Econometric methods for the estimation of the dynamic risk profile of non-linear financial portfolios

Don’t put all your eggs in one basket, is the basic premise of diversification in portfolio theory. By combining investments spread among assets, we can tame asset-specific shocks. In recent years, there has been a growing popularity of synthetic financial products, which achieve additional diversification through active management strategies and exposure to investible risk factors.

There is a need for adequate econometric methodology to evaluate the risk profile of such a dynamic fund. In a single-period analysis, the traditional methods still apply. However, at longer time intervals the complication arises that, within the risk evaluation window, the allocation

scheme and asset return dynamics change abruptly, leading to a time-varying, non-linear exposure to risk factors.

This research answers the methodological need for reliable risk profiling in the context of active management. We ask the following questions: How to deal with sluggishness in the availability of prices for instruments? How to decompose a risk measure into component risk contributions? How to estimate accurately the intraday variation in factor exposure? Importantly, the research also quantifies the economic importance: How different are risk estimates when properly accounting for non-linearity? What is the reduction in losses by non-linear diversification beyond correlation analysis? And, how should investors and policy makers react to sudden disruptions in asset prices during impactful events