Why are there so many US statistical agencies?
Posted by plaguedbyfoibles@reddit | AskAnAmerican | View on Reddit | 80 comments
It feels like a mess tbh. I feel like most countries have a single national statistics agency, whereas the US seems to have various government agencies that provide datasets / stats. You have BEA (Bureau of Economic Analysis), BLS (Bureau of Labor Statistics), BJS (Bureau of Justice Statistics), Bureau of Transportation Statistics (BTS), Economic Research Service (ERS) etc - mostly listed on https://guides.library.oregonstate.edu/c.php?g=285918&p=1906887 Why do you have so many statistics agencies when one would suffice?
CalmRip@reddit
Most countries aren't the size of the US, whether talking about population numbers or the scope and complexity of the economy.
Monte_Cristos_Count@reddit
Because statistics affects business, science, education, and literally every part of life. It's better to have an expert in a given field conduct a statistical analysis rather than some random dude
El_Polio_Loco@reddit
Just remember that the US government deals with more money than 10 or 15 of the worlds largest companies combined.
It's crushingly massive.
When you consider the size of each bureau, it simply makes sense for them to handle their own statistics.
Abdelsauron@reddit
And is somehow less effective than any one of them.
El_Polio_Loco@reddit
I’d argue not many companies can put a fleet of nuclear powered aircraft carriers anywhere in the world while also spending trillions elsewhere.
Abdelsauron@reddit
What government agency is building aircraft carriers?
El_Polio_Loco@reddit
Good question, they must just show up out of nowhere.
Thereelgerg@reddit
No.
His point is that our government doesn't build aircraft carriers. They're produced by the private sector.
El_Polio_Loco@reddit
And all the operations behind that (such as tracking, funding, specifications, etc etc etc) are functionally run through federal bureaucracy.
Abdelsauron@reddit
More like in spite of federal bureaucracy.
The bureaucracy makes everything more expensive than it needs to be. See: literally every weapons program of the past 40 years.
El_Polio_Loco@reddit
Perhaps, things are certainly run differently when the pocketbook seems unlimited.
But it doesn’t change the scope of the whole of the government.
Even if half of all federal employees went away it would still be the largest employer in the US by a decent margin.
And the level of work and pay is a bit different than Walmart stockers.
Abdelsauron@reddit
Walmart stockers work hard and that work provides a tangible benefit to the people in their community every single day.
Many federal employees are just kinda there.
The comparison isn’t really relevant but I couldn’t resist calling out this undeserved elitism.
El_Polio_Loco@reddit
Pointing out the difference between entry level work and the significantly higher number of higher paid and higher educated workers isn’t elitism.
Saying half of federal workers do nothing is certainly a jab.
Abdelsauron@reddit
It is elitism because there’s pretty much no reason to bring it up here other than to punch down.
I can jab at Federal workers all I want. They’re our employees. I am forced to pay their salary. I can vote to help decide on who their supervisor is and what they work on. If they are underperforming I am more than entitled to bring that up.
El_Polio_Loco@reddit
It's not punching down, it's making a realistic observation about their individual impact.
Put a simple dollars to it.
The average government worker (total money moved vs workers) is "worth"
Abdelsauron@reddit
This is a genuinely delusional analysis. Government workers are compensated based entirely on a budget set by the legislature, funded by the extortion of the population. It’s not a fair market value.
Moving money and the value of labor are also two different things. Most of what the government does is move money, but it doesn’t generate value. Nothing the government does generates value, only reallocating the value private citizens already generated.
Shelf stockers generate value. If nobody stocked shelves, goods would not sell. If goods don’t sell, nobody would bother making them.
I’m not arguing that one should be paid more less or the same as the other. Only that making that comparison in the first place shows how little you understand
Thereelgerg@reddit
Right. The quip about them showing up out of nowhere is just dumb.
El_Polio_Loco@reddit
So is questioning who made them.
Its like trying to belittle Amazon because someone else makes their picking robots or shipping trucks.
Thereelgerg@reddit
Yes, y'all both posted some dumb shit.
