If you want to understand geology, study earthquakes. If you want to understand the economy, study the Depression.
Under reasonable circumstances, how much can you expect to lose? This is the quantity that the financial statistic Value at Risk (VaR) seeks to measure. Since its development soon after the stock market crash of 1987, VaR has seen widespread use across financial services organizations. The statistic plays a vital role in the management of these institutions—it helps determine how much cash they must hold to meet the credit ratings that they seek. In addition, some use it to more broadly understand the risk characteristics of large portfolios, and others compute it before executing trades to help inform immediate decisions.
Many of the most sophisticated approaches to estimating this statistic rely on computationally intensive simulation of markets under random conditions. The technique behind these approaches, called Monte Carlo simulation, involves posing thousands or millions of random market scenarios and observing how they tend to affect a portfolio. Spark is an ideal tool for Monte Carlo simulation, because the technique is naturally massively parallelizable. Spark can leverage thousands of cores to run random trials and aggregate their results. As a general-purpose data transformation engine, it is also adept at performing the pre- and post-processing steps that surround ...