This is a very special episode of the PolicyViz Podcast. I’m joined by Erica Groshen, former Commissioner of the Bureau of Labor Statistics to talk about the latest jobs report and what the numbers really mean for the state of the economy, including why revisions are an essential part of getting the data right. Dr. Groshen explains how the BLS produces its trusted statistics, the commissioner’s role, and what signals she watches for to spot potential recessions. We also discuss the importance of protecting federal statistical agencies, ensuring trust in their data, and what the future might hold for the BLS. It’s a fascinating conversation that connects technical detail with big-picture implications.
Resources
Check out the BLS website.
Guest Bio
Dr. Groshen is Senior Economic Advisor at Cornell University—ILR and Research Fellow at the Upjohn Institute for Employment Research. From 2013 to 2017, she served as 14th Commissioner of the US Bureau of Labor Statistics, the principal federal agency responsible for measuring labor market activity, working conditions, and inflation. She currently serves as a member of the Federal Economic Statistics Advisory Council and the Committee on National Statistics of the National Academies of Science, Engineering and Medicine, and as chair of the STARs Insights Advisory Committee for Opportunity@Work. Before that she was Vice President in the Research and Statistics Group of the Federal Reserve Bank of New York. Her research centers on employers’ roles in labor market outcomes. She coedited Improving Employment and Earnings in Twenty-First Century Labor Markets, from RSF: The Russell Sage Foundation Journal of the Social Science, co-authored How New is the “New Employment Contract”? from W.E. Upjohn Institute Press and co-edited Structural Changes in U.S. Labor Markets: Causes and Consequences, from M.E. Sharpe, Inc. Dr. Groshen received the 2017 Susan C. Eaton Outstanding Scholar-Practitioner Award from the Labor and Employment Relations Association and was appointed a Fellow of the American Statistical Association in 2020 and a Fellow of the National Association for Business Economics in 2024.She holds a Ph.D. in economics from Harvard University and a B.S. in mathematics and economics from the University of Wisconsin-Madison.
Different Ways to Listen to the Show!
New Ways to Support the Show!
With more than 250 guests and 12 seasons of episodes, the PolicyViz Podcast is one of the longest-running data visualization podcasts around. You can support the show by downloading and listening, following the work of my guests, and sharing the show with your networks. If you’re interested in financially supporting the show, you can sign up for my Patreon platform, make a one-time payment via PayPal, or shop one of the show’s sponsors.
Transcript
00:01.96
Jon
Hi, Erica. Great to see you again. how are things?
00:05.65
Erica Groshen
Things are good, John. How are we how are things with you
00:09.21
Jon
Well, um it’s early September ah in DC. You’d think it’d be a million degrees, but actually the weather has been actually delightful. It’s been delightful for like three weeks.
00:19.22
Erica Groshen
ah
00:20.82
Jon
It’s it’s amazing, like 70 degrees. ah Yeah, i don’t I don’t know what’s happening, but we’re we’re enjoying the weather. how is ah How’s the Northwest?
00:29.22
Erica Groshen
Great. And the sum here is
00:30.19
Jon
Yeah.
00:32.67
Erica Groshen
amazing 70 degrees and beautiful almost every day
00:37.88
Jon
Wow. I don’t know if anyone’s ever said that about Seattle before. so so
00:42.93
Erica Groshen
this summer is a closely guarded secret
00:48.77
Jon
Okay. Now you’ve let the, now you’re going to let people know are you’re going to have too many tourists wandering around. Um, um Well, thanks for coming on the show. We ended up recording on um ah sort of an interesting day. So it’s it’s Friday, September 5th. The ah BLS job numbers were just released for August of 25.
01:08.96
Jon
um And so what we saw was… ah fairly, I’d say anemic growth in August of 22,000 jobs added. ah We had an upward revision in July up to 79,000. And then maybe the biggest part of that puzzle is the downward revision in June to 13,000 jobs lost nationwide. So my first question for you is just a reaction to kind of all these pieces, including these revisions, which of course is a big part of why the BLS is such an important part of the story these days.
01:42.27
Jon
um But what’s your reaction to the to the job numbers?
01:46.46
Erica Groshen
Well, it continues the impression that that that we got last time. So that was August 1st, right? um And there we we got a clear, very strong indication that job creation had stalled.
02:04.46
Jon
Mm-hmm.
02:04.73
Erica Groshen
yeah um And this month reinforces that. ah And so, yes, the estimate high.
02:16.96
Erica Groshen
is just a job creation of 22,000 jobs. And that’s zero. I mean, that’s zero. i mean that’s
02:23.95
Jon
Right. Yeah.
02:25.23
Erica Groshen
that’s zero
02:26.20
Jon
Yeah.
02:26.39
Erica Groshen
um and And the three-month movie moving average, which I like to look at because it includes the benefits of the past revisions and smooths out some of the randomness that you can get in this, ah gives us a three-month moving average ah job creation of $29,000.
02:37.57
Jon
isn it
02:47.82
Erica Groshen
So about what we got this month.
02:50.66
Jon
Right.
02:50.66
Erica Groshen
Excuse me. Um, and so that says that we’re in a new phase of this expansion, um, which is pretty close to stalling stalled basically.
03:05.30
Jon
Yeah.
03:06.29
Erica Groshen
And that happens sometimes, but often that’s the pathway to the beginning of a recession. So I’d say we’re flashing yellow now. That’s what I see.
03:17.52
Erica Groshen
um Not definitely, but but it’s flashing yellow.
03:19.86
Jon
Right.
03:22.20
Erica Groshen
And this is this could be yeah the sign that we are turning the corner into a recession. And there have been a lot of policy changes that might turn out to be the ah explanations for this.
03:36.62
Jon
Right. Now, when you both today and in your in your previous roles as commissioner and and and at the Fed, what when you dive deeper into the data, what do you look for in the sort of signposting either into expansion into recession? Like what are the what are the things that you’re digging into?
03:55.86
Erica Groshen
So, well, one of the things I look at first actually is what’s happened to temporary help jobs. Because companies that are in a hiring mode will often be adding temp workers as well, either before they turn into permanent jobs or to try out the workers or whatever.
