AIB's investigation of the $750 million (€864 million) foreign exchange loss at its Allfirst subsidiary is now focused on establishing whether there was collusion within the bank and/or by external parties with its alleged rogue trader, Mr John Rusnak.

Senior bank sources feel the trader must have had help from others to bypass the Allfirst internal control systems and rack up the losses involved.

With markets still reeling from the scale of the loss at a relatively small treasury operation, AIB still looking into the "how and why" and the trader stating he did not steal anything, there are more questions than answers about the debacle.

How were losses of this scale - about $3 million per trading day - accumulated by an ordinary trader without apparently triggering the Allfirst alarm bells?

Was he operating outside his daily and overnight deal limits and, if so, did anyone else in the division know this and help him conceal it or give him time or assistance to try to trade out of his losses?

Were the daily and overnight limits properly monitored? Who was reconciling contract confirmations with the details of the deals supplied by the trader? Who was reviewing documentation?

Were these functions properly separated from the dealing and settlement functions at Allfirst - segregation of duties is seen as critically important to reduce the potential for collusion between staff?

Did inadequate segregation or controls facilitate possible collusion between the trader, administration staff and some external parties, who could have provided the documentation required to confirm what turned out to be bogus options contracts?

Some sources speculated there could have been collusion with persons who had access to systems or documentation at the offices of the counterparties to the bogus options deals.

The counterparties to what were mostly Japanese yen trades are understood to have been Asian finance houses.

There must be major question marks over how one trader could have got a series of false option contracts up on the bank's systems without triggering any alarms.

Did Allfirst have controls specifying who were acceptable deal counterparties? Were these overridden and why was this not spotted by compliance staff at Allfirst?

Why did the bank's paper audit trails over a 12-month period not cast any doubt on the bogus options contracts?

What really happened at Allfirst will only emerge fully as the evidence is examined. Evidence will include recordings of dealing desk telephone conversations, deal documentation, which is being verified, including faxes and telex notes, and notes and conversations between treasury staff.

Treasury divisions are about risk - taking risks to generate profits for the bank and managing that risk.

They are operations in which the stakes are high - huge profits or losses can be generated from betting heavily on relatively small movements in the values of different currencies on international markets (foreign exchange and interest rates). For this reason and because the potential profits, involved can tempt traders to make money for themselves, controls are vital.

Treasury operations are structured broadly in three divisions:

dealing, where the traders doing the buying and selling operate;

deal processing, where staff process the documentation underlying and confirming the deals;

back-office settlement, which looks after invoicing and payments.

Clear separation of these functions is seen as crucial to making it difficult for bank staff to act together to defraud the bank or its clients.

Allfirst said Mr Rusnak used artificially created trades (options contracts) to mask his losses on real trades.

He was betting on yen against the dollar - buying yen betting that it would have risen in value by specified dates in the future (forward trades). When this did not happen and the yen fell, his contracts, which were due to be settled up to 360 days ahead, started to record losses.

On one side of his trading book, the trader had these real currency contracts on which in recent months his book balances were showing increasing losses. On the other side of his book, there appeared to be options contracts that would offset some of these losses.

But the recorded options contracts were bogus, so there was no cover for the crystalling losses and he was calling for more and more cash to settle these deals. This increasing demand for cash eventually triggered the alarms.

The failure of internal controls brought down Barings.

The latest failure of controls will not bring down AIB but the market has been shocked by the apparent ease with which an ordinary foreign exchange trader could rack up massive losses at a small subsidiary.