Posted by: Josh Lehner | May 14, 2014

Estimating Historical U-6 in Oregon

A lot of economic commentary in recent years has focused on economic measures beyond the headline unemployment rate, the so-called U-3 or official unemployment rate, with good reason. Hours worked were cut, the job ladder is/was broken, job polarization pushes some into underemployment and the like. Besides, the unemployment rate can go up for good reasons and decline for bad reasons. One popular metric to follow has been U-6 which not only includes those unemployed but also the marginally attached workers (who want a job, but are not currently looking so technically not in the labor force) and also those working part-time for economic reasons. U-6 is one of BLS’ Alternative Measures of Labor Underutilization and nationally goes back to 1994. At the state level, U-6 only goes back to 2002.

What follows is predominantly a utility post that will be leveraged a few different places moving forward, both from our office and at least one other state program. While hopefully it is of interest to you as well, the following provides both a quick overview along with methodological details in terms of creating historical estimates of U-6 in Oregon back to 1976.

First and foremost it cannot be repeated enough, these are just historical estimates made by our office and are not official figures, nor sanctioned by BLS or the Oregon Employment Department. BLS redesigned the Current Population Survey (CPS) in 1994 and also redesigned their alternative measures of labor underutilization at that time, rendering historical comparisons challenging, at best. See this 1995 BLS article for more details (HT: Tracy Morrissette at Employment). With that being said, these various measures of unemployment or labor underuitlization do tend to move closely over time, so while the absolute level of unemployment and/or underemployment changes, the general trends do not. In fact, as that 1995 BLS article says:

…it is important to note that there is little “value added” analytically in tracking the alternative measures over time. While it is true that each indicator provides a different point estimate of “unemployment” all seven measures have essentially moved in lockstep across the business cycle.

Pre-1994 redesign, BLS actually produced a U-1 through U-7, which are fairly similar to the more familiar U-1 through U-6 today. However these measures do differ in a few important ways, particularly in the way BLS measures working part-time for economic reasons and also the differences between marginally attached workers and discouraged workers. Nevertheless, based on careful analysis of the underlying CPS sample data, here are our office’s unofficial estimates of historical U-6 in Oregon. See below for methodological notes.



The official U-3 and U-6 figures are from the Oregon Employment Department. The data used for the historical U-6 estimates are from the Integrated Public Use Microdata Series, Current Population Survey made available by the University of Minnesota. Specifically the data are from the Annual Social and Economic Supplement which occurs each March.

Calculating U-3 based on the sample data is straightforward in terms of those categorized as unemployed and the total labor force over time. The more complicated estimates pertain to those working part-time for economic reasons and the marginally attached workers. Cross referencing the question of whether an individual was working part-time for economic reasons with the question of why an individual worked part-time yields that the part-time for economic reasons is the sum of those who could only find part-time work, job started/ended during the week, job terminated, material shortage, new job started, plant or machine repairs, seasonal work and slack work or business conditions. Applying this sum historically provides an estimate needed for creating U-6. To estimate marginally attached workers, I filtered the data such that I was looking at individuals not currently in the labor force but wanted a job and then cross-referenced these results with the question of how many weeks in the past year the individual had looked for work. Matching estimates of the marginally attached historically is the most challenging and open to the largest estimation errors in this work. Thankfully, in terms of importance in the overall U-6 measures, those working part-time for economic reasons are about 5 times larger than the marginally attached workers.

Lastly, given that these are sample data for just one month each year, the data can be noisy. To smooth the numbers a bit further, I used the ratio of sample based U-3 to actual published U-3 and applied that to the sample based U-6 estimates.


  1. […] Right now the Oregon economy continues to improve and is even a bit above our office’s forecast. As we continue to work on the next outlook, it can be useful to look back and compare business cycles to see what, if anything, we can learn from past experiences. Oregon has had two severe recessions since World War II, the early 1980s when the timber industry restructured and the Great Recession. In terms of employment, today the state largely tracks the experience seen during the 1980s, and also in terms of unemployment, as shown below (more on historical U-6). […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s


%d bloggers like this: