Is the BLS Grossly Underestimating Our Unemployment Rate?

Published on

During the decade-long Great Depression, friends often greeted each other by asking, “Are you working?” After all, anywhere from ten to twenty-five percent of the labor force was officially out of work – and the true monthly unemployment rate was probably substantially higher.

Get The Full Ray Dalio Series in PDF

Get the entire 10-part series on Ray Dalio in PDF. Save it to your desktop, read it on your tablet, or email to your colleagues

Q1 2020 hedge fund letters, conferences and more

The Actual Monthly Unemployment Rate Is Much Higher

We are now beginning to see equally high unemployment rates. And even though the Bureau of Labor Statistics has long enjoyed a well-earned reputation for providing accurate estimates of our unemployment rate, it is becoming very clear that our actual monthly unemployment rate is much higher than the official rate.

Let’s take, for example, the BLS estimate April estimate of just 14.7%. To come up with that number, the BLS conducted a phone survey of 60,000 families in mid-April. The respondents were asked up to three questions: (1) Are you currently working? If the answer is “yes”, then that person is counted as employed.  If the answer is “no,” (2) Did you work at all in the last week, even one day? If the answer is “yes,” then that person is counted as employed.

For those who have not been working, the BLS has one more question: Did you look for work during the last four weeks? That might include going to an employment agency, sending out resumes, going on job interviews, and doing computer searches for job openings. If your answer is “yes,” then you’re classified as unemployed. But if your answer is “no,” then you’re not counted in the labor force.

How The BLS Calculates The Unemployment Rate

As we shall see, how people answer this third question is extremely crucial to how the BLS calculates the monthly unemployment rate. But to lend some perspective, over the last eight decades, almost no one has ever accused – or even suspected – that this statistics-gathering agency has not played by the rules.

In fact, all they do is gather the data each month, and plug it into this simple formula:  The unemployment rate = the number of people employed divided by the number of people in the labor force.

We’ve seen how the BLS got the data for the unemployed and the number of employed. And the labor force is the sum of the employed and the unemployed.

In February of this year, the BLS reported an unemployment rate of just 3.5%. It estimated that there were 5.7 million unemployed and a labor force of 165 million. The labor force was the total of the employed (159.3 million) and the unemployed (5.7 million).

So in February, BLS calculated our unemployment rate by dividing the number of people unemployed by the number of people in our labor force:

5,700,000/165,000,000 = 3.5%

The BLS Is Using Really Old Standards

The reason the BLS measure of our unemployment rate is so much lower than we might have expected in April and May is because of the millions of people without jobs  who are not considered unemployed because they had not actively sought employment in recent weeks. And yet, the BLS is applying the same standards they have been applying for the last eight decades.

During relatively normal economic times, it is reasonable to consider a person not working is not in the labor force if she or he has not actively sought employment during the last four weeks.  But these are hardly normal times.

Millions of people who are ready, willing, and able to work are afraid of exposing themselves to the pandemic. And millions of others have not been actively seeking work because they have little hope of of actually finding a satisfactory job.  And then too, millions more are receiving as much – or even more – in unemployment benefits than they could earn at a new job.

The folks running the BLS are well aware of these arguments. But just imagine the political reaction if that agency changed the criteria for classifying people not actively seeking work as “unemployed.”  More than ten million people – instead of not being counted in the labor force – would now be classified as unemployed. This would push up the unemployment rate by at least seven percentage points.

Changes In The Measurement Of Unemployment

While a very high unemployment rate would certainly better reflect these unusual economic times, the federal government bureaucracy (aka the Deep State) is too well attuned to the political winds to even consider making any major changes in its measurement of unemployment, however necessary.

Please consider that we have a president who has consistently credited himself for having created “the greatest economy in history.” Can you imagine his reaction if the BLS were to upwardly adjust our monthly unemployment rates to better reflect current labor market conditions?

Here’s my own very rough estimate of the May unemployment rate. There were almost 40 million unemployment insurance benefits claims filed between mid-March and mid-May. At least another 10 million unemployed people were ineligible, did not bother to apply, or were unable to file. There were also 5.7 million people who had been unemployed before mid-March.

That’s a total of close to 60 million out of work in mid-May. How many of them should actually be counted as “unemployed”? Let’s assume that about 5 million either were called back to work or found new jobs between mid-April and mid-May. And that an additional 10 million were not really interested in going back to work.

So that leaves us with about 45 million unemployed in mid-May. If 115 million people are working, then our labor force totals 160 million.

We can now calculate the unemployment rate in May:

Unemployment rate = Unemployed/Labor Force = 45,000,000/160,000,000 = 28.1%

Again, this is just a very rough estimate of May’s unemployment rate. At 8:30 on the morning of Friday, June 5th, the BLS will announce its official estimate ( It will surely be substantially lower than 28.1. In fact, it will probably be around 20%.