Ortec Finance's Hens Steehouwer and Guus Boender take a closer look at economic scenarios for strategic risk management.

The recent and ongoing financial crisis has illustrated to all of us that having a good long-term financial strategy alone is by far not enough to guarantee that, in the end, we will actually meet the objectives we set out to achieve with this strategy. This holds both at the macro (country) level, for institutions like pension plans and insurance companies, as well as for individuals.

Instead, the formulation of a strategy has to be followed by an implementation that is aligned with this strategy. And once such an implementation is in place, regular monitoring needs to take place to assess if the strategy is still the right one based on the most recent information. The combination of these three steps is what we call Strategic Risk Management (SRM). This concept goes further than, but needs to be consistent with, checking if asset managers stay within the risk limits that have been assigned to them. The SRM process is illustrated below.
 


For SRM to work, it is indispensible to have frequent and integral information on both long-term and short-term risks (and returns) that lie ahead of us. Focusing on just the long term or just the short term can both be equally harmful for the extent to which (long-term) objectives will be met.

If the objective is, for example, for the Titanic to arrive in New York, then she will very likely never arrive there if the strategy (heading) is continuously adjusted on the basis of only short-term information (avoiding icebergs). On the other hand, she will also probably not arrive in New York if the initial strategy (heading) is the right one, but short-term risk management (avoiding icebergs) is missing.

In each of the three steps of a SRM process -strategy, implementation and monitoring - decisions need to be made taking into account the uncertainty of the financial and economic world that lies ahead of us. For reasons of flexibility and transparency, this uncertainty is often represented by considering large numbers (thousands) of scenarios of how financial and economic variables will evolve into the future, which can then be translated into the consequences in terms of the actual objectives.

Traditionally, we see that different types of models and approaches are used to generate such scenarios for different applications. More specifically, scenarios for long-term strategic decision-making are typically different from scenarios for short-term risk management. This can have very practical reasons, but also more fundamental ones because specialised models have been developed for specific applications.

This is fine in itself as long as these different applications are not related. However, within an SRM framework, that is exactly what is needed. If, instead, different scenario models are used in different steps of the SRM process, consistency is lost, and the foundations under the whole concept get very weak and difficult to repair.

A practical example can be found in the upcoming Solvency II regulatory framework for insurance companies. There, on the one hand, short-term scenario models with a one-year horizon can be used in internal models for calculating the Solvency Capital Requirement (SCR). On the other hand, Solvency II will also require insurance companies to make dynamic balance-sheet projections for horizons between three and five years to see how the required (and available) capital will evolve over longer horizons. If for the latter application another scenario model has to be used, the question becomes how to reconcile the results from the short-term SCR calculations with those from the longer-term analysis in a consistent way.

This leads us to conclude that SRM requires consistent economic scenarios for the different steps in the process. But why are such scenarios then not extensively used in practice? The short answer to this question is that it is very difficult to arrive at such scenarios. As EM Varnell puts it, again in the context of Solvency II, "it is useful to have a common set of real-world ESG scenarios used throughout the enterprise, against which all decisions would be based. This can put a lot of demand on the ESG model to capture many features". And "if the ESG model cannot capture enough of the features of the economy, [this] increases the risk of inconsistent decisions being taken."

To see where this complexity is located exactly, it is useful to look at explicit requirements that are put on the economic scenarios by SRM. A first group of requirements in itself does not pose a direct problem. These are the requirements of the scenario being (I) multi horizon (for example, ranging from one month, a few years to possibly decades), (II) multi frequency (for example, monthly or annual time steps) and (III) multi dimension (ranging from a few financial and economic variables to hundreds of specific portfolios).

The true challenge starts when we require that the resulting scenarios are also (IV) realistic and plausible. As one unified and accepted theory that describes the behaviour of economic and financial variables does not exist (and probably never will), in practice, this realism beholds that the behaviour of the scenarios should be consistent with a wide range of so-called 'empirical laws' that are known to characterise the financial and economic world we live in.

Unfortunately, different empirical laws are typically important at different horizons and frequencies. A long-term example of such an empirical law is what is called the 'term structure of risk and return', which is about the notion that, for example, correlations between asset classes and economic variables vary with the investment horizon. A more medium-term example is how variables are related to each other over the course of business cycles. Finally, shorter-term examples are about how volatility varies over time and how correlations go up in bad market conditions (tail risk).

Integrating these (and other) important empirical laws in one set of economic scenarios is what constitutes the biggest challenge. Leaving out some of the empirical laws to simplify matters is unfortunately not an option, as it is well known that the properties of the scenarios have a large impact on the result of models that run on these scenarios and hence also on the important financial decisions that need to be made.

Therefore, having such economic scenarios and applying them within a true SRM framework is what we all need to manoeuvre our way though times of crisis, but at the same time remain on track for meeting our ultimate objectives.

Hens Steehouwer is head of research at Ortec Finance. Guus Boender is professor of asset-liability management studies at the Free University of Amsterdam and co-founder and board member at Ortec Finance.