The Failure of Risk Management summary

It could be said that organizational risk management began when some king or leader first made the decision to fortify their city’s walls or put away extra provisions in case of a tough winter. But risk management has come a long way since its beginnings; the advent of computers in particular has dramatically improved the field. But even prior to digitization, the risk of nuclear power and oil exploration in the 1940’s increased the sophistication of risk management, a major reason why it’s in the minds of most large organizations today, regardless of their field.

So how did risk management evolve into such an important field?

So it’s clear that risk management is important. But there’s an issue: the most commonly used methods are flawed.

How?

The opinions of esteemed experts are respected and relied on by people for all manner of things, and risk-management methods are no exception. But expert opinion may not actually be so helpful for assessing risk.

Why not?

While the opinions of experts are biased like those of everyone else, even the most quantitatively accurate methods require some sort of expert advice in order to effectively identify risks. The good news is there’s a way to improve expert opinion. It’s called calibration training and here’s how it works:

The main purpose of calibration training is to give people an accurate picture of their uncertainties. While the most important parts of these trainings are repetition and feedback, there are many ways to calibrate.

The absolute best method for assessing risks is called the Monte Carlo Simulation. It’s been used to evaluate some of the biggest risks imaginable from nuclear power safety to oil exploration and even environmental policies.

How does it work?

One of the most common criticisms of quantitative probability methods like the Monte Carlo simulation is that there’s simply not enough data to simulate specific events. The skeptics agree that methods like these are preferable to soft methods such as scoring, but maintain that there’s no way to get the data necessary for a simulation to be effective, using this claim to justify the use of expert opinion and scoring methods. But the truth is, there’s a better option.

Institutions like insurance companies and the nuclear-power industry compute the odds of hypothetical events all the time. For instance, nuclear power companies have run several simulations of events that are likely to occur once every 500 years – longer than the industry has even existed.

So the accuracy of any probability assessment is based on the quality of the model that’s used and the precision of the parameters and assumptions that are fed into it. Knowing this, you can ensure your model is accurate by comparing your estimates with the facts on the ground.

Why?

Building, utilizing and maintaining the appropriate risk-management tools is essential to mitigating risk, but even with the right tools there are still plenty of barriers to managing risk – for instance, things like organizational silos, a type of structural problem resulting in the blockage of information, resources and authority.

Consider the fact that all managers are in part risk managers because every manager is required, to some degree, to measure risk and return. In most cases, managers can effectively monitor the risk of their own area. But when it comes to assessing the risk of major decisions that have implications for multiple departments – like the construction of a new production facility, for example – you’ll need a more comprehensive approach to analysis.