The market seems to have 'destabilized' back in February 2007. The plots above show a measure of the stationarity of the statistics for SPY and XLE. The plots show the stationarity for a rolling one year, two year, and three year period. When the colored lines are below the critical value (horizontal gray line), the statistics for the corresponding period are stationary.
The SPY statistics have continued to be nonstationary since February. Prior to February, the SPY had been in an eight plus month period where its statistics were stationary. In contrast, the statistics for XLE have tended to remain stationary across the entire period. The XLE was not badly destabilized in February 2007 like SPY.
The moral of the story? You must consider the stationarity of the data you are examining. If the time period you examine is not stationary, any patterns you find will likely be coincidental. The best approach is to trade something that has stationary statistics. If your usual trading method no longer seems to work it may be that the statistics of the instrument you are trading are no longer stationary.
No comments:
Post a Comment