For nearly two years, the job market Americans were told about has been significantly stronger than the one that actually existed, and the Federal Reserve made a major policy decision based on numbers that have since been quietly rewritten.
This has absolutely nothing to do with that, and the commissioner also has absolutely no control on any numbers that come out of the BLS, only the language in final reports. There’s not some magic “cook the books” button the commissioner can push, anyway. This is an org with thousands of employees and hundreds of automated programs and processes that can’t just be changed willy nilly, with monitored databases and backups upon failsafes and legions of bureaucrats who would absolutely scream to high heaven if there were even a whisper of a hint of manipulation. You can’t have a conspiracy with thousands of statistics nerds who swear an oath to be impartial. You just can’t.
And anyway, there are ALWAYS jobs revision numbers. In times of expansion, they’re usually revised upwards, and in times of contraction they’re usually revised downwards. This is only (financial) news because it happens, not because it’s a surprise. Everyone could have predicted this, and indeed when you read articles before ethe revisions come out they often mention the upcoming revision. The surprise is usually the magnitude, not usually the direction and certainly not the fact that there is a revision.
This is just caused by the structure of the employment survey, where bigger firms answer the questions faster but smaller firms usually take longer to get their answers in due to not having dedicated staff, ease of gathering the numbers, or long standing procedures - take your pick. Those same smaller firms are also the ones most likely to be affected by swings in the economy, so in times of expansion are more likely to be hiring, and in uncertain or downswing times, are more likely to have lower hiring or lower growth. But the initial numbers assume things are more or less steady until they get official responses in, and then revise to the more accurate numbers. So big nothing burger from a reliability of statistics standpoint, this is just the choice made between having accurate statistics three months late, or kinda accurate statistics a month late that are then revised to be be more accurate.
As for this report, the only news here is that job growth is still sluggish, confirmed by stats - not that there’s some grand conspiracy because people don’t take the time to understand how things actually work. Being complicated or obtuse doesn’t mean it’s corrupt. It means you’re out of your element.
I am. Absolutely.
But since I don’t know anything about it, I have to take what the press tells me at face value for now (that’s the crux of the matter with the media and their power). My impression, based on press reports, is that the data has become less reliable and, in particular, runs counter to Trump’s promises. As I said, I don’t know anything about the subject and, like everyone else, form my opinion based on the information I receive. It’s true that we’re in a political bubble here, and that probably leads to a bias on my part.
Since you seem to know more about this: Please take a look at the figures from recent years and see if there’s a trend in the magnitude of the difference between preliminary and final figures, if you’d like. That would provide a better answer to the question than my opinion alone.
I’d like to disagree with part of the rest of your comment: It’s true that bureaucratic systems are, and should be, somewhat resilient to direct political influence from the current administration. But here’s the thing: At the start of Trump II, Trump fired a lot of federal officials and replaced them with loyalists (again, this is just my opinion based on the news I’ve been following). If you systematically fill key positions in the system with your own people, you can influence the results (this is once again very evident in my country right now, as the current Ministry of Economic Affairs is now publishing “scientifically sound” reports that in some cases directly contradict the previous ones). Statistical analyses can be unambiguous, but as soon as interpretation comes into play, there’s room for maneuver.
Second, leadership can directly influence the methods used. The results are still statistically sound, but, for example, a modified data set or method creates room for maneuver (whether this constitutes manipulation actually depends on the research question and, if applicable, the political stance). If I’m not mistaken, for example, the agency that calculates the inflation rate in the U.S. changed its methodology during Trump’s second term and now uses fewer actual price data and more modeling (please correct me if I’m wrong). The selection of the consumer basket is another factor that can be adjusted.
I therefore cannot agree with your statement (as I understand it) that bureaucratic agencies and institutions are, by their very structure, fundamentally resilient and immune to manipulation.
Secondly, Trump did not fire anyone other than the commissioner, and did not replace anyone at the BLS. They lost a ton due to DOGE and the DRP, but were under a hiring freeze until recently. You can check USA jobs for past listings.
And yes, while leadership could theoretically push certain measurements or methodologies that might skew things in a certain direction, again, at the BLS that functionally can’t be done without a large hue and cry. The commissioner has very little say in program methodologies. Erica Groshen, a former commissioner, is extremely active on social media and often mentions how if there were anything out of the ordinary rank and file would go nuts. You can imagine that’s true given the org has people on both political sides, and gets accused by the other party’s supporters of cooking the books no matter who is in charge, even though methodology changes extremely slowly and is always published online.
To that matter, I am unaware of any changes to inflation measurements during Trump 2, other than I think alternative data sources (e.g. getting datasets from companies/third parties rather than measuring prices at a store), but I’d have to go dig through their methodology publications, which are all wide open and public in the website as well.
The takeaway here is that there is so much sunshine on all of these processes and data it’s functionally impossible for the stats to be a lie. Even earlier reporting this year that imputations (estimating data they can’t physically measure) were way up was still way below the margin of error, and that alone was a huge “scandal”. That was down to the Trump admin starving them of money and staff, and guess what happened a few months later? Suddenly they’re hiring again, because I assume Republicans behind the scenes got pissy about less accurate data, even if it didn’t actually change anything except the measures of accuracy.
