Menu
ETFs are becoming a preferred investment vehicle as some of them outperform most investment strategies while giving a clear picture of what you are investing into, both in terms of capitalization, industry, risk, and region of exposure. Our client was an asset manager who wanted to combine financial products such as ETFs to diversify their portfolio along capitalization, profile, region, sector, etc. while keeping the performance similar to some of the underlying ETFs.
As well as they are performing, ETFs do not provide investors with a lot of margins to maneuver. For instance, one might want to hedge against certain sectors of the economy or regions of the world. One can combine them to diversify their investment strategy, but it quickly becomes a challenge to manipulate combinations of several ETFs, as each of them can contain more than 1000 companies. Combining ETFs is a computational challenge that certainly cannot be addressed by handmade computations, as it is multifactor, and the companies involved are in the thousands.
Python
Shiny
The deliverable of such a project is not the modeling, but rather the results it yields. In this context, the resulting portfolios should be practical tools and not abstract construction. Through approximation, the resulting portfolios were perfectly feasible in our client’s setting. Optimizing computations that would otherwise have been impossible to run was an important challenge. Our team of scientists overcame this challenge through their experience with optimization in various settings (finance, physics, machine learning, etc.). Our team built custom visualizations in order to better render the optimized portfolios and to visualize the diversification in several dimensions.
Newsletter
