The ISM manufacturing index dropped from 54.9 in April to 53.2 for May, well below the 55.5 economists had been expecting.
But analysts at Stone McCarthy Research Associates spotted a huge flaw with how adjustments to the survey data were made, which distorted the headline result.
ISM manufacturing survey head Bradly Holcomb said that seasonal factory were applied to May's data that had not been intended. Stone McCarthy suggested that the ISM used April's seasonals, rather than May's. After making the correct seasonal revisions, the manufacturing sector looked a lot more promising this May, as ISM published a corrected 56 reading.
Even that however, turned out to be inaccurate. ISM have since published a second correction, with a headline reading of 55.4, in-line with what Stone McCarthy suggested the ISM result should look like. Rob Wood, senior economist at Berenberg, describes the situation as a "mess" and suggests ISM "should invest in some new quality control procedures."
Any purchasing managers' index result above 50 implies that the manufacturing sector was expanding, but the initial number signalled a much weaker expansion in the sector. The S&P500 has recovered its losses after the ISM corrected its data, restoring investor confidence.
Seasonal adjustment is a vital process to make monthly data points comparable, but things went badly wrong in this instance. Capital Economics' Paul Ashworth says "it looks like Thomas Piketty isn't the only one prone to making calculation errors."
He notes that ISM have made such mistakes before. They once released the non-manufacturing survey rather than the manufacturing one, "leading to widespread confusion". After the corrections, "it appears that manufacturing activity is still gathering momentum after the winter slowdown," says Ashworth.