Market-Moving US Economic Data Is Under Threat

Bloomberg Audio Studios, podcasts, radio news.

Every month, the US government releases a lot of data to the public.

The unemployment rate four point two for shit, a CPI report with a two handle two point nine percent year over year.

It's a big day for the markets.

We get the release of the September jobs report.

From inflation numbers to unemployment rates, it's the kind of data that's meant to give a snapshot of how the country is doing economically. When that data gets released, it can move markets. Politicians use it to tout their progress or criticize their opponents. The Federal Reserve keeps tabs on it as it sets interest rates, and local governments use it to assess things like how much new housing they might need.

You can't make good decisions with how good data. Like when you're thinking about what you're going to wear, most people check the weather.

That's Bloomberg editor Molly Smith. She's spent the past six months looking into the agencies that collect this data. She says they're under threat. Agencies like the Bureau of Labor Statistics and the Census Bureau are already struggling to operate with limited funding, and looming ahead is President elect Donald Trump's promise to slash federal bureaucracy and find cuts across agencies.

Statistical agencies from the federal government are funded by Congress, and there are people at every step who are looking to cut.

Today on the show What's at Stake if the government scales back its investment in good data from Bloomberg's Washington Bureau. This is the Big Take DC podcast. I'm salaiamosen to get a handle on how the economy is doing at any given moment. The US government tries to collect data through things like the current Population Survey.

A lot of people colloquially refer to it as the household survey in the Jobs Report.

That survey is run by the Bureau of Labor Statistics or BLS, one of the main agencies in charge of data collection for the government.

This is where we get the unemployment rate, the labor force participation rate, information about households broken down by race, gender, age, disability status, veteran status, you name it.

The BLS gets that data by contacting a representative sample of the population and then weights those results to reflect the broader population. The trouble is that sample is at risk of getting smaller, and one key reason for that is how much money the BLS has to conduct those surveys.

The BLS commissioner announced back in June that they were going to have to cut the sample size, and that was really a very historic step and alerted a lot of people to just have how serious these budgetary issues are.

Since twenty ten, the BLS's budget has fallen by almost twenty percent when adjusted for inflation, and those trends are playing out at other agencies like the Census Bureau and the Bureau of Economic Analysis. A six million dollars boost from a stopgap spending bill that Congress passed in September staved off immediate cuts to the survey sample size of sixty thousand households, but the bureaus will need more money in twenty twenty five, and beyond money that could be

It makes it perhaps less representative of the population you're trying to capture. If there are patterns between who answers and who doesn't, that also can introduce bias into the data, and that really raises a lot of data integrity questions.

And bad data can have big consequences. Mollie cited an example from two thousand and eight, at the height of the financial crisis.

When the FED was trying to understand what was happening to the economy in real time, it seemed like they were pretty unaware of just the extent to which the economy was contracting.

In January of two thousand and nine, the Bureau of Economic Analysis initially estimated that gross GDP in the US shrank by three point eight percent from the third to the fourth quarter of two thousand and eight. In twenty eleven, it revised that number to eight point nine percent, but that was after the Federal Reserve had looked at those initial numbers and decided not to cut rates.

They didn't really seem to see a huge threat to the financial system at the time. That just wasn't clear to them.

Another example occurred in the early months of the pandemic. When response rates dipped dramatically.

In April of twenty twenty, the bureau vastly underestimated the number of people who were unemployed.

That's William Beach, who was Commissioner of the BLS at the time.

Same thing with prices. We really didn't know what the inflation rate was, and so we got a good vision of what it would mean in the future. If our regular surveys during normal times had really bad response rates, we would be making policy in the dark. We would not making informed decisions, and that's a good prescription for making policy mistakes that could have great consequences for the economy.

Beach says the BLS has struggled with response rates for decades now. He says that people just aren't as willing to pick up the phone or let a census officer into their home as they used to be.

We're getting saturated by surveys. Everybody wants to have our opinion, and so the government, which needs your response to do these really essential things, is victimized by the fact that households are getting less and less tolerant of answering questions.

The BLS doesn't need every person they contact to respond to their survey for the results to hold. But the smaller the response rate gets, the less reliable the data is, especially when you get down to the more granular demographic breakdowns. Beach says that more funding could help agencies level up their data collection, but what's getting lost in the meantime is crucial for painting a clear picture of the country.

If the response rate gets too small, it'll be difficult for us to represent the whole population. We'll have to drop certain questions, like we'll probably have to drop the question about are you Asian American? Are you Pacific Islander? Are you Native American? Because we'll have too few households that are actually that way. So that's one of the reasons we're very concerned about it. We have ways we can fix that, but they're going to cost money.