04:17.57
Erica Groshen
So when you see temp jobs increasing, that often goes along with permanent job creation or as a leading indicator permanent job creation. And it’s the same thing on the downside.
04:27.22
Jon
Right.
04:28.72
Erica Groshen
Companies that are seeing a decline in demand for their goods and services are going to be getting rid of temp workers.
04:36.70
Jon
Right.
04:36.61
Erica Groshen
not adding them on, right?
04:38.53
Jon
right
04:39.00
Erica Groshen
And they’re going to do that before they lay off workers, their own workers, right?
04:42.08
Jon
Right, right.
04:43.58
Erica Groshen
And so temp help jobs has been a ah pretty reliable leading economic indicator. That said, it’s been acting weird since the end of the pandemic.
04:53.85
Erica Groshen
So, you know,
04:54.58
Jon
Yeah, it right.
04:55.14
Erica Groshen
oh
04:57.04
Jon
and and And the other question is when you look at this, there are, you know for folks who are not fully immersed in these data, and we’ll talk more about them in a bit, there’s the household survey and there’s the establishment survey.
05:07.69
Jon
and and how do you… balance the two? How do you think about the two in terms of the trends? Because they don’t always line up at a very micro level, you know, when you get sort of into states and cities.
05:21.40
Jon
So how do you think about them as you’re sort of wading through all the data?
05:26.38
Erica Groshen
Well, one thing that I definitely do when I look is when i look at the payroll survey, then I’m mostly looking at actual numbers, numbers of jobs, whatever.
05:40.74
Jon
Right.
05:40.76
Erica Groshen
And I know that that the precision is maybe… on the change, it’s not terribly high. I mean, it’s still standard error of like a hundred thousand jobs.
05:51.72
Erica Groshen
So, right.
05:52.33
Jon
Yeah. Yeah,
05:53.77
Erica Groshen
Um, but, but for the total number of jobs, the position is amazing.
05:58.79
Jon
yeah, yeah, yeah. yeah
05:59.77
Erica Groshen
um And so so it’s okay to look at numbers, ah actual numbers of jobs in the payroll survey. When you get to the household survey, you only want to be looking at percentages and rates.
06:15.44
Jon
Mm-hmm.
06:15.47
Erica Groshen
Because this is a survey of 60,000 households, and your the the only acceptable precision really is is in the rates, the unemployment rate, the
06:26.80
Jon
Mm-hmm.
06:27.93
Erica Groshen
payroll, that the labor force participation rate, EPOP, that those kinds of things.
06:33.47
Jon
Right.
06:33.83
Erica Groshen
Numbers are, those standard errors are so large that, you know, in the long run, they’re going to look kind of like the payroll numbers, but only after you adjust a lot of different ways and the payroll numbers fully, anyway.
06:44.82
Jon
Mm-hmm.
06:50.90
Erica Groshen
So, anyway.
06:51.37
Jon
Yeah. So i want i want I want to come to the adjustments because I think that’s a big part of, again, this kind of national conversation about the VLS. Before we get there, I want to i want to ask you about…
07:04.72
Jon
ah The revisions, because I think the revisions not the and last month was one of the big reasons the president just decided to fire the commissioner, the BLS commissioner.
07:06.92
Erica Groshen
Mm-hmm.
07:16.86
Jon
um I haven’t seen statement from the White House yet today. I’m sure it’s it there’ll be something. um
07:22.69
Erica Groshen
Mm-hmm.
07:23.24
Jon
But can you explain to people why these revisions exist and why, in this case at least, we’ve now had two revisions, right? The June numbers were revised both in the June, the month of June revised both in the August release and now in the September release. And so what should people know or think about to help them understand why these numbers keep getting revised?
07:46.16
Jon
Yeah.
07:48.24
Erica Groshen
So there are, when people want a measure of what’s going on in the labor market, They want two things. they They want the numbers to be accurate.
08:00.57
Erica Groshen
Right. Don’t bother telling me something that isn’t accurate.
08:03.24
Jon
Yeah, doesn’t mean anything. Yeah.
08:05.29
Erica Groshen
That’s right. And they want it to be timely. So don’t wait for three years to tell me what’s going on right now.
08:12.21
Jon
yeah
08:12.31
Erica Groshen
Right. So they want both things. And the statistical agencies try to give them both things, but you reach a limit when you try to get something that’s both really timely and really accurate, where there’s a trade-off between the two of them, right?
08:30.28
Jon
Mm-hmm.
08:31.70
Erica Groshen
And so the the the payroll survey, officially it’s called the ah current employment statistics ah program, payroll survey says, well,
08:45.14
Erica Groshen
We want to know, the way we’re going to count to jobs is how many people were on the payrolls for each employer during the pay period that contains the 12th of the month.
09:00.24
Erica Groshen
that’s That’s the question, right?
09:00.38
Jon
Okay.
09:03.27
Erica Groshen
And you and it so it asks that question at the end of the month, like the last week of the month. said It says to all the companies that have ah agreed to participate in the survey, says, okay, you choose your mode for reporting that, and you send that us that information during this window at the end of the month.
09:28.43
Jon
okay
09:29.87
Erica Groshen
And some of these employers can’t answer that question because they haven’t finished their pay period for the 12th of the month.
09:36.86
Jon
Mm-hmm. Right, yeah.
09:37.89
Erica Groshen
right? monthly or the pay periods, you know, don’t align.
09:42.58
Jon
Mm-hmm.
09:42.63
Erica Groshen
Right.
09:42.88
Jon
Mm-hmm.
09:43.31
Erica Groshen
And some of them are in some kind of turmoil or Sally is sick or whatever. Right.
09:48.45
Jon
Right.
09:49.01
Erica Groshen
And they can’t get it. But two thirds of them do.
09:53.65
Jon
Mm-hmm.
09:53.94
Erica Groshen
And BLS over time is determined that that is actually a useful, it’s it’s a useful preliminary number.
10:03.71
Jon
Mm-hmm.
10:04.11
Erica Groshen
It’s not the final number because A third of them haven’t reported yet, but a third of them have.
10:09.84
Jon
Right.
10:11.99
Erica Groshen
And so BLS doesn’t want to sit on that.
10:12.19
Jon
right
10:14.07
Erica Groshen
And it puts out the preliminary estimate using those numbers. Now, the companies then have another window that they can report for.
10:27.27
Erica Groshen
So, you know, let’s say we’re talking about June, right?
10:30.16
Jon
Mm-hmm.
10:30.09
Erica Groshen
So, yeah.
10:30.72
Jon
Yep.
10:30.89
Erica Groshen
so ah So at the end of June, they were asked that. And then again at the end of July, they were asked the same question. And the companies who didn’t report before have a chance to report again. And by then, but like around 90 percent of them report.
10:48.53
Erica Groshen
And then they get one more chance, you know, in August.
10:50.65
Jon
Right. so they So they have they have in total three opportunities to report for each month of the year.
10:59.84
Erica Groshen
Correct. That’s right.
11:01.26
Jon
and they And so for like the July…
11:01.30
Erica Groshen
And.
11:05.20
Erica Groshen
Mm-hmm.
11:06.00
Jon
form or questionnaire that they have to they have to fill out. There is a ah there’s some section of that that says, do you have updates to the June numbers or you didn’t fill it in? So is it, does every, I guess my question is, does every ah firm get the same questionnaire so that even if you did report your June numbers, you still have the opportunity to even change your own numbers as opposed to a firm that did not submit it at all because you know, Sally wasn’t in the office that week.
11:37.00
Jon
um They can fill it in. So everybody has the opportunity to revise the numbers.
11:41.41
Erica Groshen
Yes. And sometimes firms discover they’ve made a mistake and they’ll correct it.
11:42.38
Jon
Okay.
11:45.97
Jon
me Okay. Yeah.
11:49.02
Erica Groshen
ah so So they each have these three opportunities. and the So the first revision ah the of the preliminary number is mostly driven by late reporting firms.
12:05.48
Jon
Okay.
12:07.17
Erica Groshen
a Yeah.
12:07.51
Jon
So that’s probably the the pay period lag non-alignment. Yeah. Okay.
12:13.19
Erica Groshen
right um these
12:13.51
Jon
Yeah.
12:15.96
Erica Groshen
Right. The second revision that gets you to the third number, right?
12:19.84
Jon
Mm-hmm. Right.
12:21.45
Erica Groshen
That one will have the late the last stragglers, right?
12:25.64
Jon
Mm-hmm.
12:26.38
Erica Groshen
So you get you go from around 60%, 70% for the prelimp first preliminary one. You get up to close to 90%.
12:34.24
Jon
Mm-hmm.
12:36.76
Erica Groshen
for the second preliminary one.
12:39.51
Jon
Okay.
12:40.19
Erica Groshen
And then, hang on one second, Chris, we can hear that.
12:43.25
Jon
Yeah.
12:44.48
Erica Groshen
then you to, yeah, pretty close to 90% for the second one. And for the third one, you’re talking 94%, something like Okay, now, pretty close to ninety percent for the for the second one and for the third one you’re talking ninety four percent something like that
13:02.03
Jon
okay yeah
13:02.75
Erica Groshen
okay now Also between the second and the third one is when you’re most likely to get a change in the seasonal adjustment factors.
13:13.76
Jon
e
13:15.01
Erica Groshen
And the seasonal adjustment factors are what we all need to be able to detect trends because there’s a lot of seasonality in employment. And if you don’t take out the seasonal trend, you could be thinking that jobs are really growing like crazy in September.
13:33.24
Jon
Right.
13:33.90
Erica Groshen
when when they’re, you know, when this is just the future is coming back.
13:36.22
Jon
Teachers are coming back. Yeah, right.
13:38.67
Erica Groshen
Right.
13:38.94
Jon
Yeah, right.
13:39.86
Erica Groshen
So, um but those do move over time and you and BLS wants to keep up to date with seasonals. And so every month they recalculate the seasonals. And so most of the change from the the second to the third is usually seasonals.
13:55.02
Jon
Right.
13:55.75
Erica Groshen
Right.
13:55.97
Jon
Now on the household side, and we don’t need to get into into depth about it, but on the household side, it and I think just broader household individual surveys, response rates have been falling over time.
14:07.53
Jon
Is that the same on the establishment side or are those response rates fairly steady or steadier, I should say?
14:07.52
Erica Groshen
Mm hmm.
14:16.99
Erica Groshen
I would have told you that they were steadier.
14:19.84
Jon
Okay.
14:20.29
Erica Groshen
When I was at BLS, they were fairly steady. And it turned out a reason, an important reason for that was that BLS was really innovating on response modes and making it easier and easier for employers to respond.
14:23.87
Jon
Okay.
14:33.74
Erica Groshen
And that helped them to sustain those response rates for more or less the time that I was there. Since then, they have declined and they took a big dip during the pandemic and have not recovered very much since then.
14:44.06
Jon
okay
14:50.20
Erica Groshen
So now we’re talking about employer response rates that are half of what they were when I was at BLS, right?
14:58.81
Jon
Wow. Yeah.
15:00.61
Erica Groshen
um Now that’s just the ones that are agreeing to ah participate in the survey at all. and that’s important. And that’s going to affect the size of the benchmarking in particular, right?
15:13.48
Jon
Right.
15:13.99
Erica Groshen
But a volatility, right?
15:14.34
Jon
Right.
15:16.04
Erica Groshen
but But there’s also been a decline in companies getting in the survey for the first closing ah for the preliminaries. So the two thirds that I talked about used to be more like three quarters.
15:30.81
Jon
I see. Okay.
15:31.95
Erica Groshen
Right. And so that also that feeds into that would feed into bigger revisions.
15:39.49
Jon
Mm-hmm.
15:39.97
Erica Groshen
But there has been ah something else that the BLS has done to help keep those revisions and back to shrink those revisions.
15:50.43
Jon
Mm-hmm.
