to trade cryptocurrencies, you can also deposit cryptocurrencies into Celsius to earn interest, and you can also borrow money from Celsius. Although Celsius doesn't tend to describe itself that way, you can think of its business model as a bank. It has a retail-oriented commercial banking unit, through which it takes deposits, and a retail- and wholesale-oriented investment banking unit, through which it lends.
Unlike traditional banks, where most banks earn money by earning the difference between deposit rates and borrowing rates, Celsius stands out by offering depositors a yield far above the market and lending at low interest rates to borrowers. How does it manage to provide both at the same time?
The simple answer is that Celsius is losing money, and the gains are being paid in part by taking on new depositors. However, the more complete answer is remortgaging and leverage.
Basically like all financial institutions, Celsius lends money from depositors to others. Re-mortgaging means that Celsius also lends out the collateral obtained from the borrower. In particular, Celsius has been depositing UST on the Anchor protocol, yield farming ETH/stETH using Lido and Aave, and yield farming stETH/bETH using Lido and Anchor. This means that Celsius has exposure to both UST and stETH.
So Celsius took a pretty big hit when UST was decoupled. That alone might not be enough to knock Celsius down. However, Celsius has also lost hundreds of millions of dollars in various hacks over the past few months, and worst of all, depositors started panicking to withdraw their money.
At the same time, stETH also suffered an unanchoring, creating a huge maturity mismatch, whereby Celsius' ETH-denominated liabilities (depositors' deposits) had to be paid immediately, but its ETH-denominated assets were incredibly illiquid (Most of these assets are either stETH or directly locked in Ethereum 2.0). The problem is, everyone is deleveraging at the same time, so the secondary market for selling stETH is drying up.
Centralized exchanges (CEXs) don’t have enough liquidity, and neither do DEXs like Curve. Liquidity Providers (LPs) don't want to be trapped and hold large amounts of stETH, so they exit; the stETH/ETH trading pool on Curve becomes 80% composed of stETH, which means there is not enough ETH to make Celsius sold all stETH even though it was willing to sell at a loss. The only way out is to have the broker go over-the-counter (OTC), but that means accepting a bigger discount. All of this causes the value of Celsius' ETH-denominated assets to fall at market prices, driving more users to make withdrawals.
lever
lever
Oh, and leverage.
To further increase its rate of return, Celsius borrows from the MakerDAO protocol. Much like Terra, MakerDAO has two key currencies: the stake token Maker (MKR) and the dollar-pegged stablecoin DAI. But unlike UST, DAI is an overcollateralized cryptocurrency-backed stablecoin. This means that for every $1 of DAI borrowed by Celsius, $1.50 worth of collateral needs to be put into the MakerDAO loan agreement, and if the loan-to-collateral value ratio exceeds a certain threshold, the collateral will be seized (liquidated ).
The problem with gaining leverage through an over-collateralized stablecoin is that the stablecoin is a natural liquidity consumer. Just as the spread of UST led to a sell-off in the rest of the cryptocurrency market, Celsius's collateral value began to drop, bringing it closer to liquidation. With a significant cash outflow, Celsius had to use limited resources to prevent default on its loans.
Double whammy.
Traders could smell the blood in the water as Celsius continued to be battered. If they sell enough asset types that Celsius is using as collateral, they can force Celsius to be liquidated. This will allow them to buy Celsius collateral cheaply, and since all of this is publicly visible on the blockchain, short sellers know exactly what to target and how much to sell.
triple whammy.
How is the result? Celsius limits withdrawals, preventing depositors from withdrawing their money. The ensuing fear, uncertainty, and doubt cause asset prices to plummet, making collateral worth less. At this point, Celsius can attempt to repay its loan, or replenish its collateral, to prevent liquidation.
For every $1 borrowed, Celsius needs to provide $1 to repay, or $1.5 to replenish its collateral. On the face of it, the former is cheaper. Celsius, however, chose the latter. Why? Because it knows that it will definitely not be able to pay its debts, but if it is to replenish enough collateral to last for a period of time, the cost will be very low. Therefore, the small amount of liquid ETH remaining in Celsius is locked up as collateral.
In short, Celsius took depositors' money and borrowers' collateral, leveraged it up, and bet it all on the roulette table. When UST and stETH decoupled and other assets fell with it, this "house of cards" began to crumble. Instead of honoring customers' redemption requests, Celsius locked their funds. Then they doubled down on the bet and made a last-ditch effort, hoping that the price of their asset would bounce back so they could make it all back.
first level title
unhedged hedge funds
It's not just Celsius, the shadow bank, that is under pressure. Another victim of this vicious circle is Three Arrows Capital (3AC), one of the largest hedge funds in the crypto space. Like Celsius, 3AC held exposure to UST and stETH, so when losses mounted and its collateral dropped to zero, 3AC ran into trouble. In fact, 3AC is even more exposed because it is one of the main organizations that help support Terra's LFG Foundation, and 3AC has used over $500 million to buy a large amount of Luna, making LFG's crypto reserves capitalized for various other encrypted assets. But that $500 million was wiped out almost overnight.
