Tuesday, 23 September 2008

Systemic risk and social media

Presumably there's nobody who hasn't heard of the recent, er, problems in the finance industry. Whilst the wailing and gnashing of teeth are going on, presumably the FSA, SEC and other regulators are pondering how to ensure it 'never happens again'. Targets in the firing line seem to include bonuses and shorting, plus that rather vague term 'transparency' (or rather, opacity, since that's the bogey).

But is there anything much that regulators can do, without throwing the baby out with the bathwater (ie over-regulating and thus stifling enterprise)? And if so is there any connection with social software in the enterprise, thus justifying this post?

I think there might be, but it's not a nice thought. Let me step back a moment. The credit crunch and recent investment bank failures stem from the fact that a duff security - sub-prime mortgages - was wrapped up into opaque financial instruments such as CDOs (collateralised debt obligations). If I understand it correctly, this opacity made these instruments easier to sell than the underlying assets alone would have been, and it became harder for banks to know the true systemic risk they were running. When the sh*t hit the fan, ie the mortgages became worthless because of falling property prices and consequent defaults, not only were the CDOs etc devalued but putting a price on the devaluation became very hard, as did knowing which bank was exposed to what. So the problem was a combination of devaluation and ignorance. It was the ignorance as much as the loss of value that led to the credit crunch and the failures of Bear Stearns, Lehmans and (nearly) Merrill Lynch.

So where am I going with this? The regulators need to get a better, and earlier, hold on systemic risk in future, no matter what gives rise to it. This is not easy to do. Theoretically it can be done by external observation of economic indicators. It might have been possible to deduce that property prices were about to crash and that mortgage-related securities would go with them. But it's always hard to guess when a market has reached its peak, otherwise we'd all be rich. And probability is only one element of risk, the other being consequences or impact. The latter was probably very hard to measure, again because of opacity of the instruments. It all reminds me a bit of the problems at Lloyd's of London in the 1980s with the so-called LMX Spiral: risk accumulated through a chain of reinsurance contracts and no-one knew (until the s*it hit the fan) that it had ended up with a small number of syndicates, who were left holding the parcel when the music stopped.

But is it really true that no-one knows? I emphatically think not. The people who know are those doing the business. You can bet that there were people who knew how potentially toxic those CDOs and what-have-you were. I'm not talking about fraud here, although that did happen also. I mean people doing a relatively honest (by investment banking standards) job, who could see the risk, but had no incentive to do anything about it. In fact they had a big incentive not to - their bonus.

So how does a regulator find out about these risks? S/he needs to tap into what the traders are saying. Once upon a time they just said it in pubs and on street corners. Then, when email came along, some indiscreetly wrote things down, thinking it would remain private. But the regulators could order discovery of it as evidence when something went wrong and it appeared that rules had been broken.

Finally I'm getting to the point. Just as young people often blab about everything they are doing and thinking on social networking sites, then get embarrassed (and maybe risk their employment prospects) when they realise who might be reading it, we can perhaps expect such indiscretion inevitably to occur on social media within the enterprise, when its use becomes widespread. It might come to include traders talking about the latest securitisation wheeze, and how it's gonna be a great little earner for a couple of years until the sh*t hits the fan when x happens. Regulators would be very interested in seeing this material. We may even come to see it as their duty to obtain it. And perhaps to obtain it not just after the horse has bolted, but on an ongoing basis. Not a nice thought, but perhaps inevitable?