Trader’s Narrative

The pundits are blaming the market’s stumble on Wednesday on the surprising reduction in the US manufacturing base - as measured by the Institute for Supply Management’s (ISM).

The ISM index fell to 47.7 - its lowest reading since April 2003. The “line in the sand” is 50, so anything above that is expansion and anything below contraction.

If you believe the reporters and the media, the market tanked because of the negative surprise. I think there were other, more technical, issues that caused the weakness.

Predictive or Contrarian?
You have to remember that reporters have to fill thousands of inches of column space by describing the minutia of the market everyday. I don’t envy them the task. But the truth is much more boring: the market does what it does. Sometimes it reacts positively to a negative shock, and sometimes it doesn’t.

Now there is more and more talk of an imminent recession. But in reality, the ISM report and the stock market are negatively correlated.

Like the Michigan Consumer Sentiment Index, the ISM report is a much better a contrarian tool than a forecasting one. On average, whenever the ISM data has dipped below 50, the stock market has been positive in the following months. And when the ISM data is very very high, the market usually slips (or at best meanders nowhere).

The last time this indicator was around here was spring 2003. Ring any bells? The start of the bull market? Of course, back then there was a confluence of many factors. The ISM report wasn’t a cause of the bull market, just another marker that was flashing back then.

But Why?

For one, the stock market is a forward discounting mechanism. So measuring economic activity is like looking in the rear view mirror.

Also, the ISM data is not exclusively quantitative. There is a significant qualitative component as purchasing and supply executives are interviewed and asked about their take on the immediate future.

So, it is a hybrid indicator. Part hard data and part, soft human emotion. Which explains why it is a good contrarian measure.

Take a look at the graph to see what I mean (click to enlarge):