15 April 2020
‘Prediction is very difficult, especially if it is about the future’ – Niels Bohr, Danish Physicist, Nobel Laureate 1922
Although N. Bohr’s quote meant to address a seminar question about his prediction of the influence of Quantum Physics on the world in the future, it sets a base for the difficulty that theoretical and empirical sciences have in consistently relying on models of variable complexity in order to make meaningful predictions about future events. Our modern, information-thirsty and quick-results-oriented, world often assigns greater value to a popular conviction of future forecasts than to the intrinsic knowledge acquired by realized adverse events. And there lies one of the most common reasons of collective herding behavior and cognitive fallacies.
Financial decision-making is a part of a complex ecosystem that blends behavioral psychology and investment management acumen. Should decisions be consistently judged by the process by which they were derived or simply by their outcome? Although the answer relies on the definition of decision quality, it largely depends on the compatibility between the decision maker and the judge (who performs a second-order assessment).
In Finance, a decision maker continuously faces various (known and unknown) risks that could drastically and speedily affect the value of a portfolio’s holdings. The day-to-day process of evaluating the portfolio’s risk exposure to Normal or Black Swan market conditions cannot (and should not) be adequately covered by a single risk approach and its variations (let alone by no risk approach, as often observed in the field). Instead, a systematic decision-making process of applying a full set of risk methodologies should be applied to capture the adherence or the divergence of a portfolio’s probability distribution of returns from normality.
As recent markets vividly displayed, a robust risk management framework demands the implementation of scenario simulations where the distribution is extremely skewed towards tail events, situations that happen rarely. Such shocks could be caused by various macro-economic or idiosyncratic events, which can consequently spread widely to previously thought of as uncorrelated choices of assets (Systematic or Undiversified risk). Examples of historical crises that resulted to large losses of invested capital within a certain period of time (varying from days to months) include the Black Monday of 1987, the Gulf War of 1990, the Asian Crisis of 1997, the Russia Devaluation of 1998, the Global Financial Crisis of 2008 and the (so far developing) Global COVID-19 Health Crisis.
Despite the fact that a large portion of such losses are often due to excessive leverage, high asset valuations and over-concentration of positions, one should seriously consider the use of the factors underling such extreme divergences from normality, to stress-test their often seemingly well-diversified multi-asset investment portfolios.
We should certainly not rely on the assumption that history repeats itself, since the background conditions, driving factors and collective investor sentiment often differ vastly between distant periods of economic and market activity. However, to assess and verify that adequate capital is preserved to cover unexpected losses, investors should attempt to estimate the impact that the re-occurrence of such damaging historical events could have on the portfolio performance.
This way, stress-testing would give us an idea of how stretched the loss-tolerance levels of an investment strategy may turn out to be during a crisis of historical precedence. To conduct such analysis, one needs to select a historical crisis of relevance (a subjective choice) and to apply changes to the risk factors driving the price of various asset classes (such as equities, bonds, credit, commodities, foreign exchange) accordingly, in order to assess the impact to the current portfolio if an ‘identical’ market condition occurred. Such market shocks might be global or local in nature, so geographical disparities in valuation changes should be incorporated, while date ranges could largely match those of the referenced historical event, with the return change being the cumulative one over the entire testing period.
To adjust historical scenarios to modern frameworks, a risk management process should offer the functionality for stress modelling based on a combination of extreme past crises and the customization of factors to the current correlation dislocations, since each investment strategy may be subject to different set of risk factors. Value-at-Risk-based methods provide a decomposition of risk exposure into its core sources, thus identifying over-concentration or risk-adjusted under-performance pockets.
Risk measurement cannot be put on ice until market conditions dictate its sudden use. Extreme market events are evidently more frequent and violent than commonly thought of and their effect on portfolio performance should be diligently and continuously assessed. The ability to implement a multi-faceted portfolio risk analysis will enhance a manager’s confidence to the capital adequacy a strategy or a firm needs to retain in order to cover significant losses in utmost detrimental market conditions.
Yannis Sardis, PhD | Director, Finvent Software Solutions
FINVENT Software Solutions (www.finvent.com) is a trusted provider of financial software applications and custom engineering services. The award-winning KlarityRisk platform specializes in investment risk analytics and fixed income performance attribution reporting, offered to financial institutions in European and African countries. Finvent is the sole SS&C Advent distributor worldwide and its products are natively integrated with those of SS&C Advent, and a Partner of FactSet.