From IndustryWeek.

by Steve Minter

Feb 03, 2014

There is no shortage of data sets to track if you follow U.S. manufacturing, but one of the most widely quoted (just check today’s headlines) is the Institute for Supply Management’s monthly PMI report. Today’s PMI registered 51.3%, a decrease of 5.2 percentage points from December’s reading of 56.5%. Any reading above 50 indicates expansion in the sector.

Along with the broad main measure, the index for new orders dropped sharply by 13.2 percentage points to 51.2% and the production index fell to 54.8%, a decrease of 6.9 percentage points compared to December’s reading of 61.7%.

A number of the supply chain managers quoted by ISM indicated that weather may be at least partly to blame for the dropoff.

But Alan Tonelson, an economist with the U.S. Business and Industry Council, said the PMI suffers from a more fundamental problem – its forecasts don’t line up with the actual data that follows from the Federal Reserve and the Bureau of Labor Statistics.

In, Tonelson says recent ISM reports have painted a picture of robust manufacturing expansion when the corresponding production data from the Federal Reserve has been more modest. At other times, the ISM reports have been out of step with the government data.

“For example, in 2011 and 2012, the Fed reports that U.S. manufacturing output rose in inflation-adjusted terms by a nearly identical 3.675 and 3.68%, respectively. But during that period, the annual ISM headline fell from 55.2 to 51.69. More puzzling, manufacturing’s real output growth last year fell to 2.77%. But the ISM headline actually rose to 53.7.” []

Tonelson writes that ISM’s forecasts on manufacturing employment are no better.

“According to the BLS, American industry began adding net new jobs on an annual basis in 2010, with the 119,000 increase being accompanied by an ISM employment reading of 57.5. The following year, however, net job creation in manufacturing rose to 207,000. But the ISM employment reading fell to 52.9.”[]

In Tonelson’s view, ISM does a good job of capturing “huge swings” in production and employment but is much less accurate when it comes to smaller changes.