Thereelgerg@reddit
Do you have any evidence to support that position?
plaguedbyfoibles@reddit (OP)
Recognised national statistical institutes: https://ukdataservice.ac.uk/help/other-data-providers/international-data/national-statistics-institutes/
Thereelgerg@reddit
That doesn't say that "one would suffice."
TheBimpo@reddit
Perhaps there are some I don’t know, statistics, that would support this.
dcgrey@reddit
You're suggesting something sort of arbitrary. What about housing them all together would be better? As they are, they specialize in different data for different uses, with big implications for things like methodology.
Certainly one could find some duplicated efforts,.since agencies tend to expand without regard to each other. But why statisticians studying labor and others studying land usage should be housed together instead of in the department of labor and department of agriculture just sounds like some arbitrary rationalization.
Bastiat_sea@reddit
You struck on the best reason. If you have multiple agencies collecting statistics you will have them spending resources on redundant reports, simply because they both need it, and either don't share it, or aren't even aware of what one another are doing.
If it's consolidated then you can just have the Department of Survey and Census collect everything everyone needs, and every other department can just refer to that.
Also, it's just more efficient to have a department that can employ plenty of statisticians for studies then to have them scattered in small agencies where they won't always be needed.
plaguedbyfoibles@reddit (OP)
I was more thinking along the lines of government agencies sharing their statistics with a central body like the Census Bureau, whilst retaining their own copies, but you also raise a good point about data sharing, if we consider why the Department of Homeland Security (DHS) was originally established, to foster the exchange of information between the likes of the FBI and the CIA, both of whom held partial information on the 9/11 attacks which, if combined, could potentially have allowed them to stop 9/11 from happening.
plaguedbyfoibles@reddit (OP)
Why not have statisticians from these different departments feed into a single agency whose website publishes all of their statistics? That way, along with initiatives like data.gov, it makes data discovery much easier.
lala_lavalamp@reddit
Those all fall under the jurisdiction of larger agencies. Why do you think that one agency focused on statistics would be better than letting each agency handle its own research and statistics?
El_Polio_Loco@reddit
Makes sense to have one large group of statisticians under one umbrella so that they can be redirected as needed more efficiently.
If one bureau has a slow period, or a massive influx of data, then existing, trained, resources can be moved over much more quickly.
Bastiat_sea@reddit
Also 100% if you have multiple agencies collecting statistics you will have them spending resources on redundant reports, simply because they both need it. If it's consolidated then you can just have the Department of Survey and Census collect everything everyone relies one once, and every other department can just refer to that.
cbrooks97@reddit
Why do you think one agency can be equally thorough at analyzing all of these different data from all of those different angles?
plaguedbyfoibles@reddit (OP)
I didn't say that, what I was talking about is different agencies publishing / sharing their datasets / data sources with a de facto statistical agency like the United States Census Bureau.
But I think shoring up support for the GSA with their data.gov initiative would work better.
Thereelgerg@reddit
You didn't say that.
Abdelsauron@reddit
The same way a supermarket can sell meat, fish, diary, fruits, veggies, and a million different packaged foods under the same roof.
You just do it lol.
cbrooks97@reddit
Wow, not even remotely similar. Similar would be a manufacturer producing meat, fish, diary, fruits, veggies, and a million different packaged foods under the same roof. Which doesn't happen.
Abdelsauron@reddit
Actually it does.
Hoosier_Jedi@reddit
Prove to me that one would suffice. Just saying that doesn’t make it true.
plaguedbyfoibles@reddit (OP)
UK's ONS.
Thereelgerg@reddit
The ONS is not the only agency in the UK that compiles or analyses statistics. You don't know what you're talking about.
plaguedbyfoibles@reddit (OP)
Firstly, I am talking about something being the recognised national statistical institute, much like the United States Census Bureau is considered the de facto statistical agency of the federal government.
Secondly, the Office for National Statistics falls under the UK Statistics Authority, and compiles statistics related to the economy, population and society at national, regional and local levels. It also sources datasets from other relevant governmental bodies like Ofcom, the telecommunications regulator for the UK (much like the FCC).
As you can see from https://www.ons.gov.uk/aboutus/whatwedo/statistics/requestingstatistics/secureresearchservice/researchexcellenceandpartnerships/ourpartnerships, ONS also makes it easier to access datasets published by other government agencies / public bodies.
