Sornette’s Critical market crashes

D. Sornette Critical market crashes 2003


Positive feedbacks provide the fuel for the development of speculative bubbles, preparing the instability for a major crash. … The most important message is the discovery of robust and universal signatures of the approach to crashes. … [We] describe the empirical evidence of the universal nature of the critical log-periodic precursory signature of crashes. …

2. Financial crashes: what, how, why and when?

Sornette eloquently builds on the insights of Keynes, firstly on the type of ‘cause’ of bubbles:

Most approaches to explain crashes search for possible mechanisms or effects that operate at very short time scales (hours, days or weeks at most). We propose here a radically different view: the underlying cause of the crash must be searched months and years before it, in the progressive increasing build-up of market cooperativity or effective interactions between investors, often translated into accelerating ascent of the market price (the bubble). According to this “critical” point of view … a crash occurs because the market has entered an unstable phase and any small disturbance or process may have triggered the instability.

4. Positive feedbacks

4.2. It is optimal to imitate when lacking information

On the actual cause:

Keynes (1936) argued that stock prices are not only determined by the firm’s fundamental value, but, in addition, mass psychology and investors’ expectations influence financial markets significantly. It was his opinion that professional investors prefer to devote their energy, not to estimating fundamental values but rather, to analyzing how the crowd of investors is likely to behave in the future. As a result, he said, most persons are largely concerned, not with making superior long-term forecasts of the probable yield of an investment over its whole life but, with foreseeing changes in the conventional basis of valuation a short time ahead of the general public.

5. Modelling financial bubbles and market crashes

5.1. The risk-driven model

On the risk of collapse:

Nowhere is Keynes’s … analogy more relevant than in the characterization of the crash hazard rate, because the survival of the bubble rests on the overall confidence of investors in the market bullish trend.

5.4. Imitation and contrarian behavior: hyperbolic bubbles, crashes and chaos

Sornette develops a specific model in which the crash is triggered by contrarianism:

The specific feature of the model is to combine these two Keynesian aspects of speculation and enterprise 

The model of imitative and contrarian behavior leads to accelerating bubble prices following finite-time singularity trajectories aborting into a crash. The accelerating phase is due to imitation. The crash is due to the contrarian behavior reinforced later by the imitation behavior.

8. Synthesis

Sornette’s conclusions:

8.1. “Emergent” behavior of the stock market

In this paper, we have synthesized a large body of evidence in favor of the hypothesis that large stock market crashes are analogous to critical points studied in the statistical physics community in relation to magnetism, melting, and so on. Our main assumption is the existence of a cooperative behavior of traders imitating each other … . A general result of the theory is the existence of log-periodic structures decorating the time evolution of the system. The main point is that the market anticipates the crash in a subtle self-organized and cooperative fashion, hence releasing precursory “fingerprints” observable in the stock market prices.

Instead of the usual interpretation of the efficient market hypothesis in which traders extract and incorporate consciously (by their action) all information contained in the market prices, we propose that the market as a whole can exhibit an “emergent” behavior not shared by any of its constituents.

8.2. Implications for mitigations of crises

The mantra of the free-market purists requiring that markets should be totally free may not always be the best solution, because it overlooks two key problems: (1) the tendency of investors to develop strategies that may destabilize markets in a fundamental way and (2) the noninstantaneous adjustment of possible imbalance between countries. Financier George Soros has argued that real world international financial markets are inherently volatile and unstable since “market participants are trying to discount a future that is itself shaped by market expectations”.

…  [R]ecovery after an endogeneous shock is in general slower at early times and can be at long times either slower or faster than after an exogeneous perturbation.

We believe however that increased awareness of the potential for market instabilities, offered in particular by our approach, will help in constructing a more stable and efficient stock market.

My Comments

Sornette’s conclusions seem quite hopeful. Firstly if the collapse of bubbles is really due to contrarian investments then one would expect “increased awareness of the potential for market instabilities” – however arrived at – to lead to increasing pessimism about speculative investments and hence smaller bubbles, leading to a closer approximation to the long-term trend of sustained exponential growth. If Keynes’ insights are not enough, then Sornette’s should help.

Sornette’s approach to spotting bubbles is based on “precursory “fingerprints” observable in the stock market prices.” This seems to be more a product of his general view of life as statistics than any considered analysis. One contrarian approach is to try to exploit opportunities that most fund managers have missed or discounted, to bail out as the main fund managers get in, and to hedge against when ‘the man on the Clapham omnibus’ is talking about this sure-thing investment. It would seem better to have some such contrarian approach informed by Sornette type analysis.

In Keynes’ general model a crash can occur when some limit is being neared, whatever it is. Sornette uses this as a step towards his model (E.g. a Sornette video . ) But if contrarianism is the cause of collapses, then instead of watching for signs of excessive speculation, as above, shouldn’t we be watching for signs of excessive contrarianism, such as when Soros is hedging? Then our strategy would simply be to imitate contrarians when they become significant. But it would seem prudent to look out for other limits, such as when governments are seen to be failing, with too much austerity or too little growth.

While Sornette goes beyond Keynes in developing his specific model, Keynes broader insights are still relevant. For example:

  • If Sornette’s model is correct and predictive and financiers understood it, wouldn’t they behave differently, thus invalidating the model?
  • The complex mechanisms and behaviours in the model are incompatible with conventional notions of probability and information, so that if there is no quick fix it may be necessary to develop reformed concepts.
  • The issue addressed by Sornette is one of mid-term behaviour: short-term and long-term behaviours are straightforward. But theories of decision-making tend to consider either immediate or long-run decisions, not mid-term. Some theoretical innovation may be needed.
  • Sornette does not consider how changes aimed at limiting bubbles might impact on long-term growth. But (in video) he recognises that many railways might not have been built had it not been for ‘irrational’ speculation. Bubbles can leave legacies. Similarly, crashes remove ‘dead-wood’ and revitalise real economies: if we reduce bubbles will economies have to carry more dead-weight? Could this lead to a crash of a different kind?
  • Sornette seems to assume that the current cycle of bubbles and crashes is ‘a bad thing’ and should and can be avoided. But it is not clear that attempts to avoid crashes would necesarily improve the situation.
  • In Sornette’s model recoveries always follow crashes, albeit slowly. But maybe they don’t.

It seem to me that Sornette’s model is excellent as a counter to the efficient market hypothesis and provides some good inputs to both investment decision-making and policy-making.

[I, DJM] believe … that increased awareness of the potential for market instabilities [including, but not limited to, Sornette’s approach] will help in constructing a more stable and efficient stock market.

Sornette’s work should be a valued input, but so simple a model could hardly be sufficient for so broad a problem and, like Keynes, I doubt that any model could. Instead, we may need a broader understanding.

My criticisms above are not really about the paper considered as such – with due (scientific) common-sense – but with how I would expect less thoughtful (busy?) readers to take it.

See Also

Whitehead and Smuts (colleagues of Keynes’).

Dave Marsay

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: