Greenspan does not do it with models

Or maybe he does. In any case this speech (circa 2008) is a wonderful clarification of what we can and cannot do with models:

The essential problem is that our models – both risk models and econometric models – as complex as they have become, are still too simple to capture the full array of governing variables that drive global economic reality. A model, of necessity, is an abstraction from the full detail of the real world. ... The most credible explanation of why risk management based on state-of-the-art statistical models can perform so poorly is that the underlying data used to estimate a model’s structure are drawn generally from both periods of euphoria and periods of fear, that is, from regimes with importantly different dynamics. ... The contraction phase of credit and business cycles, driven by fear, have historically been far shorter and far more abrupt than the expansion phase, which is driven by a slow but cumulative build-up of euphoria. ... Negative correlations among asset classes, so evident during an expansion, can collapse as all asset prices fall together, undermining the strategy of improving risk/reward trade-offs through diversification.

But these models do not fully capture what I believe has been, to date, only a peripheral addendum to business-cycle and financial modelling – the innate human responses that result in swings between euphoria and fear that repeat themselves generation after generation with little evidence of a learning curve. Asset-price bubbles build and burst today as they have since the early 18th century, when modern competitive markets evolved. To be sure, we tend to label such behavioural responses as non-rational. But forecasters’ concerns should be not whether human response is rational or irrational, only that it is observable and systematic.

This, to me, is the large missing “explanatory variable” in both risk-management and macroeconometric models. Current practice is to introduce notions of “animal spirits”, as John Maynard Keynes put it, through “add factors”. That is, we arbitrarily change the outcome of our model’s equations. Add-factoring, however, is an implicit recognition that models, as we currently employ them, are structurally deficient; it does not sufficiently address the problem of the missing variable.

We will never be able to anticipate all discontinuities in financial markets. Discontinuities are, of necessity, a surprise. Anticipated events are arbitraged away. But if, as I strongly suspect, periods of euphoria are very difficult to suppress as they build, they will not collapse until the speculative fever breaks on its own. Paradoxically, to the extent risk management succeeds in identifying such episodes, it can prolong and enlarge the period of euphoria. But risk management can never reach perfection. It will eventually fail and a disturbing reality will be laid bare, prompting an unexpected and sharp discontinuous response.