When we come back. What agencies think they could do with more funding, and why getting that funding under the Trump administration might just be a pipe dream. Government agencies collect crucial data about everything from unemployment to inflation are struggling with limited funding that can sometimes lead to bad data. William Beach, the former BLS commissioner, is now co chair of an organization called the Friends of BLS, which independently

We have all of these internet resources and cell phone resources that we could employ. People would be given a little cash inducement to answer a survey during the month on their cell phone or through email, and we could send that survey out to a million households and if only twenty percent answered, that's two hundred thousand households that would give us a response. We would then combine those responses with the survey responses, and that's called blended data.

But Bloomberg editor Mollie Smith says Congress might not see it that way.

Some Republican senators they almost insinuated in this recent letter that they sent to the BLS that the agency is perhaps prioritizing speed over accuracy. They raised an interesting point of why does the Jobs Report have to come out on the first Friday of every month? Is it a matter of that you need more time? Like what if it came out a week later.

These senators may be raising these questions because of a few high profile incidents that occurred at the BLS over the past year.

There was an early release of inflation data back in the spring, and then there was massive data revisions back over the summer to jobs data. There was one economist who had been exchanging emails with data users about information that wasn't public at the.

Time, and Molly says, these are the kinds of incidents that could make the case for fun even harder.

It seems like, you know, they keep messing up, that these data releases are getting botched. They're like, well, you have some other problems to fix, and like maybe who knows if that enters their mind when they're making appropriations, But I would think it's probably not a great look.

As it exists today. Agencies like the Bureau of Labor Statistics are a minor slice of the federal government's budget.

Pie the BLS budget in this past fiscal year was about seven hundred million dollars. You combine that with what the Bureau of Economic Analysis and the Census Bureau have together that's a little over two billion dollars, and that equates to roughly zero point zero three percent of total federal spending.

The incoming Trump administration hasn't explicitly indicated that these agencies' budgets are on the chopping block. But Elon Musk, who Trump has tasked with running a new Department of Government Efficiency, has promised to cut as much as two trillion dollars from federal spending, and even small cuts to agencies like the BLS could have a big impact.

You're not going to get anywhere near two trillion dollars by looking at these agencies alone. This is certainly not the area where you're going to get massive savings from. But as a lot of people who I taught to pointed out, you will see massive consequences.

One of those people is Zach Brandon.

I am the president of the Greater Madison Chamber of Commerce in Madison, Wisconsin.

Brandon's job is to use data to make good economic investments in his community.

What kind of roads do you need to build, what type of transit infrastructure do you need? How many schools are you going to need?

He makes those decisions by looking at trends in the data.

If you're wrong right, if you undercount, you're going to create a shortage, and shortages generally create cost increases. And if you overbuild then there's also certainly risk to that too.

One person I got in touch with was an urban planner in the Chicago area and she had a client earlier this year who was trying to build a senior living facility in the Cleveland area. So she went to census data to find what the population was of people who were at least seventy five years old with a self care disability in the Cleveland area was and she

Private data collection companies have tried to pick up some of the slack, but there are drawbacks there too.

Private data certainly has its place, but anyone who I spoke to pretty universally acknowledged that the government data is the gold standard. The government is able to do this in a way that doesn't have any other interest but serving a public good. That's truly what these numbers are here to do. Versus when it's then introduced used by a private company, who knows if there's some other kind of incentive.

And for local organizations like the one Zach Brandon runs in Wisconsin, there's another benefit to public data.

There's generally no cost to it, and so when you're a nonprofit and you're trying to pinch every penny, that certainly helps to be able to have resources that are available at no charge.

And as former BLS Commissioner William Beach pointed out, part of the value of this government data isn't any one data point. It's about the trends they reveal.

Give you an example, we ask people, are you working or are you looking for work in the past four weeks? That is a very important question, which we've been asking the same way since the late nineteen forties, and so it has produced a wonderful time series across all that period of time, like seventy five years, in which we

Thanks for listening to The Big Take DC podcast from Bloomberg News. I'm Salaia Moosim. This episode was produced by Julia Press. It was mixed by Alex Sugia, in fact check by Audrey and Atapia. It was edited by Aaron Edward and Wendy Benjaminson. Naomi Shaven is our senior producer. Our senior editor is Elizabeth Ponso. Nicole Beamster Bower is our executive producer. Stage Bauman is Bloomberg's head of Podcasts. Please follow and review The Big Take DC wherever you