Thereelgerg@reddit
None of that addresses the fact that the ONS is not the "single national statistics agency" in the UK.
TheBimpo@reddit
Ok, show us evidence that this agency is superior to those in the US. Why is consolidation the goal? What is improved by having a single office? What challenges does the ONS face?
Raving_Lunatic69@reddit
You'd do better to compare the US to the EU rather than a single individual country in Europe. That particular entity is much more comparable to the US in scale and scope. Thinking of us as 50 individual countries rather than one single country may help, as we are largely governed that way on many levels.
o93mink@reddit
The UK has 1/5 the annual budget of the US, 1/5 the population of the US, and 2.4% of the land area. It’s just not remotely comparable.
ViewtifulGene@reddit
The BEA and Census Bureau are actually part of the same parent agency- the US Department of Commerce.
The Department of Labor is closer to labor market information than the Dept of Commerce. They actually manage the Unemployment Insurance Program that underpins our most comprehensive job counts and wage records. It wouldn't make sense to force BLS under the Commerce umbrella. And Commerce is large enough as is.
plaguedbyfoibles@reddit (OP)
Maybe shoring up support for data.gov, which has a statutory requirement for federal agencies to publish their datasets to, would be better.
Konigwork@reddit
This feels like either an upper level executive’s point of view “it’s all data, have X analyst pull it!” Or, more likely, a completely elementary/juvenile understanding of what statisticians do.
Statistics aren’t data that tell you anything on their own. They’re numbers, datapoints, and often are incorrect. You want somebody who knows what they’re looking for to translate it to a valuable readable format for decision makers.
A transportation data analyst wouldn’t be able to quickly or effectively tell you what BLS data says. You can cross train them of course! But that takes time and money, and still need to cover their day to day.
plaguedbyfoibles@reddit (OP)
Why not have statisticians from these different departments feed into a single agency, say the USCB, whose website publishes all of their statistics? That way, along with initiatives like data.gov, it makes data discovery much easier.
I'm looking at it from a data discovery perspective.
Konigwork@reddit
You’re asking why individuals agencies aren’t in favor of having a non-insignificant amount of their support staff taken from them (where they manage the day to day of the individual) and reassigned to a completely different agency where they will have to now go through multiple layers of approval to pull a single dataset?
Two things I will point out:
First, there are at least some people working in the government that actually want to get stuff done rather than just spending money for no reason. Adding layers of red tape makes things harder, and increases the public’s disapproval with government work.
Second, what makes you think the government wants data discovery to be easier?
plaguedbyfoibles@reddit (OP)
It's legally required due to an act of law passed by Congress and signed into effect by POTUS in 2019, in reference to your second point, with the General Services Administration (GSA) serving as the empowered agency.
Perhaps I didn't explain myself clearly, for which I apologise. What I am saying is have the statisticians publish their information / data dumps / datasets / data releases to multiple publicly available portals (when signed off for release to the public) rather than just one.
I thought the USCB (United States Census Bureau) made sense as the country's de facto statistical agency.
So perhaps it is more about a culture shift within the civil service branch of government, both on the part of GSA (in terms of having the personnel capable of developing a coherent, comprehensive data governance framework / strategy) and on the part of the other federal agencies, that they have people there willing to work better with the GSA open data evangelists.
Crayshack@reddit
A key part to being good at data analytics is to understand the qualitative aspects of whatever is being studied. That way, they know how to appropriately group the data, what kind of unusual patterns can be explained by simple things if you know just want to cross reference, what kind of data can be ignored as irrelevant, etc. You don't just throw a random statistician at something and tell them to do numbers.
As an example, in college, I did some research into owl migration. Something that emerged in the data was there was a dip in how many owls were caught in our sampling period roughly once a month. Because the people involved with the project understood owls and spent some time out in the field collecting the data, it was known that these dips lined up with the full moon and that the owls could see our nets. And, once that was speculated as a cause, someone paid close attention during a lunar eclipse and noted that the night as a whole saw few owls in the net but there were a bunch during the eclipse. This was a realization that a statistician analyzing the data with no idea of the context, field condition, astrological conditions, etc. would have never connected and there would have just been an unexplained dip in the data.
