Monte Carlo Methods

Monte Carlo methods, also called Monte Carlo simulations, are a category of mathematical algorithms usually run by computers due to the significant amount of math involved. In finance, these methods are often used to simulate reality in order to calculate the likelihood of risk and returns.

Monte Carlo algorithms are designed to use repeated random samplings to predict future behavior.

There's no risk. It's free for 30 days.

Start Now to Get More Out of Your Retirement Account

Some complicated calculations require mathematicians to slot in hypothetical numbers because there are so many unknown variables in the equations. So they compute the same calculations hundreds or thousands of times – always using random numbers in the places that require unknown numbers to be input. Over time, they can use these simulations to more accurately predict probable outcomes.

Monte Carlo methods background and uses

Monte Carlo methods were developed by Manhattan Project scientists working on atomic weapons at the dawn of the computer age. They’re used in virtually every category of science and math. For example, they’re helpful in predicting the flow of liquids and the behavior of cells. Geologists use them to find minerals and oil and ecologists use them to simulate ecosystem changes. NASA even applies them to space exploration. Economists and financial analysts can use the methods to simulate possible risk and returns on all levels – from a single fund to an individual’s portfolio to the overall economy.

Recent Criticisms and Responses

Recently some experts have criticized the use of Monte Carlo simulations in the field of finance and economics. For example, some Monte Carlo simulations determined that the probability of an event like the fall-2008-to-spring-2009 market drop and recession was extremely low.

However, nearly all financial industry experts now agree that Monte Carlo simulations were not the problem. The problem was that some firms and researchers were using the simulations incorrectly. For example, when calculations should have been done thousands of times to find an accurate prediction, some simulations were only run a few hundred times – they literally needed to do thousands more calculations to get more accurate information. Additionally, these complicated algorithms require a lot of numbers, and some firms were not using realistic numbers in their calculations.

Rather than moving away from the use of Monte Carlo simulations, a consensus is building that the industry standards should be higher – with more accurate and conservative numbers combined with a higher number of calculations for each simulation.