In the wake of the 2008 financial crisis several financial theories have come under scrutiny. The contributions of these models are important, but the crisis has shed light on their pervasive misuse in both finance and economics and has given us evidence of the detrimental impact that can occur as a result.
In Lecturing Birds on Flying: Can Mathematical Theories Destroy the Financial Markets? Pablo Triana fully critiques the development of financial models, their creators, the implementation of such models in practice, and the negative impact they can have on financial markets. He fearlessly asserts strong opinions questioning whether these models do more harm than good and proposes the complete rejection of financial models moving forward.
The book begins with a discussion of whether financial models should have been developed in the first place. Triana postulates that finance cannot be quantified and wonders to what extent we have caused damage to our financial system by assuming markets can be modeled and attempting to do so. Triana examines this question in three parts: Essentials, Critique, and Conclusions. In Essentials, he discusses the difficulty in mathematically defining financial markets, gives a harsh critique of the evolution of financial economics, and explains the rise of and need for ‘quants’ associated with financial innovations. He then provides a more detailed analysis of how the attempt to measure and quantify risk has created havoc in the markets. The Critique section focuses on the impact of three theories: the Gaussian copula with regard to dynamic hedging, the Value at Risk (VaR) measure during the financial crisis, and the effect of the Black-Scholes (BS) option pricing model. The book ends with a discussion of society's need for precision and a suggestion that common sense rule financial markets, instead of being tied to mathematical models.
The book makes convincing arguments for problems caused by financial models and their assumptions. The applications to the recent credit crisis are insightful and the examples are enlightening. For instance, Triana goes into deep detail about problems associated with the Gaussian copula which quantifies and predicts the interdependence of different variables. This model was implemented in dynamic hedging strategies and one main assumption of the model is that correlations are constant. Using information on the correlation between various tranches in credit default obligations, traders created hedging strategies in which they went long on one tranche (equity) and short on another (mezzanine). When a market disruption caused credit spreads to increase, however, markets moved in opposite directions to the hedge. The risky equity tranche decreased in value and the safer mezzanine tranche increased. The correlations between the two tranches changed, and traders continued to reposition their hedges even though their formulas were no longer accurate; correlations changed and the assumptions of the model did not hold at this time. As the problem exacerbated, traders needed more cash to reposition their strategies and ended up selling equities to do so. This further contributed to the tumbling of the equity markets and caused many pension plans to plummet in value.
Because of the flaws associated with financial models, Triana argues that common sense should govern financial markets instead. It is certainly true that common sense cannot be forgotten. Following a model's formula verbatim can be detrimental if the user does not also bring to bear his intuitive knowledge of financial markets. At the same time, however, applying constant assumptions in the models does ensure their consistent use. Triana does not touch upon this topic, but it is certainly an area for improvement in our markets. Reports from the recent crisis show that at least one rating agency purposely changed their credit models so that mortgage-related securities would obtain higher ratings and consequently the agency would increase its business. The agency knew the rating was inaccurate. Such actions exacerbated the problems associated with the crisis, but the models themselves are not really at fault here.
Model users should also have a thorough understanding of the assumptions being made with models and to what extent they hold in practice. Otherwise, there could be a misguided sense of confidence in the model resulting in an inaccurate estimate of risk exposure. Triana clearly shows that high risk events happen more often than the normality assumption would imply; as a result, VaR, which assumes normality, underestimates risk. Individuals took VaR as truth leading them to believe their risk was lower than it was in reality and consequently, they took on more risk than desired. They did not evaluate the validity of the model's assumptions and the impact that it might have on the resulting risk measure. Triana blames the models and their creators solely for this problem, but it is not clear they should bear the blame alone. Financial economists should have been clear about how their models should be interpreted and should have shouted more loudly when the models were being misused. And, users of financial models should have done a better job at understanding the intuition behind a model, the validity of the model's assumptions and the impact those assumptions have on the model's implications.
Triana's reasoning that common sense should rule is that one can never perfectly measure risk so we should not attempt to measure it at all. If a model's assumptions do not hold in practice then the model is pointless, and continuing to implement such a model can only be detrimental. The drawbacks to financial models which Triana brings to light are valid. Does that really mean all models should be thrown out the window? Can markets run efficiently with everyone operating on common sense alone without any risk measure?
There are some practical dilemmas with this suggestion. For instance, how will markets be regulated? Regulators need to have some measure of risk so they can monitor the risk-taking of financial institutions. Without a risk measure, how do they effectively oversee institutions so as to protect investors? Are regulators to trust that institutions are operating on ‘common sense’? Given the regulatory arbitrage that the recent crisis revealed, it seems hard to believe that this would be the case.
Furthermore what are the implications for related fields if we no longer measure risk? Not measuring risk implies we stop attempting to estimate the probability of extreme events which would lead to a breakdown in the catastrophic insurance market. Similarly, the author's thoughts imply that we should not try to model human mortality. What would the impact be in the life insurance, annuity, and pension market then? Would these changes really improve society?
Lecturing Birds offers a detailed analysis of financial models and their impact during the recent financial crisis, but I caution readers that the book is very one-sided. It clearly portrays the drawbacks associated with financial models, and the analysis on this part is very insightful. On the other hand, there is no indication of the positive externalities that financial models have created. Financial models have increased investor confidence which most likely improved market participation and efficiency overall. Developments in assessing risk have also led to innovations in other fields. These points are never relayed in the book. Triana's solution to the negative externalities created by financial models is to not develop them. Yet, there are negative externalities associated with not creating models as well. Triana effectively provides a controversial critique of financial models, but we need to analyze both their costs and benefits before we decide to embark upon a world without them.
