The Silo Effect

by Gillian Tett

  They identified two main problems. One was the state of the bank’s corporate lending book, and its loans to potentially risky companies. A decade earlier, some of UBS’s rivals, such as Credit Suisse, had suffered big losses from the collapse of the Internet bubble, since that had caused a host of technology companies to default on their loans.43 As it happened, UBS itself had not been hit too badly during that particular episode. But the headquarters of Credit Suisse sat a few yards away on Zurich’s Bahnhofstrasse, in another sober gray building, and the UBS bankers were determined not to suffer the same fate. So in early 2007, the bank asked Ernst & Young to conduct an audit of all of UBS’s risky corporate loans.44 The results of the exercise were fairly reassuring: after looking at the corporate loans that UBS held on its books, Ernst & Young declared that the Swiss bank had managed its risks prudently. The bank had stayed out of most of the potentially dangerous leveraged buyouts and private equity loans in 2006, and where it was involved in making risky bets, it had charged a high fee that helped to compensate it for the danger.45 “We had a conservative stance toward leveraged finance,” Phil Lofts, a UBS banker who was running the credit risk section (the part of the bank that oversaw risky corporate loans), later recalled. “We were [only] in one or two deals that ran into trouble, like LyondellBasell [a deal to fund a chemical company merger].”

  The second source of concern lay with a hedge fund that UBS was running, called Dillon Read Capital Management. This had been created back in 2005, when some of the Wall Street traders that Costas and Ospel had hired at the start of their expansion campaign had become fed up with the conservative style of the bank and threatened to leave. To stop this, Costas persuaded the UBS management to create a dedicated hedge fund, and then gave up his job as head of UBS Investment Bank to run this unit.46 DRCM was free to take risky trades, but it was ring-fenced from the rest of the bank. This was initially done to ensure that DRCM could attract funds from outside investors without violating any regulations. However, ring-fencing the hedge fund had another benefit: many of the more conservative Swiss bankers hated the idea of using UBS money to take big proprietary bets, and wanted to keep the hedge fund at a distance. They liked to think of themselves as a sober, risk-averse bank.

  During 2006, DRCM expanded and its proprietary bets appeared to perform well.47 But by early 2007, DRCM was producing losses, partly because the hedge fund traders had misjudged the direction of the U.S. housing market.48 By the spring, the losses were estimated to be heading towards $300 million. The UBS managers were horrified. The losses confirmed all their worst fears about the dangers of proprietary trading.

  The UBS bankers conducted an audit of the hedge fund’s portfolio, and forced the fund to sell some of its assets. But these were so illiquid that when the fund tried to sell $100 million of specialized housing bonds, this created a $50 million loss in just one day.49 Bitter fights erupted between Costas and UBS management in the Zurich headquarters. The top managers knew it would be embarrassing to shut the fund down so quickly after creating it. Costas was a formidable name. But the UBS board could not stand the idea of suffering these losses. So in May 2007, after months of wrangling, the bank announced that it would close down DRCM.50 It was such a humiliating retreat that in the following month Peter Wuffli, the chairman of the board, was forced to resign. But the closure also underscored a crucial message that UBS wanted to project to the outside world (and itself): it considered itself risk-averse.

  Yet, even as the UBS board fretted about Costas’s hedge fund, and conducted endless surveys of its corporate loans, there was one issue that it did not debate: the mortgage-linked CDOs that were sitting in the CDO warehouse. By the spring of 2007, it was widely known that America’s long housing boom was coming to an end. House prices were no longer rising as quickly as before, and even falling in some states. There were reports of mounting subprime defaults. At some banks, this news had prompted the top management to clamp down on businesses that dealt with U.S. mortgage bonds. Just down the Bahnhofstrasse in Zurich, Credit Suisse was cutting its exposure to the market. Brady Dougan, the CEO of Credit Suisse Investment Bank, was a seasoned bond trader51 who had lived through violent swings of the credit cycle on Wall Street before, and he could sense that the financial climate was changing.

  At UBS, however, the top managers did not feel any need to act. That was partly because the men running the investment bank had less experience of bond markets. The man who replaced Costas as head of UBS Investment Bank, Huw Jenkins, had risen through the ranks of the equity department.52 But the other reason that the UBS leadership were more relaxed than Credit Suisse was that they did not think the bank was running any risk from its exposure to U.S. mortgage products at all.

  Back in 2005, when the bank had first started creating CDOs, its own internal risk managers had conducted a brief debate about how they should classify these new instruments in the bank’s accounts. Should they be placed in the category of “mortgages,” which were considered moderately risky? Or should they be simply treated as AAA assets, since they were considered ultra-safe? It was not an easy issue to resolve, since nobody had created a AAA bond with risky subprime mortgage loans before. The UBS risk managers were thus like botanists in a jungle, confronted with a new plant that did not fit into the existing taxonomies. But in the end, they simply chose to focus on the AAA tag, and placed these instruments in that bucket, in the accounts. That meant that the bank did not need to post a big reserve against possible losses if it held these assets on its books. But it also meant that the banks’ own risk managers tended to ignore the CDOs. When they tallied up all the assets that the bank held in internal reports to the board, the risk managers and auditors generally did not even split the CDOs into a separate category in the accounts, but simply lumped them together with other AAA assets, such as treasuries. Due to this system of classification, the assets in the CDO warehouse in New York had vanished in plain sight.

