Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Economic Data Under Fire: Accuracy, Trust, and Transparency with David Wessel
312 Plays3 hours ago
In this episode of the show, I talk with David Wessel from the Brookings Institution about the state of the US economy, the reliability of government statistics, and why trust in data matters more than ever. We cover the latest job numbers and what downward revisions mean, the pressures facing agencies like the Bureau of Labor Statistics and Census Bureau, and how politics threatens the credibility of official data. David shares insights on the risks of eroding confidence, what businesses and governments might do in response, and why high-quality, transparent data is essential for policymaking and for the public. It’s a timely and important conversation about the role of data in shaping our economic future.
Keywords: David Wessel, Jon Schwabish, PolicyViz Podcast, Bureau of Labor Statistics, BLS, Census Bureau, U.S. economy, economic data, data revisions, trust in data, survey fatigue, government statistics, job market, unemployment, economic indicators, federal data, data transparency, data integrity, public trust
Subscribe to the PolicyViz Podcast wherever you get your podcasts.
Become a patron of the PolicyViz Podcast for as little as a buck a month
Follow Giorgia on Twitter and find her book “This is Me and Only Me” on Amazon
Follow me on Instagram, LinkedIn, Substack, Twitter, Website, YouTube
Email: jon@policyviz.com
Recommended
Transcript
Introduction and Episode Context
00:00:13
Speaker
Welcome back to the Policy Viz Podcast. I'm your host, John Schwabisch. Yes, another episode of the show. I know you're surprised. You've gotten used to this every other week cadence for the last few years and things have sped up.
00:00:25
Speaker
I've had a bunch of episodes coming out one after the other. Well, there's a lot to respond to in the world of data and statistics and data visualization. So I am working fast and furiously to get that timely information to
Challenges in the American Job Market
00:00:40
Speaker
you.
00:00:40
Speaker
And on that note, today, i welcome David Wessel, director of the Hutchins Center on Fiscal Monetary Policy at the Brookings Institution. David spent a long time at the Wall Street Journal before going over to Brookings.
00:00:53
Speaker
He and I have worked together for many years. most recently on the data user advisory committee at the BLS, which has been canceled under the new administration in the White House.
00:01:03
Speaker
But I invited David to join me on the show to talk about the state of the American job market. And also importantly, I think for people who are listening to this show, the state of the federal statistical and data agency ecosystem. What does it mean to have revisions to BLS or census data?
00:01:23
Speaker
What does it mean to have attacks by members of the political class on the BLS, the Census Bureau, the Bureau of Economic Statistics? So we talk about all these different issues. We talk about what private businesses may be doing to fill in the gaps, what our state and local government agencies doing to fill in the gaps, all the sorts of things that are important.
00:01:44
Speaker
when it comes to the bedrock of data visualization, which is obviously the primary focal point of this show, but the baseline, the bedrock of good data visualization is good data. And if we don't have good data, then we are just painting pictures.
Importance of Quality Data for Visualization
00:01:56
Speaker
So David is an expert in the field and he has been talking about and thinking about these issues for a long time. So I was very pleased to have him join me on the show for this week's episode of the policy, of this podcast.
00:02:09
Speaker
And so here's my interview with David Wessel from on the Brookings Institute.
00:02:15
Speaker
Hey, David. Good to see you again. Good to see you, John. It's been a while. think maybe last time I saw you was one of the BLS advisory council May it rest in peace. Yeah.
00:02:25
Speaker
Unfortunate. Just one of the other things that we now lose, I think. um So there's a lot to talk about. I want to start with economics and then talk about economic data, if we could. So...
00:02:37
Speaker
Job numbers have been coming out fast and furious the last couple of months showing, I think kind of a weakening or slowing job market. And I'm curious what your take
Economic Uncertainty and Job Market Revisions
00:02:45
Speaker
is. It's sort of mid September now. So kind of in between the two numbers, but what do you see happening? What's your take?
00:02:53
Speaker
Well, it seems to me that the job market has is weakening. i i think the recently revised numbers underscore that the pace of hiring has slowed.