This has absolutely nothing to do with that, and the commissioner also has absolutely no control on any numbers that come out of the BLS, only the language in final reports. There’s not some magic “cook the books” button the commissioner can push, anyway. This is an org with thousands of employees and hundreds of automated programs and processes that can’t just be changed willy nilly, with monitored databases and backups upon failsafes and legions of bureaucrats who would absolutely scream to high heaven if there were even a whisper of a hint of manipulation. You can’t have a conspiracy with thousands of statistics nerds who swear an oath to be impartial. You just can’t.
And anyway, there are ALWAYS jobs revision numbers. In times of expansion, they’re usually revised upwards, and in times of contraction they’re usually revised downwards. This is only (financial) news because it happens, not because it’s a surprise. Everyone could have predicted this, and indeed when you read articles before ethe revisions come out they often mention the upcoming revision. The surprise is usually the magnitude, not usually the direction and certainly not the fact that there is a revision.
This is just caused by the structure of the employment survey, where bigger firms answer the questions faster but smaller firms usually take longer to get their answers in due to not having dedicated staff, ease of gathering the numbers, or long standing procedures - take your pick. Those same smaller firms are also the ones most likely to be affected by swings in the economy, so in times of expansion are more likely to be hiring, and in uncertain or downswing times, are more likely to have lower hiring or lower growth. But the initial numbers assume things are more or less steady until they get official responses in, and then revise to the more accurate numbers. So big nothing burger from a reliability of statistics standpoint, this is just the choice made between having accurate statistics three months late, or kinda accurate statistics a month late that are then revised to be be more accurate.
As for this report, the only news here is that job growth is still sluggish, confirmed by stats - not that there’s some grand conspiracy because people don’t take the time to understand how things actually work. Being complicated or obtuse doesn’t mean it’s corrupt. It means you’re out of your element.
I am. Absolutely. But since I don’t know anything about it, I have to take what the press tells me at face value for now (that’s the crux of the matter with the media and their power). My impression, based on press reports, is that the data has become less reliable and, in particular, runs counter to Trump’s promises. As I said, I don’t know anything about the subject and, like everyone else, form my opinion based on the information I receive. It’s true that we’re in a political bubble here, and that probably leads to a bias on my part. Since you seem to know more about this: Please take a look at the figures from recent years and see if there’s a trend in the magnitude of the difference between preliminary and final figures, if you’d like. That would provide a better answer to the question than my opinion alone.
I’d like to disagree with part of the rest of your comment: It’s true that bureaucratic systems are, and should be, somewhat resilient to direct political influence from the current administration. But here’s the thing: At the start of Trump II, Trump fired a lot of federal officials and replaced them with loyalists (again, this is just my opinion based on the news I’ve been following). If you systematically fill key positions in the system with your own people, you can influence the results (this is once again very evident in my country right now, as the current Ministry of Economic Affairs is now publishing “scientifically sound” reports that in some cases directly contradict the previous ones). Statistical analyses can be unambiguous, but as soon as interpretation comes into play, there’s room for maneuver. Second, leadership can directly influence the methods used. The results are still statistically sound, but, for example, a modified data set or method creates room for maneuver (whether this constitutes manipulation actually depends on the research question and, if applicable, the political stance). If I’m not mistaken, for example, the agency that calculates the inflation rate in the U.S. changed its methodology during Trump’s second term and now uses fewer actual price data and more modeling (please correct me if I’m wrong). The selection of the consumer basket is another factor that can be adjusted. I therefore cannot agree with your statement (as I understand it) that bureaucratic agencies and institutions are, by their very structure, fundamentally resilient and immune to manipulation.
Well to your first point, you can go yourself to their website. Everything is public, and they even have a handy page just on revisions: https://www.bls.gov/web/empsit/cesnaicsrev.htm
Secondly, Trump did not fire anyone other than the commissioner, and did not replace anyone at the BLS. They lost a ton due to DOGE and the DRP, but were under a hiring freeze until recently. You can check USA jobs for past listings.
And yes, while leadership could theoretically push certain measurements or methodologies that might skew things in a certain direction, again, at the BLS that functionally can’t be done without a large hue and cry. The commissioner has very little say in program methodologies. Erica Groshen, a former commissioner, is extremely active on social media and often mentions how if there were anything out of the ordinary rank and file would go nuts. You can imagine that’s true given the org has people on both political sides, and gets accused by the other party’s supporters of cooking the books no matter who is in charge, even though methodology changes extremely slowly and is always published online.
To that matter, I am unaware of any changes to inflation measurements during Trump 2, other than I think alternative data sources (e.g. getting datasets from companies/third parties rather than measuring prices at a store), but I’d have to go dig through their methodology publications, which are all wide open and public in the website as well.
The takeaway here is that there is so much sunshine on all of these processes and data it’s functionally impossible for the stats to be a lie. Even earlier reporting this year that imputations (estimating data they can’t physically measure) were way up was still way below the margin of error, and that alone was a huge “scandal”. That was down to the Trump admin starving them of money and staff, and guess what happened a few months later? Suddenly they’re hiring again, because I assume Republicans behind the scenes got pissy about less accurate data, even if it didn’t actually change anything except the measures of accuracy.