Today's U.S. employment report is in many ways eerily similar to last months, with continued decline in nonfarm payrolls despite continued government job growth and despite the fact that many jobs are simply assumed through the faulty birth/death model (astonishingly, they actually increased the number of assumed jobs, from 53,000 per month last month to 64,000 per month this month). Also, like last month, job growth in previous months was downwardly revised. Excluding the phantom jobs assumed by the birth death model, private sector job growth is now negative on a year over basis and is shrinking at a rate of 150,000 to 200,000 per month, considering that published private sector employment growth was a negative 100,000 (roughly) both of the previous months and considering the 64,000 phantom jobs assumed. This estimate is also confirmed by the household survey which shows negative job growth on a year over year basis, including
government jobs. Excluding government employment, the decline is of course even bigger. For a more detailed reasoning around these issues, see the analysis I did after the previous month's release
The main difference between these two releases was that the unemployment rate rose sharply in Mars, after declining slightly in February. But that is of lesser importance as employment fell during both months, with the difference in the change of the unemployment rate being caused by a falling participation rate in February and a rising one in Mars.
The overall conclusion is that the employment numbers confirm a U.S. recession, particularly in manufacturing and construction, while the government sector seems to be the only one that is growing.
UPDATE: Menzie Chinn at Econbrowser has a good illustration
of the effects of the revisions on employment levels and changes.