00:03:05
Speaker
And I'm also a bit influenced by the historical pattern that when there are a lot of downward revisions, that often happens at turning points in the economy. um So my story seems to be that there was a lot of uncertainty we've worn out two words, uncertainty and unprecedented.
00:03:26
Speaker
um There was a lot of uncertainty and there was some hesitation and people weren't sure what was going to happen to tariffs and stuff, but it feels to me like we're past the waiting period and employers are beginning to plan on a slowing economy and that means less hiring.
Post-COVID Recovery in Various Sectors
00:03:45
Speaker
Yeah. Are you looking at this by sector? Are you seeing, for example, auto sector being slower than others? Or is it too early to tell, too soon to tell? Well, for a long time, it looked like we were in a period of post-COVID catch-up.
00:04:02
Speaker
So like ah one of my colleagues here has a nice chart showing employment by sector compared to where it was today. just before the pandemic. And you could see that hospitality was taking a long time to get back to pre-COVID levels. Maybe it never will.
00:04:19
Speaker
But basically, I'm not seeing much. I haven't looked much sector specific. ah I'm thinking more of the general zeitgeist of the economy. Yeah. So you you mentioned the downward revisions, and maybe I could ask you two questions here.
00:04:34
Speaker
One, to explain from your perspective outside BLS ah why these downward revisions happen and what you take from them and what people who are not familiar with these numbers, with these revisions, what they should...
00:04:48
Speaker
think about because I think we've seen attacks on the BLS and other statistical agencies about changes to data that in many cases are probably unfounded. And and so what should people know about these revisions that mean that they can still trust numbers they are coming out of BLS and census?
Understanding BLS Data Revisions
00:05:05
Speaker
So I know this may sound counterintuitive, but I think people should have more confidence in the Bureau of Labor Statistics because they issue revisions as they get more information.
00:05:17
Speaker
It would be nice if they got it all right the first time. But on the payroll jobs numbers each month, for instance, they have a survey. ah They ask 120,000 businesses, how many people are you employing?
00:05:32
Speaker
And a lot of people don't respond. on the first go round. So they get more responses on the second go round and they revise the numbers. And then on the third go round, they get still more. So as if I would rather they were getting the numbers right the first time, I have confidence that if they get the numbers wrong or as they get more information, they're not afraid to revise them.
00:05:53
Speaker
um As you know, there was this big annual benchmark revision where they revealed that we had 911,000 fewer jobs in March 2025 than they had previously estimated. That's about six tenths of a percent is the total number of jobs in the economy.
00:06:10
Speaker
And that's because every year they take all the unemployment insurance records and they see from a much bigger set of companies, nearly all employers, how many people they had on their payrolls.
00:06:21
Speaker
So clearly the first set of numbers are a survey. The survey is flawed. they um They're having bad response rates, particularly in the first round, and they're updating them.
00:06:33
Speaker
So um i I know that it's upsetting to people to see like, wait a minute, couldn't you get it right? um I'm kind of reminded there's this great Federal Reserve transcript where the staff reports that the Bureau of Economic Statistics has reduced their estimates of GDP growth.
00:06:52
Speaker
And then Fed Governor Ben Bernanke says, well, if they can restate GDP growth retroactively, can we do the same thing on interest rates?
00:07:03
Speaker
so So let me put it this way then. So now the BLS... Another commissioner has been nominated, there'll be hearings, but even with this confidence that you think people should have in the agency, what do you think they should do?
Enhancing Data Collection Processes
00:07:19
Speaker
Or is there anything they should do to A, increase that confidence and B, just generally speaking, going forward, like what are the steps, what are the procedures you think that they need to implement to you know make sure that people are confident in the data and that the data are still high quality?
00:07:32
Speaker
Right. So I think there's a bit of uneasiness among the people in the outside the statistical agencies who have for a long time said, you know, you could do better.
00:07:45
Speaker
yeah There are things you could do better. And then the agencies say, yeah, well, we need the resources and um etct etc, etc. So I think it's important that there be a process of continual improvement in the way they collect data, particularly given how technology makes it possible. And they've taken steps in this direction to get stuff at lower cost by scraping the web and so forth.
