U.S. Factory Activity Slowed in April. Or Maybe It Picked Up.

Factory activity in the United States slowed in April. Or maybe it picked up a bit. It depends on who you ask.

The Institute for Supply Management’s manufacturing purchasing manager index came in at 52.8, the weakest report in nearly two years and far lower than the 55 reading economists expected. IHS Markit’s manufacturing purchasing managers index came in at 52.6, an unexpected improvement from last month’s reading.

Importantly, both indexes indicate that manufacturing continued to expand in April. And both paint a similar picture of slower growth than last year, indicating that the manufacturing sector is weighing down economic growth at the start of the second quarter.

“Although the PMI ticked higher in April, the survey remains consistent with manufacturing acting as a drag on the economy at the start of the second quarter, albeit with the rate of contraction easing,” IHS Markit economist Chris Williamson said in a statement. “Historical comparisons indicate that the survey’s output gauge needs to rise above 53.5 to signal growth of factory production. As such, the data add to signs that the economy looks set to slow after the stronger than expected start to the year”

On Twitter, Williamson indicated that what is going on in the conflicting reports is that the Institute for Supply Management’s survey is simply playing catch-up with the slowdown his own survey detected earlier this year.

Both the indexes are considered “soft data” because they are based on questions asked of manufacturing executives–the so-called purchasing managers–rather than hard data on actual production, sales, prices, and hiring.

The divergence between the two is in their direction: the ISM survey shows slowing growth and the IHS Markit survey shows a bit of a rebound in growth.

The Institute for Supply Management (ISM) index is older and more influential. It is used by many US economic nowcast and forecast models, including those of the New York and Atlanta Federal Reserve banks. The Atlanta Fed’s nowcast for second quarter GDP growth, for example, fell to 1.2 percent from 1.3 percent growth after the release of the ISM figure.

The IHS Markit index, however, may have the better track record. Last year, IHS Markit’s Williamson found that the IHS Markit data over the past ten years had an 89 percent correlation with the official “hard” data while the ISM’s correlation is just 81 percent.

The two indexes diverged at the end of last year, with the IHS Markit index signaling a relatively robust pace of expansion while the ISM survey showed an abrupt slowdown. As it turned out, the IHS data was sending a misleading signal about the strength of the economy and manufacturing in the fourth quarter.

Last month, however, they switched places. The IHS Market index dropped to 52.4, down from 53.0 in February and below the 55.6 reported a year prior. The Institute for Supply Management reported that its index jumped from a one year low of 54.2 in February to 55.3 in March.

The divergences can be confusing or even frustrating for analysts. The two indexes are based on more or less identical data and are composed in very similar manners. They are composite indicators derived by polling executives at manufacturing companies. These folks get asked five questions about output, new orders, employment, inventories and suppliers’ delivery times.

But they weight the answers differently. The ISM index is a straight average of its five component series while the IHS Markit index weights the average so that forward-looking components carry a higher weight.

Another difference may be in the size of the companies surveyed.  According to Williamson, evidence suggests that the divergence in indexes is most likely due to the exclusion of smaller companies in the ISM panel.  In other words, the ISM data has a bias toward large companies serving global markets so that it is gets pushed around by economic trends outside of the U.S.

Williamson also argues that equity market trends may weigh on the ISM figures, since the purchasing managers of big companies may feel better or worse about economic conditions depending on how their own company’s stock is doing. And while both indexes use seasonal adjustments to their data, these adjustments are not identical and can cause the results to diverge.

One of the strengths of the IHS Markit numbers over the past decade has been in identifying turning points in the economy. According to Williamson, “the IHS Markit data have identified every major turning point in the US manufacturing cycle over the past ten years (based on using three-month moving averages of the survey and official data to identify turning points). In contrast, the ISM survey data have missed recent turning points.”

Of course, Williamson works for IHS Markit and that might weigh on his opinion that his company’s survey is more accurate. Nonetheless, his analysis of the figures and data appears to be sound.

It was not just the headline numbers that diverged in the most recent reports. The more established ISM survey showed a steep drop in new orders and exports. The IHS Markit survey showed new orders and exports improving.

Both indicated a slowdown in hiring, however. That is consistent with Wednesday’s report from ADP and Moody’s Analytics, which showed manufacturing adding just 5,000 jobs in April.

Importantly for those watching the effects of tariffs and trade disputes, both surveys also showed that price pressures eased, indicating that tariffs are not raising prices for consumers. The IHS Markit survey also showed that input cost inflation eased further.