Is this global financial crisis just old stuff amplified or have we seen the burst of the first “rational bubble” in economics?

Three months and a half into the revelation of the size of the iceberg, the global financial crisis gives me a sense of “déjà vu” at a much larger scale (this is “all past crises combined” at a power of X), somehow mixed with a feeling of something new, “recursive”, like Borges’ non-linear, labyrinthic, fractal/Mandelbrotian novels, snapshots of many possible contingent pasts and futures: each with its crystal-clear logical unfolding path.

On the “déjà vu” side, we can confidently scale-up all the past ingredients/explanations of twin financial/currency crises, the inter-twined pro-cyclical macro-financial-prudential (rediscovered) Minskyian processes through which financial vulnerabilities are created and amplified, the need for Krugmanian trigger events (Lehman) setting self-fulfilling panic selling, the (reverse, non-monetary) Bernankian accelerated deleveraging process, in the midst of Stiglitzian informational imperfections, etc. Conventional wisdom, trying to make real and credit multipliers work again, learning from other proxy great crises (including the US Great Depression and Japan’s Lost Decade) posits an aggressive globally coordinated, large, lasting, expectations-boosting, fiscal-monetary counter-cyclical package (putting aside its global political economy difficulty and country specificities). All this combined with (urgent, decisive) fixing for the credit markets through recapitalization and ring-fencing bad assets.  OK, we all pray it will work.

On the “recursive” side, the feeling of being in a labyrinth comes from the existence of bifurcations (policy-path-dependent outcomes for 2009-2010), Saramago’s blindness (the shut down of any long-term planning horizon), fear of the intractable (uncertainty/dilemmas for policy action in both mature and developing economies, etc.) and trial and error behavior with U-turns (the TARP, the massive but ad hoc reliance on the Government’s balance sheet credibility).  OK, we pray it’s just a temporary hangover feeling.

Now how come that despite so many Roubinian warnings and the blatant evidence of (macro) global imbalances since 2005-2006, individual (micro) risk-taking behavior did not fundamentally change until (macro) aggregate leverage and asset prices reached excessive, proportions?  Obviously, the “bubble” literature pops-up, mostly in its “irrational exuberance” (Greenspanian, Kindlebergerian) version of non-rational behavior, the “unknown” on top of fundamentals that reconciles standard asset price models with rational expectations. But what if this crisis is a first example of a “rational” bubble despite its empirical rejection by the literature?  For that exotic equilibrium to hold, we would need bubble values of assets to equal the discounted, risk-adjusted, expected value of streams of (rising, fundamentals’ driven) income. Risk needs also to be estimated with “quasi-certainty” as bounded and computable.

The hypothesis submitted here is that the new micro financial tools, incentives, practices used throughout the last boom fueled that feeling, providing some elements for the possibility of a rational bubble.  In that case, rational agents need not to be receptive to macro data signaling potential un-sustainability and/or the likelihood of a Talebian “Black Swan” event. This feeling could explain the spreading of a (micro) perception of sustainable (macro) global imbalances above the level that otherwise would have triggered an automatic market correction (e.g., the external –not internal as in reality happened- predicted external financing limits for the US CAD that, when crossed for too long, would produce a typical “emerging market” confidence crisis in the US). Why did this not happen? Because of a (micro-macro) disconnect where improved fundamentals were explaining satisfactorily rapidly rising asset prices[1]. That resulted from the combination of the workings of asset/risk-pricing models used at an agent’s level, the financial industry’s institutional and incentives set-up and the confidence effect brought by risk diversification. We created for ourselves a rational feeling of “computable control” over the randomness of the future, a pseudo-certainty anchored on the assumption of constant volatility used in asset/risk pricing models. Extreme events did not need to be taken into account. Add to that the (macro prudential-regulatory) rules where portfolio diversification is good because risk was essentially uncorrelated, stationary and exogenous. In fact, it seems to me that the way the financial services industry has been working since financial globalization, tends to make risk more correlated and hence quickly systemic. In parallel, it might be argued that risk prices can be subject to extreme (over 5-6 sigma) events and, as a consequence of the techniques describe above, under-pricing may be endogenous.

