Six-sigma standard deviations are supposed to happen every 10,000 years, and yet they seem to be happening every month in the markets.

Such standard deviation moves can't be set aside as once-in-a-millennium occurrences, but while an institution may have the multi-dimensional models to evaluate them, if the system is decayed and senior personnel are focused on other career objectives, then it isnÆt going to work.

Jerome KervielÆs activities at Societe Generale were initially characterised as a ôfraudö, though we now know he didnÆt do it for personal gain. When he moved over from operations to fixed income, the bank never took away his access card. He was good at hacking computers; his accounts looked a bit suspicious from time to time, but he always had a good explanation.

Contrary to preliminary speculation, it was not the hedge funds that held the ticking time bomb of subprime exposure. It was, and is, the banks. What is more, their systems can't even count how much of it the own.

ôIn 2007 the hedge fund industry demonstrated conclusively that its risk management processes were much more effective than those of the banking system,ö says Peter Douglas of hedge fund consultant GFIA. ôThe global banking system wrote off $100 billion of shareholder value, while the global hedge fund industry, assuming it is about $2 trillion in size, and applying the HFR composite fund index return of 10.2%, created more than $200 billion of investor value.ö

What about Amaranth -- the hedge fund that blew up in late-2006? Yes, it was exposed but the blow up wasn't due to ignorance. The fund was aware of its positions, and so was the market. Transparency with Amaranth wasnÆt the problem. In fact, the transparency of the positions alerted the market that someone was making huge bets and so took opposing positions.

A hedge fund could certainly lose as much money trading, but it would be many times more difficult to hide. Why? It's not the systems, because banks have much better systems and risk control than hedge funds do. In Societe GeneraleÆs case, the key was that a trader was able to conceal a proprietary position by falsely calling it a client position.

Large, especially large-losing, bets require cash margin as they go along, and so the whole position cannot be hidden for long periods of time as the trader always needs someone to move cash to cover his bets. Since you cannot fake cash, you need to have a reason to move it, and the reason usually is that this is a position taken on behalf of a client who will ultimately reimburse the bank, or alternatively that the positions are meant to hedge an exposure the bank has to a client who is taking the opposite bet, and who will ultimately owe the bank the money they are spending.

Hedge funds have no clients in this fashion and so are not able to mistake a proprietary position for a client one. A bank's trader can conceal a position entirely, by trading at off-market rates which don't require large cash up-fronts, and then refraining from entering it into the systems.

ôThis is much more difficult for a hedge fund as their prime broker is going to shut them down or demand collateral if an unknown trade shows up, and the independent administrator is going to get involved as well,ö says a Hong Kong-based hedge-fund manager. ôIn addition, it's much less likely that a counterparty lets a hedge fund slide without margin issues than that let a bank. Everyone assumes that a major bank will be creditworthy, and that's where counterparties get sloppy.ö

Broadly speaking, hedge funds at least know what is going on in their books, whereas banks sometimes do not. For example, risk management people in the banks do not set limits on the amount of bond inventory or proprietary asset purchases by credit analysts who want them to curb their excesses. That duty is fiercely retained by traders who prefer not to be controlled. What the credit risk departments take a view on is the two-day broker settlement risk for buying those instruments from a broker.

So where to next? Is it conceivable that banks will learn from the mistakes of others? The complexity of products and systems is now such that you have to be automatically sceptical. For example, by signing up to Basel II, the directors of Japanese banks must state that they understand the capital-weighted status of their exposures. Should anyone believe them?

The balloon of epiphany has gone up again and shows no signs of coming back down.