Ten years after the failure of Lehman Brothers, the conventional narrative is that the bank failed because of the US real estate bubble and the collapse in the market for complex, property-related financial derivative contracts.
These—and other factors like dormant regulators, corrupted credit rating agencies and the ludicrous bonuses paid out by banks to bankers from imaginary profits—undoubtedly helped contribute to the 2008 crash.
But what actually pushed Lehman over the edge?
To answer that question, we must rewind the narrative to the 1960s paper crisis in the eurobond market, described in our previous article on settlement.
The 1968-69 New York crisis showed the risks inherent in a certificate-based share and bond trading system, requiring the movement of paper from each seller to each buyer.
Over the next two decades, paper certificates were replaced by entries in electronic ledgers, and the responsibility for processing trades was passed to a series of trusted intermediaries. But the new system generated risks of its own—and much larger ones.
JP Morgan’s greatest trade
Concerned at the late 1960s settlement logjam and its potential impact on their livelihoods, eurobond traders sought a way to shift the processing of trades away from a system reliant on messenger boys running between banks’ vaults. Instead, they created a centralised institution, responsible for all market counterparties’ settlement and safe custody.
In 1968 Morgan Guaranty Trust Company (now JP Morgan Chase) set up an entity called Euroclear to handle the settlement and safe custody of clients’ bonds.
“Euroclear was so successful that it almost became an embarrassment”
Euroclear, owned by JP Morgan for over thirty years before being sold to its users, turned out to be one of the bank’s greatest investments, repaying the initial outlay many times over.
According to the late Ian Kerr, a banker in the 1960s and 1970s, the new settlement system was a huge cash cow.
“Euroclear was so successful that it almost became an embarrassment,” Kerr said in 2003.
“[But] I felt that it behaved like a rich, schoolboy bully and could have led the way towards much lower clearing and settlement costs much earlier. When anyone asked me about them, my attempted neutral response was, ‘At least they give quite a good cocktail party at the IMF’.”
Euroclear—and its owner—were a one-stop shop. As well as settling trades via Euroclear and looking after clients’ assets, JP Morgan provided credit lines to bond trading firms. This put the bank in the position of receiving privileged information about other firms’ inventories and trading flows.
It also provided an early indication that the new trusted intermediaries in the settlement and custody business could have an almost unassailable competitive position.
Between the 1960s and the 1980s, securities settlement around the world switched from the bilateral processing of paper certificates and cash to a system based on virtual records, vertically integrated service providers and tie-ins to central banks.
The key entities in the new system were central securities depositaries (CSDs) and central counterparty clearing houses (CCPs).
Under the new post-trade securities infrastructure, the first exchange of information after a stock or bond transaction involved the passing of details to the CCP, which steps in as the buyer to each seller and the seller to each buyer.
CCPs match trades and net them, helping reduce the number of financial flows between individual counterparties. Based on the netted financial flows calculated by the CCP, CSDs then settle trades, adjusting ownership records in exchange for payment.
CSDs may tie banks and brokers into the national monetary system, since the settlement of debits and credits between CSD member firms often takes place in accounts at the central bank.
In many markets the CSD does not recognise the individual clients of a bank or broker. Instead, client assets are held at the CSD in the form of an ‘omnibus’ client account held by the bank at the depositary.
“Things were digitised in the 1960s and 1970s under a mainframe technology architecture”
Unsurprisingly, CCPs and CSDs are seen as systemically important institutions, meriting special attention from supranational bodies like the Bank for International Settlements (BIS) and the G20 Financial Stability Board.
In technological terms, they are also very much the product of their time—which was forty or fifty years ago.
“The current model of market structure was set up in the late 1960s and early 1970s,” Robert Sams, chief executive of Clearmatics, a company working on distributed ledger technology, told New Money Review.
“Things were digitised during that period under a mainframe technology architecture. Even the legal form of dematerialised assets reflected the background assumptions of the technology. Much of that now needs to change,” says Sams.
A dependence on bucket accounts
Repeated cases of misuse of client assets point to a particular weak point in the current system: custodian banks’ use of omnibus (bucket) accounts.
Tim Reucroft, director of research at Thomas Murray Limited, a London-based advisory firm, explains.
“Omnibus accounts were introduced as a direct response to the late 1960s paper jam,” Reucroft told New Money Review.
“Instead of settling trades across thousands of accounts, the idea was to put everything in one bucket called an omnibus account and settle it from there.”