15:50.59
Erica Groshen
And that’s. Most of the time, and I’ll tell you, I think I can do this pretty succinctly. ah The way that BLS treats most late reports is just there’s a missing value in there, right?
16:05.69
Jon
Yeah.
16:06.13
Erica Groshen
And so you’re imputing the same growth rate to the ones who don’t report as to the ones who did report, right? There are some very large companies who sometimes report late.
16:22.79
Erica Groshen
They don’t act like small companies and they’re important.
16:25.15
Jon
Yeah. Right.
16:27.63
Erica Groshen
They’re big, right?
16:27.81
Jon
Mm-hmm.
16:29.11
Erica Groshen
And so BLS has some firm specific imputations that it uses for those very large companies.
16:37.91
Jon
I gotcha. Right.
16:38.34
Erica Groshen
And that has reduced revisions because yourre your, your, your, extrapolating from their recent trends and their past behavior in other ways.
16:50.50
Jon
right So I want to i want to ask um i want to ask in ah in a moment what you sort of see coming in the future and what people should be looking out for in the data. But I want to um just sort of put a bow on this part about quality of BLS data.
17:08.00
Jon
And maybe you can tell folks…
17:08.05
Erica Groshen
Mm-hmm.
17:12.38
Jon
what the commissioner’s role is in this. I mean, you know, the, one of the reasons that commissioner McIntyre for was, was fired by the president because, you know, he claimed the jobs were rigged and they were falsified and this sort of thing.
17:15.80
Erica Groshen
Mm-hmm.
17:24.94
Jon
And ah it’s the other thing that i think is important for people to know is what does the commissioner do and what is the commissioner’s role in these monthly numbers? I’m guessing that you are not working in whatever coding language BLS folks are doing, working on the birth-death model or the the you know seasonal adjustment model. So while you were there, what was your role in the in this monthly data collection and release?
17:54.89
Erica Groshen
So the way you want to think about this is it’s really a finely oiled machine, right? This is much more like a manufacturing process than anything else.
18:04.19
Jon
Mm-hmm.
18:05.26
Erica Groshen
It’s not like consulting.
18:09.90
Jon
Yeah.
18:10.58
Erica Groshen
and um The numbers come in and they are processed and then the tables are spit out, right?
18:17.61
Jon
Mm-hmm.
18:18.81
Erica Groshen
And all of the work and the investment goes into the design of that system, not into the any ability to manipulate them on a monthly basis.
18:25.39
Jon
Mm-hmm.
18:32.14
Erica Groshen
It’s just not in there. In fact, it’s designed so that that if you tried to manipulate it, you’d break it.
18:38.55
Jon
Mm-hmm.
18:39.50
Erica Groshen
Right? um You have teams of experts doing quality control for specific industries just to make sure there’s nothing crazy in there because some company put in a million workers when they met 10, you know.
18:53.67
Jon
Mm-hmm.
18:54.45
Erica Groshen
um And ah but they’re just looking at that particular industry when they’re they’re fine with that.
19:02.25
Jon
Mm-hmm.
19:02.77
Erica Groshen
Then they get added up in a small number of people. Just check those things to make sure there’s nothing crazy going on there.
19:11.60
Jon
Right.
19:11.77
Erica Groshen
And then the numbers are final. And that’s it.
19:13.81
Jon
Mm-hmm.
19:14.60
Erica Groshen
And the commissioner gets them after they’re final. The only thing that the commissioner sees that’s at all preliminary is the draft of the narrative.
19:25.81
Erica Groshen
Those are the paragraphs.
19:26.13
Jon
Okay. Right.
19:27.58
Erica Groshen
And the narrative I learned in my first few months there when I tried to make them lively is they’re not supposed to be lively. There are no adjectives in there.
19:39.01
Erica Groshen
Up, down, that’s it, right?
19:41.88
Jon
Yeah.
19:42.83
Erica Groshen
The BLS answer is the glass half empty or glass half full is this is an eight ounce vessel with four ounces of liquid.
19:47.17
Jon
Right.
19:53.02
Erica Groshen
you know, or and that, that’s, that’s, to right.
19:53.45
Jon
Yeah. And that’s it. I mean, even the visual of the employment situation report is like plain text.
19:57.12
Erica Groshen
That’s.
20:02.51
Jon
It looks like it was published on a typewriter. I mean, it definitely has this as as as a neutral look and feel as you could get something on on on the web in 2025.
20:15.18
Erica Groshen
There are, there are words they use and words they don’t use.
20:18.56
Jon
Right.
20:18.56
Erica Groshen
And that’s that, you know,
20:20.47
Jon
And so do all of these teams, well, I guess I’ll ah make it a little bit little bit more detailed.
20:23.22
Erica Groshen
No,
20:26.98
Jon
So, Are the, for the jobs, and I’m sure differs by different data collection, but do but but for jobs, are people assigned are they split into different roles?
20:32.46
Erica Groshen
totally
20:38.82
Jon
So there’s the group that does large businesses, the group that does small businesses, are they distinguished by region of the country and and and and they and they kind of don’t talk to each other?
20:49.83
Jon
And then there is some other group that talks to each other to sort of pull it all together.
20:55.54
Erica Groshen
Yeah. um So course, there are the people who work on data collection, they’re mostly out on the field, and they’re convincing companies to participate.
21:03.20
Jon
Okay.
21:05.61
Jon
it
21:05.56
Erica Groshen
And they don’t have much of a role in processing the numbers as they come in, although maybe someone from an industry group might reach out to them about some particular company’s report that doesn’t make any sense, right?
21:19.93
Erica Groshen
But that would be on the one off basis in some certain circumstance, right?
21:20.48
Jon
Okay.
21:24.86
Jon
Right. Okay.
21:25.79
Erica Groshen
like So once once the data have all been collected and they the and the collection period is over, then you have the different industry groups in the main office looking at them separately from each other and only looking at their industry and not communicating with each other until they’re added up together and only a few people
21:30.58
Jon
Mm-hmm.
21:46.29
Jon
you Right. Right.
21:52.11
Erica Groshen
have seen that.
21:54.09
Jon
Right.
21:54.16
Erica Groshen
And those very tightly choreographed, very short turnaround. Right.
22:01.17
Jon
Yeah.
22:02.10
Erica Groshen
And yeah, because that the deadline for reporting the numbers is typically the Monday before the Friday when the numbers are reported.
22:13.80
Jon
Yeah, so that was that’s an interesting that’s an interesting point. So what does the calendar look like? The firms are reporting in the last week of the month. so So that’s when it’s being collected in that last week.
22:26.62
Jon
And then it takes, I’m assuming those next, it’s two full weeks of the analysis and assembling.
22:34.09
Erica Groshen
No, no, no, no. It’s, uh, it’s not, it’s, it’s three days, two, three days.
22:38.88
Jon
Oh, wow. Wow.
22:39.69
Erica Groshen
Yeah. Yeah. Um, yeah. So yeah, it’s very, very fast, but there, there is time, you know, to look at it and, and the, the, um,
22:46.61
Jon
Yeah.
22:51.97
Erica Groshen
That’s why the the the programs are really important because they they flag the issues and then the people look at them and address them and most and it and it runs smoothly.
23:00.22
Jon
isn it
23:02.83
Erica Groshen
But but you know sometimes I got it on Tuesday night and sometimes I got it Wednesday morning. but
23:09.38
Jon
Right. Right. Right.
23:10.75
Erica Groshen
think
23:10.96
Jon
To go out on Friday, right.
23:12.02
Erica Groshen
close
23:12.16
Jon
It goes out Friday at very specific time. Right.
23:14.76
Erica Groshen
That’s right. That’s right.
23:15.90
Jon
um Okay, so with that is background and how and in your role as commissioner, there have been a…
23:21.31
Erica Groshen
um Oh, you didn’t ask me those.
23:22.86
Jon
Oh, yeah.
23:23.43
Erica Groshen
So then what does a commissioner do?
23:25.50
Jon
Oh, right, right.
23:26.09
Erica Groshen
Which
23:26.37
Jon
So what does the commissioner do? Right, yes.
23:29.85
Erica Groshen
you don’t do.
23:30.54
Jon
Yeah, so far we’ve established you haven’t really done anything.
23:32.02
Erica Groshen
yeah it which you don do
23:33.62
Jon
right Right, yeah.
23:37.90
Erica Groshen
So, um, on, on the numbers day, I prepared to brief the council of economic advisors and the secretary of labor, and also to answer questions from journalists or whatever, know, that might’ve come up afterwards on that release.
23:48.26
Jon
Mm
23:54.23
Erica Groshen
So I was just educating myself on what the numbers said.
23:55.05
Jon
hmm.
23:57.21
Jon
Right.
23:59.22
Erica Groshen
Right.
23:59.42
Jon
Right.
24:00.61
Erica Groshen
Uh, and i So there there are two two changes that I made to the release, this boring release, right?
24:11.78
Jon
Yeah. yeah
24:12.65
Erica Groshen
oh i um If you look at the very last paragraph of the narrative every month, when it talks about the revisions, there is a clause there that says, um this time in that first, ah the second text,
24:30.36
Erica Groshen
the second sentence actually with these revisions, employment in June and July combined is at this point, 21,000 lower than previously reported.
24:41.24
Erica Groshen
Right.
24:41.72
Jon
Yes. Yes.
24:42.41
Erica Groshen
That’s there. Cause I said, let’s add it up for people.
24:43.79
Jon
ah I, I noticed that sentence this morning.
24:46.33
Erica Groshen
and
24:48.26
Jon
i was like, i don’t think I’ve ever like paid that much attention to it, but yes, I saw that this morning.
24:48.91
Erica Groshen
and
24:52.92
Erica Groshen
You are welcome. You are welcome.
24:58.00
Jon
Okay.
24:58.23
Erica Groshen
And,
24:58.36
Jon
So we have you to thank for that. Like just like, yeah.
25:00.48
Erica Groshen
that and and that Now that one I could do just say, okay, let’s put that in and I and i want that in going forward, right?
25:06.08
Jon
Yeah.
25:08.34
Jon
Right.
25:08.34
Erica Groshen
So that that could be done because that’s in the narrative.
25:09.02
Jon
Right.
25:11.86
Jon
Right.
25:11.86
Erica Groshen
The other thing that I did was if you look at a table B,
25:21.09
Erica Groshen
There is a line there. There is a row there, a couple of rows there that says three month average change in thousands.
25:29.28
Jon
Mm-hmm.
25:31.91
Erica Groshen
I suggested that that be in there, too. Again, anybody could do that math, but.
25:34.76
Jon
Yeah. yeah Right.
25:39.22
Erica Groshen
My time in the Fed, we were often looking at the three-month average because that incorporated the recent revisions and helped to change our view of what we thought was going on before.
25:41.50
Jon
Mm-hmm.
25:51.98
Erica Groshen
And it smoothed out some of the randomness. And so we looked at it there and I said, well, I think it should be there because i think that people should be looking at that number.
25:59.89
Jon
Yeah.
26:02.69
Jon
Right.
26:03.42
Erica Groshen
Therefore, let’s put it in.
26:05.10
Jon
Now, how many… how long, once you were like, we should put this in, how many conversations, meetings, how long was the process to get it through sort of like the, especially the career folks who had been there for a long time to say, yeah, this is a good idea. We should put this in.
26:19.94
Jon
Mm-hmm.
26:20.83
Erica Groshen
So the the probably one or two meetings for that. and then
26:25.06
Jon
Mm-hmm.
26:26.17
Erica Groshen
um and then i saw a draft of it as they worked it through. And I don’t remember how many months took.
26:33.63
Jon
here
26:34.59
Erica Groshen
At least two months, maybe three months to get that change into…
26:36.33
Jon
Yeah. Right.
26:41.57
Jon
But, but both of these are, and, and they are both taking data that are already produced and just disseminating them basically additional dissemination.
26:49.93
Erica Groshen
Yes.
26:54.54
Erica Groshen
Right. it that I was not changing any methodology.
26:55.62
Jon
Yeah.
