Empirical Simulation Analytics in Financial Engineering

The financial industry has always been data-intensive. Nowadays, the amount of financial information available to market participants is huge, leading to enormous challenges on how to efficiently use it. Computer capacity is also at a stage when simulations can be easily performed and any resulting data stored without problems. Still, some financial services seem not to have adapted fast enough to the new data related facilities available. This study focuses on the financial engineer service – the development of investment products and/or strategies. Most of these products and/or strategies are still developed relying on theoretical models that, no matter their degree of sophistication are, by definition, a simplification of reality. They are sold to investors relying on model-based analytics. Most these products are only taken to test in a “real life” context after they are sold to investors.In today's Big Data context this is hard to understand. We advocate that financial engineers, should back-test their proposals' design using empirical (past) information, or empirically simulated data. Risk analytics statistics resulting from empirical data are both realistic and intuitive, from the common investor point of view. We use a well-known portfolio insurance strategy – the constant proportion portfolio insurance (CPPI) – to illustrate our point of view