|
On October 7 in Florence, the plumbers of the world's financial markets will attempt heroically to put aside their differences and cooperate. In theory they are all trying to avert the same disaster. In practice, for the last few years, they have been pulling in different directions and pursuing at least five different solutions.
The disaster is Herstatt risk, something the world has come close to only once, in 1974, when Bankhaus Herstatt in Cologne was shut down by the German banking supervisors. The closure, at 3.30pm German time, left all of Herstatt's dollar foreign exchange transactions half completed in New York, where it was still only 10.30am. The panic caused, even by that small number of failed payments, resulted in dislocation and gridlock in the foreign exchange markets.
Leading bankers and central bankers vowed it should never happen again. And it hasn't. In 22 years, including the collapse of several banks Continental Illinois, Drexel Burnham Lambert, Barings and disruptions such as the 1990 invasion of Kuwait and the 1991 attempted Russian coup, there has been no gridlock of Herstatt proportions.
But the only thing preventing it has been fast-thinking central bankers. For 22 years very little has been done to the plumbing of financial markets to reduce the risk of gridlock.
Ask those leading bankers and central bankers what is their worst nightmare and it is still the "gridlock caused by sudden failure of a major international bank". But it is important not to forget other potential causes of gridlock political chaos in a major country, the imposition of exchange controls on a reserve currency, the default of a big exchange, or the collapse of even a medium-size bank.
Former New York Federal Reserve Bank chairman Gerald Corrigan says that for years he was a "voice in the wilderness" on payments risk. "In the last four or five years there's been much greater willingness to work for the common good," he says. "Payment systems were taken for granted. Now the issue has come out of the back-room into the board-room." In January, as chairman of the international advisory committee at Goldman Sachs, he put together a symposium on the subject. "This should be a concern not just of commercial banks," says Corrigan, "but also of investment banks and the world's top 100 globally active industrial companies."
Because of the new risk management culture among banks, and because central banks have decided they must retreat from being the markets' ultimate punchbag, a crunch has come. In March, the G10 group of leading central banks in a report now known as the "orange book" gave market practitioners two years to reach a solution. Failure to do so, they implied, would lead central banks to impose a solution themselves, in the form of penalties or a capital charge.
So banks, some willing, some unwilling, are being forced to concentrate and spend money on this risk, which some of them believe is so small it's hardly worth thinking about. Even in the Herstatt case hardly anyone lost money.
But recent studies suggest the risk is far greater than previously thought. Foreign exchange market volume has grown to an estimated $1.4 trillion a day that means $2.8 trillion in payments on both sides of the transaction. Around 80% of that volume is concentrated between the 20-or-so major players. Detailed surveys of the length of time at risk between payments of one leg and receipt of the other leg of a transaction, often via correspondent banks, show that the gap can be up to three days. During that time more than the entire principal of one leg of a transaction can be at risk. Given today's technology there appears to be no excuse for this sloppy risk management. The strongest message from the G10 orange book is that banks can do a lot simply by tightening up their own procedures for pursuing settlement and tracking their exposures.
Widening the net
To be fair, there have been several private-sector initiatives since the mid-1980s to address Herstatt risk more precisely, foreign exchange settlement risk. But they have been fragmented by politics and prejudice. Citibank and a club of major international banks created FXNet in 1984. That nets foreign exchange payments bilaterally between roughly 30 international banks in 16 currencies, reducing the physical payments and hence their mutual settlement exposure. According to FXNet, daily transactions of around $170 billion are netted down to flows of around half that number. That accounts for about 8% of the world's daily foreign exchange volume.
Swift (Society for Worldwide Interbank Financial Communication) created Swift Accord in 1990, which logs members' foreign exchange transactions and allows them to net bilaterally. A group of European banks last year went live with Echo (Exchange Clearing House), a multilateral netting system, after two years of trials followed by one awaiting approval. Multilateral netting reduces the physical flows still further, since members transfer their deals to a single counterparty, the Echo clearing house. A single daily payment in each currency, covering all its transactions with other members, is all that is necessary between each member and the clearing house.
But Echo, which accepts deals through Swift Accord, cost £50 million to set up and is still losing money, say rivals who believe it has too few members logging too few deals to achieve critical mass. Echo officials say it is a capital-intensive business that will take time to see returns the system continues to add members and volume. It is seeing flows of $6 billion gross a day (which could represent as much as $18 billion in principal at risk, say Echo officials) netted to around $2 billion.