26:58.22
Erica Groshen
This was all communication.
26:58.88
Jon
Right.
27:01.41
Erica Groshen
all right and And I wasn’t doing it just for this month because X, Y, Z.
27:01.60
Jon
Mm-hmm. Yeah.
27:07.09
Jon
Yeah.
27:08.07
Erica Groshen
I thought an ongoing basis, it would be helpful to the users
27:08.65
Jon
right
27:13.07
Jon
Right. Right.
27:13.64
Erica Groshen
and and i and the staff. Agreed with me.
27:17.53
Jon
yeah Right.
27:18.30
Erica Groshen
and
27:18.80
Jon
Yeah.
27:19.29
Erica Groshen
um And so there it is. and And it has lived on. So no one know one has decided it was useless. I was used to it.
27:26.55
Jon
People are actually using it and now it’s there. that’s Well, there you go.
27:29.10
Erica Groshen
from these used to it
27:30.55
Jon
um So, yeah.
27:32.21
Erica Groshen
But there’s that.
27:33.31
Jon
Right.
27:34.27
Erica Groshen
And then, but then there’s the big picture, right?
27:34.48
Jon
Right.
27:37.22
Erica Groshen
And this was what really, uh, was, I considered the most important job, which was looking over the horizon.
27:37.38
Jon
Mm-hmm. Mm-hmm.
27:44.98
Erica Groshen
Where should we be going?
27:45.20
Jon
Mm-hmm.
27:46.06
Erica Groshen
What should we be prioritizing of all the things that BLS could be doing? What should we be but doing? And being the public face of the agency, fending off, uh,
28:01.40
Erica Groshen
ah Requests that were inappropriate for the agency, encouraging ones that were appropriate, that kind of thing. So, you know, CEO type things.
28:11.84
Jon
Right. right
28:13.56
Erica Groshen
But the CEO isn’t down there producing yeah in the factory. They’re not running the machines.
28:19.51
Jon
Right. right
28:21.26
Erica Groshen
I wasn’t I wasn’t programming anything. are right. Right.
28:25.88
Jon
Right. Probably in Fortran if I know anything about, if I know anything about government work, but yeah.
28:32.32
Erica Groshen
ah
28:32.53
Jon
um
28:33.64
Erica Groshen
but hey what That wasn’t, no, it wasn’t Fortran.
28:36.90
Jon
It wasn’t for that. Okay. um Okay. So with that in mind of, of the commissioner’s role, there’ve been a couple of names floated um to replace the the commissioner, ah Commissioner McIntyre for, if you were on the, I guess it’d be a Senate, ah Senate panel. If you’re on the Senate panel, what are you looking for in a nominee or what questions are you asking a nominee for the position?
29:06.77
Erica Groshen
Yeah. so some management leadership management experience, if you’re going to do this for twenty five hundred people across the country and you’re producing seven of the principal federal economic indicators for the entire country and therefore for the world, really, um you know, leadership management is something you have to be ready for.
29:30.44
Jon
Yeah.
29:30.51
Erica Groshen
um Expertise in economic statistics, even if you’re not running the calculator yourself, you understand what they mean and be able to brief people, answer questions.
29:37.06
Jon
Yeah.
29:41.24
Erica Groshen
um ah yeah A good relationship with experts, you know,
29:46.09
Jon
Mm-hmm.
29:46.95
Erica Groshen
um ah And knowing enough about the agencies, mean, you’re drinking from the fire hose anyway, but you gotta have some kind of grounding.
29:55.58
Jon
Right. Right. Right.
29:59.78
Erica Groshen
ah Ability to interact effectively with Congress, with the media, with senior Department of Labor staff, all of that’s important.
30:09.05
Jon
Yeah.
30:11.68
Erica Groshen
And then you get to something that sounds squishy, but is fundamental. which is that trust is mission critical for a statistical agency.
30:27.07
Erica Groshen
You might as well not produce statistics if they’re not trusted, right?
30:31.25
Jon
yeah
30:32.05
Erica Groshen
Because if they’re not trusted, people won’t use them to make their most important decisions. like In which case, why produce the data? That’s only reason to produce them is so that they will be used and they’re not going to be used if they’re not trusted.
30:46.83
Jon
Right.
30:47.28
Erica Groshen
so And also, ah people, respondents,
30:52.59
Jon
Mm-hmm.
30:52.49
Erica Groshen
or ah aren’t going to assume the burden and the risks of responding to your surveys if they don’t think that they are going to create trustworthy statistics, right?
31:01.79
Jon
Mm-hmm.
31:07.87
Jon
So, so, in
31:08.21
Erica Groshen
Why
31:09.25
Jon
yeah So in your mind, what does the commissioner what can a commissioner do to create that trust from the, from the public, not just the, I mean, the public receiving, as as you mentioned, the public receiving the information, but also the people who are being asked to contribute to the, to the data.
31:18.38
Erica Groshen
Mm-hmm.
31:28.33
Jon
What, what can, can what can the commissioner do?
31:31.57
Erica Groshen
ah There are some really important guidelines that are put out.
31:36.80
Jon
Mm-hmm.
31:38.01
Erica Groshen
OMB Statistical Policy Directive number one is full of all of the things that an agency should be doing and the commissioner should be ensuring that the agency does.
31:42.78
Jon
yeah
31:52.04
Erica Groshen
i And Statistical Policy Directive three is all about the principal federal economic indicators in particular. and what should be done with how you put them together, how you disseminate them, all of that.
32:04.79
Jon
Right.
32:04.83
Erica Groshen
So saying that you are going to prioritize all of those things. And then there’s a publication called the Principles and Practices for a Federal Statistical Agency that’s put out by the National Academies goes through all of these practices.
32:19.37
Erica Groshen
So um
32:22.61
Erica Groshen
pledging that these are going to be The standards for your behavior and those of the agency is critical, right? And then saying that if anybody superior to you tells you that you should violate them, that you are going to refuse to do that.
32:40.96
Jon
Right, right. Now, well, to that point,
32:42.91
Erica Groshen
Right.
