Fat Tailed Distribution and the 2008 Financial Crisis

Fat tailed distributions are statistical functions that look like a Gaussian (Normal) distribution, but actually have more events farther from the mean than expected. This makes the histograms with ends being “fatter” than normal, thus the term “fat-tailed distributions”. Why is this important? When models assume a normal distribution of a variable in their predictions, most of the time the prediction works very well, but on rare occasions the prediction is wildly off. In many cases, that may not be important, but when you are modeling large, important things (like the mortgage market), a prediction that is wildly off can have huge impacts on people’s lives. This is one reason why many people were surprised during the 2008 financial crisis.