How can investors, large or small, overcome what appear to be obvious shortcomings in risk management, as occasionally indicated by portfolio crashes? In his November 2016 paper entitled “Managing Risks in Institutional Portfolios”, Andrea Malagoli critiques conventional investment portfolio risk management methodologies and offers precepts for robust risk management. He relies on a few empirical observations rather than abstract theoretical principles. Based on these observations, he concludes that:
- Risk factors embedded in the economy change over time, occasionally triggering crises, but more often spawning secular trends spanning multiple shorter business cycles (such as inflation/disinflation trends).
- Changes in these risk factors impact long-term portfolio performance by affecting risk/return behaviors of common asset classes. Risk management requires understanding how statistics (such as asset return volatilities and correlations) and risk factors interact.
- “New” risks are not new, but simply risks not observed during the professional life of most active investment managers. Understanding secular risk dynamics requires consideration of very long historical data samples.
- Conventional risk management methodologies obscure:
- Assumptions built into models that may be invalid.
- Risk dynamics lost in approximations.
- Effective risk management is less about developing sophisticated mathematical tools and more about: (1) challenging assumptions used to construct portfolios and (2) creating processes to deal with surprises.
- Operationally, risk management involves buying insurance when and where it is most likely needed. Diagnosis/prevention is a better approach than prediction.
In summary, robust risk management involves challenging the adequacy of conventional models and data and planning what to do when their inadequacies materialize.
Cautions regarding conclusions include:
- The author assumes that “robust risk management” is worth the effort without demonstration.
- The author refers to assessment of important risks that “are hard if not impossible to observe directly let alone to quantify objectively” as “art,” undermining any argument that risk management is a reliably executable discipline (distinguishable from lucky calls).