In North America there is a similar multilateral clearing house, Multinet not yet in operation being established by six Canadian and two US banks. Multinet is going through the same tortuous legal and regulatory process that Echo endured, in order to clear deals in seven currencies in several jurisdictions. Multinet has an agreement with FXNet which, according to FXNet chairman Peter Bartko, could reduce FXNet's net daily flows of $85 billion to just over $30 billion, and further to a mere $6 billion if instantaneous payment-versus-payment is applied.
Multilateral netting is a good interim solution, since it reduces the total payment flows. Echo has added a payment-versus-payment feature, and Multinet intends to do the same, to reduce the time-lag between payments as well as the volume. But they are both clearing houses, which means they become the counterparty to every deal done through them.
Many of the biggest players dislike the idea of a clearing house because it undermines their competitive advantage. A clearing house gives all members' deals the same credit rating. It also stands in the way of inter-dealer relationships and opportunities for horse-trading and re-negotiating forward deals during their life. JP Morgan and Bankers Trust, among others, see that as an unnecessary sacrifice.
So Multinet and Echo have failed to attract the biggest banks, with a handful of exceptions. Chase is a member of Multinet, for historical reasons the banks merged with Chase, Chemical and Manufacturers Hanover, were not members. ABN Amro, BNP and HSBC are among the bigger members of Echo. ABN Amro, Barclays, HSBC and Lloyds Bank are members of Echo and FXNet (see diagram, page 70).
There are technical differences between Echo and Multinet and differing philosophies on risk management. But the biggest gulf between them is historical and tribal. Each side recognizes that it makes little sense for two systems to compete to become what is close to being a public utility they rob each other of critical mass. For five years there have been attempts by big players to bring Echo and Multinet together and perhaps to modify the clearing house concept, but to no avail.
The G10 central banks' committee on payments produced a series of papers between 1989 and 1993, encouraging efforts to reduce Herstatt risk, deliberately falling short of backing any one private-sector solution. Yet one report in 1990, known after its chairman as the Lamfalussy report, has set minimum standards for any netting system seeking approval from a central bank. The Lamfalussy minimum standards demand, among other things, legal certainty in the relevant jurisdictions, technical soundness, fair access, and in the case of multilateral netting the ability to ride out the default of the one counterparty with the biggest overall position.
Most bankers accept the Lamfalussy standards, although there has been some murmuring about supervisors wanting to go to "Lamfalussy plus" demanding enough liquidity to cover default by the two biggest position-holders.
In 1994 major New York banks, members of the New York Clearing Houses Association (NYCH), formed a global payments committee to examine progress on netting schemes so far, found it wanting, and tried to redirect efforts tothe core of Herstatt risk: that is, to reduce to zero the time-gap between one payment and its counterpart, by means of simultaneous release of funds, so-called payment-versus-payment (PVP). Many domestic payment systems in national currencies are already capable of supporting this. Real-time gross settlement (RTGS) exists in most of the G10 currencies, but much work is needed to reconcile national idiosyncracies, as Echo's experience with cross-border PVP bears out. With payment systems extending their opening hours so that they overlap across time-zones, payment-versus-payment across continents becomes more feasible. In 1997 the US Fedwire system plans to be open from half past midnight until 6.30pm.
Major international banks seized on the NYCH initiative and formed a group of 17 members in June 1994, calling themselves the G20. In March this year the G20 went public, announcing that it was concentrating on a model for payment-versus-payment through "continuous linked settlement". This would be real-time settlement whereby matched deals between members that satisfy risk management limits become irrevocable and are settled instantaneously through the passing of entries on the books of a special purpose bank. The initial focus is on G10 currencies (which account for the bulk of foreign exchange volume). But the aim, according to programme director Rob Close of Barclays, is eventually to handle "all the world's FX settlement". The criterion for admitting a currency for settlement by the continuous linked settlement bank (CLSB) would be that it has legal certainty based on a national real-time gross settlement system providing payment from central bank accounts. And, to safeguard liquidity, banks using the system would be of a minimum size and standing. The system would also take netted payments from bilateral or multilateral netting systems, such as Echo or Multinet, or netted flows from pairs of banks.
