At Marketplace, we do the numbers every weekday. Much of the data we report on is released by an independent agency under the Department of Labor: the Bureau of Labor Statistics. The consumer price index, the producer price index, the jobs report — they all come from BLS. 

Marketplace isn’t the only organization that relies on this data — it’s utilized by researchers, policymakers and businesses to make informed decisions. But now some are raising flags about how a proposed rule making it easier to fire federal workers could potentially impact the agency that releases this critical economic data.

What if a jobs report doesn’t paint such a rosy picture of employment, for example? Would BLS economists feel comfortable reporting those figures without fear of being fired? Would they feel pressure to delay a data release or withhold some of it?

Those are some of the concerns raised by Erica Groshen, a former BLS commissioner and current senior economics advisor at Cornell University. She recently spoke with Marketplace’s Nancy Marshall-Genzer. The following is an edited transcript of their conversation. 

Nancy Marshall-Genzer: We interviewed you last October on this and so I was looking back at that story, and you were talking about how when you were Commissioner, did you see any numbers ahead of time?

Erica Groshen: I did. I saw, say, most of the big, principal federal economic indicators I would see a few days in advance, up to possibly a week in advance, and was shown the narrative for any kind of comments I might have on the narrative and was briefed. I could be briefed on anything that I wanted to see in advance, but I didn't see any numbers before they were final. I would only see the final numbers and then could be briefed on what was in them. And I was, of course, subject, as it was everybody at BLS to the laws that would have put me in jail up to five years or fine me $250,000 for breaching the security of that information before the release time.

Marshall-Genzer: And were you able to change those numbers if you didn't like them?

Groshen: No. That was the deal. I had no part in constructing them and no part in changing them. I suppose I could have flagged something and said, "That really doesn't look right to me," and that would have been well within my responsibilities, and then they would have come back and explained why it was the way it was. But it didn't happen any time when I was there.

Marshall-Genzer: I mean, there was a firewall right between you and the economist?

Groshen: Right. I had no access to the underlying data. All I had was the tables that were getting ready for release and then, in the briefings, I would sometimes be shown some associated information, but anything that I was shown would be public upon release. It was not subject to manipulation by me. I had no access to the underlying programs or microdata or that sort of thing.

Marshall-Genzer: So what is the Trump administration proposing, then?

Groshen: To create a new classification of civil servants where those civil servants could be fired by the president for not helping the president to fulfill the president's policy agenda … who is supposed to be affected is anybody who is in a policy-influencing position. And the whole purpose of federal statistics is to inform policy. So you could imagine a new spurt of inflation or a drop in inflation, and the administration wants to play it up or downplay it, or a spurt of employment growth or a drop in the number of jobs or rising unemployment rate, and the administration wants to take full advantage of it for political purposes. And then, I mean, I'm not saying that they would, but this opens the door to that right, in a way that the door was not opened before. 

Marshall-Genzer: So just getting back to the timing of releases, I mean, how could that be affected or changed or manipulated for political purposes?

Groshen: OK. So there's an OMB [Office of Management and Budget] rule that says that principal federal economic indicators have to have a schedule announced a year in advance for all of the releases that are going to happen over the year, and the agencies are required to hold to that schedule unless they can't. And then there’s a process for how to change it if that happens. So how would they change it? 

First of all, it's a new OMB. OMB could change that policy directive, but assuming that that policy directive isn't changed, then the agency could discover some glitch in the way the numbers were calculated, and say, "We have to withhold it until we're sure that the numbers are right." So, they could hold off, you know, something came up right. They could say the numbers weren't high enough quality, and they're skipping this month's numbers or something. You know, I don't know, but that's the kind of thing that could happen. In terms of moving them up, I don't know the language very clearly, but reporting them early is as much or more problematic, actually, than reporting them late, because that gives essentially advance notice to those who notice that they're reported early, who see that they're reported early. And the goal with federal statistics is to give everybody equal access to important information.

Marshall-Genzer: Do you think BLS economists could be pressured to manipulate statistics to make the administration look better?

Groshen: Well, I think this opens the door to that kind of pressure. I believe that BLS economists would fight that in every way they could. 

Marshall-Genzer: Tell me more about undermining trust. I mean, what would that look like?

Groshen: Trust is mission-critical for at least two reasons. If people don't trust the data, then they don't use it for the intended purpose, which is to guide their most important decisions. So you might as well not be producing it. And then who knows what they're going to rely on? They'll make worse decisions, and so our economy won't work as well. They'll be more hesitant to make decisions, and they won't be as good, right. Now also, trust is important for the quality of the data itself, because so much of the data are collected by surveys that are voluntary. So, if you participate in a federal survey, you are performing an important public service. And if you don't believe that the numbers that the survey responses you're giving are going to be held confidentially, if you don't believe that they are going to be used to make trustworthy numbers, then there goes your incentive to participate in the surveys. You get even lower response rates. You might even have people relying on the surveys, and then the numbers are not as good as they were before. So statistical agencies live and die by trust. Data can get worse, and all of the decisions — the monetary policy decisions, the personal decisions, the business decisions — that rely on the data will not be as good.

Marshall-Genzer: So a business might make a decision based on faulty data and then get hurt pretty badly?

Groshen: Yes. Their investment decisions could be not nearly as good. They could move to the wrong part of country, they could wrong side of town. They could decide to invest in the wrong product.

Marshall-Genzer: How else could the economic data be manipulated to make the Trump administration look good? How else could economists be pressured?

Groshen: Well, there's also avoiding bad news, right? If certain sectors of the economy are suffering more than others, you might want to suppress that. Let’s say some favorite group — I don't know, let's say older voters. They care about older voters, and it turns out that the unemployment rate for older voters goes up really high. Then they might decide, “Well, we're just not going to publish that detail anymore,” right? Or some other group whose support they don't want to lose.

Marshall-Genzer: Proponents of the say that the government should be aligned with the policy priorities of the president. I mean, is there something to that?

Groshen: Well, of course. Elections matter. We have elections to set policy for the country. So, let's focus on the statistical agencies, [they] have an obligation to provide trustworthy information to the public and to the administration on the issues that they are dealing with. Put it this way: Any administration that wants to pursue a policy, to do it well needs trustworthy data so that it can pursue its policy to the best effect. And so agencies need to be fearless in the production of that information to actually help produce the best version of the policy that the administration wants, and that is a discipline that has served our country well since the beginning of the statistical system.