瑞银丑闻真相：倒霉与流氓交易仅一线之隔

    到底什么样的人算“流氓”？我们到thefreedictionary.com网站查了查，第一个定义是“目无法纪、欺诈成性且靠不住的人”。根据这个定义，瑞银集团（UBS）雇员中符合这个描述的恐怕不止奎库•阿多博利一个。据指控，阿多博利的违规交易导致瑞银亏损了23亿美元左右。显然，他的举动之所以被归入“流氓”行为并不单单因为他亏损了20多亿美元，而是因为他亏掉的是银行自己的钱。

    你或许会觉得我们没有必要这样愤世嫉俗，且听我们解释：

    去年，世界证券交易所联合会（the World Federation of Exchanges）根据国别进行了一项针对同日确认（SDA）的调查。交易确认是一只毛茸茸的老鼠，蜗居在华尔街一个毫不起眼的角落，也就是运营与后台部门（Operations and Back Office）阴暗的墙角。人们现在才发现，这个部门时至今日还在靠回形针和橡皮筋工作，而高频交易员们使用的电脑则排成了行，足以使《太空堡垒卡拉狄加》（Battlestar Galactica）里的首席武器官垂涎三尺。

    这个行业盈利的部门，即销售与交易，霸占了所有的荣耀、所有风头和所有的资产。而交易处理部门几乎总是跟在后面望尘莫及。瑞银集团的这起丑闻却使负责交易后事务的部门进入了公众的视线。根据《金融时报》（the Financial Times）的报道，这位被指违规操作的瑞银交易员非常熟悉确认程序，了解“已经完成交易的确认程序……可以在交割完成后再进行”。公司可能在卖家向买家确认某笔交易之前就获得该交易的支付款项。尽管在卖家的证券交付之前取出买家的现金并非易事，但卖家仍可以在自己的账簿上把这笔现金记录为已经收到的款项，甚至可以将其用于进一步交易。

    《金融时报》解释称，这种风险在场外交易市场尤其普遍，而交易所开放式指数基金（ETF）、外汇期权和多种大宗商品衍生品的通常进行场外交易。货币市场工具也是如此交易的，万一价值几千亿美元的短期商业票据出现交割违约，结果将不堪设想，尽管我们的交易账户清单里仍会将这些票据称为“现金”余额，但其实这个数字非常具有误导性。

    降低此种风险的方法之一是要求双方在同一天确认交易，即使正式的交易确认文件（纸质也好，电子版也好）要到一两天后才能出具也无所谓。上述世界证券交易所联合会调查显示，日本、印度和香港的同日确认率最高，均超过90%；二线国家，即同日确认率80%出头的国家包括德国、法国、英国和中国；在全球领先的市场中，美国远远落在最后，只有46.9%的交易是在同一天确认的。或许有人会说，从实际交易数量来说，美国同日确认的交易量仍比其他市场多，可这个事实不容忽视：我们在全球最大的十个市场中排名垫底。巴西总统罗塞夫的拥趸想必会注意到，我们甚至比巴西还差了15个百分点。目前有人倡议，在全球三十个左右的国家协调交割惯例，尽管它们各自的监管体制、银行和市场惯例以及现行后台惯例各不相同。在此，我们只能祝他们一切顺利。

    瑞银违规交易丑闻曝光前不久，《金融时报》曾报道过被业内人士称为“交割违约”（fails to deliver）的现象。最近几个月来，交割违约现象大增，也就是交割日到了，买家却没有交付现金，或者卖家没有交付证券。《金融时报》数据显示，目前仅在美国市场，这种违约的规模就达到每天2,000亿美元。在美国，我们至少还知道这个数字呢，而欧洲根本就没有相应的数据。

为生存而违约

    部分观察人士认定，许多投行是蓄意违约，以此“应对财务压力”。在一段时间内，比方说两个会计周期之间或者两次部门审计之间，交割违约使一位交易员、一个交易柜台乃至整个金融机构得以在其账簿上同时纳入其已经出售但尚未交付的证券的价值，以及已经收录但尚未到账的现金的价值。根据惯例，保证金规则和交易系统通常允许某次出售证券的对价实际到账之前就投入用于买入新的证券；如果在太多下游交易交割之前，现金支付违约问题就能够解决，这个问题或许不会引起注意。否则就可能引起连锁反应，造成随后的一系列交易都必须中止，那么这家公司就不但要承担自己的亏损，还得承担所涉交易的对手方的亏损。看起来，这就是瑞银集团面临的情况。