32:46.49
Jon
Should people be worried? And given your what we’ve talked about with what the commissioner can and can’t do when it comes to some of these monthly numbers, should people be concerned about a new commissioner being able to modify, manipulate, change the numbers in any way?
33:06.75
Jon
Or is it this sort of bigger picture, broader… perspective that is that is important, um but should people be maybe a little less worried about someone coming in and being able to change the the numbers on a month-to-month basis?
33:25.25
Erica Groshen
if the commissioner comes in alone, right, then they’re not going to be able to do something very much, very quickly, right?
33:30.19
Jon
Right.
33:37.05
Jon
Mm-hmm. Right.
33:37.49
Erica Groshen
and They don’t have access to the underlying data. They don’t have access to the programs, right?
33:44.40
Jon
Right.
33:46.10
Erica Groshen
It’s not like they can just go and change that unless they want to retype the tables themselves, right?
33:52.23
Jon
right On a typewriter.
33:54.41
Erica Groshen
and and And there are thousands of the tables that are all put out at the same time in the same processes that have to add up.
33:54.70
Jon
Yeah. Yeah. Right. Right.
34:01.17
Erica Groshen
Right.
34:01.98
Jon
Right.
34:01.86
Erica Groshen
This is a finely oiled machine. Everybody knows what they’re doing and the commissioner is not running it on a moment to moment, moment basis.
34:10.15
Jon
see
34:10.27
Erica Groshen
And they’re the only appointee in the whole agency. Right.
34:13.71
Jon
right
34:14.96
Erica Groshen
ah So. What would then be the signs that they tried to do this?
34:21.73
Jon
Yeah.
34:21.71
Erica Groshen
Right.
34:22.54
Jon
Right.
34:22.64
Erica Groshen
Well, they might bring in a posse, you know, ah well right.
34:26.17
Jon
Yeah. Their own, their own people. Right.
34:27.89
Erica Groshen
But even they they would have to figure out what to do and how to do it.
34:31.87
Jon
Yeah.
34:32.08
Erica Groshen
Right. And how to do it on the fly. All right.
34:36.69
Jon
Right.
34:37.29
Erica Groshen
Right.
34:37.58
Jon
And very complicated stuff on the fly. Like just, just, that just so people understand, like, this isn’t like, you know, changing a tab in an Excel file.
34:39.88
Erica Groshen
That’s right.
34:47.00
Jon
Like this is, this is complicated stuff with lots of programs and lots of math going on. It’s not like you just toggle some switch.
34:54.09
Erica Groshen
that’s
34:54.93
Jon
Yeah.
34:55.57
Erica Groshen
right That has been tested to make sure it all adds up, right?
34:57.94
Jon
Right. Right. Yeah. Right.
34:59.54
Erica Groshen
yeah um Okay, so so what would you see? Well, you’d probably see some delays in releases.
35:06.48
Jon
Mm-hmm.
35:07.39
Erica Groshen
If they’re trying to do it, you know it would they’d have to be superhuman to not screw it up in some way.
35:12.94
Jon
Yeah, right, right.
35:13.56
Erica Groshen
Right. um You you’d see some unannounced or unexplained changes in methodology that and that would violate principles of transparency that the agencies are supposed to follow.
35:21.12
Jon
Mm-hmm.
35:26.33
Jon
Yeah.
35:29.11
Jon
Mm-hmm.
35:29.19
Erica Groshen
Right. um You would ah and and you’d you just see mistakes, you know, in them that aren’t typical. Right.
35:39.57
Jon
Right.
35:40.27
Erica Groshen
um you’d see injection of policy spin into the release narrative. so all a sudden you’d get adjectives.
35:47.97
Jon
Yeah.
35:47.85
Erica Groshen
and
35:48.94
Jon
Right. So just the language itself.
35:49.34
Erica Groshen
right That’s right.
35:52.05
Jon
Interesting.
35:52.09
Erica Groshen
um
35:52.87
Jon
Yeah.
35:53.04
Erica Groshen
yeah um and And then you’d have the signs of staff resistance because remember the culture of the BLS is definitely apolitical data nerds.
36:05.02
Jon
Yeah. yeah
36:05.59
Erica Groshen
Totally dedicated apolitical data nerds. And you would get
36:12.25
Erica Groshen
whistleblowers, resignations, leaks, crazy transfers, things like that.
36:20.02
Jon
Okay.
36:20.08
Erica Groshen
So i think I think it would be visible.
36:24.26
Jon
Okay. All right. So that that’s the things to look for. So then um what do you think Congress or…
36:36.23
Jon
you know, other policy leaders, policymakers can or should be doing in this, we could almost call it a transition period from one commissioner to the next and then and then going forward.
36:47.20
Jon
I’m going to assume funding is part of that. um we can We can put that to the, maybe to the side if you want, but what should ah what should policymakers role be in in this entire process?
37:01.08
Erica Groshen
So a rapid and thorough vetting of any candidate.
37:05.24
Jon
Mm-hmm.
37:05.62
Erica Groshen
Right. The kinds of things that we talked about. So make sure that that that this candidate is worthy.
37:08.76
Jon
Uh-huh. Uh-huh.
37:13.12
Erica Groshen
Right. And and will uphold the trust. um Then there’s some other important things. ah There is a ah policy that the administration is pushing forward called schedule policy career to convert, reclassify um civil servants who deal in policy influencing jobs, who have policy influencing jobs,
37:46.94
Erica Groshen
ah so that they they’ll be reclassified so that they can be fired at the pleasure of the president, essentially.
37:53.78
Jon
Right. Right.
37:55.29
Erica Groshen
That converts these civil service jobs into much more like political appointments. That should not be applied to anybody in the statistical agencies because the numbers influence policy, but the but but the statistical processes are not about influencing policy.
38:07.26
Jon
Mm-hmm.
38:11.82
Jon
Stop. Right. right Yeah.
38:16.53
Erica Groshen
ah That distinction is not clear in anything written so far about these positions.
38:20.80
Jon
Mm-hmm.
38:22.92
Erica Groshen
ah ah We talked about investing in improving and modernizing BLS methods. Every statistical agency needs to be modernizing all the time. BLS has not been able to because it’s been so strapped for funds.
