At Rain we talk a lot about what we call ‘information days,’ brief periods of time that are extremely rich with information. We find that evaluating narrow time periods of isolated stress (or euphoria) can be just as informative about manager positioning and underlying risks as longer-term statistics, if not more. The days for instance between May 3 and June 25 yielded a tremendous amount of information about the downside behavior of different assets during a period of fairly extreme interest rate volatility in the middle part of the yield curve.
The typical correlation relationship that investors had become accustomed to during the preceding years – the US dollar strengthens as US yields fall, while credit spreads widen and stocks sell off – completely inverted during the May-June period because risk was driven by the stronger underlying economic integration of these variables, rather than the more spurious relationship between asset classes that persisted during less turbulent markets. Likewise on September 18, the day the Fed announced it would not yet slow its bond purchases, assets with any direct or indirect interest rate risk rallied strongly. Were an entire portfolio to have done well that day (as most traditional asset allocation approaches did), it would have likely had a large slug of underlying interest rate risk running through it.
These small and isolated stress tests quickly get lost in the longer-term numbers analysts typically rely on. Or worse, they’re not even measured because standard analysis of asset classes tends to focus on largely unrelated variables. For us, however, the micro stress tests help fill out more of the texture of risk that becomes important when trying to identify how different assets might be economically integrated with each other. Most importantly, this helps inform how different assets are likely to behave in a crisis environment.
This is an excerpt from our article “Information Days: What scatters, what sticks together, and why?” published October, 2013.