Saturday 21 Jul 2018 | 19:37 | SYDNEY
Saturday 21 Jul 2018 | 19:37 | SYDNEY

Reader riposte: Still more on complex systems


Sam Roggeveen


18 November 2008 13:22

Will Clegg writes on how to regulate a financial system that Ken Henry described as 'so complex it defies understanding':

It would appear that 'emergent behaviours' (inherently unpredictable behaviour emerging from within very complex systems) are often the source of systemic volatility in financial markets. The mechanisms which translated an increased default rate in the sub-prime mortgage market into a systemic global liquidity and credit crisis, able to undermine international aggregate demand and the solvency of sovereign entities, were not predicted by people capable of (1) constructing a compelling analysis and (2) gaining adequate official attention for their views.

Similarly, the implications of any particular set of rules that might have prevented the sub-prime crisis cannot easily be estimated. The very complexity of our financial system, and the disaggregation of power within it, is one of the key reasons global leaders are emerging from the G20 meeting with such modest reform agendas. Will new rules produce public failures? What will their effects be on allocative efficiency? Will they generate new, disruptive incentives for regulatory arbitrage? Far easier, it would seem, for governments to defend the status quo and underpin an inherently volatile system with state guarantees.

It is the unpredictable nature of this present crisis that should be remembered by the leaders returning from the G20 summit. This is because it is an important reminder of the limitations of regulation in highly liquid, globally integrated, over the counter capital and money markets. Regulation must be precise to be effective (never mind tolerable to politically influential groups of financial entrepreneurs).

But how can one be precise in specifying and regulating that which one cannot plausibly predict? Further, how can one convince society and financial market actors that the costs of a particular risk management technique are not outweighed by the outcome of the risk eventuating, if the risk cannot be precisely or plausibly specified? Perhaps we could make our financial systems simpler. But, would we be better off for it?

A further note: Chris Skinner, who wrote to us on this issue yesterday, also sends along the following links about complexity theory and emergent behaviour. Thanks Chris: