In that interval of a few hours between the release of the Friday jobs report and President Trump’s decision to fire of the Commissioner of the Bureau of Labor Statistics, I had a few people ask me whether I thought it was possible that the books for May and June had initially been “cooked,” since they ended up being revised dramatically downward. That question seemed a bit quaint after the subsequent firing of Erika McEntarfer. But the answer I gave is relevant in a few ways to the situation going forward.
What I said was that in the Trump era we can’t really rule anything out. (More than cooking, I noted just a few days ago that DOGE-cuts have forced BLS to rely more on estimates relative to data collection in its inflation calculations.) But we should go in with a strong assumption that that is not the case — that there isn’t any cooking — for a number of important reasons.
For me, trust figures very little into this judgment. The first of those reasons is that it would simply be very hard to do. BLS is staffed by career government economists and statisticians, very apolitical people in their work, who are just not the kind of people who are going to go along with anything like that. To the extent they were ordered to do so or Trump found a compliant statistician willing to cook for him, that fact would almost certainly leak out in short order, either through leaks to the press or people resigning.
But there’s a related and deeper reason why this would be difficult to pull off. And it goes to the deeper cultural and methodological conflict that in many ways typifies our age — the conflict between data, expertise, and method on the one hand and power, bullshit and personalist rule on the other. The numbers out of BLS (and other federal government data custodians) aren’t like calls from a referee, which are in-the-moment judgment calls. A bad call from a ref might look fishy. But it’s difficult to say conclusively that it’s dishonest. They’re not even like Federal Reserve decisions which, while dependent on data, still rely on subjective judgements which could go either way. Those are decisions which could be “cooked” or, more relevantly now, be swayed or pressured by the threat of more firings. At least as I understand the BLS and Commerce Department economic data, it’s just not like that. These are very complex models into which large amounts of data are poured and then those are the numbers. There are probably some edge-case roles for professional judgment. But in general, unlike our sports refs and Fed decision-makers, it’s a show-your-work enterprise. The models are the models and the data are the data. In the short term, they are what they are. In the medium term, changes to the models or collection methods are also a show-your-work thing.
Even to the extent some government economist at BLS might want or feel pressured to give Trump better numbers, I’m not sure they’d know how to do it. And here I’m referring to what I think is a basic culture clash. Career government economists or statisticians are a certain kind of person. They’re not deal makers. I wouldn’t imagine they’re big risk-takers or — perhaps “risk takers” isn’t quite the right word — not people with a lot of life experience drawing wildly outside the lines. Systematically goosing the data would involve both having the technical expertise to work the models and data together and then also figure out the best ways to distort them in the most plausible ways. Obviously anything is possible. We’ve seen that in the last six months. The most absurd things end up happening. But there is a basic cultural clash there that I think makes it genuinely difficult, I think, to cook the books in a plausible or even half-credible way. I think it’s too much cognitive dissonance for most of the current people in question.
Needless to Trump could appoint Stephen Miller or — given his ever-expanding portfolio — Marco Rubio to take over and they could just put 400,000 new jobs in a report every month going forward (without any of the supporting data). But we’ll know that’s bogus.
Which brings us to the “what now?” As I was writing this morning, I saw that Trump says he’ll nominate a new person in the next three or four days. That choice will tell us a lot about where we’re going. Whatever actually happens with the numbers, there’s no question to me that trust in government economic data has already taken a huge blow. How can it not? One bad jobs data release and the Senate-confirmed career government economist in charge of the agency was fired. How can that fact not be top of mind for whoever Trump appoints to replace her? That’s assuming the new person is someone who defaults to trying produce accurate data as opposed to some NewsMax host who is committed to only producing “America First” numbers. Any legit commissioner knows their job is safe only as long as the numbers are good for Trump.
That bell can’t be unrung. But the state of BLS’s credibility is different from the mechanics of what’s actually happening inside the agency. It’s not impossible that this will be like Trump’s firing of James Comey in 2017 — an arbitrary and impulsive decision but one in which Comey was replaced by a more or less normal career government prosecutor. Perhaps Trump will nominate a more or less normal person to succeed McEntarfer. Maybe? I’m not holding my breath. That brings us to another basic question: would a normal person take the job under these circumstances? And how normal could they be if they do?
But even that doesn’t address the full facts of the matter. McEntarfer can’t have been the one actually crunching those numbers. That was done by people who were are still in place and presumably will be for next month’s numbers. The deputy commissioner is William J. Wiatrowski. Presumably he is now acting commissioner. He’s been deputy commissioner since 2015 and he’s already been acting commissioner twice before — for two years during Trump’s first term (2017-2019) and again in 2023-24 when McEntarfer was awaiting confirmation. If his LinkedIn bio is accurate, he’s been at BLS for literally 45 years!
It’s difficult to capture the magnitude of the importance of government economic data (and other data in different realms). Employment numbers and all government economic data are, taken together, like the instrument control panel for the national economy — for policy makers, markets, corporate decision-makers. Trump’s action injects uncertainty into every decision-making process throughout the economy. It’s not too much to say that credible and consistent economic data is a big driver of prosperity and growth. It allows more informed risk-taking, better informed decisions. But it’s not only these in-the-moment decisions. The revised and final data become the canonical record of what happened. And not just for government. Histories of this era will be written based on that data, economic research, etc.
The impact of this decision is great for the credibility of government financial data which, for most of the last century, has been deemed basically beyond reproach as a systemic, professional and non-politicized record of economic fact. The impact on credibility is already there. What actually happens and how those subsequent actions impact the economy going forward remains to be seen.