  This did not mean that the bankers entirely ignored mortgage risk. From time to time, regulators would visit the different teams in New York and London, and ask them about the bank’s exposure to the American housing market. The senior managers from UBS headquarters did the same. But the team that worked in London did not have much information about what was happening in New York, and vice versa; each team only knew about its own books.

  In London, for example, one team of UBS traders was trading mortgage bonds, and during the course of 2006 and early 2007 they placed big trading bets that would produce profits as American house prices fell. So when the regulators from Switzerland visited the London offices, the UBS risk officer told them that the bank was “short” housing risk. But the CDO operation in New York was “long” the market, because the warehouse was creating CDOs. The bank also had exposure to mortgage bonds as a result of trades it had made with a category of insurance companies known as “monolines.”53 The New York long exposures were dramatically bigger than the short positions that the London team held. But nobody normally tallied these positions up, or presented them to the UBS board that way. “There were many formal reports [in the bank] which sought to present a portfolio view of UBS’s risks, including reports that sought to capture real estate securities and loan exposure,” UBS later admitted in a report that it wrote for shareholders.54 “However, there was no comprehensive view available . . . due to incomplete data capture.” What was true on a micro level was not correct on a macro level.

  Inside UBS, there were plenty of people who could have spotted that discrepancy—if they had chosen to do so. The traders who were handling mortgage bonds in London, for example, could have asked hard questions about what was happening in the CDO warehouse. The financiers in New York who were running the CDO warehouse could have told the investment bankers in London to take a look at the swelling size of their books. But none of the teams had an incentive to share much information with each other. The CDO traders in New York did not want anybody in London meddling with their profit streams. Th
e traders in London did not have much incentive to speculate in public about what might be happening in New York, since that was unlikely to help them get paid. As Upton Sinclair the novelist once observed, “It is difficult to get a man to understand something, when his salary depends upon his not understanding it!”55 In theory, UBS liked to say that all these departments were joined together on a single matrix. In practice, though, the operations looked more akin to a collection of competing tribes.

  The risk officers were also fragmented. In theory, the bank had 3,000 risk officers who were supposed to monitor in a holistic way what the bankers were doing,. But these risk officers worked in three different departments that tracked different types of risk (credit risk, market risk, and operational risk56). These teams did not talk to each other very much. Nor did they swap information. That made a mockery of the idea of risk management, since the risk department was supposed to look at the dangers facing the entire bank. However, the bankers rarely noticed this; like every other part of the classification system, it was simply taken for granted. Generally, the bankers at UBS did not have any incentive to question it.

  In the spring of 2007, when the Swiss regulators went to London to conduct their usual examination, they asked the UBS risk managers to describe the bank’s exposure to the U.S. housing market. The risk officers talked about the shorts that the London team had taken out. They also spoke about some of the positions held by the Costas hedge fund. However, they did not mention the New York positions at all. “The super senior CDO positions . . . were not included in the risk reports,”57 the regulators later observed, adding that the “Chief Risk Officer of the Investment Bank was not [even] aware of the existence” of the super senior CDO warehouse.

  The regulators did not see any reason to query the picture presented by the chief risk officer, who seemed genuinely convinced that UBS was short housing market risk. So the regulators passed this message back to the UBS management, who were relieved that they had been given a clean bill of health.58 “If [those CDOs] had been included, the internal calculations would not have shown any short exposure at that time . . . [but] this incorrect assessment resulting from the incomplete data was also passed on to the [UBS] Corporate Center [in Zurich],” the regulators noted. “From this point on the bank’s management placed its trust in the supposed short positions and shifted its attention to other, seemingly bigger risks.”

  As the spring wore on, the managers in the granite building on Bahnhofstrasse kept fretting endlessly about the dangers of risky corporate loans. But they barely talked about mortgages at all. “The Group Senior Management was alert to general issues concerning the deteriorating US housing market [but] they did not demand a holistic presentation of UBS’s exposure to securities referencing US real estate assets,” the UBS shareholder report later admitted. “[This] was in contrast with the attention that Group Senior Management gave, for instance, to leveraged finance transactions which were subject to extensive debate.”59 It was a situation akin to the operators of a nuclear plant fretting endlessly about how to control the dangers arising from ultra-complex fission processes, while ignoring deadly cracks in the concrete walls of the building that were expanding under their very nose, hidden in plain sight.

  ON AUGUST 6, 2007, the UBS edifice started to crack.60 By then, the party in the global financial markets had turned sour. In the early summer of that year it had emerged that a fatal problem was besetting the business of securitization. This had boomed between 2002 and 2007, as bankers repackaged mortgage loans and corporate debt and sold these to new investors. But when the U.S. housing market started to deteriorate, some mortgage borrowers began to default, creating a contagious sense of fear. To return to the analogy of the butcher again, the financial markets faced the equivalent of a food poisoning scare: as investors realized that some bad mortgage loans (or rotten meat) might be sitting in the financial sausages that bankers had been selling into the market, they became worried about the health of complex products such as CDOs. Nobody could quite identify where those losses were likely to be sitting, because the packages of remixed loans were so complex. But as fear mounted, investors played safe and stopped buying all mortgage assets. Prices plunged and the markets froze.

