In the process of testing out and selecting a new color scheme for displaying complex data on a Microsoft Excel chart, we somehow managed to accurately forecast the overall trajectory of stock prices last week. Over a week in advance.
Here's how we did that. The complex data we were displaying was generated from our standard model of how stock prices work, which combines future-oriented data related to the amount of dividends expected in future quarters with historic stock prices, which our model uses as base reference points for projecting stock prices into the future.
Specifically, our standard model for forecasting stock prices incorporates the historic value of the S&P 500 from 13 months earlier, 12 months earlier and one month earlier in projecting a particular day's most likely stock prices. The chart below shows the trajectories for each that apply in 2015 (the heavy black line represents current day stock prices).
If you pay close attention to our chart, you'll find that the current day S&P 500 is almost perfectly following the trajectory that stock prices did exactly one month earlier, when investors were weighing the potential negative impact of falling oil prices on the growth prospects upon the U.S. stock market in the future. That dynamic is what our model has picked up upon.
The fact that it would appear to be repeating is a phenomenon that's largely driven by the U.S. bond market. Our thinking is that the investment decisions that were made in December 2014 are being repeated again in January 2014 with the maturation of one-month U.S. Treasuries. It's kind of like an aftershock after an earthquake, or in this case, a noise event in the U.S. stock market, where a past event is actually driving stock prices in the current day, to the extent that such an event can.
The degree to which events in the past might drive today's aftershocks or to more often dissipate as yesterday's echoes would really appear to depend greatly upon what investors expect for the future as the maturity and option expiration dates associated with their previous investment choices come to pass and provide the means for executing new decisions. If the future doesn't change in the interim, why should the investment decisions of investors change?
So it's actually luck that is behind that our model's ability to have accurately projected the trajectory that stock prices followed last week. It's just fun to see that it's possible to build a particular kind of luck into a mathematical model!