Predicting The Unthinkable In Financial Markets

The nuclear strategist Herman Kahn coined the phrase ‘thinking about the unthinkable’ in a series of black comic Air Force briefings that became On Thermonuclear War (Princeton University Press, 1960).  Faced with a year-long crisis in US credit markets analysts have embraced similar imagery in their forecasts of catastrophic risk.

Several different players in the financial ecosystem rely on the forecasts for multiple payoffs, one for their target audience and the other for themselves:

  • Research Analysts: (1) Provide clients with guidance and metrics to the market turbulence; (2) stand out in the pecking order of research firms and competing industry/sectoral analysts to remain relevant.
  • Investment Media: (1) Catastrophes as the source of drama and headlines to keep consumers engaged; (2) Financial and operational synergies of convergent media production.
  • Fund Managers: (1) An external input to valuation models for visiting potential firms to invest in; (2) A parameter for deciding on the asset classes, diversification and hedging for investment portfolios.

Some questions to ask in evaluating any catastrophic forecasts that predict the unthinkable:

  • What is the source, type and timeframe of the evidence presented?  The source may be company interviews, earnings calls, investment calls and trade seminars.  The type may be firsthand observation, market rumour, financial model, computer simulation or analyst conjecture.  The timeframe may be historical simulation of past data, quarterly forecasts or a longer time horizon for capital financing, global market entry, innovation pipelines or sustainability projects.  The source enables you to filter any possible agendas, the type refers to the information structure, whilst the timeframe often has embedded assumptions about cause-effect relationships, impacts, and the actions of others.
  • Why is the analyst making this forecast and could there be other agendas? Analysts have biases and personal theories that an attention economy might amplify.  At a group level this becomes self-reinforcing collective wisdom that may turn out to be flawed.  In embracing a current meme in a true believer stance analysts create a cognitive frame prevents them from considering alternative outcomes, options and possibilities.  At its most cynical this question is a reminder that forecasts are not objective or value-neutral, especially if the analyst is under pressure to generate earnings revenue or has a different private opinion to their public view.
  • What is the analyst’s track record in accurate forecasting?  This focuses on the analyst’s patterns of thinking and rhetoric in forecasts; how their performance compares to an industry, market or sectoral baseline; and the margin of error in their past forecasts.  This can be used to construct a brains syndicate, to filter out media reports and noise, to surface hidden assumptions and how they affect performance, and as a quality assurance check.
  • How might the catastrophic forecast be hedged? This shifts the focus from optimistic versus pessimistic views to the risk management focus on mitigative strategies and action planning.  To be effective, this requires an understanding of your risk profile and risk-return needs (risk averse, neutral or seeking), your time horizon, and the nature of the financial instruments, investment portfolio and markets to be used.

Subprime Winners: Rational Herds & Decision Researchers

US capital and derivatives markets in mid-2008 provide a real-time laboratory for behavioural finance analysts who want to understand the madness and wisdom of crowds.  The past week’s case studies include the implosion of the US bank IndyMac and the market volatility triggered by fears that Fannie Mae & Freddie Mac are highly exposed to liquidity risk.

As financial reporter Michael S. Rosenwald notes in The New York Times, these recent events appear to fit the behavioural finance hypothesis that individual investors who make fear-driven and risk-averse decisions can trigger pricing shifts as an aggregate rational herd.  Guillermo A. Calvo and Enrique Mendoza found in a 1997 paper that globalisation counteracts the emergence of rumour markets based on imperfect information and country-specific knowledge, although not in emerging markets due to uncertainties.

However the recent events have different conditions that set delimits on Calvo and Mendoza’s model: the United States is the epicentre of the bear market triggered by the 2007 subprime crisis, Fannie Mae and Freddie Mac have psychological primacy as major financial institutions with US Federal Government backing, and investment media firms such as Bloomberg and CNBC use globalisation to create de facto rumour markets amongst day-traders and others.

Readers interested in rational herds should also check out Christopher P. Chamley’s book Rational Herds: Economic Models of Social Learning (Cambridge University Press, Cambridge UK, 2004), excerpt here.

Decision researchers are the other early winners of the 2007 subprime crisis, due to the failure of many quantitative models to predict the Black Swan event.  Rosenwald mentions Harvard University’s new Bio-Behavioral Laboratory for Decision Science which conducts ‘conducts research on the mechanisms through which emotional and social factors influence judgment and decision making.’  He also refers to the Oregon-based nonprofit group Decision Research.  An Australian-based counterpart might be the Capital Markets CRC, an R&D consortia that focuses on ‘new technologies and improvements in market design’.

Investment analysts still have divergent opinions on recent events.  However the research agenda above prompts several new questions:  What happens to rational herds and rumour markets when bio-behavioural methods of decision-making are no longer ‘imperfect information’ but are widely understood and integrated into investment choices?  How will markets be redesigned to cope with this eventuality, and who will take on this responsibility?  What new financial instruments, markets and products will emerge generativity?

Errors In Quantitative Models & Forecasting

Could the roots of the 2007 subprime crisis in collateralised debt obligations (CDOs) and residential mortgage-backed securities (RMBS) lie in financial analysts who all used similar assumptions and forecasts in their quantitative models?

Barron’s Bill Alpert argues so
, pointing to a shift of investment styles after the 2000 dotcom crash from sector-specific, momentum and growth stocks to value investing.  Investment managers who prefer the value approach then constructed their portfolios with ‘stocks that were cheap relative to their book value.’  In other words, the value investors exploited several factors — the gaps in asset valuation, asymmetries in public and private information sources, price discovery mechanisms and market participants — which contributed to mispriced stocks compared to their true value.

However, the value investing strategy had a blindspot: many of the stocks selected for investment portfolios also had a high exposure to credit and default risk.  The 2007 subprime crisis exposed this blindspot, which adversely affected value investors whose portfolios had stocks with a high degree of positive covariance.

Alpert quotes hedge fund manager Rick Bookstaber who believes that financial engineers have accelerated crises and systemic risks via the complex dynamics of new futures contracts, exotic options and swaps.  These new financial instruments create interlocking markets (capital, commodities, debt, equity, treasuries) which have the second-order effects of larger yield curve spreads and trading volatility.  Alpert and Bookstaber’s views echo Susan Strange‘s warnings a decade ago of ‘casino capitalism’  and ‘mad money’ as unconstrained forces in the international political economy.

Quantitative models also failed to foresee the 2007 subprime crisis due to excessive leverage, difficulties to achieve ‘alpha’ or above-market returns in market volatility, and the separation of risk management from the modelling process and testing.  Other commentators have raised the first two errors, which have led to changes in portfolio construction and market monitoring.  Nassim Nicholas Taleb has built a second career on the third error, with his Black Swan conjecture of high-impact events, randomness and uncertainty (see Taleb’s Long Now Foundation lecture The Future Has Always Been Crazier Than We Thought).

Alpert hints that these three errors may lead to several outcomes: (1) a new ‘arms race’ between investment managers to find the new ‘factors’ in order to construct resilient investment portfolios; (2) the integration of Taleb’s second-order creative thinking and risk management in the construction of financial models, in new companies and markets such as George Friedman’s risk boutique Stratfor; and (3) a new ‘best of breed’ manager who can make investment decisions in a global and macroeconomic environment of correlated and integrated financial markets.