Bruce Krasting puts his finger on what bugs me about the employment numbers released dutifully every week by the Bureau of Labor Statistics:
This morning we hear that total jobs increased by 80K. The reported number represents a change of 0.06%. This extraordinarily precise estimate is made based on a survey of only 1/8% of households. The error rate on the calculation is easily +/-250,000.
or, said another way, the tools that are being used to produce the numbers don’t have the accuracy, precision, or even the calibration needed. You don’t measure the distance between an electron and a nucleus or the distance between the earth and the nearest star with a yardstick.
When you look at results that are a heckuva lot more empirical, say, Social Security’s revenues from payroll deductions, the picture is somewhat different. Yeah, the values are increasing. They’re increasing very, very slowly.
Perhaps the government could save money with a smaller survey sample.
Has anyone here ever been in the household survey, or do they know anyone who has been? I haven’t and I don’t.