Skip to content


Caroline Rivers and Parker Skinner

While the terms employment and unemployment may seem straightforward, not all without a job are counted as unemployed. To be considered unemployed, one must not only be without a job, but also seeking a job and available to work. For example, the retired and disabled are not counted as unemployed-- one group is not seeking jobs, while the latter is not available to work.

Additional groups represent how the unemployment rate can be distorted by either overstatement or understatement. The first is discouraged workers, jobless individuals who are not counted as unemployed because they have not been actively seeking a job in the past 4 weeks, typically due to a lack of success in their recent job searches. These workers fall under the umbrella of marginally attached workers, those who want a job and have searched for one in the recently past, but are not currently looking. Underemployed workers, those who want a full-time job but are working part-time due to a lack of full-time positions, are also not counted in the unemployment rate.

Below is a graph representing the difference between the reported unemployment rate versus the unemployment rate when these uncounted groups of marginally attached workers and underemployed workers are included.

When compared to the chart of Civilian Unemployment rate shown below, you can see the rate is much higher when these groups are considered, an example of how the unemployment rate can inaccurately reflect the state of the economy. For example, in June 2009 the civilian unemployment rate is 9.5%, while it is 16.5% when marginally unattached and underemployed workers are accounted for.

Unemployment tracked from the end of 1989 until December 2017. This graph shows the fluctuation of the Unemployment Rate.

Here the ever-changing rate of unemployment is presented using data from the end of 1989 until December 2017. The concept of natural rate of unemployment is also shown as the rate has never and will never fall to 0% (in fact over this time period it rarely drops below 4%).

This next graph presents the fluctuating GDP growth rate over the same time period as the above graph. As you can see the growth rate of GDP is inverse to the unemployment rate, i.e. when the GDP change is negative the unemployment rate rises. When the GDP change is positive yet unemployment still rises it is known as a jobless recovery. This phenomenon was evident in 1992, a period of economic growth also known as a growth recession.

17 thoughts on “Unemployment

  1. ingramk20

    The recession experienced at the beginning of the 2000s was a result of the dot-com bubble, correct? Could the high unemployment rates that continued despite recovery of the economy be attributed to structural unemployment? The book describes a phenomenon known as a skills mismatch, meaning in this case that there were considerably fewer jobs in tech while there was a large supply of skilled tech workers looking for jobs. I don't know a lot about this recession so this is just speculation but I would be curious as to the driving factors behind the jobless recovery.

    1. the prof

      To get at your suggestion we could look at whether localized [and thus presumably structural] unemployment remained high after 2001. I think the answer is "yes" - you can go to FRED Graphs to do an informal check.

  2. bullr20

    Its interesting to think of "jobless recovery" in the context of the budding tech industry and the possibility of automation. From all the articles I've read, the "Will a robot take your job?" question seems to be rooted in actual industry trends. Some economists say that in an increasingly automated economy, we have the opportunity to restructure our workweek so as to increase free time while maintaining a living salary for individuals and a healthy economy on the whole... Not sure if this would be possible, but I would love to hear more.

    1. minsong20

      Increasingly automated jobs, I can only believe would reduce employment. I think, in the short term, this would be very apparent in the underemployed chart as people become less satisfied with their jobs.

      1. the prof

        Firms surely automate to save costs, which lowers product price and increases quantity. That offsets the job loss component, while potentially at the same time upgrading the skill content: setting up machines, versus feeding parts into a machine.

  3. bernsteinl20

    I would be interested to see how free lance work could be factored into unemployment rates and whether it is factored in correctly. Do people who do some short term work for companies consider themselves employed, even when they do not actively have a client? I'm curious to see if this leads to an under or over representing of the unemployment numbers. This is especially pertinent because this style of employment appears to be growing.

    1. skinnerf20

      I'm not sure where short term workers fall in terms of unemployment, however, if they are considered the same as part time workers, they are fully employed at least for the time that they are working. Free lance workers, such as writers, would be deemed independent contractors and therefore fall under the employed category.

    2. Molly Mann

      Additionally, I would also be curious in how having multiple jobs to make ends meet or how side jobs with under the table payment influence unemployment and underemployment rates. I think that employment rates try to put a black and white answer on a grey area. While there are many types of ways to make a living, I don't think the employment rate accurately represents what's really going on in the job market.

  4. dodsonm20

    Is taking account for the groups of marginally attached workers really helpful? I get that it more accurate depicts the unemployment rate in one sense but at the same time it changes almost at the exact same rate as the civilian unemployment rate. If we're trying to identify changes in the economy and both rates change with the same magnitude what is the true advantage of using one or the other?

    1. calihanj20

      I think including marginally attached workers and underemployed workers can be beneficial because it can show how close a country is to labor equilibrium. For example, a 5% unemployment rate is considered to be close to the equilibrium, and you could have two countries that both have 5% unemployment rates. When we add in marginally attached workers and underemployed workers, if one country is at 12% and the other is at 7%, we can see that the 7% country is closer to labor equilibrium.

      1. dodsonm20

        I understand what you mean. I was more wondering why one would be preferred over the other when considering changes in unemployment. Since both change at similar rates I don't think the difference is that important when trying analyze the change specifically.

  5. dugganj20

    I think it is important to include workers who are classified as marginally attached or underemployed. Including these workers gives a more accurate statistic of who is without a job and wants one, whether or not they have given up looking. I think the difference between the unemployment rate with excluding vs. including the marginally attached workers is substantial enough to always recognize.

  6. prochniaka20

    The unemployment rate is often cited when talking about the economic position of our country. However, it is interesting to look at all the different factors considered when determining what it really means to be unemployed. It shows that everything is not always what it seems when it comes to reporting about the economy.

  7. samuelm20

    One issue I have with unemployment measures is that they will never accurately describe situations- individual stories of peoples- in numbers. In our attempts to simplify our economy to quantifiable data, we create pockets of excluded populations. The common example would be “unemployed” workers who are not interested in finding a job, but rather continue to go in for interviews to continue receiving unemployment checks. Although this sheds a negative light on unemployment checks, I give this example to shed light on other groups of people which may be excluded, especially those who may need it. Where is undocumented work accounted for?

  8. scottm20

    The debate of how to properly quantify the unemployment rate is an interesting case. As explained in the blog, the exclusion of certain groups—discouraged workers, who have given up searching for a job, marginally attached workers, who would like to work but are not currently searching for a job, and underemployed workers, who work part-time because they cannot find full-time jobs—can distort the true value of the unemployment rate. However, to keep this in check, the Bureau of Labor Statistics calculates U-1 through U-6, six measurements of unemployment, to quantify different aspects of unemployment. While the aforementioned exclusions occur with U-3, the most commonly used rate, other unemployment measures take into account these exclusions. There is no "correct" way to define the unemployment rate, its vast range of variables and volatility make for a complex value regardless of how you calculate it.

  9. the prof

    A good set of comments! Indeed any single number will provide a partial picture, even if we stick to the aggregate US economy. And the aggregate is a sum of regions in which we live and work: it's quite different to live in a local area that's booming versus a city in decline.


Leave a Reply