No economic indicator — whether of price levels, employment, output or whatever — is flawless.
No indicator, by itself, gives a full picture of what it is measuring. All indicators involve some assumptions that to the uninitiated seem stupid. The same probably is true about indicators of a human’s health. But in economics as in medicine, it is vital to be able to measure variables, even if imperfectly, that inform on the state of things.
All this is in response to communication from readers about last week’s column, which focused on the low unemployment rate in our state, which is coupled with perplexing changes in Minnesota’s labor force. Today’s column was to deal with two issues: on the difference between two differing sets of unemployment numbers and on changes in the fraction of all adults who either have jobs or are looking for them.
However, so many emails raised questions about the basics of measuring unemployment that it seems best to leave that second question for later.
The first point to make is that the employment rate is not, as many apparently believe, dictated by the White House. After decades of hearing this assertion, it is wearying. It is discouraging that such a large fraction of the U.S. population believes that such labor statistics and other indicators, such as price indexes, are corrupt.
They are not. The statisticians in the Bureau of Labor Statistics take pride in what they do. If someone tried to force them to cook the books, it would be leaked to the media almost instantaneously. The system is even set up so that those surveying households about who has jobs are insulated from the group surveying employers about how many people are on payrolls.
But nevertheless, during Republican administrations I hear people on the left complaining that the president is dictating phony numbers from the Oval Office. When a Democrat occupies the White House, it is those on the right who see a scam.
So let’s proceed to the limitations of the most commonly cited unemployment rate. This is called the “U3 rate” because that is the table in which it appears in the monthly statistical release.
Remember that every person in the country past his or her 16th birthday falls into one of three categories. If they have a job they are “employed.” If they don’t have a job, but are taking concrete steps to get one, then they are “unemployed.” These two rubrics combined compose the “labor force.”
The third category is those who do not have jobs and are not trying to get them. Typically, it is because they are retired, disabled, in school full time or working only in the home. These people are “out of the labor force.”
One technical caveat is that the 16-and-over group is also reduced by those in the military and those who are institutionalized, which most commonly means being in prison or a mental hospital. Take these away and you have the “civilian, non-institutionalized population 16 and over.”
Assigning specific criteria to these general definitions is not easy, however. Just what constitutes being employed? If you worked only one hour for pay in the week in question, were you really employed? If not, then what should be the threshold? Five hours? Ten? And why ask about only one “reference week” out of a month? Why not two? If you covered a multi-week span within a month, you would get somewhat different results.
What if you have a master’s degree in engineering but are driving for Uber? Or what if you would like to work full time, but your manager only gives you two six-hour shifts per week?
And what constitutes looking for a job? What if you filled out applications a few weeks ago or are going to a job fair soon, but did nothing during the reference week? What if you really want to work as a systems analyst for Target, but you don’t interview well so you do hacking and play video games while living in your parents’ basement?
For any of these variables, one can come up with some example that most people would agree should be covered and another that is ridiculous. And there are many possible circumstances that fall in between.
Uniformity over space and time in making such economic measurements is vital. One cannot leave it to the subjective judgement of individual interviewers of whether someone is employed or is actively seeking work or not. Arbitrary, but clear and consistent, is better than subjective if you want to see how things progress.
EMTs know that you can get different blood pressure readings on a patient using different cuffs. Which cuff is giving the “correct” pressure? Take readings at intervals with the same cuff, however imperfectly, and you get more useful information on what is happening to the patient than if you switch from one cuff to another and another.
The most commonly cited U3 unemployment rate has strict criteria for what constitutes being employed and what constitutes looking for work. There have been refinements in survey and tabulation techniques since the series was started in 1913, but the resulting numbers are comparable back over the whole range of data.
If one is surveying thousands of households for this data, it is possible to ask additional questions to determine if people are getting as many hours as they want or are using skills in which they have previous experience or education. And one can identify those who genuinely want a job, but did not meet the specific criteria deemed to be trying to get work.
So the BLS also tabulates and publishes data on these “discouraged workers” and on those who are “underemployed” and even those who are “marginally attached to the labor force.”
Add these to the headline U3 unemployment rate and one gets the “U6 rate.” It is always higher than the U3 rate. And the gap between them varies with the business cycle. In a recession, the U6 rate is much higher than the U3 rate. During a boom, when employers are avidly seeking new hires, the gap shrinks.
One often reads breathless announcements that the “real” unemployment rate is much higher than the headline one. But which really is “correct”? The answer is that they give complementary information about labor conditions. Each in itself gives an incomplete picture. Moreover, just knowing those two gives less information than if one also knows the specific breakdown of those with “fewer hours,” “less skills” or who are “discouraged workers.”
We would get different results if we set the age threshold at 18 rather than 16. The first is a better reflection of the division between youth and adulthood in 2016 than it would have been in 1913. And should we change it to 20 or 22 since many people now get post-secondary education?
The answer is that it is better to keep the same standard in place and collect enough information about age and education so that we can make additional tabulations that answer key questions. The BLS’s monthly Labor Situation report contains a wealth of information, but few people ever look at it, especially those prone to see incompetence and corruption in everything government does.
The difference between the narrow U3 rate and the broader U6 rate is a part of what is going on in Minnesota labor markets right now, but a small one. This difference is largely “cyclical.” That is, it depends on changes in the business cycle of economic expansion and recession. The “participation rate” or percentage of the total population that is “in the labor force” is a “structural” question, one that flows from longer-term changes in societal values about education and women’s roles or from demographic shifts like the baby boom.