Many may think it would be strange to undertake scenario analysis if you run systematic or quantitative shop, but in many cases, it is all the more important because it sets the agenda for futures research and may identify alternative environments when model performance will be an outlier either up or down.
While many view that systematic decisions are mainly run from a model, there are a large number of qualitative decisions that have to be made from the model-maker. The choice of models, the weights for risks, portfolio exposures, leverage, the introduction of new models, the research agenda all require some human decision based on the preferences of the manager. Even attempts to systematize as many decisions as possible require some initial modeler preferences.
We think a regular scenario analysis review can help with the process. At one extreme, scenario analysis can just be used to identify areas for new research. Alternatively, scenario analysis may help tilt the exposures of models, markets and sectors.
The objective is not to develop scenarios of actual prices, but form scenarios of what the environment may look like. What are the alternative investment worlds that we may inhabit. For example, with the BREXIT vote, the UK will be at best a more uncertain investment or one that shows increasing downside risk for domestic companies. This may change the quality of signals relative to historic back-testing.
We divide the scenarios for this type of analysis for systematic traders into three major areas or focal points for discussion: volatility/ uncertainty, structural, and policy scenarios.
For volatility and uncertainty, the scenario question is simple. Are we going to be in a high volatility or low volatility environment? Similarly, is the markets going to face more or less uncertainty? These questions will lead to cases for what will have to happen to make the market remain or move to a new state. Walking through what would have to happen to change the state and the likelihood of that event is critical for good portfolio management.
Structural scenarios are usually more long-term. What is the state of trading fin-tech? Are there regulations that will change the competitive landscape for trading? Will there be a change in the major players in the markets traded? Will competition for alpha increase over time? If the costs of trading increase, it will impact some strategies more than others. Structures matter.
Policy scenarios are always a key focus for any research. Will central banks change their monetary policies? Given the radical changes over the last eight years, the impact on models and systematic behavior will be strong. Will there be changes in fiscal policy or regulation that will change the macro environment? Certainly, a move to fiscal stimulus will change economic growth prospects and affect relative sector performance.
Models are sensitive to the environment on which decisions were tested, so preparing for changes in the environment is a very useful risk management tool independent of price forecasting. The nature of how this is done can be highly variable and individualized to the culture of the firm, but undertaking future scenarios seems to be a critical responsibility of the manager.