00:08:09
Speaker
um But as you know, John, they've sort of been starved for resources and that makes it, it's great to say, oh, you should do this. I know there, I've talked to a number of business people who say, why don't they just use private sector data? And I've pointed out to them two things. One is, you know, some of these private sector data providers to charge the government for them.
00:08:31
Speaker
It's not free. And secondly, the data isn't collected. ah with measuring the economy as its objective. It's collected for business purposes and it's not always a representative sample and the historical ah continuity is not always there.
00:08:47
Speaker
um Third, I think that um the the problem we have is that the president has created the impression that he's going to fire people who don't give him the numbers they want.
00:08:59
Speaker
um I like I've likened this to your team is losing the baseball game. So you fire the scorekeeper, but your team is still losing the game. Right. And so I think that he's eroded people's confidence that these are hardworking technicians who are doing their best, making mistakes, but making them not because of some political agenda. and I think it's important that that people in the agency stand up if they're being pressured so the rest of us know it's happening.
00:09:29
Speaker
And it's important that people outside the agencies emphasize the importance of having good, unbiased data, even if it's imperfect.
Political Influence on Statistical Data
00:09:40
Speaker
What, from your perspective, is the long-term risks of those continued attacks, be it on management or on the on the analyst level?
00:09:48
Speaker
Well, they've already lost a lot of people at the BLS. yeah um And I worry that they they're going to have trouble attracting competent people. Also, if people don't believe the numbers, then they're going to at least open the possibility that ah other numbers, which may not be ah as well founded as the BLS numbers, will take over.
00:10:11
Speaker
mean, one thing that's just a good example of how dangerous this is, the the Treasury sells a lot of bonds that are inflation protected, and your return on the bonds is pegged to the consumer price index.
00:10:25
Speaker
And if people think that the consumer price index is being monkeyed with for political reasons, are they going to be willing to buy those bonds? Or if they do, are they going to demand a premium, which means the government has to pay more for them?
00:10:39
Speaker
So I think and I think also that it's not just what's going on at the BLS. It's what's going on at the Federal Reserve. It's what's going on in the firing people at independent agencies.
00:10:52
Speaker
it's the attacks on higher education. So the whole thing tends to erode people's confidence or trust in the government, which hasn't been that high anyways. And so that does two things. One is it makes it really hard for the government to actually fix things that need to be fixed that require the trust of the people.
00:11:10
Speaker
And secondly, the United States is the world's largest borrower. And if people start to suspect that our numbers are being cooked, our central bank isn't independent, they're going to demand slightly higher rates when they lend us money, and that's going to cost us all.
00:11:25
Speaker
Right. Now, in your conversations with businesses, have you heard businesses making any sort of pivot in their reading
Shift Towards Private Data and Local Governments
00:11:35
Speaker
of data? Are they relying more on their own internal data or data that they can buy from other private sector organizations?
00:11:42
Speaker
Well, I think, first of all, businesses in general on this question have been profiles of no courage. It's very hard to get ah the chief economist at a Fortune 500 company to come on the stage at Brookings and say, good economic data is good for business, good for the economy, good for America.
00:12:02
Speaker
Secondly, i think everybody is thinking about, well, if we, mean, what happens if we wake up one morning and ah the Bureau of Labor Statistics says, well, we don't like all these provisions, so we're only going to issue numbers every other month or something.
00:12:18
Speaker
I think people will turn to alternative. There are some alternative data sources out there. um It's uncomfortable because a lot of them are a benchmark to the BLS or the census data. But yes, I mean, I haven't heard anybody who's any more relying on private data.
00:12:35
Speaker
I mean, as you know, a lot of businesses seem to rely more on anecdotes than data when it comes to judging the economy anyways. But I think people are beginning to say, what what are the alternatives? And I think they willll have we have some options. I mean, you know, there's this...
00:12:50
Speaker
um people who did the Billion Price Project who scraped retail prices at five major retailers, they got going because Argentina was cooking the books on its inflation numbers. So they tried to come up with an alternative. And I think that people will be, if the government numbers start to differ substantially from what privately gathered data suggests, and there starts to be reports from insiders at the agencies that they're not allowed to do their jobs, then people will begin to rely more on private Yeah.