Let us assume that observed macro-financial pro-cyclicality is an aggregated reflection of individual agents each acting under the underlying assumptions of the dynamic stochastic models for asset-pricing commonly used by the financial services industry.  Remember that 5-6-sigma events fall under an exponential convergence to zero probability of occurrence, under a Gaussian view of financial data.  This (dominant) view, was embedded in conventional wisdom/practice and risk-pricing tools, and contributed to boost financial industry’s outreach/ambitions. “Fat tail” distributions might have been considered by individual trading strategies but excluded at an aggregated strategic decision-making level. Why? Individuals are pre-disposed to look for data that confirms their priors, thus their empirics will tend to mimic their own assumptions. Thus, the assumption of a normal distribution remains (very) practical and attractive because it allows a nice tractability of asset-pricing models[2]. Add to that (1) the elegance (and Nobel recognition) of Modern Portfolio Theory, the beauty of the Black-Scholes-Merton equation giving you insights of an implicit pricing mechanism for stocks, options, etc. derived from observed data in actual trading; and (2), the homogeneity of the educational structure (e.g., teaching of finance in colleges, MBAs, etc.). You will get a financial technology that favors both the ability to expand the volume of “well-priced” complex securitized assets and the emergence of new products. This (endogenous) feature is possibly more important than (exogenous) deregulation and poor supervision.

Hence, our “technical confidence” boosted the “production function of financial services”: lower user cost (more computer power and broader availability) in the supply of specific financial services (tractable tools for pricing of risk, hedging, etc.) resulted in wider-than-otherwise dissemination of asset/risk pricing techniques across the industry and beyond. It may have been a significant contributing factor to the significant growth of assets under management, to the discovery of new risk mitigating products/structures and to the emergence of conduits to sell them. Contrary to common storylines, perhaps low savings in the US actually “caused” the need for more leverage and financial creativity (e.g., many forms of ABS, special funds/investment vehicles, etc.) more than the excess supply of non-US foreign savings. Given the available new financial technology, the intuitive traders’ implicit “Gaussian” feeling/culture in the financial services industry amplified balance sheet increases, hedging strategies, high leverage, etc. Accidents occurred (Jerome Kerviel) but were quickly dismissed as idiosyncratic not (potentially) systemic.  The LTCM crisis could be interpreted as a “containable” event. The success in solving it with traditional policy instruments enhanced confidence and possibly allowed the dissemination of the new financial technology to continue. There was no irrational behavior here but rational deepening of the financial markets.

There could be more endogeneity: institutions and incentives in financial services industry primarily in the US (including remuneration rules, under a highly competitive environment, less regulated with idiosyncratic agents like investment banks –before the crisis–, etc.) also contributed to produce excessive leverage. It is possible that the above market-average returns of many “initiators/pioneers” (e.g., investment banks, hedge-funds, etc. excluding –ex-post–Madoff’s type of scheme) created irresistible incentives for herd behavior in the whole industry. The business model and very short time-spam for rewards to these “initiators/pioneers” (short-term results yielding huge bonuses) contributed to eliminate “contrarians” (institutions and individuals). Infinite horizon is not a practical trading assumption. Thus, financial institutions could “prove” that the new analytical rigor in pricing assets/risk indeed delivered “high”, “predicted” returns for a long enough period of time.  The icing on the cake was the critical legitimizing/comfort role played by ratings agencies that amplified micro behavior through their own pro-cyclicality, allowing diversification and securitization techniques to price risk at a low (under-estimated) level[3].

Finally, uncorrelated and non-systemic risk is hard to reconcile with interconnected, integrated/globalized balance sheets of global financial services, and “intricate linkage” of assets/liabilities especially in the deep/liquid US market. The type of new financial products required extensive balance sheet diversification[4] to spread risk and special purpose vehicles. “Opacity” and “lack of transparency” in balance sheets were the logical result of widespread global financial securitization. Being a portfolio risk backed by pooled cash flows, the credit quality of securitized assets is by construction non-stationary due to possible changes in volatility that are time, and structure-dependent. Credit derivatives were also used to enhance the credit quality of the underlying portfolio to make it acceptable to final investors. The dominant (micro) behavior assumed that securitized products were properly structured, “insured”, and that the pool would always perform in a Gaussian world. However, under “extreme events” the affected tranches will experience dramatic credit deterioration and loss. We know the end of this movie: after hitting the iceberg, (i) a small “suspicion” of presence of any “toxic” assets in just one small market segment (sub-prime) translated into contagion of virtually all global financial institutions; (ii) when the “odd event” indeed occurred, the mood changed abruptly and global panic hit within a single trading session; (iii) amplifier-cyclical prudential/regulatory rules and practices (mark-to-market, CARs, etc.), led to (iv) the astonishing speed of global transmission; and (v) the still present standstill and/or contraction in international/local credit markets.

