For seven years from 2012 to 2018 the monthly payroll employment showed a solid trend of around 200,000 gains each and every month. If it was much above or below this trend, analysts found some excuse to explain the difference and expected the off-trend observation to be quickly reversed. So far this year most analysts continued to act as if this pattern was being repeated.
However, in August the Bureau of labor Statistics (BLS) rebenchmarked the data to more recent Census data. They announced that it would lower reported employment growth but they did not release the revised data until the August employment report last Friday. The new data shows that 2019 payroll employment was significantly weaker than originally reported. The January monthly increase was still around the old 200,000 trend. But in the seven months from February, 2019 to August,2019 the average monthly gain was only 123,000, roughly 60% of the old trend. Moreover, the 12 month moving average fell from 225,000 in January to 165,000 in August and every observation from February to August was below the still declining 12 month moving average.
Seven consecutive months should be enough to clearly demonstrate that trend payroll employment growth has fallen to a new, significantly lower trend than the old 200,000 trend.
Analysts tend to focus on the month over month change so it was understandable that they did not notice that the revised data was showing much more weakness than the old data. Remember,guessing monthly economic releases was a game created by the brokerage house, especially the bond houses, to generate volume because their earnings was much more sensitive to volume changes than other variables.