    在美国，还有一种称为“未经清算”（ex clearing）的神奇所在，未能完成交割的交易可以进入此种永久性的中间状态。它们成了互为对手的经纪人之间的账外合同，双方都同意不再追究此事。完全没有任何办法查清按这种方式处理的未完成交割的交易到底有多大金额，因此也就无法弄清在银行的资产负债表上到底有多少钱属于永远不可能真正追回幽灵资本。

    《金融时报》报道称，直到2009年5月雷曼兄弟公司破产后，美国国债市场才引入了违约罚金制度。仅在当月，每日国债违约额就达到了5,690亿美元。美国现在计划在抵押担保证券交易中也引入违约罚金制。交易人士指出，此举可能会迫使投行将违约转入市场的其他门。《金融时报》称，ETF交易的违约率现在比普通证券交易还高。

    让我们回到流氓交易员一事上来。阿多博利在瑞银集团的证券部门工作，根据《华尔街日报》（the Wall Street Journal）的说法，他的职责是进行相当“安静的交易”。据报道，其亏损源于他使用未经对冲的指数期货，一边倒地对市场走势下注；。据报道，阿多博利还通过虚构的相抵消的交易来掩盖亏损。据称，他利用ETF创造了虚构的交易，因为ETF的交割周期比实际造成他亏损的金融工具更长，这些所谓交易的对手方是欧洲公司，后者所在市场的规则并不要求确认ETF交易。

    上述《华尔街日报》报道的评论意味深长，它指出瑞银的内部风险监控主要聚焦于固定收益证券的自营交易——因为瑞银在2008-2009年期间正是因该业务而减记了500亿美元资产，当时巨亏的部分原因在于太过依赖某一类证券。“瑞银对其风险控制体系的全面检查未能防止后台员工找出伪造交易的办法。”公平的将，如果某个雇员一心想要违反法律，他总会找出办法。可是，总的来说证券交易业务的监管力度不够，“因为它通常是为客户打理资产，风险较小”。亏掉客户20亿美元，可以称为一系列坏运气，亏掉银行自己的20亿美元，那就叫欺诈。

    瑞银首席执行官郭儒博已经为此辞职。阿多博利正关在英国监狱里，等待接受审判。与法兴银行的热罗姆•凯维埃尔不同，迄今似乎还没有迹象表明他会号称得到了上司的默许。我们不认为这件事上还会曝出什么重大的新变局。瑞银的风险控制很成问题。现在回过头来看，人人都会高呼：对雇员的欺诈行为不加防备简直是愚蠢至极。在瑞银的交易柜台上，交易权被交给了一个头脑发热、丧失理智的傻瓜；监管他的人显然也是傻瓜，他们的所谓管理程序就是每天打印海量交易记录，而不是真正去了解雇员的交易风格。有鉴于此，我们能给瑞银的最好的建议也只能是：以后别再雇傻瓜干活了。

    译者：小宇

    What exactly makes a person a "rogue"? We checked at thefreedictionary.com and found the first definition as "an unprincipled, deceitful, and unreliable person." Using this definition, we are not sure that Kweku Adoboli, the UBS employee alleged to have fumbled some $2.3 billion of the bank's money, would be the only employee on his desk to merit the characterization. Clearly, what makes Adoboli's actions "rogue" activity is not the fact that he allegedly lost over $2 billion, but that he lost the bank's own money.

    Lest you think us unnecessarily cynical, we offer the following:

    Last year the World Federation of Exchanges did a study of same-day affirmation (SDA) by country. Trade affirmation is a small furry rodent that inhabits the musty shadows of an un-sexy corner of Wall Street known as Operations and Back Office. This is the part that, as people are only now finding out, is held together by paper clips and rubber bands, while the high frequency traders run computer arrays that would make the chief weapons officer of Battlestar Galactica drool.

    The revenue generating part of the business – sales and trading – has glommed all the glamour, all the publicity, and all the assets. The transaction processing piece, meanwhile, has almost always had to play catch-up. The UBS scandal has brought the post-trade piece of the business to public attention. According to the Financial Times, the accused UBS (UBS) trader had intimate knowledge "of how confirmation that trades have been done… can happen after settlement." It is possible for a firm to receive payment for a trade before the seller has confirmed the transaction to the buyer. And while it is not so simple to take the buyer's cash out before seller's the securities have been delivered, the seller can show the cash on their own books as having been received, and can even spend it in further transactions.