“Robbing Peter to pay Paul”
But some custodians have been caught out, in the UK financial regulator’s words, “using Peter’s money to pay for Paul’s transactions”.
In April 2015, for example, the Financial Conduct Authority (FCA) fined BNY Mellon, the world’s largest custodian bank, £126 million for failure to adequately record, reconcile and protect safe custody assets.
Hinting at a widespread industry problem, this was the eighteenth penalty levied in four years on UK financial institutions for breaches of the country’s custody rules.
BNY Mellon, said the UK regulator, had committed two cardinal sins.
It had mixed clients’ assets with assets belonging to the bank in omnibus accounts held with third parties. And it had borrowed assets from some clients’ omnibus accounts to settle the transactions of other clients, before the proceeds of covering trades undertaken by those other clients had been received. Both practices put client money at serious risk.
In some markets, says Thomas Murray’s Reucroft, clients can now ask for their assets to be kept separate from those of other clients. Under Europe’s CSD Regulation, passed in 2014, for example, individual segregation is available from CSDs on demand. But this often comes at a cost, Reucroft says.
“The US custodian banks are very reluctant to offer anything other than an omnibus account,” he told New Money Review.
“There’s a price differential when people ask for individual segregation. The custodians are not embracing it because it would be a massive cost to their system. And if the CSDs themselves don’t offer it, what do you do?”
Custodian banks’ preference for omnibus accounts has also helped them develop other lucrative financial market businesses, such as securities lending, in which clients’ shares or bonds are lent in bulk to borrowers, often hedge funds, for a fee and in return for collateral.
This fee is then split between the bank and the asset owner—but the risk of something going wrong usually rests with the latter.
New—and scarier—types of risk
In an era where the average mobile phone has many times the processing power of a 1960s mainframe computer, a securities settlement system where individual ownership records are still seen as too difficult to introduce and maintain seems odd.
Some markets have already leapfrogged the legacy post-trade infrastructure of Europe and the US. Many emerging markets, for example, now insist on individual segregation of clients’ securities holdings at local CSDs as a matter of course.
But the risks in the centralised system of CCPs, CSDs and custodian banks go beyond the potential for another settlement logjam, and the costs of the current system are not just those of anti-competitive behaviour.
The post-1960s centralisation of clearing and settlement at trusted intermediaries had unpredicted consequences, blindsiding regulators and arguably helping trigger the 2008 financial crisis.
The withdrawal of repo funding pushed both Bear Stearns and Lehman Brothers over the edge
Above all, the 2008 crisis was a run on the repo market, where financial institutions could obtain short-term funding by pledging their debt holdings. JP Morgan and BNY Mellon, though private institutions, acted as the main clearing platforms for the whole repo market, extending nearly $3 trillion in credit on an intraday basis to repo market counterparties at one point during 2008.
It was the withdrawal of repo funding that pushed both Bear Stearns and Lehman Brothers over the edge into bankruptcy that year: Bear Stearns had $50bn in overnight repo loans when it failed, while Lehman had $200bn outstanding.
As the parties responsible for extending and rolling over repo loans, JP Morgan and BNY Mellon had an effective power of executioner over the failing brokerage firms—and this was a power they eventually exercised.
“CCPs are super-systemic”
According to Paul Tucker, former deputy governor of the Bank of England, who now runs the Systemic Risk Council think-tank, the centralisation of risk at clearing houses is now the biggest fault-line in finance.
“CCPs—especially those clearing globally traded instruments—can be thought of as super-systemic. It is striking that political leaders and economic commentators are not more focused on this,” says Tucker.
A reminder of the potential risk embedded in CCPs came earlier this month when Einar Aas, a Norwegian energy trader, blew a big hole in the mutualised default fund of Nasdaq Clearing, having eaten up the initial and variation margin supposed to protect other CCP members against individual default risk.
Though JP Morgan has now quit its clearing role in tri-party repo clearing, BNY Mellon, as well as being the largest custodian bank in the world, remains the central institution for that market. Its importance as a potential single point of failure has arguably increased.
Can blockchain help?
Proponents of distributed ledger technology argue that their solutions can address the problems of moral hazard and the risks that are inherent in the post-trade infrastructure. But can blockchain-like technology really replace the most protected part of the financial system?
In the third and final article of this series, we review the prospects for a decentralised structure of share, bond and asset ownership.