TheBimpo@reddit
How would a single office suffice? Why would that be empirically better? This is really shallow reasoning. Have you read even a single sentence that explains the origin and purpose of each office?
Honestly, this is the kind of “common sense” reasoning that is being used to dismantle our institutions. One person who has an opinion based on zero research makes a grand statement and that is somehow wise and correct.
Each agency has a unique set of circumstances under which they arose. Consolidation does not mean efficiency or improvement.
plaguedbyfoibles@reddit (OP)
Why not have statisticians from these different departments feed into the USCB, whose website publishes all of their statistics? That way, along with initiatives like data.gov, it makes data discovery much easier.
But you retain all of the agencies as they stand.
TheBimpo@reddit
What problem are we trying to solve?
plaguedbyfoibles@reddit (OP)
It is difficult to access data, or it is difficult to tell which dataset / data release a news publication is referencing in an item / article.
Perhaps this is more a question of, firstly, ensuring that news publications know how to provide link(s) to the relevant data dump(s), and secondly, ensuring that https://data.gov/ remains timely, relevant and fit for purpose.
TheBimpo@reddit
So the problem is you don’t know how to use the existing databases. Are they something you need to use professionally or are you just a curious person who wants this system overhauled to suit your specific requirements?
plaguedbyfoibles@reddit (OP)
No, I'm just a non-US citizen interested in knowing which datasets a news article is referencing and conducting my own data discovery exercises.
I found my real answer: https://data.gov/open-gov/ states the following:
As far as I can see, that law has not been amended or overturned, so I think it is fair to say that it is still in effect. And https://data.gov/about/ states that this initiative was developed and is maintained by the General Services Administration (GSA), who largely self fund via commercial services they sell to different federal agencies, but their profits are remitted / returned to the Treasury Department and sit in funds, such as the Federal Buildings Fund, to potentially be reallocated to them next year via appropriation bills in Congress.
So potentially the current administration can use their legislative majority to divert funding away from their data governance / data democratisation / data cataloguing initiatives.
Chimney-Imp@reddit
Op is plagued by foibles
Thereelgerg@reddit
So your answer to having "so many US statistical agencies" is to . . . create another US statistical agency?
mazrael@reddit
Relevent xkcd
SpiritOfDefeat@reddit
The U.S. is a massive country, both in size and population. Having one agency oversee all statistics for the third largest country on earth by land area, with a population well over 300 million, isn’t really feasible. It would almost certainly need departments within the agency anyway to focus on specific areas such as crime, transportation, agriculture, labor, etc. So if you’d need separate departments anyway, why not just have full blown agencies who specialize in these areas anyway? It’d be far more of a mess for a one size fits all approach to statistics, without specialized staff focusing on their niche.
plaguedbyfoibles@reddit (OP)
Why not have statisticians from these different departments feed into a single agency whose website publishes all of their statistics? That way, along with initiatives like data.gov, it makes data discovery much easier.
CFBCoachGuy@reddit
You’re getting a lot of shit, but to an extent you are right. I work a lot with government data and each agency has these little inconsistencies about how they record data that make it really hard to try to consolidate them. Every agency records a date a different way (and it changes every few years too!). A standardized reporting guideline would be very helpful. It could also help communication between departments. The ACS reformed its data collection a couple years ago, which meant the loss of several useful resources- including many that HHS relied on.
However, centralizing data also allows for easy modification or deletion by elected parties. “Efficiency” and centralization are the same arguments Musk and co are using to justify the deletion of swathes of government data.
plaguedbyfoibles@reddit (OP)
Yes, I understand the trade off of centralisation. I am not necessarily saying that we centralise data, but sync it to a central repository.
Or whichever approach works best. I just want data to be better managed.
It seems that, as per https://data.gov/open-gov/, there is a statutory requirement (stemming from https://www.congress.gov/bill/115th-congress/house-bill/4174) for federal government agencies to publish their publicly designated information (not sure if the legislation mandates particular timescales) using non-proprietary text file formats to the Open Data government portal, along with their metadata catalogued within the portal, and they are working with other federal government agencies to push / drive this agenda.
