On 26 August 2009, IBF co-organised an industry talk with the MAS Academy featuring Professor Anil Gaba, Academic Director of Centre for Decision Making and Risk Analysis (CDMRA) and Dean of Faculty, INSEAD.

In his presentation entitled “ Uncertainty and Risk - The Paradox of Control”, Professor Gaba discussed the limits of predictability in vitally important areas such as life, medicine, investments and business.

Illusion of Control

Professor Gaba started his presentation by sharing the results of a poll conducted on 600 INSEAD MBA students and executives focusing on how much control they had in achieving their wishes. The results were as follows:

•  Happiness………………….…64%

•  Health and longevity…..........52%

•  Wealth…………………….…..53%

•  Success……………………... 63%

This poll revealed the illusion of control that people have. He elaborated on how complex forecasting models don't necessarily predict the future well, and that the KISS (Keep it Simple Statistician) Principle or the simple models are far from being usefully accurate in a prediction-making process. Even with their limitations and inaccuracies, forecast models are used because “humans tend to make predictable mistakes and are predictably irrational”.

Judgment, Risk and Uncertainty: The Traditional thinking Style

Traditional thinking styles can be categorized into the Normative Approach (i.e. How should people make decisions) and the Descriptive Approach (i.e. How do people make decisions). Professor Gaba acknowledged that there could be a prescriptive approach to formulate easy-to-use and practitioner-oriented rules for enhanced decision making.

One of the key issues involved when making a decision was the illusion of control, which made a person underestimate uncertainty, or believe that he can influence the outcome of certain events based on their actions. By downplaying or ignoring uncertainty (e.g. by failing to accept limits to predictability), people often fell victim to the illusion of control in many domains of life. Professor Gaba believed that we can often gain more control by giving up control, hence “the paradox of control ”.

Paradox of Control: Finance

There are many optimal rules for allocating wealth in a portfolio across N risky assets, some backed up by Nobel Prize-winning research such as the Classical approach that ignores estimation error, the Bayesian approach to estimation error and the Optimal combinations of portfolios.

Based on a study across 7 datasets/markets covering 1963 to 2004, a Naïve Rule¹ on wealth allocation fared better than any of these optimal rules . Instead of complacency (and assumption of control) over a complex system, one should understand and accept the true extent of uncertainty and build “safety factors”.

Forecasting: Finance

Professor Gaba felt that the current financial crisis could be attributed to greed, excessive risk taking, and insufficient regulatory oversight. “People in finance consistently underestimate uncertainty, resulting in mismanagement of simple risks such as real estate risks,” said Professor Gaba, pointing out statistical data that showed recurring “bubble bursts” throughout the decades in 1929, 1930, 1966, 1980 and 2008.

Professor Gaba quipped, “A Senior Chinese Govt. Official when asked about a possible bubble in the Chinese Market in 2007 said, ‘Financial Markets are like Champagne – if there are no bubbles, no one will drink them!'.”

Forecasting Management

Professor Gaba then categorized uncertainty into two forms: (i) “Subway Uncertainty” which can be modeled and quantified, and (ii) “Coconut Uncertainty” which is difficult if not impossible to model and quantify.

Most real-life situations have a mixture of Subway and Coconut Uncertainties, such as Price of Oil, Earthquakes, and Financial Markets.

Professor Gaba felt that the forecasting models of the socio-economic world are an abstraction, focused on technical wizardry around internal validity without adequate emphasis on external coherence, with no safety factors (for Coconuts) built in.

There was also the issue of biases in humans when forming judgments, such as the Anchoring Bias which occurs when we use an initial estimate as a starting point, and adjust our estimate from that anchor. These biases were often observed in areas such as sales forecasting, and budgeting.

The Overconfidence Bias : A bias that occurred when people were overconfident in their abilities and overconfident in assessing uncertainty (i.e. setting low guesses too high and high guesses too low).

The Loss Aversion Bias, which could be seen when investors often show the disposition effect: selling performing assets too soon and holding on to their losses too long. Most investment clients will not sell anything at a loss and are likely to throw good money after bad money.

For people to overcome uncertainty, acceptance was the first part of what Professor Gaba calls the ‘Triple A' strategy of accepting, assessing and augmenting.

Accepting

•  what we control and what we don't control and the limits to predictability

•  the true extent of uncertainty and risk and hence the paradox of control

Assess: (for Subway Uncertainty)

•  avoiding systematic (predictably irrational) cognitive biases in our judgments

•  weigh in emotional barriers, such as greed, fear, and hope

•  use simple models and averaging

•  do not use complex models unless empirically proven to outperform simple models

Augment: (for Coconut Uncertainty)

•  expand the Subway uncertainty

•  turn uncertainty into an opportunity and where possible, build resilience, agility, and “emergency exits”

Professor Gaba concluded the lively seminar by reiterating the following:

(a) we have to accept that there are some things in life that we can't predict; and

(b) the notion that we can control risk is but an illusion.

¹ A Naïve Rule is the division of wealth equally across N assets available with no estimation of parameters or optimization.