In the process of mitigating risks and preventing the possibility of future crises, the ability to accurately estimate future risk and understand the amount of uncertainty an estimate carries is essential. However, risk measurements are heavily reliable on data quality and quantity, and unfortunately insufficient data is a prominent and common issue in finance. At the same time, regulators demand all more strict regulatory requirements for institutions that further rely on access to financial data.

In this article, we evaluate the rolling window procedure to alleviate the problem of inadequate data by increasing the number of observations extracted from a limited set of data.