Odd Lots - Lots More on the Big Problem With the Monthly Jobs Report

Episode Date: September 4, 2025

We've been in a strange labor market for a while now. The unemployment rate is still nice and low at 4.2%. But the pace of job creation has been slowing markedly. And furthermore, not only has the pace of job creation been slowing, it seems almost every monthly Non-Farm Payrolls number ends up getting revised lower. Of course, this comes at a time of some big transitions in the workforce — whether we're talking immigration changes, aging demographics, or AI. As such, just understanding the monthly data has never been more difficult. And because it's so difficult, it's also challenging to get a read-through from data to policy. On this episode we speak with Steven Englander, global head of G10 FX research and North America strategy at Standard Chartered Bank. In addition to talking about the state of the labor market, we also discuss the goings-on in bond markets, and why the stress is particularly acute in Europe. Only Bloomberg.com subscribers can get the Odd Lots newsletter in their inbox — now delivered every weekday — plus unlimited access to the site and app. Subscribe at bloomberg.com/subscriptions/oddlotsSee omnystudio.com/listener for privacy information.

Bloomberg Audio Studios, Podcasts Radio News. Joe, I have an embarrassing confession. Go on. You know the birth-death adjustment? I have the same confession. I used to think it was something about, like, the population of people and the labor supply, but it's not. It's about business formation. Yes, that's right.

Oh, my confession is even more embarrassing, which is that I literally always forget what it is. I did a deadlift. I'm both the most popular. trader and most successful trader at Citadel. Fed is going viral. Barges. This is an after-school special except... I've decided I'm going to base my entire personality going forward on campaigning for a strategic pork reserve in the U.S.

Black gold! These are the important questions. Is it robots taking over the world? No, I think that like in a couple of years, the AI will do a really good job of making the oddlots podcast. One day that person will have the mandate of heaven. How do I get more popular and successful? We do have the perfect guest. You're listening to Lots More, where we catch up with friends about what's going on right now.

Because even when the odd lots is over, there's always lots more. And we really do have the perfect guest. When people are listening to this, it is Jobs Day. We're recording this September 4th, Jobs Day, September 5th. The important thing to know right now is that to some extent that monthly non-farm payrolls report that everyone depends on and relies on has never felt to me like more of a moving target. in terms of what I'm supposed to be looking at. For many years, it's like, okay, how many jobs were created this month?

Yeah, it's going to get revised a little bit, but that tells you something. Now between seasonality, post-COVID sort of normalization that's still processing. And then, of course, this sort of changes in immigration policy, which have swung dramatically in a year. These numbers, it's not clear that you actually have to put it in some work to understand these numbers. No, totally. And also, I mean, the BLS itself seems to have some difficult. with the numbers because the trend that we've seen is they put out a non-farm payroll, like an initial estimate on a Friday.

And then a few weeks later, you get the revisions, and the revisions always seem to be downwards lately. Lately, they've been mostly downwards. And then there's annual revisions, and those are coming up, I believe, next week. And, you know, we were at a time when setting aside the collection, and of course setting aside the fact that the BLS chief has been fired and there's going to be a new one at some point sitting outside all of these things. We're also at a time when there are significant questions about just the macro state of the economy

and whether the sort of low hiring, low firing mode, which is characterized for a while now, is at risk of deteriorating further. I don't know. It's a tough, it's a tough read. You know who we should ask, Steve Englander. Okay, but let me start with a broader question. Excellent. Excellent.

And the thing is that the question has changed. Like in the past, we used to say, okay, last month or, you know, last 12 months, NFP growth was 200,000, looks like 100,000 now, things have slowed down, that's all we had to know. This time around, because supply is so important, everyone is making an estimate and kind of saying, well, native-born workers, you know, maybe 70,000, if you're pessimistic, if you think we have net immigration from foreign-born workers, you have a view of 50, and if you think that there's still some legal immigration coming in, maybe you're at 100. And that means that getting the level right. is really important. And the problem with the NFP numbers that actually has two components,