But outsiders have tended to see this project as too much of an exclusive club. By concentrating on payment-versus-payment and real-time gross settlement the G20 project which has added three more members to make up the 20 seems to be trying to upstage the other efforts to reduce risk by netting. Real-time gross settlement can accept netted payments, but it doesn't need them. Certainly the promoters netting systems see the G20 as yet another competitor threatening to rob their efforts of critical mass.
In June a number of predominantly Echo banks European banks which tend to be second-tier in the foreign exchange markets formed a lobby group in protest at the G20, calling themselves the G40. The G40, thankfully, is not offering yet another mechanical solution. It simply wants to keep itself informed of what the G20 is doing, says spokesman Finn Otto Hansen, general manager at Den Norske Bank. "We want to cooperate with them," he explains. "To start a new system to cover all the risks you need more than 20 banks."
The G20 says it welcomes another 40 banks concerned about Herstatt risk. But it is keeping the details of its CLSB under wraps. "It's a very complex project," says Close. "We want to keep the industry informed, without having a wide debate until technical matters are researched further." A G20 working group colleague explains: "It's hard enough getting 20 people round a table to reach a decision."
The G40 makes the point that, although its 48-or-so member banks account for about 30% of FX trading volume, and the next 50 banks only around 15%, those 50 banks represent the highest risk of unexpected failure, and 15% is still $500 billion a day of exposure. "The risk you're taking on medium-size banks has historically been the highest," says G40 spokesman Hansen.
Political games
In Florence on October 7, nearly 2,000 delegates at Sibos, the annual Swift International Banking Operations Seminar, will have a chance to hear the latest on the G20 project. Chairman of the G20 steering committee, Mark Urkowitz of Chase, and Close of Barclays will outline progress so far on the continuous linked settlement system and how it might affect the audience. Representatives of G40, Echo, FXNet, Multinet will be there. If there is going to be a watershed, marking a new era of cooperation on Herstatt risk, this should be it. But don't hold your breath.
The G20 itself is riven with politics, not only between banks and groups of banks, but also within the banks themselves, between payments people, risk managers, and FX dealers. Those close to the G20 admit that the initiative came from payments people. The cynical view is that payments divisions favour gross settlement because of the opportunity to squeeze revenue from each transaction. FX people favour netting because it reduces the number of transactions, which to them represents a cost. Risk managers protest that all bankers welcome risk reduction as a common goal. G20 spokesman Close, who is payments strategy director at Barclays, explains that payments revenue will drop significantly with continuous linked settlement (CLS). "The transactions will be processed through CLS by book transfer, which is far different from making a charge on a payment." But if there's a loss, "the payments people's books would be hit". Reducing risk, says Close, means reducing revenue "but that's a small price to pay".
Others see the G20 as a political game being played across the Atlantic. Some fear the aim is to create a cartel through which all FX payments must pass. A G20 source says pragmatically: "We'll try to make money, but not at the expense of eliminating settlement risk." The G10's orange book makes it clear that central banks will not favour a solution that seeks to maximize profit, or that ventures into central bank territory. Ernie Patrikis, first vice president at the New York Federal Reserve, cautions: "We don't want to see the CLS bank creating money or making one-sided clear payments. That should be done by domestic payments systems, which in the US means Fedwire and Chips [Clearing House Interbank Payments System]."
The most cynical see G20 as an exercise in prevarication kidding central banks that something is being done while ensuring that none of the netting schemes gains critical mass. "The last thing a triple-A bank wants is a level playing-field," admits an FX manager at one top-rated bank.
Bartko at FXNet is certainly frustrated by the lack of decision: "I wish they'd pick one [solution] so I don't have to compete with things that don't exist."
A G20 member describes the exercise as "almost more a corporate finance problem than an intellectual problem". But to avoid further dissipation of effort the G20, FXNet, Multinet, Echo and G40 factions now have to talk to each other. "It's a challenge to bring them together," says a G20 source, "with their different shareholding structures". It will be difficult for the Sibos delegates to strip away their tribalism and reach dispassionately for a common solution.
