Gorge-wide Labor Market Information

spac2008 Monthly Labor Market Information (Excel files only)

spac2007 Monthly Labor Market Information (Excel files only)

spac2006 Monthly Labor Market Information (Excel files only)

          2005 Monthly Labor Market Information (Excel files only)

          2004 Monthly Labor Market Information (Excel files only)

          2003 Monthly Labor Market Information (Excel files only)

Determining County Unemployment Rates
from Scott Bailey, Washington State Economist
The national unemployment rate comes from a monthly survey of 60,000 households. The households are a random sample based on Census data. With so many households, the survey is very accurate, with the exception that it is slightly biased because the Census is slightly biased, undercounting low-income and non-white households.

An important thing about the survey is that it does not ask anything about unemployment claims. Someone is counted as unemployed if they didn't work in the past month, and they were available for work and engaged in some kind of job-seeking behavior, like sending in a job application. Only about one-third to one-half of the unemployed are receiving benefits.

A second important thing is that if unemployed workers quit looking for work, they aren't counted as being unemployed, and so the unemployment rate drops. Nationally, about 1.3 percent of the labor force has dropped out of the labor market since the recession of 2001. If they were all still looking for work, the unemployment rate would be 6.7 percent, not 5.4 percent.

In the state of Washington, there are 1,300 households in the monthly survey. Statistically, this is good enough to determine the unemployment rate over the year, but not on a monthly basis--there is too much fluctuation from one month to the next. So the monthly numbers are "smoothed" in a complex statistical process. Employment is smoothed using estimated nonfarm employment from our survey of businesses, and estimated agricultural employment. Unemployment is smoothed using unemployment claims data.

In 2003, Washington's unemployment rate was estimated at 7.5 percent, with an error range of plus or minus 0.7 percent. Statisticians would say, we have 90 percent confidence that the state's rate was between 6.8 percent and 8.2 percent. So the measurement is not very precise! There are times when we see over the month changes in the unemployment rate that are not supported by other economic indicators, and do not seem to represent what is actually happening in the labor market.

At the county level, there is no direct measure of unemployment. We start by pegging each county to the state rate in 2000 based on Census data. We proceed from there on a month to month basis, using factors like population, claims, nonfarm employment, and agricultural employment. There is a formula used to make an initial estimate of county labor force data, and then these data are adjusted ("forced") to add up to the state total. For smaller counties, this forcing process sometimes does interesting things to the numbers.

Return to top | Previous Page