one of which is the one that we all think about. They survey about 160,000 businesses, maybe 600,000 establishments, and basically say, okay, how many workers did you have last month? How many did you have this month? And through their statistical analysis, they say, okay, this is a change. So that's for firms that are in continual operation that gives them that number. The problem is, that they have no handle on firms that have just opened and very little handle on firms that have just closed. So they use something called the birth death adjustment to adjust for employment by those latter two categories of firms, you know, net job creation from new firms minus closing firms. And the problem is that they have a very simple model. And it was fine in the past when all you

cared about was the direction. But now when you're saying it really matters. if NFP gross is 50, having a bias in that number really affects things. And what you see is that that number has, through thick and sin, has stayed about 100,000 jobs, you know, when you seasonally adjusted, because they publish a not seasonal adjustment, but you can do a rough and ready seasonal adjustment. It's very stable at about every month. 100,000 jobs, or almost 100,000 jobs are coming from that birth, death adjustment in terms of what we see on Bloomberg page Friday morning at 830.

We have another source of information on that, which lags, but which is far more accurate. This is something called the business employment dynamics. It lags by about eight months, but it's based on the quarterly census of employment and wages. And what that shows is that in 2024, like Q4, 2024, say to four quarters, both the employment growth coming from continuing firms and continuing operation and the employment growth from newly open firms, less just closed firms, they've tanked. So you're sort of looking at a reliable source because it's not even a sample. This is the entire population.

They know what opened and what's closed, saying, hey, there just isn't any significant jobs creation from newly formed firms. And the NFP number keeps telling you that there's 100,000 jobs coming from that. This is already excellent. Next week, we are getting yet another one of these huge. QC Quarterly. Q-C-E-W. What do you call it? I call it Q-C-E-W.

What's the right? How do we call it? I call it Q-C-E-W. Okay. We get a new Q-C-E-W. That's the 9th, September 9th. Yes. We get that. Right, tell us what this is. Let's walk through this part again in terms of what is it that's high quality about this data, why we expect it to show further downward revisions, and why there continues to be the sort of downward

bias in this initial snapshot versus latter, better insight. Okay. First, NFP has a big sample, but it's a sample. They're sampling errors and people who don't report and you don't know if there's a bias and who's reporting and who's not reporting. So there's always some inherent error there. QCEW is basically the universe. It's not a sample. They get administrative data from the labor department saying how many people paid into unemployment insurance. And that's basically everybody because everybody does. Everybody who's working does pay into unemployment insurance. And so when they come up with a number, there's a bit of revision because there's sometimes some firms don't report in time, but there's not much. And it's very authoritative. And so if you sort of say, oh, QCW and the business employment dynamics tell us that in 2024, there was almost no job creation from newly opened, let's just close firms, you believe it because you're not going to get a better source. There's no other source to that. And so that's why it's used for the benchmark. It probably, should be used to re-benchmark employment more frequently in the year because they do publish it quarterly. But, you know, when it comes out, it's a big deal. And we think it's likely to tell us that somewhere between 750 and maybe 1.1 million jobs, that's the overstatement between Q1-2020 and Q1-25. Wow. And that would knock off a lot in terms of headline employment growth. The key point

is that there's no reason to believe that the bias is really shifted. One of the other unusual things that's going on at the moment is we have this new head of the BLS installed by the Trump administration. And that new head suggested initially that the BLS could just stop publishing jobs numbers altogether or maybe they could publish them less frequently, like on a quarterly basis or something. Is the quarterly idea, maybe that's reasonable given the lag with the Q-C-E-W data. I think that they can do things that are far less dramatic to improve the quality of the monthly numbers,

because we do have some information that's relevant for jobs creation by newly open firms. Going back to business employment dynamics, if you look at how job creation from existing firms, continuing firms, and job creation from new firms move, they tend to move together. The amplitude is different, but the direction is very much the same. they can use the sample data that they get from the 160,000 firms, 600,000 plus establishments that they sample, and say, look, if continuing firms are telling us that job creation is 20,000 a month, it's very unlikely that job creation from new firms is going to be 100.

