In the FT today,
Jean-Pierre Mustier, head of corporate and investment banking at SocGen, gave more details of the methods Mr Kerviel used to evade the banks’ controls.
“Every two or three days, he was changing his position. He would input a transaction that would trigger a control in three days and before that happened he would replace it with a different one,” said Mr Mustier.
He said the rogue trader was managing hundreds of thousands of concealed trades and an equal number of falsified hedges to give the appearance that any loss was offset.
Ok, so most banks are easily able to upload/import huge volumes of trades into the sales trading systems from spreadsheets and the like, simply because this is how most clients send in program trades. By keeping the value of each one relatively small, they may have kept below transaction limits. However, most traders have position limits that they have to keep within and his daily turnover would have been immense if he was rolling these positions every few days. If the control functions weren’t alerted by his activity, I’m surprised a desk head wasn’t chatting with a trader who was doing so much business.
From the reference to a “three day” control, it might be that SocGen operated a timetable such as
- Trade date – trade put on and call for margin made if insufficient funds on the account to cover the trades
- T+1 – Margin call due. Ordinarily risk teams will be monitoring sizeable owed balances and checking that funds are coming in. Likewise, middle office teams would normally be liaising with clients to ensure funds were received and applied
- T+2 – Missed margin calls from 2 day old trades escalated up the line, but positions rolled thereby pushing the ageing of the trades below a threshold and halting the investigation
Here’s where the “lone gunman” assertion looks very peculiar. It seems inconceivable that someone in one of the teams, particularly client services, had never sought to speak to the client especially one who never paid their margin calls. Whilst the trader would have been consulted, it would be highly irregular for these teams to accede to a request that all communications with a client were routed through a trader.
Kerviel would have also been sitting in a trading room surrounded by colleagues. Usually prop traders work in teams and so SocGen are effectively suggesting that Kerviel was conducting all these activities in full view of his colleagues. Admittedly a colleague sitting at a keyboard all day, playing with spreadsheets etc is completely normal behaviour for a trader.
In the Nick Leeson case, he was booking trades to an error/suspense account. Most banks have controls over opening accounts [documentation required, several internal authorisations] and classify accounts to which trades are booked as House, Exchange/Market and Client. One way round this is to make use of previously little used or dormant accounts. However, Suspense/Error accounts in each of these categories are normally closely inspected but even if they were only checked every few days, the volume of trades assigned by Kerviel to such accounts, if he was using this approach, would have been enormous. Moreover, the account balances would have been sizeable and warranted attention. I should acknowledge that, sadly, a consistent pattern of activity or balance can lull people into a comfort level that things are “normal”, whereas eractic fluctuations or activity draw suspicion.
Again, all the above is conjecture on my part, but the story coming out currently is very flimsy and it’s a risky PR strategy.