00:13:21
Speaker
And have you talked to anyone have you thought about what folks at state and local level, particularly in the government should do? I mean, ah just take, for example, you know, ah you know, maybe faith in the CDC vaccine recommendations is going to decline. and So instead of going to the CDC website, I go to the, you know, my state's Department of Health ah website, like, what are some of the solutions or actions that state and local agencies might need to take?
00:13:51
Speaker
That's a good question. think, i think To the extent that state and local governments are relying on federally gathered data, they may want to do more data gathering on their own.
00:14:03
Speaker
um And it's not as hard as it used to be. It's difficult, um especially given people have survey fatigue. But yeah, if I were running a state or a local economic development thing, I'd want to think about, is there a way that I can get a handle on what's happening in my state that doesn't depend on BLS.
00:14:24
Speaker
um And I think also that they have a responsibility to help build trust in the local government even if people are losing trust in the federal government. Because as we know, the surveys generally show people seem to trust their local government more.
00:14:37
Speaker
he um you know ah But it's it's a bit scary. And it's scary because things that we used to take for granted, we can no longer take for granted.
00:14:51
Speaker
I don't believe that the any of the numbers in the BLS or the Bureau of Economic Analysis or the Census Bureau have been cooked. um I think we would know if they had been.
00:15:02
Speaker
But am I 100% confidence that I will be able to say that in three months, six months, nine months, and 12 months? No. And that's not something that I would have thought in my 35 years in Washington, I would ever say.
00:15:15
Speaker
Right. You mentioned the Census Bureau. I'm i'm curious on your thoughts about this I'll call a proposal. I'm not sure how formal it is, but the the administration has proposed either sort of change in the census or some sort of other census that wouldn't count people who live in the country without legal status. And I'm curious what you think the impacts of that would, I mean, we can put aside whether they would actually do it, but like what the impacts would be of of that alternative would be.
00:15:46
Speaker
Well, the first thing to say is that this kind of, is an ugly reminder of some of our history, you know, the three fifths compromise where we count some people as less than other people. In that case, it was slaves and it was largely for political reasons.
00:16:04
Speaker
So obviously there are a lot of communities that have higher than concentrations of people without who are undocumented. And so they would lose census money and you know, local so federal aid to education money, and maybe even congressional representation as a result.
00:16:27
Speaker
I don't quite, i understand the president's argument, like, you know, you should only count citizens and all that stuff. um I haven't actually, it's a good question I should look, maybe you know, John, how it works out regionally. I mean, for instance, so the Republicans seem to think that Texas is a right place to elect Republicans.
00:16:47
Speaker
But my guess is that Texas has a disproportionate number of undocumented. Would that reduce the ah growth of the Texas population enough to affect one congressional seat?
00:16:59
Speaker
um So and I can't think of anything good to come out of it. And it's part of a rather unfortunate or not rather very unfortunate xenophobia and war on immigrants.
Impacts of Census Changes on Funding and Representation
00:17:09
Speaker
Yeah, I did.
00:17:11
Speaker
um just looking this up because I did some work on this and basically looked at the share of people without legal status that were calculated by Pew as a share of the total population.
00:17:25
Speaker
and I think the largest, the state with the greatest share is Nevada followed by Texas, followed by New Jersey and Florida, and then California, a little bit, a little bit lower down. So, um,
00:17:41
Speaker
Yeah, I mean, it would definitely impact certainly potentially the apportionment if that were if it were used in that way, but also the provision of of federal state grants.
00:17:53
Speaker
exactly and And do you think that that's just ah generally speaking, that people just don't understand how census data are used or people just don't care? And and it's more of a, at this point, it's up it's a political argument. And so, you know, people are okay with that.
00:18:10
Speaker
Well, first of all, i't know who these people are you're talking about. i mean I'm sure that most people have no idea that this is going on, right? There's a little bit of a Washington bubble. They can tell you that the price of coffee is up a lot. Right, yeah.
00:18:26
Speaker
You know, I know there's been all these campaigns every time we do the decennial census to help people understand that you should really cooperate because it may have to do with how much money your a school local school gets. yeah But in general, I don't think people understand.
00:18:41
Speaker
And I think it's mostly part of um President Trump picking up on the antipathy that some Americans have towards immigrants. And it's just ah just one more facet of that.