We can use a variety of statistical methods to sort of make a guesstimate. It won't be perfect, but it'll be a lot better. Speaking of the new BLS person, Joe, I saw the most worrying sentence in an analyst note this week. So the sentence was, I have it on my screen. The sentence is, because Trump has his own BLS person now, I don't necessarily see a bad number on Friday. I'm not that pessimistic. I think, I don't know. I don't know.

I don't have a view on this. But who said this? I don't want to say, because I'm not sure it's public or. Oh, but... But this is a notable person? Yes. It's someone you know, actually. Okay, fine.

And not me. And not Steve. Not Stephen. Let's talk about the conversation has focused on these levels and pace of job creation and so forth. We recently did an episode with Austin Goolsby, the Chicago Fed president, and he was like, so much as a flux, I'm not really looking at levels. And he called what he's like, he used this term very dramatic, the for horseman of truth. He's interested in rates.

He's interested in the unemployment rate, which is still at 4.2%. He's interested at the hiring rate, which we had the Joltz report come out this week. It's at 3.3%. He's interested in the firing rate, which I think has still been fairly low. Maybe he's looking at like the rate of wage growth. I don't remember what the fourth horseman was. In a time of volatility, what do you think about this idea? Like let's just forget about levels and just focus on rates because they don't get revised as much.

You go out and you ask thousands of people, are you employed or unemployed right now? And if 4.2% say they're unemployed, that is probably a reasonable proxy for whether people are employed or not. Right. But Joe, if you turned out to be the next person named to the FOMC to the board, and you see payroll number and it comes out at 75,000 like tomorrow's consensus, is that a strong number or a weak number? Well, there's some say. I just look at the, I don't know. I mean, I don't know. Let's say, so right now, just for what it's worth, the unemployment rate is at 4.2%. Economist expected to tick up to 4.3%. Let's say it comes in 4.2%. Let's say we have a weak NFP number, but the unemployment rate stays at 4.2%. Why don't I say, okay, things are still more or less fine? Well, let me say this. I like rates, but I like the employment to population ratio. Okay. Because we know that participation rate is cyclical. And if you look to the employment to population ratio, all these comments that unemployment rate is stable, no, it looks in balance, nothing's changed. They don't hold up very well. We've had pretty consistent drop, not a 2008 type of drop, but, you know, was like over 60 a year ago.

Now it's 59.6. Okay. It's telling us that there's kind of increased softening. I think you can't be choosy about which rates you look at, but I think in cyclical, periods, probably employment to population is telling you more. And this time, it's more on Waller's side than it is on Powell's side. I know we've been talking about weaknesses in the NFP estimates, but what's your expectation for the official number? Well, we're a little bit split-brained on this, left-brain-right brain, because we live in the world where we have to get the market

reaction right. So our forecast is 75,000. It's very close to consensus, you know, nothing dramatic. I'd just say this, that you've got to realize there's so much randomness in the number. It could be 125,000 and not be meaningful, but it's just the way it goes. You know, what we've argued, if you look at the range of forecasts, almost everybody is between like 40,000 and 100,000, 105,000. So you get a number like 30. I think that will be a very dramatic number in terms of... If it's outside the range of expectations. Yeah, I mean, because this range is really tight in terms of market expectation.

And I think it would put 50 on the table for the Fed because I think the argument would be that if you get a number that's so low, you probably should have cut in June or July. And so you're not saying, oh, my God, the world's coming to an end. Everything's falling apart. But you're saying, yeah, well, we kind of missed it. The data weren't there at the time. In retrospect, had we had those data, we probably would have cut in June or July. So we're just doing catch-up now.

Now, for us, the real issue is how do you interpret like $100,000? because we would say, okay, 100,000 less our bias of 70,000 means real job creation of 30,000. So you should be talking about 50. I don't think we've convinced the market yet that that's the way you should look at it. By the way, just before going on, so on the eco page, consensus is 75K, consensus is for the unemployment rate to take up to 4.3% also average our earnings of 0.3% month over month growth. So that's sort of like what we're looking at. This week, we've had some sort of soft data. We had initial claims today coming at 237K.

