Second year in a row where they've had a nearly 900,000 person fix to the monthly data.
You would think with the error range supposed to be +/-126,000 that every now and again they would err on the upside more often, or try to fix the flaws in the establishment survey that lead us to this.