00:18:52
Speaker
Yeah. um But yeah. yeah And as you say, I don't know whether it's legal under the 14th Amendment and all that stuff. That's going to be on my pay grade. Right. um Can you help people? We've talked about falling response rates just a little bit, but can you help people understand?
00:19:10
Speaker
I mean, the obvious answer is survey fatigue. People just don't like to answer surveys. But is there a ah more nuanced answer to that question of why? and Yeah. I mean, first of all, the survey fatigue is a real thing.
00:19:22
Speaker
I mean, if I go to CVS and buy a pack of gum, I get a text asking me to rate the service at CVS. I went to the dermatologist and I got something asking me to fill out. a risk I mean, it's like enough already. Right. You know? yeah yeah um um And of course, the only reason i ever filled about it is I'm pissed off. at them so yeah um So I think there is survey fatigue.
00:19:48
Speaker
I think there's a lack of trust in government. And it's um why should I give you this information? You might do something with it I don't like. And the the Trump administration has not helped on that regard.
Survey Response Rates and Trust Issues
00:20:03
Speaker
um And then i think the other thing is that. So i these things are the same. One is trust in government and the other is privacy. And people are some people.
00:20:14
Speaker
I mean, it's kind of interesting. I find the younger people are less worried about that. They assume that everybody knows what they're doing. They put it on Facebook and Instagram. So off one more thing. yeah um But I do think that people worry that.
00:20:26
Speaker
because of all the reports of government data being hacked and stuff that they're doing. And fourth, um you know, whenever you get a call or an email asking for information, you do have to ask yourself, is this some kind of phishing scam?
00:20:40
Speaker
Right. And so I think there's lots of things going on. um But, you know, the BLS problem is not limited to households or individuals like me and you. It's also businesses.
00:20:51
Speaker
And so i just don't think that people see it as their They have a lot going on there. The businesses don't see it as their patriotic duty and they don't see the value in responding to these surveys.
00:21:03
Speaker
um That's the best I can do. So do do you think that's a ah a problem that can be solved? I hope so. um You know, the Federal Reserve does this extensive survey of consumer finances and they had trouble getting people to respond, particularly rich people, and so they pay them.
00:21:27
Speaker
So um I think that we can make the surveys easier. We could probably think harder, and I know the census and BLS think about this. you know When you tell somebody, this is going to take 40 minutes your first reaction is, yeah, right. ah And then also, we do know that there are now more and more administrative data that can allow us to get a lot of information without doing the surveys.
00:21:57
Speaker
We still have to do surveys and the administrative data is not always up to date, but um More and more, we should be able to get some of the information from administrative data, government records.
Improving Survey Participation
00:22:11
Speaker
um I mean, JPMorgan Chase does a hell of a lot anonymized with the people who have accounts at JPMorgan Chase telling us a lot about the economy. So yeah, I think we could figure out how to do better. We need to make it easier for people.
00:22:25
Speaker
We need to help them understand how the data is used. I don't know about you, John, but every once in a while I get a survey from some graduate student or some economist or some public, some PR firm that wants to know you know my impression. And mostly I turn them down because I'm too busy. But I do appreciate it when, if you do participate, then they send you the results.
00:22:49
Speaker
So maybe there's a way to get people to see little bit more about what your information was used for. Yeah. I've never been in the American Community Survey, but that's always been a but always even a hope. yeah Yeah, I know.
00:23:02
Speaker
yeah I know. i guess Exactly. Yeah. I mean, like, you know, I'm one of the people who really likes it when I get the long form for the census. You're right. Exactly.
00:23:12
Speaker
This is good. I'm going to take some time and really sit down with this. Yeah, absolutely. um So we've we talked a bit about ah solutions and looking forward, but I'm curious, what do you think ah people listening to this podcast, right?
Defending Statistical Agencies and Government Data
00:23:26
Speaker
Who are working in data, they're visualizing data, they're, you know, they're analysts and communicators, like what should or can they do to think about helping protect statistical data agencies?
00:23:38
Speaker
So I think there's been a concerted effort by, people like the ones you've had on the program, the former commissioners, the BLS and stuff to defend the agencies and to say they're not perfect, but we've made their life miserable by starving them for resources.
