如何理解非农数据？非农是一个复杂的抽样调查后的结果，因为过程复杂，分成家计调查和企业问卷。而就业数据本身又包罗万象。我们需要从杂乱的信息中抽离出有用的成分。这完全是一个技术活。

Great Headlines Mislead On Jobs

ECRI | Jul 5, 2019

As the U.S. election cycle gets underway, expect much debate over just how strong the economy really is after last week marking its longest uninterrupted expansion in America’s history. After all, the jobless rate is near a half-century low and the S&P 500 Index is close to a record high. The bond market, though, is still signaling that the Federal Reserve will soon be forced to ease monetary policy to shore up the economy. How can that be?

A key part of the answer lies with jobs “growth,” which has been slowing much more than most probably realize. Certainly, that’s caused year-over-year (yoy) payroll growth, based on the Labor Department’s Establishment Survey, based on a broad survey of businesses and government agencies, to decline to its lowest reading since early 2018 (top line in chart). But yoy job growth, as measured by the separate Household Survey, based on a Labor Department survey of actual households, which is used to calculate the unemployment rate, is only a hair’s breadth from a five-and-a-half-year low (third line). (The data in the chart does not reflect Friday’s employment report.)

Growth in ECRI’s more comprehensive U.S. Coincident Employment Index (USCEI), which includes both those figures and more – is near its worst reading since late 2013 (bottom line). Because it subsumes data from both surveys, its verdict about overall job growth is more reliable than the others’.

But there’s even mo