Thursday, October 7, 2010

October Thoughts about Seasonality

Market seasonality has probably been more of a factor in our decision making of late than it should. Rick wrote about it in this space last week. To make that statement in a year where “Sell in May and go away” has proven to be a perfect timing indicator might seem a little harsh. In fact, many of the analysts we follow publish composites of past market performance based on a variety of time frames, and those composites, along with the more traditional seasonal themes like “the January effect” or “summer rallies” are always a factor in our decision making process. Of course the market still has not managed to take out the late April high from this year so “sell in May,” with perfect hindsight, was excellent advice. However, the past few months have been dominated by our seasonal concerns about September and October. September is, on average, the worst performance month of the year for the broad market, and October is well known for having more than its share of well documented market meltdowns and investor riots. Now that we’ve just closed out September with the best performance for the month since the Great Depression, a disturbing result for “seasonality gurus,” we’ve spent some time digging further into the subject of seasonality as it pertains to third-year presidential cycles.

Ned Davis Research gives us some excellent data that seems immediately relevant. Consider the following information:

The DJI (Dow Jones Industrial Average) has gained, on average, 6.4% in 9 cases where there has been a change in the Congress in the third year of a presidency. The DJI has gained on average 6.4% in the 13 cases of a third-year presidential cycle within a secular bear market. The DJI has gained 6.1% on average in the thirteen cases in the year of a capital gains tax hike, and the DJI has lost 3.6% on average in 18 cases where the market was 22-34 months after a cyclical bull market within a secular bear market.

For the most part the data seems benign except of course for the 3.6% loss 22-34 months after a cyclical bull within a secular bear. Since the current bull market began in March of 2009 that seems to increase the likelihood that 2011 would be a poor year for the market. Of course, the 6%+ gains in the other data seem to indicate a decent year might be forthcoming. Ned Davis’s market cycle composite indicates that the broad market should be rolling over right now and testing new lows before it finishes the year with a strong rally. The question is how much weight to give any of this past seasonal information? It has little to do with the market fundamentals and technicals we track so religiously. Additionally, it has nothing to do with the current global economic condition and it doesn’t speak to current market valuations. Yet, it is seductive in its premise that the past could repeat. Considering that we are experiencing a financial condition that is unprecedented – a Great Reflation following a Great Recession – market seasonality and past market cycle composites will remain a part of our process….but not overly so.