Welcome to another issue of Net Interest, my newsletter on financial sector themes. This week, the subject is interconnectedness in the financial system. It follows on from my piece a few weeks ago, How Markets Work, and looks at some of the similarities between the current crisis in crypto and the financial crisis of 2007-08. Paid subscribers also have access to additional content on breaking themes in the weekly More Net Interest section.

Today’s edition of Net Interest is brought to you by Third Bridge.

Third Bridge Forum is the biggest archive of expert interviews in the world. Just last year over 16,000 investment professionals from 1,000 firms across private equity, public equity and credit downloaded over 500,000 interviews. The coverage is extensive – covering both public and private companies, in any sector, across all major geographies. I’ve seen it for myself – the insights Forum delivers are in-depth and unique. If you want to request a free trial visit thirdbridge.com/net.

Request a free trial

The Hedge Fund, the Bank and the Broker: How It All Fits Together

Fifteen years ago, as the financial system began to crumble, I sat at my desk in the heart of London’s hedge fund district trying to make sense of it all.

By summer 2007, the subprime crisis was already in full flow – mortgage originators like New Century had gone bankrupt and investors in securities backed by their and others’ mortgages were nursing losses. Initially, policymakers thought the challenges isolated. Ben Bernanke, chairman of the Federal Reserve, testified before Congress that “the impact on the broader economy and financial markets of the problems in the subprime market seems likely to be contained.” A few months later, he put a number on it: “Some estimates are in the order of between $50 billion and $100 billion of losses associated with subprime credit problems.”

We now know of course that subprime was far from contained. First, a pair of hedge funds went down. Bear Stearns had run a ‘high-grade structured credit strategies fund’ from within its asset management division since 2003 and, in 2006, it added a higher-leveraged, higher-risk ‘enhanced fund’ alongside it. The funds’ assets were pretty illiquid – 60% of the high-grade fund comprised subprime mortgage-backed CDOs – and their managers had broad discretion around where to mark them. In April, they marked the enhanced fund down 6.6% but, with subprime securities in freefall, this didn’t withstand scrutiny and later had to be revised. “There is no market… its [sic] all academic anyway – 19% is doomsday,” emailed one of the managers privately, while announcing a 19% drop in the fund’s value and a freeze on client redemptions.

Over the next few weeks, the funds continued to lose value. In order to leverage their enhanced fund – even the high-grade fund – Bear Stearns managers had borrowed from JPMorgan, Merrill Lynch and others against the value of the funds’ holdings. After the funds froze redemptions, Merrill Lynch seized its collateral and dumped it on the market, causing valuations to plunge further. A death spiral commenced. By July, the high-grade fund was down 91% and the enhanced fund had lost everything. On 31 July, both filed for bankruptcy. 

In a June meeting, Federal Reserve governors compared the situation to the collapse of Long Term Capital Management nine years earlier. “These guys were regarded as very smart people, who knew the business well,” observed Vice Chairman Tim Geithner. “This is just a good example of how little one can know in some sense and what leverage does to your exposure to liquidity risk. In this case, people were just not willing to give them the time to realize whatever value might be in these positions, and this is just a natural consequence of leverage.”

By now, other dominoes were beginning to fall. IKB Deutsche Industriebank AG was a German bank, founded in 1924 to support medium-sized enterprises with loans, capital market services and advisory services. In the early 2000s, in a quest to generate more profit, it strayed into “securities arbitrage”. It launched an off-balance sheet vehicle called Rhineland which bought structured credit, which it financed by issuing less expensive commercial paper. By the end of June, the vehicle had €14 billion of assets, 95% of which were CDOs and CLOs (collateralized loan obligations).

Even as the Bear Stearns funds struggled, IKB put on a brave face. The bank put out a press release on 20 July, confirming profit expectations for the full year “in the light of this positive start into the financial year 2007/2008”. But Rhineland was struggling to place commercial paper with investors, who didn’t want anything to do with the mortgage assets on its balance sheet. One week later, Deutsche Bank cut off its credit lines; IKB was forced to backstop its structured credit vehicle and had to be bailed out. 

The IKB episode highlighted that toxic mortgage assets were lurking in the portfolios of even the most risk-averse investors. From then on, shoes would drop all over the place – investment funds, surviving mortgage originators, banks, brokers – all the way to Lehman Brothers over a year later. 

I didn’t succeed in building a full model of the financial system at the time but I saw that the system was a lot more interconnected than it appeared on the outside.

Hedge funds, banks, brokers – they are all linked, and the way they are linked is through cash and collateral flows. Across the system, cash is borrowed, assets are bought, and assets are used as collateral. It’s a three-layered system: assets, funding, collateral. Most of the time, it works smoothly but when the layers become too tightly knit, problems can cascade. 

What makes the overall picture hard to discern is that the cash and collateral flows change form as they move around the system. Mortgage originators took less liquid housing loans and turned them into more liquid mortgage securities. CDO structures took lower rated mortgage securities and turned them into higher rated tranches. Rhineland bought CDO tranches and spat out commercial paper. Nobody was keeping track of it all (least of all the Robbinsdale Area School District in suburban Minneapolis which bought some of the Rhineland paper). 

Since the crisis, these linkages have become better understood. Policymakers regularly map them and are alert to signs of stress. Central banks have shown a willingness to intervene when the system gums up, most recently in March 2020. 

So I never imagined seeing a replay of those events so soon. But over the past few weeks, they have been reprised within the crypto ecosystem. We’ve discussed here before how many of the same structures that evolved in traditional financial services have been replicated in crypto, sometimes unknowingly. It seems that goes for the vulnerabilities, too. 

In one way, recent events in crypto help to shine more light on the events of fifteen years ago. Everyone has their pet theory of the cause behind the financial crisis: low interest rates, government housing policy, unsustainable credit demand, weak credit underwriting standards, mixed public/private mandate of Fannie Mae and Freddie Mac, China’s accumulation of dollar reserves, failure of financial regulation, role of credit rating agencies, growth of shadow banking system, poorly designed compensation schemes, financial institution leverage, mark-to-market accounting, financial institution interconnectedness, perception of ‘too big to fail’, short-selling, and so on.

But in the time since, we’ve watched an alternative financial system evolve unfettered and it’s ended up in the same place. Maybe it’s not any individual component of the system that’s the cause; rather, instability is an inherent feature of the system.

Cryptopocalypse

So what are the key events of the current crypto crisis?

First, a hedge fund gets hit. Three Arrows Capital (3AC) is a Singapore-based hedge fund focused on crypto assets founded in 2012 by Kyle Davies and Su Zhu. In April, it reportedly had over $3 billion assets under management. The fund was heavily exposed to the crypto asset Luna and when Luna collapsed, 3AC was hit hard. It filed for bankruptcy on 1 July. 

Then the bank. Celsius Network LLC was set up in 2017 to provide banking services in crypto assets. It claimed to have 1.7 million retail customers, pulling in as much as $24 billion of assets by December last year. Its appeal to customers lay with the high interest rate it offered on deposits – up to 18% on some cryptocurrencies and 7% on stablecoins. Its business model is laid out in a suit a related company brought against Celsius this week:

Like a bank, Celsius is meant to invest those funds responsibly, earn a return, pay the depositors the interest they earned, and keep the profit. Importantly, if Celsius fails to profitably invest depositors’ funds in investments that earn more than the interest owed, they will operate at a loss. Central to consumer trust, however, is the promise that upon request, Celsius has sufficient funds to return the crypto-asset deposits for each of its users.

But Celsius didn’t invest its funds responsibly and on 12 June was forced to halt customer withdrawals. It took on a whole series of unhedged risks with customer funds:

• Exchange rate risk. According to the suit, Celsius had massive liabilities to depositors in ether but had not maintained ETH holdings equal to those liabilities. When the price of ether rallied in early 2021, Celsius was left exposed. “When Jason Stone left Celsius [in March 2021], Celsius had a $100-200 million hole on its balance sheet that it could not fully explain or resolve.”

• Liquidity risk. According to the FT, the company took its customers’ ether and locked up huge sums in a new version of the currency that is meant to track it 1:1 but became less liquid. 

• Credit risk. The Wall Street Journal describes undercollateralized lending practices: Celsius required business borrowers to post only 50% collateral on loans (and would then take that collateral to borrow more money for itself, a whole different story that also has echoes of the global financial crisis). 