    The FT explains that this risk is particularly prevalent in over the counter markets, where ETFs, foreign exchange options and a number of commodity derivatives trade. It is also where money market instruments trade, and we do not fancy the notion of a failure to deliver on a few hundred billion dollars' worth of short term commercial paper in what our brokerage statement misleadingly calls our "cash" balance.

    One way to mitigate this risk is to require both sides to affirm a trade same day, even if the formal trade confirm document – paper or electronic – is not issued until a day or two later. The industry group review found the highest rates of SDA were in Japan, India and Hong Kong, all well above 90%. A second tier – in the 80%-plus range – includes Germany, France, the UK and China. Among leading global markets, the U.S. ranked dead last, with only 46.9% of trades being same day affirmed. One could argue that this is still a larger number of actual trades than the other markets process, but the fact remains that we rank dead last among the ten leading markets in the world – fifteen percentage points behind Brazil, fans of President Rousseff will note. There is an initiative afoot to harmonize settlement practices among some thirty different nations, each with different regulatory regimes, bank and market conventions, and different current back-office practices. We wish them luck.

    Shortly before the UBS scandal broke the FT reported on a phenomenon known in the industry as fails to deliver. Fails – when the buyer does not deliver cash on settlement date, or the seller does not deliver the securities – have spiked in recent months and now average $200 billion a day in the U.S. market alone, according to the FT. At least in the U.S. the number is known. There is no equivalent data available for Europe.

Failing for survival

    Some observers are convinced that banks are deliberately failing "as a way of dealing with financial stress." For a period of time – say the time between accounting cycles, or between departmental audits – fails permit a trader, a trading desk, or even an entire institution to carry on its books both the value of the securities it has sold, but not delivered, and the value of the cash it has booked, but not received. Margin rules and trading system conventions generally allow the proceeds of a sale to be used for new purchases before the cash is actually delivered, so a fail on a cash payment may go unnoticed if it is resolved before too many downstream transactions settle. Or it can set off a cascade where successive trades have to be broken, with the firm eating both its own losses, and those of its counterparties on the trades in question. This appears to be roughly the case facing UBS.

    In the U.S., there is a further Never-Never-Land called "ex clearing," where aged fails go into permanent limbo. They become an off-book contract between the counterparty brokers, who agree not to pursue the matter. There is no way of knowing the dollar value of fails that have been thus dispatched, and consequently it is impossible to know what amounts of money on bank balance sheets are actually phantom capital that will never be recovered.

    The FT reports that only in May 2009, after the Lehman collapse, was a fail penalty introduced in the US Treasurys market. In that month alone daily Treasury fails reached $569 billion. The U.S. now plans to introduce fails charges for mortgage backed securities, which traders say will force fails into other segments of the market. The FT reports that ETFs are now more likely to fail than are normal equity trades.

    This brings us back to our rogue trader. Adoboli worked in UBS's equities division, where he was charged with running a rather "sedate trading book," according to the Wall Street Journal. The losses reportedly stemmed from one-way bets on the direction of the market, using unhedged index futures. The losses were reportedly masked with falsified entries showing offsetting trades. Adoboli reportedly created fictitious trades using ETFs, which settle over a longer cycle than the instruments on which he was actually losing money, recording them with European counterparties who were not obligated by market rules to confirm ETF trades.

    In a revealing comment, the Journal article said UBS's internal risk monitoring focuses on proprietary trading in fixed income – because that is where the bank took a $50 billion write-down in 2008-2009, partly because of heavy concentration in one class of securities. "The overhaul of its risk-control system didn't address the danger of a back-office employee finding a way to fake trades." Fair enough. If an employee is determined to break the law, he will figure out a way to do it. But the equities business is in general not closely monitored "because it had typically been a client business that carried less risk." Losing a client $2 billion is called a streak of bad luck. Losing the bank $2 billion is called fraud.

    UBS chief Oswald Grubel has resigned as head of UBS. Adobli sits in a British jail as he waits for his fate to be decided. Unlike SocGen's Jerome Kerviel, he does not yet appear to have taken the position that his superiors knew of his activities. We doubt there is another major shoe to drop in this matter. The bank did not have adequate controls. In retrospect everyone will clamor that it was idiotic not to anticipate employee fraud. On the UBS trading desk, a trading book was handed over to an idiot who got in well over his head, and who was apparently overseen by idiots whose risk management process consisted of initialing reams of printouts every day, rather than getting to know the trading style of their employees. The best advice we can give to UBS: don't hire any more idiots.