The technical differences between Echo and Multinet, although small, embody philosophical differences. Multinet is described by supporters as a "robust" clearing house. Its medium-term debt has a triple-A rating from IBCA. It accepts and nets deals only after they have been rigorously matched and tested for counterparty limits. If deals are rejected, sometimes on a mere technicality but payment-versus-payment deals are not rejected both sides, including the innocent party, have to settle that day's deals bilaterally. Echo is seen by supporters as more user-friendly, but its critics argue that it is perhaps less robust. Echo initially accepts its members' deals. If there is a technical problem, counterparties are given a deadline of around two hours to put things straight. But if there is still a problem, for example if one bank has exceeded its exposure limit which is a soft limit, not a deal-breaking limit the offending counterparty has to trade down its position or put up margin. However, the trade goes through and the innocent party is not affected.
To achieve a merger of Multinet and Echo, one of these philosophies must be modified. It seems that the big US banks favour Multinet's more rigorous, less flexible risk-management approach. But Echo's advantage is that it is up and running, and it works. It is also planning to introduce hard limits by around April 1997, so that trades can be rejected for credit reasons. "We've taken on board some of the criticism of the soft limits," says Peter Davey, deputy chief executive at Echo.
If the big banks have their way, a merged multilateral netting system would also have to lose its clearing-house status. The big banks want their counterparties to remain their counterparties. JP Morgan, which was involved in early discussions with Multinet, pulled out when it foresaw that forward foreign exchange contracts netted in the Multinet clearing house would compete with its over-the-counter currency swap business.
Evolving a common private sector solution to reduce Herstatt risk may just be a risk manager's and central banker's dream.
Bob Blower, assistant director at the British Bankers Association, is sceptical whether there is enough incentive to see the project through. The FX business is in decline, volatility has been low, costs are under pressure. This is not a time for new spending with no palpable return. Blower believes a lot can be done without seeking fancy solutions. FXNet has shown what risk reduction simple bilateral netting can achieve. Bilateral netting agreements between banks, such as Ifema (the International Foreign Exchange Master Agreement) and the master agreement of Isda (the International Swaps & Derivatives Association), address replacement cost and close-out netting. They have reduced much of the incentive to create a multilateral foreign exchange clearing house. Citibank, originally one of the founders of Multinet, pulled out to focus on its own bilateral netting agreements. "It would have been difficult getting people to sign with us bilaterally if they knew we were going for a multilateral system which would take multiple years," says Sy Rosen, vice president for payment systems at Citibank. Citi is also heavily involved in FXNet as a founder member.
For most banks, risk management means putting their own house in order. The culture of risk awareness, driven by small shocks such as Barings, Daiwa, Sumitomo, Procter & Gamble, Metallgesellschaft, and led by central banks and a handful of global players, means that banks are under pressure to upgrade their systems and risk management just to compete. "We don't have to do any benefit-selling any more," says Dajit Saund, of risk management software vendor SunGard, "that's driven by regulators and crises. The second-tier and lower banks are scrambling like mad."
For purists, that is not enough. The banks at the forefront of the risk management culture want to identify and quantify all the risks they face, and apply a business cost.
Establishing a price
Herstatt risk, unfortunately, is not one of those Greek letters that appear in value-at-risk calculations. Quantifying it requires something more than projections of historical volatility. JP Morgan tried three years ago, presenting the New York Clearing House with a paper equating Herstatt risk with a one-year unsecured loan. Morgan's peer group were not impressed, complaining that a loan is a different asset class. One bank applies a Herstatt model internally, which has "opened a few eyes in the firm", says a risk manager. "We've run the model on a shadow basis," he says, "but we intend to charge [internally]. In effect, the FX business will pay." SBC Warburg has aggregated the time between irrevocable commitment of funds release and settlement reconciliation, the amount, and the probability of default (net of recovery), and allocated capital on a portfolio basis to the foreign exchange business. That gives a figure for the consequences of counterparty settlement failure, but it doesn't address the exposure to a systemic event, when the foreign exchange market goes into gridlock. "We do calculate concentration risk, by going through the same analysis, excluding the probability of default," says Bill Mundt, managing director for global settlement risk at SBC Warburg. "We also ask ourselves 'Would an expected default on behalf of one of our counterparties impair our own ability?' so that we're not the one to create the systemic risk."
Because systemic gridlock is such a low probability event, and there is only one historical case, a Black-Scholes model or other volatility calculation would not throw up a useful number. Moreover, how do you factor in the likelihood that a central bank will intervene and take the problem away? One risk manager says his bank has introduced a "fudge factor" in its systemic risk analysis "this is a guesstimate of the likelihood of getting some kind of relief", he says.
