Surveying Current Unemployment Data (DJI) (2024)

Introduction

The unemployment rate is an often-cited figure in macroeconomics, and it's one of the mainstays of news coverage around economics. The unemployment rate is an important "lagging" indicator for the economy.

"Lagging" in this case means that unemployment follows growth with a delay. After the economy grows, we should see an improvement in the unemployment rate. If the economy shrinks, we should see a worsening of the unemployment rate.

See how the currant line below flairs up during historic recessions like in the 1980s, 2000s, and in 2020.

Historically, the unemployment rate that is reported by the Bureau of Labor Statistics ("BLS"), has averaged around 5.7%. However, this figure alone is not the whole picture. There is a ton of data to go through regarding unemployment, with many figures higher than this for certain groups.

Before we break down demographics, we need to understand how unemployment is calculated.

U-3 vs. U-6

The currant line in the chart above shows the "U-3" unemployment data. This is the "headline figure" that is reported by news agencies and is used by the Fed to make policy decisions. However, this figure has a particular definition.

The BLS defines U-3 as:

…total unemployed, as a percent of the civilian labor force...

To count as "unemployed" for that total, survey participants must not have a job and have actively looked for work in the four weeks leading up to the survey.

This means that we leave out many workers, such as those who are discouraged from searching for work, those unemployed due to having to take care of a disabled or ill relative ("non-economic reasons"), and those who are "underemployed," i.e. having to take a part-time job when they are really searching for a full-time position.

We have other measures that we could use, and they are defined by the BLS as:

U-4, total unemployed plus discouraged workers, as a percent of the civilian labor force plus discouraged workers;

U-5, total unemployed, plus discouraged workers, plus all other marginally attached workers, as a percent of the civilian labor force plus all marginally attached workers; and

U-6, total unemployed, plus all marginally attached workers, plus total employed part time for economic reasons, as a percent of the civilian labor force plus all marginally attached workers.

The one I want to focus on is U-6, which counts U-3 plus underemployed and discouraged workers.

Note: The date range is shortened due to the U-6 data being unavailable pre-1994.

Pay particular attention to the fuchsia line above, which is our fullest count of unemployed and underemployed folks. Its average is almost double that of the U-3 figure.

The spread between U-3 and U-6 has been flat from 2019 to now. We have returned to pre-pandemic levels of discouraged and marginally attached workers, despite the unemployment rate itself being lower. The lime line below shows the change in spread between these two figures.

Below, we can see the 5yr charts for U-3 and U-6, which have had a fairly strong correlation that has been waning in recent months, with a surge of underemployed workers being reported in U-6 and left out of U-3 in the beginning of this year. That trend has mostly reversed, and we've seen the spread normalize.

This spread widening and tightening is cyclical and follows seasonal unemployment changes. What's different about the recent divergence is that it is larger and sharper than past years.

The change in the spread earlier this year, a divergence of almost 50%, shows how counting underemployed and discouraged workers changes perspective on the data. While headline unemployment figures didn't report this change, U-6 did.

Discrepancies in the Data

The BLS thankfully captures an abundance of data, which helps us further understand not only how many people, or at what rate, but which people are most affected by and contributing to unemployment figures.

Note: The data for these figures reflects the breakdown of the U-3 unemployment figure.

Education

One of the stark realities of unemployment is that it is less likely that higher educated folks are going to be laid off in bad times. It is far harder to replace a worker with a doctorate than one with less than a high school education.

Zoom in to just this past year and we can see that recently, the trends have not been bucked, with a clear gap between those with and without college or university degrees. The doctoral unemployment rate is sub 1%.

So when we consider unemployment figures and how to manage the "natural rate" (which the Fed believes is about 4% in U-3), we can see that undereducated folks with less than a high school diploma, are going to need a lot more help than those with. This should inform policy decisions moving forward as we need to consider what kinds of jobs these folks are qualified for.

Age

Predictably, the older you are, the less likely you are to be unemployed. This is a very complex issue, as age ranges overlap with tons of other factors like education, race, etc.

Today, these figures are alarmingly different, with the spread between the youngest and oldest workers growing in the last few years. The unemployment rate for teenagers is over 11%.

Note: Unemployment figures excludes students, so the 16-19yr range is only counting people who are not in schooling or another kind of training program.

The difference between those under and over 25 is very stark, and highlights how badly affected younger people are by unemployment. However, teenagers account for much less of the labor force than the other groups, so even though they are the worst hit by rate, that group has the smallest number of workers of any of them.

Currently, teenagers account for the smallest amount of unemployed people, with the largest amount being in the 25-34yr range. Despite having a "natural level" at 3.8% vs teenager's 11.7%, there are almost twice as many unemployed 25-34yr folks in the US.

Gender

The gap between men and women is fairly tight, with the groups going back and forth over the past three years.

One thing to note is that there is not a clear trend in how men and women are laid off during recessions. Men were more affected by the Great Recession whereas women were more affected by the 2020 Pandemic.

Because the rates are so close to each other and to the average, there isn't too much to glean from this other than the unemployment rates of men and women are dynamic and dependent on a lot of other factors.

Race

The government only tracks a few broad categories of race and does not track ethnicities, so this data may seem arbitrarily decided by some. Note here that this data is self-reported by participants and may depend on how the people surveyed view race (e.g. some Armenian folks report as "other" while some report as "white,") which means this data is not going to be as accurate as other categories with more clear-cut definitions like education.

What we find is that there is a consistent gap between the white and Asian groups and the Hispanic group, and another consistent gap between them and the Black/African American group.

That gap has been persistent since the introduction of these statistics, with the 2020 Pandemic being the only time where the paradigm briefly flipped and Hispanic folks were more likely to be unemployed than Black folks.

Ability

In a win for equality, we've seen the rate of unemployment between disabled and non-disabled people narrow, with a current spread lower than its historic average (the data starts in 2009).

This may be in large part thanks to the work-from-home push that was strengthened over the 2020 Pandemic, which has enabled more disabled people to work and evened the playing field, so to speak, from an ability standpoint.

The Labor Force & Employment

One of the last things I want to delve into here is the employment data. It is one thing to see how many people are looking for work and who want work, but how many people are actually working?

We can look at this in a few ways.

Labor Force Participation

This metric shows us what percentage of the population counts as "in the labor force," meaning that they are either employed or are counted in U-3 unemployment.

By this metric, it looks like there are fewer people working today than there was pre-Pandemic. This isn't necessarily true, which we can see when we look at the employment level over the same timeframe.

Employment Levels

The difference between the February 2020 and April 2024 employment levels is 1.7%, meaning that the amount of employed folks grew despite the labor force participation rate declining by 1%.

The prevailing theory is that we are seeing an uptick in both employment and unemployment, as the US population has grown over that time, primarily from immigration. Immigrant employment is far more volatile than native born employment, but the spread has narrowed since the 2020 Pandemic.

In the last year, we've seen an uptick in folks taking part time jobs for economic reasons, and a steeper uptick in folks who could only find part time work.

Conclusion

Unemployment data shows us a mixed picture, with the overall figures looking very positive for the economy, but only for some individuals. We still see mixed data for several groups. A recent uptick in employment data shows us that the trend is not just that unemployment is stable, but that it is stable while jobs are being added and more folks are getting work. Despite net job creations and an uptick in employment figures, unemployment has stayed level YTD.

This data is overall positive for the US economy and shows strong resilience in the labor force, despite some cracks in the armor that we should watch. An uptick in U-6 unemployment, outside of its normal seasonal cycle, that does not align with U-3 unemployment would be startling and would make me reconsider my bullish stance on the economy.

Thanks for reading.

John Bowman