At this point, the story should be clear. 3AC was overleveraged and they were being asked for margin calls when the value of their collateral dropped. The big difference is that they basically don't communicate with anyone, so they get liquidated by the counterparty. These are some huge counterparties, including centralized exchanges like Bitmex, FTX, and Deribit, lenders like Celsius, BlockFi, and Genesis, market makers and trading firms like 8BlocksCapital, and brokers with whom they trade OTC.
first level title
end of beginning
What's next? I have no idea.
If the 3AC debacle we've seen in the past few days was a Lehman Brothers moment for crypto, it's worth remembering that the immediate aftermath of Lehman's collapse was seen as free market accountability against moral hazard. victory. But within days, it was realized that this was not a controlled defeat. The worst isn't over yet -- far from it.
Lehman Brothers did not mark the end of the 2008 crisis. This isn't even the beginning of the end. At best, this is the end of the beginning. It's not over yet. It's unclear what happens next. Maybe another stablecoin crash, like Tether or Magic Internet Money (MIM). Maybe it's a financial service provider like Nexo or Babel Finance, which has suspended withdrawals like Celsius. Perhaps one of the many exchanges that had to close out a position at a loss.
Likewise, perhaps we will meet a knight in shining armor who will inject enough liquidity into the financial system to calm everyone down. The DeFi world may not have a “lender of last resort,” but traditional finance used to be as well. Back in 1893 and the panic of 1907, there was no Fed. Instead, JPMorgan personally rescued the U.S. financial system. twice!
FTX is a promising candidate. In a way, it is large enough to reasonably create incentives to closely monitor risk: in order to make money, it needs to have solvent counterparties, and it needs to be solvent itself. What's more, as a CEX, it is less affected by a sharp downturn in the market than a lending institution. That, combined with its friendly relationship with market maker Alameda Research, could help provide emergency liquidity given that market makers tend to do well in times of market volatility.
Mamma Mia! here we go again
By now, these two stories should sound familiar.
Irrational exuberance and a savings glut spur demand for safe assets.
Financial institutions restructure and repackage new yield-generating securities until they are deemed safe.
Market participants acted as shadow banks, increasing leverage until the system became so fragile that a single howl of the “big bad wolf” could tear it down.
When that happens, there is a run on every financial institution, even those with even a tiny bit of this no longer safe asset.
This led to margin calls, liquidations, panic over counterparty risk and contagion until credit dried up completely.
In 2008, after all of this, we learned a few things.
First, the financial system transforms risk, not destroys it. This can help reduce idiosyncratic risks by creating secure assets whose information is not sensitive. Yet safe assets are at the heart of systemic risk, and even seemingly safe assets like money market mutual funds can go bust.
Second, don't mistake leverage for genius. Often, leverage is used because of greed or moral hazard, not because there is a real opportunity.
first level title
biggest trick
The biggest trick centralized DeFi institutions have ever played is convincing the world that they are not traditional financial institutions (TradFi). In fact, the word itself is a misnomer. This trend towards centralization is not surprising. In any network, there are easy-to-use entry and exit channels, especially since decentralization inherently suffers from adverse selection and lack of recourse. Can DeFi be an exception to this rule?
I think this crisis has shown that potential. While some of the major centralized DeFi institutions may have done poorly this time around, DeFi protocols and critical infrastructure have weathered the storm so far. The transparency that everything is on-chain helps sound the alarm, which is why more people know about Terra than Celsius, and more people know about Celsius than 3AC.
As long as there are transactions under the blockchain, there will be things that cannot be audited. Facing these known unknowns, panic and contagion are almost inevitable. But that's the whole point of the project: ensuring trustless and transparent transactions through blockchain and smart contracts. If everything is on-chain, you can prove beyond a doubt that you are exposed to the risks that people worry about.
Of course, there will always be some centralization in the crypto space. Among traditional finance advocates, some see this as a reflection of the sheer stupidity of DeFi projects; once you open the Pandora's box of fraud, deceit, and deceit, you can't close it; and increased transparency and tradability mean volatility increase; since there is no central bank, the interconnected nature of DeFi means that it will always be teetering on the edge. I disagree. When all is over, when the worst has happened, there is one thing left in Pandora's box: hope!