But the fudge factor may be diminishing. Central banks today are less intervention-minded than they were. "Central banking is becoming collateralized," says Blower at the BBA. "Central banks are giving fewer and fewer free overdrafts." It works both ways. The solution to the Barings crisis last year was not a quiet takeover by British banks, assisted by cheap liquidity from the Bank of England, as it might have been in earlier days. The market was told there is no safety net.
The main message of the G10 orange book was that the private sector must find its own solution to Herstatt risk. The dirigistes are not so happy with that conclusion. Researchers at the IMF have been working on a paper, to be published in mid-September, suggesting that national payment systems are a public good, and that the integration of real-time payment systems around the world should also be treated as such. Most major national payment systems can settle transactions on a real-time basis. This raises a question about the need for parallel private-sector systems. "It's a controversial point whether payment-versus-payment can work in the private sector, legally and practically," argues an IMF official. "Ultimately it's the central banks that provide the good funds."
Government-owned payment systems are beginning to make a charge for daylight overdrafts, which implicitly is a charge for Herstatt exposure. US Fedwire charges a flat fee. The Bank of England asks banks to collateralize daylight borrowing with repos (deposits of government securities). The Swiss National Bank charges an hourly overdraft fee.
These daylight charges implicitly reflect the risk that a bank will fail to repay or settle. But how should that risk be priced? The US Fedwire real-time payment system charges 55 cents a transaction, while Chips (Clearing House Interbank Payments System), which is private, charges 25 cents. Does that price differential recognize that Fedwire is charging for Herstatt risk, while Chips is getting a free ride, since the central bank implicitly stands behind it, were it to fail?
Should central banks set the price for being the ultimate insurers of Herstatt risk? Unfortunately, that elegant solution is the very one they wish to avoid. Even if they do stand between the world's financial system and perdition they don't want anyone to count on it.
Perhaps the private sector could establish a consensus price, as it has to some extent with options valuation, even if the model makes some debatable assumptions. Correspondent banks might be able to set charges that factor in Herstatt risk by counterparty, with a further levy for systemic risk.
The British Bankers Association (BBA) intends to run a round-table on this topic in the autumn. "Correspondent banks could make a market in settlement risk," says the BBA's Blower, "but this isn't like hurricane insurance, it's more like terrorism insurance." Others have compared the risk to that of meltdown in a nuclear power station. "The risk is so small, but potentially so devastating," says a central banker, "that you work to ensure it doesn't happen."
"Intellectually, [establishing a price] is the right way," says Hansen at Den Norske Bank, "but it would be complicated for the first bank to do it." Merrill Lynch and Bankers Trust already monitor and rate their counterparties for payment and settlement efficiency. But sanctions and charges against inefficient counterparties are difficult to implement.
On a counterparty-by-counterparty basis, Herstatt risk is a credit issue. "Whether it manifests itself as liquidity risk, settlement risk or insolvency risk is just a matter of timing," says Hansen. The credit risk can involve a bank or group of banks, a country, or a clearing house or payment system. The ultimate guarantor of a national payment system is the central bank. But does a multilateral payment system also need an ultimate guarantor? If the G20 implements its continuous link settlement system, is its planned group of private-sector liquidity providers enough of a defence?
These almost philosophical questions have not yet been thoroughly explored. Robert Heller, president of the International Payments Institute at Tiburon, California, has suggested that the IMF could step in, in moments of crisis, as a provider of settlement finality, using its stock of special drawing rights (SDRs). "I don't think of this as replacing current settlement schemes at all," says Heller. "I'm thinking of it in times of severe financial crisis." Citibank vice chairman Onno Ruding has supported Heller's suggestion. But IMF sources believe the SDR is not widely enough used to take on this role. Moreover, Heller's would be an unfashionable public-sector solution.
As the G20's continuous linked settlement (CLS) system is more accurately modelled, questions arise about the possible effect on domestic money markets, and the conduct of monetary policy. It will be a multicurrency payment system operating offshore, and outside national payment system hours. With extensive, large-volume payment-versus-payment, there are likely to be peaks in liquidity demand. Where will those be met: by central banks or by the CLS system's liquidity providers? What happens if they can't meet the demand. Will national central banks step in? Can central banks step in when their national payment system isn't open? The US national payment system Fedwire is planning to open from half-past midnight New York time, starting in 1997. That should mean a good overlap with Japan's settlement system which closes at 1.00am New York time. But this creates an anomaly for California: the next settlement day starts at 9.30pm the previous evening.
