ADP's Huge Miss Employment Numbers Versus Government Numbers, Can You Trust Either?

There was a big hoopla this week when the ADP numbers were released showing massive employment gains. Those gains ended up being nonexistent in the figures released by the government two days later. Who do you trust? Can you trust either? Normally, I'd tend to trust more a private organization that reports real numbers, but it turrns out that is not the case at all. Both entities estimate numbers and both make seasonal adjustments, and neither can be perfect.

Bloomberg has an article on this today. It reports economists from JPMorgan Chase & Co.’s and Goldman Sachs Group warning to beware of December

“I don’t think ADP has been particularly useful in forecasting” the government’s employment figures, Feroli (JPM) said in a telephone interview from his New York office. “ADP has historically had problems in December because of a variety of technical reasons.”

They talk about the "purge effect", when employees sometimes remain on company records until December,  regardless of when they are dismissed or quit, when businesses update, or purge, their figures with ADP. "The Roseland, New Jersey-based paycheck processor estimates this change when adjusting its data for seasonal changes and the company’s projection may have been too large this year because there were fewer firings than in previous years, economists said".

The Labor Department said that 113,000 workers were added to payroll last month. This was 184,000 fewer than ADP predicted - just two days earlier. According to Bloomberg, this is the largest ever gap ADP has had on an initial private payroll forecast for December and it was a ‘Big Miss’.

Why it matters so much? "ADP’s December estimate of 297,000 prompted 22 economists, including forecasters at Morgan Stanley and Deutsche Bank Securities Inc. in New York, to raise their forecasts".