The March 8.2% "headline unemployment rate" declined because the labor force contracted by 164,000, and although 31,000 fewer people were employed, there were 133,000 fewer unemployed persons, because the number of people not in the labor force grew by 333,000.
Is that clear enough for you?
But apart from this silly welter of overlapping statistics, large-scale changes in employment and unemployment are always most transparently measured by the employment-population ratio, as I described it here.
What percent of American adults have a job? If you squint at the last little dip in the BLS chart above, you'll see that the answer to that question got even more discouraging last month, and it hasn't ever gotten much better since December 2009.