The CLS system has a provision that if liquidity is unavailable in one currency, an equivalent amount can be delivered in another currency, provided the counterparty has enough value in its collateral pool. Some experts regard this as a potential problem "migration of liquidity problems from one currency to another," says one. "A problem in dollars causes a problem in yen." But others are relaxed: "Can an offshore system create money? No," says Eric Sepkes, vice president of global transaction services at Citibank, London. "Money has already been created by the FX trade itself. It's a debate that must be had: but I think you'll find that all this results in payments in the national system." Nevertheless, the CLS system is being designed to ensure that using one currency to settle another only occurs intra-day, not overnight, to allay concerns by some central banks.
Such questions are getting close to the theory of finance: what is the nature of a financial system and the ideal relationship between banks and central banks? This is happening at a time when the European Monetary Institute (EMI) is wrestling with the design of Target, a payment system for the euro. Yet, so far, there has been little or no contact between the G20 and the EMI. "We're assuming that if an industry utility works with European currencies it will work with Emu [European monetary union] and Target," says David Roscoe, a managing director at JP Morgan. A Bank of England source points out that even with a real-time euro payments system "there won't be a single law, so you need to understand the insolvency laws of any state where the CLS operates".
A tenet of the G10's orange book, and of all attempts to mitigate Herstatt risk, appears to be that the way today's FX markets work should not be affected. But that is a rather pious hope. The innocent introduction of Swift 19 years ago revolutionized foreign exchange markets by speeding transaction times and increasing volumes in a way that was not foreseen.
A multinational system of payment-versus-payment is likely to produce another revolution. But it remains to be seen whether it will increase the velocity of transactions and widen the number of counterparties, or whether the strain on liquidity will slow velocity right down. Instinct suggests ways will be found to produce the former result.
Essential reading
Management of operational risk in foreign exchange, the New York Foreign Exchange Committee, April 1996
Settlement risk in foreign exchange transactions, Committee on payment and settlement systems of the central banks of the G10 countries (Allsop report or "orange book"), BIS, March 1996
Papers from a symposium on risk reduction in payments, clearance and settlement systems, sponsored by Goldman Sachs, January 1996
Risk reduction and enhanced efficiency in large-value payment systems: a private sector response, New York Clearing House Association, January 1995
Reducing foreign exchange settlement risk, New York Foreign Exchange Committee, October 1994
Central bank payment and settlement services with respect to cross-border and multi-currency transactions (Noel report), BIS, September 1993
Report of the committee on interbank netting schemes (Lamfalussy report), BIS, November 1990
Report on netting schemes (Angell report), BIS, February 1989
Close shaves
Bankhaus Herstatt, June 1974. The closure of this German bank at 10.30am New York time resulted in a classic chain-reaction of mistrust, as Herstatt's dollar FX payments failed to settle.
Continental Illinois, 1982 to 1984. This could have been the big one, but the realization of this big bank's insolvent position took two years. In the words of Ernie Patrikis, first vice president of the New York Federal Reserve: "If you're going to die, that's fine. But we want you to have a nice slow death."
Bank of New York, November 1985. Computer failure at the bank threatened to disrupt payments in the New York market. The New York Federal Reserve stepped in with a $23 billion discount loan.
Drexel Burnham Lambert, February 1990. Following Drexel's failing access to liquidity because of market rumours, the Bank of England acted as honest broker to settle payments between Drexel's London subsidiary and its foreign exchange and gold trading counterparties.
Bank of Credit & Commerce International, July 1991. A UK and a Japanese foreign exchange counterparty each lost the principal amount of deals with BCCI having failed to settle before the bank went into liquidation.
Invasion of Kuwait, August 1990. Fears about the survival of the Kuwaiti banks and the dinar threatened gridlock, averted by the prompt action of the Bank of England and other central banks.
Attempted coup in the Soviet Union, August 1991. Correspondent banks in deals with Soviet banks refused to release funds to settle contracts even though countervalue amounts had been received. Some Soviet banks faced liquidity problems, but gridlock did not spread.
Barings, February 1995. Barings just after its collapse was unable to cancel receipt of an Ecu amount, threatening to cause a knock-on effect throughout the Ecu clearing. system.
|