risk controls at UBS: the case of trader Kweku Adoboli

Kweku Adoboli is the UBS trader who ran up losses of $3.2 billion through unauthorized trading in stock index futures over a three-month period without being discovered.  Both the Financial Times and the Wall Street Journal have extensive accounts of what Mr. Adoboli did.

Here are my observations:

background

Legally, traders act as agents for the institution they work for.  Once an employer introduces an employee to counterparties as being authorized to trade for the firm, the counterparties have no obligation to try to figure out what the trader is doing.  Until the employer informs them otherwise, the counterparty’s job is simply to execute the orders they receive.

Mr. Adoboli worked on a small trading desk called Delta One, that processed buy and sell orders that UBS received for ETFs.  For this story, the most important characterisitics of ETFs (see my posts on ETFs vs mutual funds for more information) are that:

–ETFs trade continuously throughout the day, in large aggregate amounts but typically in many small orders

–firms that run ETFs have no direct dealings with the investing public.  They keep their costs low by having brokers do virtually all trade processing and record keeping for them.

Brokers recoup the administrative expenses they incur through the commissions and bid-asked spreads they charge customers.  Once they amass a large net position in a given ETF, they can close their exposure out by transacting with the firm that runs the ETF.  They may also attempt to make additional gains through the timing of these transactions.

Brokers routinely hedge part or all of their ETF exposure through derivatives markets.  The name of Mr. Adoboli’s unit, Delta One, signifies that the trading desk “delta,” or the change in value of the hedges for a given change in the underlying position UBS held, should be “one.”  That is, the two should match exactly; there should be no net exposure.

Mr. Adoboli

Mr. Adoboli’s initial job at UBS appears to have been in the back office, as one of the administrative employees processing and recording the activities of the Delta One desk.  One of his unit’s jobs would have been to reconcile the desk’s accounts of the trades it made each day with the confirmation notices sent by counterparties. 

Mr. Adoboli was a good enough employee to be promoted to the much higher status job of trader.  One key fact that he learned from his back office time was, surprisingly to  me, that for a whole class of plain vanilla short-term derivative contracts, counterparty banks never sent confirmations on the day of the trade.  Apparently, standard procedure was to only to send settlement instructions a few days before the contract came due.

on the Delta One desk

Despite the name, the Delta One desk had to take risk.  And, from Mr. Adoboli’s behavior we can conclude that the desk rewarded traders for successfully taking risk.  But these risks would have been small, like:

–widening the bid-asked spread slightly, or

–delaying making a hedging transcation by five or ten minutes in hopes of getting a higher price, or maybe even

–by “anticipatory hedging,” over-hedging at a favorable price, figuring that new orders would soon come in.

three months ago

That’s when Mr. Adoboli exceeded the risk limits specified by his desk.  Who knows what happened?  He may have accidentally added an extra zero to a trade.  More likely, he may have decided he wanted to quickly make enough trading profit to get a higher bonus, or to be recognized as an astute trader and promoted to a “prop trading” desk whose principal job is to try to make trading profit (“prop” is short for proprietary, meaning it trades with the firm’s own money).

In any event, at some point Mr. Adoboli’s trading went badly and he began to make substantial losses.  Rather than reporting what he’d done to his boss, he used his back office knowledge to record fake trades that offset his losses.  He selected instruments where he knew no confirmations would be sent–buying him time until close to settlement day for him to recoup his losses and enter more, counterbalancing, fake trades to erase them from the records.

Apparently, toward the end, Mr. Adoboli was making speculative trades covering as much as $5 billion in securities, all without being detected.

What appears to have tripped Mr. Adoboli up was that the back office noticed it was not receiving settlement instructions for fake trades set to settle on September 22nd.

observations

In the mid-1980s as I was beginning to learn about bank stocks, a colleague who was an excellent bank analyst told me she had one main criterion for separating good banks from bad.  In a good bank, when someone makes a mistake and reports it, he’s rewarded; in a bad bank, mistakes are punished, so employees hide them.

It’s hard for me to believe that Mr. Adoboli was able to conceal his unauthorized trading from his direct supervisor–in a five- or six-person section–for so long.  That person must have been asleep at the switch.

It’s also surprising that there was such an unaddressed loophole in UBS’s trade reconciliation procedures–and that no one noticed that one person was doing so much unreconciled trading.