Moody's Analytics chief economist Mark Zandi said technical issues may have made ADP's report on the April job market look much stronger than it actually was.
Earlier Wednesday, ADP's report with Moody's said private sector hiring hit an eye-popping 275,000 in April, just about 100,000 more than expected.
But Zandi said on CNBC shortly after the release that technical factors may have inflated the number.
He later detailed three components that go into the ADP number that were unusually volatile or out of line. One was a very low reading of ADP's own payroll data. Another was the very low level of unemployment claims during April, which reached a 50-year low. The third was the high level of volatility in the Bureau of Labor Statistics' recent monthly jobs data.
In ADP's press release, Zandi said the number overstates the strength of the economy.
"Bottom line, I don't think it's going to come in at 275,000. My sense is its going to come in pretty close to consensus 175,000 to 200,000," he later said in a telephone interview.
ADP's payroll number sometimes tracks close to the actual government nonfarm payroll report, and sometimes not. Still, it is monitored by Wall Street as one metric to watch ahead of the monthly jobs report, typically released two days later.
Zandi said he doesn't normally qualify the number. "I did that because there are three technical issues going on that conspired to pump it up," he said.
"We take the ADP number, other economic variables and use that to estimate the BLS number. The ADP number came in unusually soft compared to what we've been seeing in recent history. Relative not only to last couple of months but last couple of years," he said.
At the same time, very low jobless claims in the survey week counterbalanced the weak ADP payroll number. And that inflated the number. Another factor is the 275,000 ADP number includes some actual history of the government reports.
"When trying to predict this months' change we also look at recent history. If you're in a period of strong job growth, that will affect your estimate. If you're in a period of weak job growth, that will affect your estimate. And if you have a period of extraordinary volatility, that's also going to show up in the number, and that pumped up the [ADP] number," he said.
Zandi said there are often technical factors surrounding the number, but this was an extreme case.
"It's an estimate. I could adjust it, but I didn't' want to do that. I didn't' think that was appropriate," he said. "I thought it was important to provide context."
Nonfarm payrolls, released by the government, have been extremely volatile this year. January job gains totaled 312,000; February was 33,000 and there were 196,000 jobs created in March.
As for the government's report Friday, Zandi said it may contain its own quirks. For instance, the government report theoretically could be inflated by census workers added to the government payrolls in April or May.
"That could be a couple hundred thousand people," he said.