That was ahead of the estimate of 230K. ADP employment, take it or leave it, 54K versus estimates of 68K down significantly from the 106K revised last month. ISM services employment, 46.5. That was a little bit shy of expectations. So soft and there's this risk of softening. I just think it's really interesting the range of possibilities for September because, you know, we were just in Jackson Hole. and you're talking about, oh, maybe there's a case for 50, but there's also still a lot of residual concern about inflation, and the inflation dragon has yet to fully be slayed. The reason I would focus on the labor market is because if it's as weak as we think it is

and correctly measured, the slack in the labor market will take care of inflation. Okay. That the, again, I observe more on Waller's side that you're not going to have any kind of power on the labor side. If I can make one comment, I mean, some people get their pleasure from banging their head against the wall.

The wall I bang my head against once or twice a year is taking all of these incoming data on labor market indicators and trying to predict non-farm payrolls. You got to have a hobby. You got to have a hobby. And this one might be the definition of insanity. And none of them work a bucket of spit. I'm looking at the Band-Aid on your forehead. When you talk about banging your head against the wall for pleasure, you're actually talking literally it would appear. Well, you should see the wall.

Okay. Okay, we would be remiss if when we have you in the studio if we did not ask about what's going on with bonds. So one of the things that's happening today is because of that softening jobs data that Joe just laid out. We are seeing a little bit of a recovery in bonds because there's more expectation that the Fed might cut. But the big story in recent days has been this huge sell-off in bonds, particularly at the long end. And it does seem- Particularly in Europe. Particularly in Europe, but also in the U.S.

It does seem kind of weird that we're talking about the economy slowing. And meanwhile, the long end of the curve just keeps going up. What's going on? Markets' attention span is maybe not as long as you give a credit for. And I think the problem is this, that you look at fiscal situations that are deteriorating globally and likely to deteriorate in the U.S., especially if anything happens with, the tariffs to pull them back. Oh, yeah.

And markets are kind of saying, look, in the long term, this doesn't look good. It looks like there's a lot of borrowing out there. But the long term, in market terms, can be six weeks and it can be six years. You don't know when those forces are going to matter. So absent anything else, the market pays attention to it. Then you sort of come in this week,

and every number seems to be soft. You say, oh, my God, the Fed's going to cut. And all of a sudden, this sort of selling of bonds, especially because the market probably got reasonably short, worrying itself about the fiscal situation, they say, oh, my God, maybe not so short, and people are buying back the bonds that they've sold, and you get the kind of dramatic movement we've seen this week. We've argued in the short term, both FX and bonds are going to be driven by the Fed and the U.S. economy, and by short term, I mean the next month.

Yeah. There's six weeks. I think once you get past that, we don't think inflation's going to disappear. And even outside of the tariff-induced inflation, the non-tariff goods, services seem to be at best state steady and even maybe even edging up a bit. There's a bit of fiscal stimulus in Trump's fiscal package. Tickets possible that the economy is not as bad as it might look based on the employment numbers that we're getting, especially if there is a productivity pickup in the data that's not really recognized. So we see a possibility, or we actually see more than a possibility. We expect all their weakness in the next couple of

weeks, but by time we get to the end of the year, we could see the dollar strengthening, positioning is kind of short dollars in our view. And same with the bond market. You know, while the market serves saying, oh, my God, last week was debating zero and 25. This week, I suspect they'll be debating 25 and 50. While the market's debating that bond yields are going to come down, once that's kind of settled in terms of market expectations, I think they'll look at the fiscal picture and kind of say, you know, low fours, maybe not. It's interesting. It occurs to me when you say how, when we talk about how bad the economy is, that there's actually two different ways of what bad could mean. So one is bad could just mean a sort of where you are on the cycle.

We're in a deceleration cycle. We're not creating as many jobs. Therefore, we have to have a lower rate of interest to get things going again. But then bad could also mean more sort of qualitatively where what's really bad is slackening growth and also a firm inflation picture such that traditional measures of policy stimulus or de-stimulus don't work as well they'd like because there is some sort of deeper rot. And it sounds like what you're saying, if I can put all this together, is that cyclically, there's a slowdown, there may be a case for 50 basis point cut very soon, a case for lower rates, but not so bad in the case of the U.S. economy is broken and therefore policy measures won't work to get it revived again. Yeah, I think that's right.