Meanwhile, the capital propping up its balance sheet was very thin. According to the Wall Street Journal, Celsius had $19 billion of assets and roughly $1 billion of equity in 2021. That gives it a capital ratio of just over 5%. At the time of the financial crisis, that would have been standard for a bank; not so any more. 1

Celsius also points at some of the interlinkages within the system. According to the Wall Street Journal, it had a credit facility for up to $1.1 billion from stablecoin issuer, Tether. We’ve talked about Tether here before. Meanwhile, Tether itself was an early investor in Celsius, with a 7.8% stake as of last spring. 

But if a bank collapse wasn’t enough, we now also have a broker. Voyager Digital was founded in 2018 (the speed with which these companies have risen and fallen is quite remarkable). Its primary operations consist of (i) brokerage services, (ii) custodial services through which customers earn interest and other rewards on stored cryptocurrency assets, and (iii) lending programs. The company pitched its services to retail investors. 

In theory, Voyager should have been unaffected by turbulence in the crypto markets. The company itself states, “As Voyager is a brokerage services provider that is not directly dependent on the price of Bitcoin or any other cryptocurrency for revenue, Voyager has limited exposure to an industry-wide selloff.”

But like others before it, Voyager wasn’t content simply to do its thing, acting as an agency broker, matching customers with counterparties to facilitate their trades. No, its focus was growth. Over the course of 2021, it grew its number of unique, verified users from 160,000 to 3.2 million. It attracted customers by offering yield on the assets they store with Voyager. By March 2022, it had $5.8 billion assets on its platform, on which it had paid out $220 million in yield over the preceding 12 months. But with only $260 million in brokerage fees to cover the yield, Voyager had to establish other revenue lines, and it ramped up its lending activities.

The trouble is, its loan book was very highly concentrated. Between March 2021 and March 2022, Voyager lent out a cumulative total of $5.15 billion to just nine different counterparties. Nor was its lending sufficiently collateralised – as at March 2022, borrowers had posted collateral of $227 million on loans of $2.0 billion. “But the people we lend to are some of the biggest names in the industry,” the company’s CEO proclaimed on his investor call in May.

And then one of them – Three Arrows Capital – went bust. The amount outstanding to 3AC was $654 million, but as at end March, Voyager only had $258 million of equity capital (equivalent to a capital ratio of 4.3%). Voyager calls it a “cryptopocalyse” [sic] but really it’s just poor risk management. 

Usually, financial companies fail because of a lack of liquidity rather than a lack of solvency. Not so here. The company did experience a run in the days after Celsius restricted withdrawals which it tried to resolve by negotiating a credit line with Alameda and putting a cap on its own customers’ withdrawals. But confirmation that it wasn’t getting its money back from Three Arrows made that moot, and this week it filed for bankruptcy protection. 

Like in the case of Celsius, the situation shows how intertwined the crypto ecosystem is. Voyager was a lender to Alameda and then took a rescue loan from Alameda. Alameda is also one of Voyager's largest shareholders, with a more than 9% stake.

The situation also brings the issue of deposit insurance into the spotlight. Voyager had mis-sold its customers on the protections afforded by deposit insurance. “In the rare event your USD funds are compromised due to the company or our banking partner’s failure, you are guaranteed a full reimbursement (up to $250,000),” the company stated in 2019. But as survivors of the global financial crisis know, deposit insurance only applies if the bank fails – and Voyager’s bank, Metropolitan Commercial Bank, is alive and well. 

Instead, Voyager’s customers – 97% of which have less than $10,000 in their account – have become creditors in a bankruptcy proceeding. For now, they are being offered a mix of tokens, new common shares in a reorganised Voyager and a claim on any recovery in the 3AC loan, but that is all up for negotiation. 

It’s been 623 days now since a bank in the US went bust, so perhaps it’s no surprise people have forgotten what deposit insurance means. The previously longest stretch was just prior to the financial crisis, when deposit insurance proved its mettle. It took multiple earlier financial crises for the safety net of both deposit insurance and the Federal Reserve as a liquidity provider of last resort to be instituted. Recent events in crypto are a fitting advert for their merits.