00:23:56
Speaker
I think we have to be um very quick to condemn any efforts that we see to manipulate the data. um You know, i think we're all counting on the whistleblowers.
00:24:08
Speaker
um But then I think we do have to prepare for if the government data is, even if it's not corrupted, but it's just being questioned, look for ways to come up with privately gathered data so we can see whether how far off the government data is and to think about how to do that.
00:24:28
Speaker
Yeah. um David, anything I missed that you think is important for people to know when it comes to these, I would say, regularly produced data from the federal side? And we've talked a lot about the BLS, the Census Bureau.
00:24:44
Speaker
I mean, there's just so many data sets out there for people to think about. but Right. So I think there's a number of different uses of these things. One thing is I think people... need to understand that some of this stuff really, really goes directly into things like escalation clauses in in leases or, ah it you know, there's productivity adjustments to what we pay doctors for Medicare.
Role of Accurate Data in Policy Making
00:25:10
Speaker
or as I pointed out earlier, the inflation index bonds. So some of this stuff, it's not just getting a nice picture of the economy. It's actually making adjust, or I should have said, adjusting the tax brackets.
00:25:23
Speaker
Right. ah Stuff. So like getting these numbers right affects us all because of all these escalators. Secondly, um look, we are we are dealing with a very...
00:25:35
Speaker
complicated and rapidly changing economy. And ah I'm enough of a nerd to think that we need to be always looking at what's happened, ah how much did the letting China into the World Trade Organization affect employment and where, and how should that inform our relationship to China?
00:25:59
Speaker
If we worry about why some children of poor people seem to have a hard time getting out of poverty themselves, which government programs, whether it's food stamps or Head Start or Medicaid or giving money to schools or whatever, have the most bang for the buck?
00:26:17
Speaker
And people have a lot of opinions about that, and it's easy on the bumper sticker level. But if you really want to make good decisions about how do we make life better for most Americans, um you do need some foundation of academic, rigorous research, which requires data.
00:26:38
Speaker
And I don't think people generally understand how much of the social programs we have are influenced by the research that people have done before. I mean, just take for an example, the Earned Income Tax Credit.
00:26:51
Speaker
It's one of our biggest anti-poverty programs. It's a bonus that we pay low-wage workers. And it has survived in part because there's good data on what effect it this has on people's willingness to work or the well-being of their children and stuff like that.
00:27:08
Speaker
So without good data that honest researchers can do, And sometimes we'll discover some programs don't work and we can get rid of them. We'll be ah driving blind. And so i think it has both short-term, the tax indexation and long-term implications if the data gets worse.
00:27:27
Speaker
Yeah. David, thanks so much for coming on the show. Really interesting talk to you again. And hopefully they'll have a user advisory group put back together at some point and we'll see each other then.
00:27:39
Speaker
I hope people appreciate that. We had to pay for our own lunch at those things. That's right. That's right. We weren't wasting any government resources. That's right. That's right. All day.
00:27:50
Speaker
All right. Thank you, David. Appreciate it. All right. Take care. Thanks, John. You too. Bye.
00:27:57
Speaker
Thanks for tuning in everybody. Hope you enjoyed that. i Hope you learned a lot. Please make sure you stay tuned to the show. I've got lots more episodes coming up. Gonna do some pivoting here and there. Gonna talk about data visualization. We're gonna talk about data visualization in the real world.
00:28:12
Speaker
We're gonna talk about maps. We're gonna talk about some books. I've got some good stuff coming your way, as well as continuing to talk about threats to the federal statistical and data agency ecosystem and threats to federal and local data.
Conclusion and Listener Call to Action
00:28:25
Speaker
Please be sure to rate, review the show wherever you get it. Hope you're enjoying it. And if you want mine, if you have one of my books, Better Presentations, Better Data Visualizations, Data Visualization in Excel, please take a moment, rate or review the book wherever you get it, Amazon, Goodreads, or wherever else you are purchasing your books.
00:28:43
Speaker
So that's all I've got for you. Until next time, this has been the PolicyViz Podcast. Thanks so much for listening.