We actually don't think they'll do much more than 50. They'll say, okay, we've caught up, time to wait and see. And our baseline, because we're not yet sure about what the employment number is, is 25 and stop. But, you know, we've been talking about the choice being between 25 and 50. Just adding one more. It really is just on the Europe situation. We're mostly talking about the U.S. Like France has also their 10-year yield has been shooting higher.

Yeah, this is the crazy thing. I think if we didn't have everything that was going on in the U.S. at the moment. That's what I'm saying. The France story would be huge in the market. Yeah, so give us your talk to us about Europe for a minute. Yeah, look, there's a France story. I'm actually scheduled to go to France, but it looks like they might be on strike next week, so I'm not sure how that's going to play out.

The structural problems that they face are kind of enormous. I mean, the French deficits about as large as the U.S. deficit, and their interest rates are way lower than U.S. interest rates. They have a very fragile government. They're trying to do some sort of fiscal consolidate. because it's like it's not just a minority, it's like a tiny government. The opposition parties are kind of saying no way. And so that's setting the U.S. aside, the French and broadly speaking the European situation isn't that good.

And in some cases you're seeing sterling trade like an emerging markets economy, in that the correlation of interest rates in the currency is not the normal G10 one where higher rates lead to a stronger currency because people look at the return. The higher rates are viewed as risk premium and they're associated with a weaker currency. I mean, yeah, you can look at levels, sorry, just to keep going, you can look at levels of debt to GDP or ratios, et cetera. But it sounds like the common thread here, at least to me, and this is the bias or this is my lens when you talk about what's going on in France. Maybe you won't even be able to visit there because of strike. When you talk about we had Liz Trust on the podcast recently, politically, these are not well-functioning policies right now, are they?

Yes. And I'd say that the ability, particularly to get through unpopular measures, like fiscal consolidation, is very limited. And that's why they're all sort of have their backs against the wall. And the markets are looking at this and kind of saying, you don't really have a source of growth and you have to do austerity, but you're not. So we're looking at these ratios and where they're going to go. I think the one advantage the U.S. has is that here, at least you can tell the story where you say, look, some of the productivity numbers look pretty good on the ground. You sense that AI is making inroads. We don't know when it's going to matter in terms of actual realized productivity, but it could.

You can tell a story that's somewhat optimistic. Whereas if you look at Europe, they kind of have high energy costs. Their capital markets aren't as well developed for financing these kind of innovative firms. And they're kind of lagging on the technology side. So the battle seems to be between the U.S. and China and other countries sort of really lagging. Which doesn't mean the U.S. outcome is going to be great, but at least you can tell the story, whereas it's much harder to tell the story, a European story of, say, private sector induced growth. What's that line again, Joe? It's like the U.S. innovates, China iterates, and Europe writes the regulation. Do you remember that?

Yeah, and they also get tons of vacation and they have this amazing life. And put out think pieces. Yeah, which sounds great. So many people would kill for a job writing think pieces. I lived in France for a number of years. And I'd say the quality of life of the median French worker is well above that of the median U.S. worker. That's what I'm saying. We need to have some, this is an important point. You observe that.

Yeah. Yeah. And the question is whether it's sustainable. It's like living on your credit card. We need to regulate more. We need to, instead of innovating, let's try regulating. No, we should go drink some wine for lunch. Yeah, sounds great.

Great suggestion. Lots more is produced. by Carmen Rodriguez and Dashel Bennett with help from Moses Andam and Kail Brooks. Our sound engineer is Blake Maples. Sage Bauman is the head of Bloomberg podcasts. Please rate, review, and subscribe to Odd Lots and Lots More on your favorite podcast platforms. And remember that Bloomberg subscribers can listen to all our podcasts ad-free by connecting through Apple Podcasts.

Thanks for listening.