38:39.83
Jon
Mm-hmm.
38:40.50
Erica Groshen
um Congress should also clarify the um the distinction between a presidential appointee in a term limited position, like the head of the BLS and appointees that serve at the pleasure of the president.
39:03.75
Erica Groshen
If they are the same, then why is there a difference?
39:08.09
Jon
Yeah.
39:08.06
Erica Groshen
And if they’re not the same, then exactly what can you fire someone in a term limited position? The understanding used to be, and when I took the job, my understanding was that I could only be fired for cause, for not doing my job or malfeasance or something like that, not
39:21.96
Jon
he
39:26.05
Erica Groshen
because I produced unwelcome news to the president.
39:29.91
Jon
yeah Right.
39:31.50
Erica Groshen
but And so if Congress wants there to be a distinction, they’re going to have to step up to the plate for that.
39:38.69
Jon
Right. which Which I’m sure, by the way, which I’m sure you did on occasion in four years there.
39:40.66
Erica Groshen
Mm-hmm.
39:44.88
Jon
I’m sure there was news that the administration that you that you worked under was not particularly happy with.
39:52.28
Erica Groshen
I was told at one point that revisions were causing presidential heartburn to President Obama. and But that’s as far as it went.
40:05.08
Jon
was right.
40:05.66
Erica Groshen
and
40:05.88
Jon
Yeah. Well, you could just send him a bottle of Tums, I guess, and and call it a, day yeah.
40:10.48
Erica Groshen
That’s right.
40:11.73
Jon
Anyway, sorry, i I interrupted. Okay, so go ahead.
40:15.08
Erica Groshen
That’s okay. So those are um those are really the main things. um I don’t know if Congress can tell the um administration to lift the hiring freeze for the statistical agencies, but they should.
40:26.57
Erica Groshen
Because BLS is down over 20% of its staff, a third of its top leadership ah positions are empty.
40:27.17
Jon
Yeah, right.
40:35.35
Erica Groshen
This is a problem for just maintaining operation.
40:35.49
Jon
Mm-hmm.
40:38.80
Jon
Yeah, right.
40:39.90
Erica Groshen
Yeah. Or one more thing Congress can do is that there are some restrictions, legal restrictions on sharing data between federal agencies and and federal and state agencies that really hamper the ability of the statistical agencies to produce modern statistics.
40:53.01
Jon
Mm-hmm. Mm-hmm.
40:59.49
Erica Groshen
Those can be lifted because statistical agencies do not endanger the privacy and the confidentiality of ah um of any of these respondents. or i mean there’s There’s no logical reason for the statistical agencies not to have access to these data to produce better statistics for us all to use.
41:22.64
Jon
Right, right. To talk to talk across with with each other.
41:26.25
Erica Groshen
To talk with each other.
41:27.54
Jon
Right, right. um Okay, so before I i let you go, are… where are where are you right now? How are you feeling about um the BLS and and these various attacks? I know you’re part of, if not leading the the friends of the BLS group, like where are you feeling about where the agency stands now? and And we’ve talked a little bit about this, but where it stands now and then where things are headed with the, um you know, we can, we can leave names of potential commissioners ah off, but like, what,
42:02.30
Jon
Where are you, how how are you sort of viewing the world these days from, from, ah from your perspective?
42:09.07
Erica Groshen
ah I have two conflicting emotions. One is i am I’m very worried about the continued collateral damage and intentional damage to the statistical agency for all the things we talked about.
42:22.54
Jon
Mm-hmm.
42:22.52
Erica Groshen
And i I understand that things could get much worse.
42:27.60
Jon
Mm-hmm.
42:27.98
Erica Groshen
And I think that would be really bad for our country as a whole.
42:32.23
Jon
Mm-hmm.
42:32.35
Erica Groshen
in in so many ways. ah you know ah Countries that don’t have reliable statistics ah lose stature quickly.
42:42.75
Jon
Mm-hmm.
42:44.56
Erica Groshen
right and And we would inhibit the ability of our businesses to make good decisions and inhibit the ability of policymakers to make good decisions.
42:57.06
Jon
Mm-hmm.
42:58.21
Erica Groshen
Chaos, all of those things. So that I’m very worried about that. At the same time, I think that the uproar that we have seen since August 1st has been very, it’s been very heartening to me.
43:16.66
Erica Groshen
All the sectors, bipartisan, public, it’s been vigorous, it’s been sustained. And this signals an appreciation um ah for this long neglected public good.
43:28.16
Jon
Yeah.
43:28.24
Erica Groshen
Right.
43:28.38
Jon
Yeah.
43:28.67
Erica Groshen
You don’t know what you got till it’s gone or about to be gone.
43:32.12
Jon
yeah
43:33.96
Erica Groshen
So is this going to be a turning point for federal statistics, at least, you know, in in the next years? So ah this recognition could lead to support from Congress and administrations that that has been lacking in the past few decades.
43:51.67
Jon
Yeah.
43:53.46
Erica Groshen
um and And here’s some real irony. If some of the reason that we haven’t had improvement to our statistical system has been fear of disruption or people say, well, it’s not necessary to disrupt things.
44:08.22
Erica Groshen
Well, this administration is not worried about disruption.
44:11.09
Jon
yeah
44:11.47
Erica Groshen
So maybe either by this administration or in the aftermath, of the disruption caused by this administration, there’ll be an opportunity, a moment for really modernizing the statistical system the way it should be.
44:27.54
Erica Groshen
And then that calls on all of us to help decide what that plan should look like and get it in place so that this administration or the next one has something good to work from, not just cobbled together by some congressmen at the last minute.
44:43.67
Jon
Right. All right. Well, on that note, um we shall see what happens. Erica, thanks so much for coming on. Joy, getting to see you again, getting to chat. And I appreciate you taking the time out of your day to talk with me.
44:56.18
Erica Groshen
Oh, you’re welcome, John. And thank you for inviting me and giving me a chance to talk to you and to your audience. This was fun.
45:02.98
Jon
Great.