In my previous blog, I focused on analytics used for execution risk. Here are some other common analytical techniques that are often used in enterprises.

  1. Risk maps: A favorite with firms, risk maps are suited for prioritizing discrete risks. The likelihood and impact of each event is mapped on two axes to identify high likelihood, high impact events so that they can be closely monitored. The disadvantage is that it does not take into account any correlation between risks.
  2.  Discrete scenario analysis: Companies use this to determine riskiness of decisions. They define scenarios for each key discreet risk, estimate likelihood of each scenario, determine cash impact, and probability adjust it to obtain risk values. The higher the range between the best and worst case scenario, the more risky is the decision. Firms can also use it to minimize the impact of the worst case scenario.
  3. Decision trees: This is used for sequential risk where discrete outcomes can be defined. Each sequential decision leads to a probability adjusted outcome till an end node is reached. This helps to be aware of the outcomes, prepare for them and take steps to mitigate risks at each stage.
  4. Simulations: Used mostly for continuous risk, simulations can be run using analytical tools. Variables and their probability distributions are determined, and each deterministic model is run thousands of times with different input values chosen from the distributions. Simulations can also factor in correlations between variables and constraints. The most common simulation technique is the Monte Carlo method.
  5. Gain/loss curves and tornado charts: Used for decision making, the gain/loss curve is used to determine how much money a company can gain / loose from a particular risk over a range of probabilities. Similarly, tornado charts show what impact a risk has on a particular metric e.g. revenue or earnings per share.
  6. Financial techniques: Some commonly used financial computations of risk are Value at Risk (VaR), Earnings at Risk (EaR), and CaR (Cash at Risk) that help to evaluate the financial amount at risk in different scenarios.
  7. Game Theory: A more sophisticated technique; it mathematically models conflict and cooperation between rational decision-makers. It is particularly used for strategic decision making where external parties are involved.
  8. Benford law: Frequently used for fraud detection, this helps to identify statistically unlikely occurrences of specific digits in randomly occurring data. The law states that in any set of large data, more numbers start with lower digits than with higher digits and follow a certain definite pattern. Running a fitment to this curve can throw up anomalies in the data especially those that are fraud related.
  9. Continuous Transaction Monitoring: Data is run through statistical or other logic to identify outliers on a continuous basis. This can be best used for prevention of occurrences both from an execution as well as a fraud standpoint.
  10. Algorithmic techniques: The beauty of analytics is that once the data is available, it can be run through any type of algorithm depending on objective. These could include different types of pattern recognition, statistical distribution matching, gap testing in sequential data, and other heuristic algorithms.

In my discussions with CFOs, I find that they are increasingly incorporating the first six techniques as a regular part of their financial planning process. Different kinds of stress testing and incorporation of black swans enables plans to be more realistic and highlights best and worst case scenarios upfront to increase preparedness.

While analytics provides an objective view of data related to risk, a word of caution – these have to be tailored to organizations and individuals – no two persons look at risk in the same way. These results are thus best viewed in conjunction with studies from behavioral economics, so that they can be used most effectively.