  Initially, the UBS directors did not think that this panic would hurt their bank. After all, it was corporate loans, not mortgage loans, that UBS managers were worried about. “The board of directors and group executive board were convinced, up until the end of July 2007, that their investments in the subprime market were secure,”61 as Straumann, the Swiss professor who was asked to write a report on the UBS, debacle later observed. “All risk reports, as well as the internal and external audits, had arrived at the conclusion that UBS would be able to deal with declining [U.S.] real estate prices without any difficulty.” But when the board gathered for its regular meeting in Zurich in early August, they received a shock. They were told that though the traders in London had gone short the U.S. housing market, the bank had a CDO warehouse in New York that was holding more than $20 billion super senior CDOs.62 The board members were stunned. “Most of us had no idea what ‘super senior’ meant—we had never heard this word before,” one of the UBS directors confessed.63 Or as the regulators noted in a report: “Many senior managers claimed that they only found out about the super senior CDOs when the crisis broke in August 2007.”64

  Initially, the board members did not panic. After all, they reasoned, those bonds were called “super senior” because they were super safe. It was client business, not the type of proprietary activity that had caused such havoc at the hedge fund. “People were probably more worried about the leveraged loans,” Lofts observed. So after the UBS board finished its meeting in early August it issued a bland warning to shareholders that indicated there might be some modest future losses from the mortgage markets—and asked the top managers to investigate further.

  What this report revealed, however, was alarming: mortgage borrowers were defaulting on their debts on a large scale and scared investors were dumping bonds linked to those mortgages. That had caused the price of the super senior CDO tranches to plunge by 30 percent or more. That was alarming in itself. But what was doubly worrying was that UBS bankers had never made any preparations for this type of scenario. The problem, once again, lay with the classification system, and the UBS bankers’ failure to question it. Back in 2005, when the UBS bankers started buying those super senior instruments on a large scale, they had classified them as “marketable” instruments (or products traded in the markets) rather than “credit” or “banking book” assets (loans). The distinction sounded complex and theoretical. However, it had a practical implication: when instruments were classified as marketable, banks were not required to take a large reserve of capital onto its book. Thus UBS had never posted big reserves of capital against the risk that the CDOs might lose value. The models that the bankers use to measure the risks of these instruments suggested these should never lose more than 2 percent of their value. Thus UBS was only equipped to deal with losses up to that level. Now the market price of these CDOs had tumbled 30 percent, blowing a hole in the bank’s accounts.

  As the losses mounted, the UBS chief risk officer for the bank was unceremoniously fired. A new man, Joseph Scoby, who hailed from the asset management business, put in charge.65 He promptly set about changing the entire classification system. For the first time, CDOs were ripped out of the “safe” category in the bank’s accounts. Instead of being lumped together with treasuries and other AAA assets, they were moved into an accounting category of their own. Then the risk managers tallied up the different pieces of the bank’s exposure to mortgages for the first time. The picture shocked them. “We would sit in these meetings, and go: ‘Whaaaat the fuck?’ ” one risk officer from that period later explained. “People just could not believe it.” Almost overnight, the bankers realized that they had completely misunderstood the world: the “risky” hedge funds had actually been less dangerous than the “sa
fe” client business. The entire classification system had been upside down. “It suddenly hit me that we’d been worrying about that $300 million that [the hedge fund] dropped—but we had ignored ten times that in the CDO book!” another senior banker observed.

  Frantically, the UBS board scrambled to repair the damage. But the losses on the mortgage CDOs kept swelling as the price of the bonds dropped. First, the bank admitted to $10 billion in losses. Then, this swelled to $18.7 billion.66 By the spring of 2009 the bank had revealed a loss of more than $30 billion.67 As the pain spread the anger of politicians, regulators—and Swiss voters—intensified. From time to time, the UBS managers tried to point out that they were certainly not the only bankers suffering from these losses. Banks such as Citigroup and Merrill Lynch announced similar devastating write-offs.68 So did other institutions, ranging from insurance giant AIG to Allied Bank in Ireland. As the scandals spread, the same themes kept reemerging over and over again: almost anywhere you looked, banks, insurance companies, and asset managers had failed to spot the risks building up in separate desks and departments, because different silos of gigantic institutions did not communicate with each other and nobody at the top could see the entire picture. “All of these leadership flaws are virtually a hallmark of large banks,” Straumann observed in his report, noting that “Citigroup was compelled to undertake write-downs in even greater amounts” than UBS.69 But that UBS was in a widening club was not much comfort. Or not when the reputation of the bank lay in tatters. Nobody was surprised when UBS announced in the middle of 2009 that Ospel would resign. By then many of the other senior managers from the bank were gone too.70