Confidence levels, holding periods and their effect on risk data modelling

  • Subscribe to updates

  • Privacy
  • This field is for validation purposes and should be left unchanged.

Confidence levels, holding periods and their effect on risk data modelling 

Our third and final blog on risk management modelling where Tony and John discuss the effects of confidence and holding period. Using Operational Risk Software can be key.

Taken from: Mastering Risk Management 

The confidence level used in modelling represents the level at or below which the economic capital figure given by the model will be sufficient. Another way of putting this is that the economic capital required will be insufficient only above that level. For example, the 99th centile is the point at which 99 out of 100 results (on average) will be at or below that level of capital. It is often said to be 99 years out of 100 years or the ‘worst year’ of events in 100 years. It should be remembered that the worst year in 100 years could be this year or next year and not 99 years away! Clearly the higher the confidence level, the higher the amount of economic capital required. 

The holding period (sometimes known as the liquidity horizon) is the time that it takes to neutralise or dispose of a risk or asset. This can vary considerably depending on the industry and the risk/product. For example, a busy plumber having bought a wash basin may be able to sell it to a customer in days. However, a large supermarket chain may suffer for many months from the results of a cyber attack. Clearly, if it only takes a few days to dispose of a risk the amount of economic capital required is relatively low. The holding period can therefore very significantly affect the economic capital that the firm requires to support its risk profile. 

Is correlation and causation different?

There is much debate about the correlations of different risks, event types and business lines in risk management. The correlations are less relevant than the accuracy of the primary data. 

A correlation of two events means that a high value for one risk is associated with the high value for the second risk. They both happen together and it is not an indication of which happened first. 

Causation is where there is a linkage between two risks such that one risk causes another to happen. So, an earthquake may cause a fire (but a fire will not cause an earthquake). In other words, an earthquake and a fire are correlated, but only a fire is causally related to an earthquake. 

The quantity of additional capital required if an earthquake occurs to support the risk of fire is determined by how many times a fire occurs flowing an earthquake. If a fire occurs in only one earthquake in five there is said to be a 20% correlation. If a fire occurs every time there is an earthquake (a 100% correlation) the capital required for an earthquake must also include the capital required for a fire. But for a 20% corelation a smaller amount of additional capital is required to add to the capital required for an earthquake. 

What is diversification?

Diversification of a firm’s risks allows a reduction in the economic capital required. Anything less than a 100% correlation creates a diversification benefit i.e. a reduction in the capital. 

As an example, Risk A has an impact of 3 and Risk B has an impact of 4. At 100% correlation, impact of both together is 7 (simple sum). The fully independent impact of both together is 5, which is the square root of the sum of the squares of each risk. (the square root of the sum of the squares is an acknowledged method of calculating a fully independent impact of any number of risks). The diversification benefit is therefore between 2 and 0. In practice, you would investigate the diversification benefit of several pairs. 

Summary

Modelling gives a further step to challenging the often poor data and to ensuring that the risk environment and control environment match the extent of exposure that the firm is willing to take. Business benefits can come from the least expected direction. 

In our next blog Tony and John discuss confidence levels and holding periods and their effect on modelling.      

Mastering Risk Management by Tony Blunden and John Thirlwell is published by FT International. Order your copy here: https://www.pearson.com/en-gb/subject-catalog/p/mastering-risk-management/P200000003761/9781292331317    

For more information contact us today on sales@risklogix-solutions.com 

  • Download Now

    Whitepaper: Proving Corporate Integrity with ESG
  • Privacy Policy
  • This field is for validation purposes and should be left unchanged.

Related Posts

11 Parameters for modelling qualitative risk data
11 Parameters for modelling qualitative risk data  In the second in our series of blogs on challenging risk models, Tony and John look at the parameters relating to the modelling of qualitative data. Operational Risk Software can be key to supporting this discipline.    Taken from: Mastering Risk Management  These parameters relate to the challenges that …

11 Parameters for modelling qualitative risk data Read More »