It is the best of times; the stock market has had a fabulous bull run and the ISM index is rising and pointing to even stronger economic growth. It is the worst of times; the market (SPY) is poised for a great fall and the new calculation of the ISM index appears to have a different relationship to the economy, suggesting weakening growth with perhaps a recession bottoming in 2014.
In January 2008 the Institute of Supply Management changed itsr calculation of the Purchasing Managers Index, commonly referred to as the "ISM index." If the new calculation had the same relationship to economic growth as the old one the ISM would be signaling an economy gaining strength. The new relationship is not yet fully clear, but appears to suggest the strength will come in 2015 and we may have a recession before that.
The chart below shows the historical relationship between the ISM index and GDP growth. The black line is annual GDP growth. The red line is the 3-month moving average of the old ISM index with a six-month lead time. The violet line is the 3-month moving average of the new calculation also with a six-month lead time.
You may notice there is a two-month gap between the old calculation and the new calculation. I leave out the two data points that average a month by one calculation and and two months by the other calculation.
Neither way of calculating indicates a bottom for the great recession. The old way of calculating ends pointing down. Its last point, December 2007, is weaker than November 2007. The new way of calculating begins pointing up; February 2008 is stronger than January 2008.
In its new way of calculating, the ISM index appears to have its strongest linear fit to growth with a 21-month lead time. If the chart below is a fair representation of the relationship ISM suggests annual GDP will drop to 0.9% in April 2014.
Such weak annual GDP would almost certainly mean recession. In the 2001 recession annual GDP only got down to 0.7%. The 1960 recession low was 1.1%. Usually when annual GDP drops below 2% we are either in a recession or start one within a few months. The only exception to this in the last 65 years was the 4th quarter of 2011, when annual growth touched 1.9% and then strengthened rather than going into recession. The previous 11 times annual growth was above and then dropped below 2% were associated with recessions.
The ISM index may be calling for a mild recession, but there is not enough history with its new calculation to have great confidence in the call. If it does now reliably indicate what happens in the economy 21 months ahead of time it would be one of the most important economic indicators ever.
We will have a much better idea if the lead time is really 21 months when we can look at the correlation over a complete business cycle. Specifically it will be important to see how the bottom in the ISM index lines up with the low point in the growth rate of GDP during a recession. In addition to the lead time the new calculation appears to have a different calibration: In the old calculation an ISM reading of 50 suggested annual growth would be about 2.8%. In the new calculation a reading of 50 appears to be consistent with annual GDP growth of 1%. In other words an ISM reading of 50 with the new calculation may be as bad as a reading of 42.3 was on the old calculation.
On the other hand, the new ISM calculation could have the same relationship to growth that the old calculation had with other factors accounting for the apparent discrepancy. We may not know the actual relationship between ISM and GDP until we have two or more business cycles to examine.
My best guess is that the new calculation of ISM has a substantially different relationship with growth. The three and a half years of correlation has so far had a fairly consistent relationship. The implication of a weakening economy is consistent with my outlook. So when the new ISM number is announced Tuesday morning I expect it will be more of an indication of the economy in 2015 rather than in the next few months.
Additional disclosure: There is no guarantee analysis of historical data their trends and correlations enable accurate forecasts. The data presented is from sources believed to be reliable, but its accuracy cannot be guaranteed. Past performance does not indicate future results. This is not a recommendation to buy or sell specific securities. This is not an offer to manage money.