YouSaid · the spoken record

Guy Berger

lines on the record: 66
first: 2025-08-13
most recent: 2025-08-13
sittings or episodes: 1
sources: podcast

Every line below is reproduced as it was said and linked to the record it came from. Nothing here is summarised or generated. Directory · Search · Corrections

“This should be a resource to explore us and to improve because I think it is worthwhile. I don't want to say that I'm glad to have these revisions. I just recognize this is not an easy problem to solve. And by the way, that they've implemented solutions in the past, trying to deal with this stuff. It's just hard. I think that my point being that the faculty county has been cooling sort of semi-continuously for like three and a half years, maybe stabilizing recently, maybe stabilize recently now starting to cool again, like makes catching these terms harder. And we just essentially been like there have been like three turns in a row in one direction instead of what typically happens, which is maybe you have one or two years where we get negative versions, you fell into recession, and the next year is like a”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“The reason I mentioned this is like it sets up that you don't usually get this very long extended sequence of downward visions. Instead, what we've had, by the way, over the last two and a half, really three and a half years, because I think the peak lay market heat was like sometime in very early 22, late 21, is just a very slow gradual cooling. So essentially every year like the underlying mechanism of job creation is a little colder and the BLS essentially is like failing to anticipate that. Should they be anticipating that? I mean, at some point, like, we can debate on maybe they've been too optimistic about their assumptions. I mean, if all of a sudden they became way more pessimistic and the economy stabilized, maybe it does. I mean, some indicators just so the economy is stable, then it might be that in fact they start undercounting job growth. Like my feeling is like my sense on this is it'd be great if they had better indicators to improve this. This should be.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“There's just churn. There are a lot more businesses dropping out of your sample because they just don't exist anymore. And there are a lot of businesses that are joining, should be joining your sample, but we don't know about them yet. You're going to start kind of having to, for the modeling chunk of this is going to increase. That's happened. So if you are going to want to fix that, you are going to need to be able to model in a balanced way what's going on with the data better. The second part, and this is sort of one direction of it, we've been in an environment where like that's been, that's actually quite unusual by historical standards. Usually the way business cycles work is the economy falls off a cliff in a recession. It can be a mild recession like in 2001, or it can be a serious one like 2008, or it can be something bizarre like COVID, where essentially the economy shut down for most of the economy shut down for two weeks, then reopen.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“Revised up a year or year and a half later in the annual revision. And when the economy is slowing down, they tend to be revised down. Now, I think there's two parts of this that I think are particularly relevant to this. So when there's a description of why we've had negative revisions in the annual revisions, or we're getting more. And the other one is just in general in terms of the size of these revisions. One thing that changed post-pandemic, we talk about low response rates. The amount of churn due to birth and death has gone up since the pandemic. If you think about this in terms of what parts the BLS is effectively collecting data on in real time, so what parts it essentially has to model or guess what is going on, it's gone harder. Like the more of your employment, all of your employment in real life and like the real true number, if all of it's coming from businesses that exist in a very undynamic economy, you might actually, BLS might actually do a really good job and revisions would be minimal if all of a sudden”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“Even if you were to do it to do that, you would still potentially substantial revisions later on. There's the issues with the BLS as SAMP, like for the established survey, the BLS as sample. Businesses die. So, you know, you hear about the net birth death model. So it's like with some sort of demographic thing. No, it's about businesses. Like businesses shut down and they stop responding to the BLS. Businesses created and the BLS is slow to add them to the sample. I mean, just you don't, the business is created. You don't know about it right away. As a result, they try to model that. What's happened? And this, by the way, historically has led to revisions over the years. You can look at a history of revision. Some have been very large. Some have been smaller. The revisions tend to be pro-cyclical in the sense that when the labor market is, which is not surprising if you think about the distribution of births and deaths, like the BLS will tend to be slow to register accelerations in the economy. So when the economy is starting to accelerate, perils typically get”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“Numbers. The other thing that happens that is less discussed, and this is really sort of an inside baseball thing, is the BLS recalculates seasonal adjustment factors every month. And sometimes the revision comes from that too. I mean, if you look at it, you would see there's a revision to the non-seasonal adjusted number, and there's recalculated, there's the revision to the seasonally adjusted number, and they're not quite the same thing. And in the end, employment level of like 160 or whatever million it is, like these things from that perspective look tiny. They seem big for the monthly change. But the reason I'm throwing it out there is like these are things that normally a lot of other indicators just wouldn't even necessarily cross your mind that much. And here the numbers just the way you're used to looking at them, the changes in them seem big. But so that's one set of revisions is like what happens to the most recent data. Then after like three months, they freeze them. There's a second set. And by the way, this is why delaying the report by like two months, which I think is a bad idea.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“Responses, and as a result, you're just getting less accurate numbers early on. Now, I think one argument should be let's just wait. I mean, and there are going to be issues that come up with even this approach, but let's say you wait two months. I'm not a big fan because generally I think they're obviously revisions, but there's usually a pretty good correlation of the early number and the eventual number. So essentially, like, I think it is valuable to have a blurry signal. We already know, by the way, since people invest a lot of energy in things like other wonderful providers like ADP and Indeed. And, you know, my old employer, LinkedIn, that does this awesome LinkedIn hiring rate, we know that people are really interested in this data. If suddenly the BLS sat on it for three months, it's not, it's not like we'd be like, okay, we finally reduced this imperfect noisy signal. We were replacing it with an even worse set of signals. I mean, all these are wonderful, but they're not substitutes. And so I think that's sort of my perspective on that. Now, you get these late responses. They changed.