>I have question for the economists on the list. According to my local
>newspaper there are 2 unemployment surveys used, one surveys businesses to
>payrolls the other is a household survey asking families if they have an
>unemployed member. If I'm not mistaken, the employer survey is the one thats
>taken seriously and used by the department of labor. But recently the other
>is being used by republicancs because it gives them better employment
>numbers. They say that this is the accurate count because residential
>surveyes also count the number of people not looking for work because they
>have started their own bussinesses. Which survey is most accurate and why?
Minor correctioN: there are two employment surveys, not two unemployment surveys. Unemployment is measured as part of the household survey.
I've posted this to the list before, but it never hurts to do it again. There are to surveys, one of about 60,000 households (known as the household survey) and another of about 300,000 employers (known as the establishment or payroll survey). Though 60,000 is a very large sample for a monthly survey, it's still small compared to the 100+ million households in the U.S. the establishment survey covers a much larger share of its universe - the establishments surveyed are responsible for about a third of U.S. employment. Also complicating the household survey is the fact that the raw survey figures have to be converted to national numbers using demographic weights and population estimates derived from the Census, In early 2004, employment estimates in the employment survey were revised down, because Census decided that it had been overestimating immigration in the much less welcoming post-9/11 world. For all these reasons, the Bureau of Labor Statistics always says that the establishment survey is a much more reliable estimate. It's also "benchmarked" every year against unemployment insurance records, which have a near-100% coverage of the employment universe. The benchmark revisions are generally quite small, suggesting that the monthly figures are quite accurate.
The establishment survey doesn't include the self-employed. That can be a problem, but it's usually thought that lots of un- or semi-employed people call themselves self-employed to avoid the stigma of joblessness.
You're right that the Reps have been using the household employment numbers because they look better. But most serious analysts use the establishment survey to measure employment trends (overall and by industry) and the household numbers for demographics (employment by sex, race, educational level, broad job category.
Doug