DJIA forecast (September 24, 2008)
This dramatic September is very interesting in respect to working forecasting models. I can state that this situation can be described as: "Dow Jones is backing to its old patterns". A couple years ago I worked with models that have "short memory", in other words I considered only several years of price history to create the forecast models � because at that time the stock market structure was changing very fast. Annual cycles did not work; there were no September drop in 2006 and Christmas rally in 2007 (though these are the most reliable patterns in Annual cycle).
But in winter 2008 something has changed in the air. Dow somehow began to remember its old patterns. Not perfectly of course, but a lot better than it was last few years. This is how it looks together with the forecast based on 120 years history (I practically did not use earlier such a long price history):
I understand that it is not a popular statement, but IMHO we are around the bottom now. Of course if no "force major" occurs anymore and there will be no new "skeletons in the cupboard� revealed as we have seen recently.
However one uncertainty still presents. For the forecast above, we have used 120 years of Dow�s history. But if we recalculate the same Annual cycle taking into account only those periods when Democrats have been in power, we get another Annual pattern:
I can also try to find the most similar year to what we have now.
Just look at these two price charts: 2002-2003 yy - red chart and nowadays (September 2008) - black chart. It is very close to the first Annual cycle model: