AILSA CHANG, HOST:
President Trump has fired Erika McEntarfer, the head of the Bureau of Labor Statistics, accusing her, with no evidence, of faking a jobs report that showed fewer gains than expected. The report last week showed that the jobs market weakened sharply during the late spring and early summer as Trump's tariffs began to take effect, with the unemployment rate inching up to 4.2%. We were curious, amid President Trump's accusations, how exactly that data is gathered and analyzed for jobs reports at the BLS. So we called up Michael Horrigan, who oversaw the bureau's employment measurement programs from 2014 to 2019. He's now president of the nonprofit W.E. Upjohn Institute for Employment Research and joins us now. Welcome.
MICHAEL HORRIGAN: Thank you.
CHANG: So explain to us, just very basically, how are monthly employment numbers generally compiled?
HORRIGAN: Well, just to be clear, there are two surveys that are reported each month. One is the household survey, the current population survey. That's the one that generates the unemployment rate. And those numbers are actually received from the Census Bureau. Those household interviews end up becoming the unemployment rate.
CHANG: OK.
HORRIGAN: For the payroll survey, this is a data collection that is done by BLS of hundreds of thousands of firms every month, turned into an estimate of the number of payroll jobs in the economy. And one of the really sort of important numbers that people look at is what is the level of employment this month compared to the level of employment from payrolls from the prior month?
CHANG: And those numbers are revised eventually?
HORRIGAN: Yeah. So, you know, one of the things that happens in data collection, not just in this survey but in surveys across the government, is that it takes time to get the data. So when they get the data for the first month, they get a sample of the reports, and then they use that to estimate what the total level of employment will be. A month later, they have additional reports that are then also incorporated into the number, and then that becomes the final number. They're also revised for a slightly different reason, which is called concurrent seasonal adjustment. With an additional month of data, they also have additional information that they use in order to revise the seasonal adjustment levels.
CHANG: OK. So much of this math goes way above my head, but let me ask you the main question I have...
HORRIGAN: (Laughter) Sorry about that.
CHANG: ...Listening to you. What is the trade-off between getting numbers that are timely and getting numbers that are considered truly complete? Like, do you think it would be better for BLS to wait to release the numbers until more of the survey responses and other adjustments are made?
HORRIGAN: Well, you know, these are very vital economic data. And the Federal Reserve, for example, policymakers in general, often have to make decisions based on the data for, say, in this case, July with those numbers coming out in August. And if we waited three months for the July numbers, it may be too late to make critical decisions in terms of the movement of the economy.
CHANG: Right. Well, let me ask you, what did you personally make of the latest jobs report, the one that upset President Trump so much?
HORRIGAN: Well, I've seen revisions like this before, and it pointed out to me that we're in a shift. And, of course, everyone's looking at that in terms of the new tariff policies and other aspects of the Trump economic plan. And I sort of saw the revisions that occurred for May and June as evidence that the economy has slowed down quite a bit.
CHANG: Given all of the deliberate analysis that you just described, given all of the hard work that goes into a published jobs report, how do you think that careful work could be affected if staff at BLS are working under a threat of a president saying he will fire people if he doesn't like the data.
HORRIGAN: All right, a couple of things - first, if Bill Wiatrowski, who's now the acting commissioner, if he had been the acting commissioner last Friday, he would have released the same numbers. The staff at BLS are dedicated, nonpartisan federal employees. Their main mission is to put out numbers as accurately - as accurate as possible that help to inform policymakers on the direction of the economy. In doing so, they go to great lengths to ensure data security. Only those with a need-to-know in the employment situation release have access to the data, and it is later in the week that the commissioner sees the numbers in the analysis that the staff has put together.
It would be very difficult for a commissioner to come in and say, I don't like that number. I don't like that top-side number. I think it's not reflective of really what's going on. There would be just no circumstance under which a commissioner should or could change the top-side number based on the statistics. There are things - in response to your question - there are things that a commissioner could do that would shade things. So, for example, the BLS puts out a press release. If they didn't like the numbers that came out of the BLS, they could say, just put out the numbers on the website, and we'll have DOL do an analysis of the numbers, which to me would be a very, very large mistake.
CHANG: Well, what went through your mind when you heard that President Trump fired Commissioner McEntarfer, accusing her of basically faking the latest jobs report?
HORRIGAN: Well, as a 33-year veteran of the federal service, and she is a veteran of the federal service, my heart went out to her. I mean, I was stunned. It was 100% unjustified. It was politically motivated. And, quite frankly, I think the fear that people have is that the reason he gave, which was false, is actually something that he's going to try to do in the new appointee going to BLS. He accused her of falsifying the numbers, which, as I said, is very difficult to do. But I think people are fearing, what is this new appointee that's going to head up the BLS going to do if they have fealty to President Trump?
CHANG: Well, that's the thing. What happens if there is an unusually good jobs report next month? Do you think people are actually going to believe it?
HORRIGAN: Well, I hope that people will understand that the bureau has maintained the integrity and quality of the data, and that a good jobs report, if then followed by a poor or a mediocre jobs report, has been done in the same professional way that has been done in the past and that they can testify to the fact that any new commissioner has not interfered with that process.
CHANG: But if the credibility of the jobs report erodes over time, what's the fallout if people don't have confidence in the official economic readings from the U.S. government? Who would suffer?
HORRIGAN: Everyone suffers. And it's not just BLS numbers. If there is an attack, a political attack, on the credibility of the numbers coming out of the federal system, that attack would not be limited to BLS. It would include GDP, population estimates from the Census Bureau, just about every kind of statistic that comes out. Having objective, nonpartisan data, just boring data that is put together with high quality and integrity - that is needed for our economic and political debates. And to threaten that just makes us poor as a country.
CHANG: That is Michael Horrigan, former head of employment measurement programs for the Bureau of Labor Statistics and now of the Upjohn Institute. Thank you very much.
HORRIGAN: You're more than welcome. Thank you.
(SOUNDBITE OF BADBADNOTGOOD AND GHOSTFACE KILLAH SONG, "STREET KNOWLEDGE (FEAT. TREE)") Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.