I am not sure of course that we have seen the burst of a rational bubble but if so, the policy implications to fix the crisis might be more complex.  Let us list some puzzles. First, turning around expectations and re-pricing assets seems to me easier when your tools and institutions are not plagued with a (now perceived) built-in bias to under-price risk. The observed volatility and downward over-shooting of so many asset prices seem to point to a long and pro-tracted period before market confidence is re-established, longer than if the feeling about the crisis was one of a simple “irrational” bubble. Second, we may have an institutional design problem in the financial services industry and its regulators: perhaps even careful supervision and regulatory enforcement (which did not happen) may not have had anticipated systemic risk by looking in detail at each individual/institution’s accounts and exposure. The systemic risk could only have been detected under a more global, fully transparent, comprehensive framework for registration and clearing of transactions. In its absence, uncertainty could remain unaddressed, again making the recovery more difficult. Under these circumstances, our traditional (national, monetary and fiscal) policy instruments might find unusual difficulties to fight deflationary pressures. Finally and third, there might be a longer than usual period to reset benchmarks (e.g., risk-free asset and interest rate). It looks like a crisis in the global reserve currency country and largest market of the global economy is obviously perceived as global. Hence, it causes a “flight to (relative) quality” to the crisis-country currency’s Treasuries, the current “safest heaven”. Irrespective of any long-term view of the US future fundamentals, US Treasuries are now the best “parking lot” of global liquidity with customers glad to pay just to be there. But since we know that this appreciation of the crisis-country currency is temporary, we can project that its “vacuum cleaner”, “black hole” effect on international capital markets is temporary.  However, not knowing its exact duration makes the re-establishment of financial trading benchmarks more difficult and subject to political economy factors regarding policy reactions in other currency areas.

Luiz Awazu Pereira da Silva

References:

Bernanke, B. S. “Nonmonetary effects of the financial crisis in the propagation of the Great Depression”, AER 73 (June 1983): 257-276

Bernanke, B. S. and Blinder A.S. “Credit, Money and Aggregate Demand”, AER 78 (May 1988): 435-439 Borges, J.L. “Ficciones”, Enencé Editores, Buenos Aires, (1956)

Krugman P., “A Model of Balance of Payments Crises” Journal of Money, Credit and Banking 11 (August): 311-25

Meltzer, A.H., “Rational and Irrational Bubbles” in “Asset Price Bubbles, The Implications for Monetary, Regulatory and International Policies” ed. By William C. Hunter, George G. Kaufman and Michael Pormerleano, MIT Press, (2003)

Minsky, H. “Stabilizing and unstable economy”, MacGraw Hill, (2008 ed.)

Saramago, “Ensaio sobre a Cegueira”Editora Caminho, Lisboa (1995)

Stiglitz J.E. and Weiss, A., “Credit Rationing in Markets with Imperfect Information”, AER 71 (June 1981) 393-410

Taleb, N.N., “The Black Swan”, Random House, (2007)


[1] Numerous studies by the US Commodity and Futures Trading Commission (CFTC) could not find robust evidence of a role for speculative activity (position changes by non-commercial traders) in the explanation of rising commodity (e.g., oil) prices, driven essentially by commercial activity observing the evolution of stocks, demand and supply.
[2] Practioners know that real world prices differ from the simplifying assumptions of this new financial technology, do not follow a strict stationary log-normal process and that the risk-free interest is unknown and varies. Pricing discrepancies between empirical data and models’ results have been observed in options, corresponding to the likelihood of extreme price changes. Trading strategies use that “feeling” for big pay-offs. Despite all that and although volatility “smiles” and is not constant, results from asset-pricing models are often used to set up strategies to minimize risk. Despite their inaccuracy they serve as a good pragmatic approximation.
[3] Confidence ran so high that basic common sense (e.g., approving NINJA loans) became old dispensable since even this class of risk could be priced and diversified
[4] The legal national limits to regulation/supervision created more incentives for cross-border, cross-institutional investment diversification and off-balance sheet operations.