- People who are concerned about a recession in the United States are particularly focused on the ISM index that has been in a contraction area in recent months.
- But the ISM index is poor at predicting economic downturns and typically overestimates manufacturing sector weaknesses.
- ISM is also in line with other strong economic indicators.
- Neil Dutta is Chief Financial Officer at Renaissance Macro Research.
- Visit the Business Insider website for more stories.
Economic concerns are constantly finding new reasons to be nervous about the US economy. First, it was the weakness of the housing market. Next time was the inversion of the Treasury yield curve. Now is the favorite data for the recession ISM Manufacturing PMI.
The ISM index is a survey-based measure that asks around 300 purchasing executives across the industry whether activity is up, down or flat compared to the previous month. A reading above 50 in the index indicates that the sector is expanding, while a reading below 50 indicates a contracting manufacturing industry.
Since March, the index has dropped from 55.3 to 48.3, and below 50 over the past three months, triggering worries about an economic downturn or even a recession.
As with all past freaks, economic pessimists focus largely on a single data point while missing the larger picture of continued economic growth.
ISM is bad for measuring the market
As the recent downward trend picked up, the ISM production index began to gain more and more attention. A simple query in Google Trends shows search interests increasing after a disappointing September reading.
While the public's attention is recent, the ISM has a near-mythical reputation in the financial markets, and spawns a cottage industry of market-time models based on its direction. But the utility of ISM to evaluate the markets is dramatically oversold.
ISM, like the financial markets, can be considered as a growth momentum barometer. As a result, financial market indicators go a long way to explaining the ups and downs of ISM.
Think of a simple four-variable model that uses the S&P 500, 1
But if the financial markets explain a good part of the movement in ISM, it seems rather risky to use ISM to make sweeping market calls to begin with.
This is the rough equivalent of saying "I'm bearish on the US stock market because the market has sunk" – like a second-rate form of technical analysis, a forecast for the market based largely on the assumption that the market is usually only responds to itself with decline that begins even more slowly.
To the extent that a market timing model using ISM works, it does so only at the extremes, when the index is particularly low or particularly high. In the middle, the ISM signal for stocks is mostly just noise.
ISM is a poor economic indicator
While ISM is a deficient timing tool, it is almost as bad as an indicator of business cycle timing. The ISM index sent as many false signals, defined as a dip below 50, into "constricting" territory, as correct.
There have been two recessions in the last 20 years, and these recessions have been led by 10 different peaks in the ISM and just as many falls below 50.
Before the latest measurements, the ISM production index dipped below 50 twice since 2010: in 2012 and 2015. It is clear that a recession did not follow any of these signals. In short, over the last decades we have seen several false signals for all right.
Furthermore, while the ISM reflects the broad movements in factory activity, it can often exceed relative to actual activity both on the up and on the down. Based on a chart showing ISM's expected production activity compared to actual growth in production activity, we can see that ISM follows production fairly closely, but that there are notable periods in which it surpasses.
As an example, from 2017 to 2018, the ISM was in line with an annual production growth of 6% to 7%. Actual growth was about half as strong.
Or what about 2008? ISM maintained its break-even mark of 50 even when actual factory production collapsed.
Recently, it is important to highlight the fact that the ISM appears to be in step with other survey-based measures by the US manufacturing industry.
Markit PMI, a study with more than double the ISM, has recently improved to 51.3, the best level since April. The Markit measure has also been less volatile, tracking production more closely than ISM in 2017 and 2018. Similarly, regional production PMIs such as the New York Empire and the Philly Fed have also held up a little better.
While the decline in the ISM index is worrying to some, we believe they are overselling their cause. ISM is a nice rough and clear indicator, but using it as a basis for economic and market demand is full of risk.