What is Monte Carlo simulation?

  • Subscribe to updates

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

What is Monte Carlo simulation – and how did it get its name? 

A question that often crops up – exactly what is Monte Carlo simulation and can it be managed by operational risk software, and how is it useful to the business? Here Tony and John give an explanation.  

Taken from: Mastering Risk Management 

Monte Carlo simulation is used where there are no or insufficient actual data or where further data are required. It is a method of creating artificial data based around clear and commonly accepted guidelines. The artificial data are created through randomly making new data many thousands or even millions of times using a random number generator. This is a mathematical formula for producing seemingly random numbers. There are many such formulae available in modelling although most modellers use just a few of the better known and well tried and tested formulae, such as the Mersenne Twister. This is because all random number generators are not completely random but some of the better known ones have managed to pass most of the tests for checking that the numbers largely seem to be random. 

Once numbers have been generated, they are then related to the data. For risk management modelling, this means that controls can be randomly failed and then risks can randomly occur (or not, as is the case in real life). Also, the frequency of losses can be randomly generated (although this is often guided by actual data). Generation of simulated frequency then enables the generation of the correct value of random impacts for losses (again, often guided by actual data). And, yes, it is named Monto Carlo after the casino, which of course generates seemingly random numbers. 

Clearly many random occurrences should be used to generate any results which will be used. This enables results to be averaged and therefore, hopefully, to be closer to the expected value. Indeed, the result should tend to become close to the expected value as more simulations are made. When all the results are within say, 1% of each other the result is said to be stable. Another way of saying this is that the results converge on the expected value of that convergence has been achieved. Monte Carlo simulation is a way of getting more data, albeit artificial, as one of the most intractable problems in non-financial risk is simply the lack of data on which to act. 

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