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“Think once you start thinking about that, you start raising questions of how realistic would this be. So let's back up because I think that's sort of it. The idea of having the early data is you have a survey of just a certain number of establishments instead of making everybody do it to lower cost of collection. There are things you could improve in general which we can talk about. But I guess I want to start with that. Now I think that realistically when revisions come about There's two sets of origins to think about. The first is that what we think of as the early revisions, the BLS sends out these forms. Some firms respond promptly. Some firms take longer to respond. They will tell you, you know, if we're collecting, we just had the June report. We're coming up on the July report. They might not tell you their payroll look like in June until August, in which case it's going to come out with a September report. It's like responses. Responses have been falling over time. And particularly earlier.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“Yeah, sure. So the first thing that's think about this way one thing that could be done theoretically for Tongence Fiction is that in fact, you know, somehow we set up every business, including like pretty lo-fi businesses. I'm going to pick something. Like there's like a food truck in the corner and there's like that you probably operate. Some of them still operate with like physical cash registers. And we set them up with some sort of like super automated system and either it regardless of what they choose to do just sends information to DC. And by the way, this would not just have to do with how people they work, but also how much they pay them, how many hours these people work. And like, and just sends it automatically. And that would be, and then by the way, if they don't, or if it gives us some leeway, but if they choose to opt out.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“So, okay, I think that catalyzes the first burning question for a lot of folks, which is 2025. We have AI, we have all these technological innovations. Why do we have revisions in the first place? And to answer that, I would love to just better understand like Statistical process involved with this survey data? And why is it that the initial pass, you know, it's expected to have revisions? And then the second question associated with that is, yeah, what are normal revisions, quote unquote? And why is it that it gets more accurate over time? And why is it that we get this more blunt number up front and then it gets more accurate as we go forth? What does that process all look like?”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“There may be ways we can improve that that I'm glad to discuss. But part of it is just trying to get a more accurate picture. I mean, would we rather have that there were, you know, in the case the numbers are wrong, they just stay wrong? Like I don't, you know, I think it's sort of the nature of timeliness versus accuracy that exists”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“That's itself a complicated thing that we're probably going to talk about later today. Now, I think what's sort of bugging people with these revisions, and we're also going to talk about this, is like, first, the annual revisions have been negative. They were negative last year. I think they might have been negative the year before. And we're probably due for another big negative one this year. And on top of it, the monthly revisions recently have tended to be negative in the sense a number of you released and on average it'll be lower two months later. I think that in particular the most recent was it was very large in absolute terms. I mean, there were some charts that were saying it was the biggest since date X that didn't adjust for the size of employment. I'm not really a big fan of that kind of approach. I think you do want to scale it, but they were in the sense that they did change people's, I mean, there's no denying that they changed people's perspective on what employment growth had been recently. And unfortunately, that happens.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“Amazon is like it'd have like a bunch of establishments to pick one company. And there's a good chance, by the way, that in the case of Amazon, different establishments are in different industries because there may be people that are sort of information services because they handle servers. And then there's a part that's maybe non-store retailer and then transportation. Like just giving a sense that this is like a pretty rich thing. And those companies tell them this is how many employees we have, how many hours they worked, what we paid them in average, you know, in pay and which allows you to calculate things like the average work week and the average hourly earnings. And then zooming out, like I think the key thing, and I think what's controversial, and the non-form payroll employment estimate comes out of that survey. What happens with the establishment survey, but not really with the household survey, is you get these revisions, first the monthly ones, two months after one and two months after the initial release, and then on top of that, every year you get a big annual revision.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“Yeah, so you covered the first half, which is there's the household serving the establishment service where they go in the jobs report. The household surveys where you get data like the unemployment rate, the participation rate, prime work age employment population ratio, there's some other stuff in there. It comes from called the household survey, but it's like the official names, the current population survey. They call people households on the phone and they just ask them like, you know, how old are you how people you have in your household? They ask a bunch of other questions, but they also ask like, do you have a job? If you don't have a job, you know, you're actively looking for one, et cetera. And a lot of these data come out of that. And then there's a second survey that's the establishment survey where essentially businesses fill out forms and they tell people, and when you say businesses, we sort of think about firms, but these are establishments. The idea is like it's like a worksite. So it can be an office complex or it can be a construction site. So for example,”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“Super excited. So, okay, we are talking at a time where it feels like there's a lot of cross currents happening in terms of labor markets, labor market data, how it's gathered. We're recording this Tuesday, August 12th. We just had the announcement of the new BLS commissioner that Trump is nominating to be confirmed by the Senate, EJ Antony. So it's a very timely moment right now to really set the stage for a one-on-one style of just really understanding labor market data. How does it work? How does the sequences of all work? So yeah, nobody better to help us walk through all of this than yourself. So, okay, with that said, I want to start our journey with the NFP, otherwise known as the Drops Report. So obviously a lot of this is done with surveys. There's the household survey, the establishment survey. Then there's also the Challenger survey over there as well. So we'd love to just start by just getting us to understand the process of.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT
“The annual revisions have been negative. They were negative last year. I think they might have been negative the year before and were probably due for another big negative one this year. I don't want to say that like I'm glad to have these revisions. I just recognize this is not an easy problem to solve. There's all these huge things happening and they're pushing against each other. Maybe it is that the demand side effects of the tariffs are going to really start kicking in. That's one possibility. The other one is, by the way, the immigration squeeze is ongoing. My God is still that the weakening impulses will eventually kick in.”
2025-08-13 · Forward Guidance · Should We Be Worried About Jobs Report Revisions? | Guy Berger · IDENTIFIED FROM THE TRANSCRIPT