It seems that they provide a list of methodologies, best practices to use when collating and publishing their datasets / data releases.
Maybe they need to drive more work on this front, although it probably will not be prioritised under the current administration. But they could definitely consider aspects like automated integrations into the portal so that, when a government agency publishes their dataset on their public facing website, it is also published to https://data.gov/open-gov/
Katskit89@reddit
I think Europeans forget this.
anclwar@reddit
Sometimes I'm like "But Canada can do it!" And then I remember that Canada has 300 million fewer people living in it than the USA and I sigh, defeated.
theother1there@reddit
It is just how the US Federal Government was built. The US has a very big aversion to creating new agencies so anytime a new task needs to be done, the task is often given to an existing agency rather than creating a new one. Building a new statistical government agency was seen as too big brother so the task of creating statistics was given to various agencies to do themselves.
In rare situations do the US government consolidate separate efforts it is almost always in the realm of national security. That happened after WW2 when the Department of Defense was created and likewise after 911 with the creation of the Department of Homeland Security.
A classic example is the IRS. Although it is tasked with collecting taxes, the federal government loves outsourcing difficult problems for the IRS to determine. What is a church? Who have sufficient healthcare insurance? What is a legitimate EV? All given to the IRS.
Another example is the US Secret Service. Almost everyone thinks they solely were created to protect important politicians but that was not true. They were created by President Lincoln to deal with the issue of combating counterfeiting currency. It was not until 1901 after the assassination of President McKinney that Congress gave them the additional task of protecting the US President. Why? Because they were the burliest armed dudes around with spare capacity. So even till this day, they have both mandates which if you think about it is ridiculous.
plaguedbyfoibles@reddit (OP)
I think the reason why, ostensibly, DHS was established was to encourage better communication / cooperation between the FBI and the CIA, because both the CIA and FBI had partial information which, if combined, would have allowed them to tackle the 9/11 hijackings, right (without seeking to delve too far into conspiracy theories)?
As for Lincoln, I think he occasionally used Pinkertons for his security detail. And you thought they would have taken the protection of US presidents a little more seriously after the murders of both Lincoln and Garfield.
I would imagine too, that there is a tendency for the purposes of the different government agencies to be codified in law by Congress, like with the establishment of the dual mandate for the Fed as defined in the Federal Reserve Act. Maybe this can also stymy efforts to consolidate departments / agencies, as we have a tendency in the UK to merge government agencies together, but I would imagine this does not happen as much in the US.
I appreciate the stance that, where duties / responsibilities overlap, it is more concerned with division of labour and specialisms, for instance how there is the Federal Trade Commission (FTC) which pursues companies for breaches of antitrust law, but there is also the Department of Justice's (DOJ) antitrust division, which also executes this function, as they specialise in different segments of the economy and seek to complement, rather than undermine, each other's work.
theother1there@reddit
It is just how the US Federal Government was built. The US has a very big aversion to creating new agencies so anytime a new task needs to be done, the task is often given to an existing agency rather than creating a new one. Building a new statistical government agency was seen as too big brother so the task of creating statistics was given to various agencies to do themselves.
In rare situations do the US government consolidate separate efforts it is almost always in the realm of national security. That happened after WW2 when the Department of Defense was created and likewise after 911 with the creation of the Department of Homeland Security.
A classic example is the IRS. Although it is tasked with collecting taxes, the federal government loves outsourcing difficult problems for the IRS to determine. What is a church? Who have sufficient healthcare insurance? What is a legitimate EV? All given to the IRS.
Another example is the US Secret Service. Almost everyone thinks they solely were created to protect important politicians but that was not true. They were created by President Lincoln to deal with the issue of combating counterfeiting currency. It was not until 1901 after the assassination of President McKinney that Congress gave them the additional task of protecting the US President. Why? Because they were the burliest armed dudes around with spare capacity. So even till this day, they have both mandates which if you think about it is ridiculous.
Abdelsauron@reddit
A lot of federal agencies are basically just jobs programs.
Brother_To_Coyotes@reddit
Everyone wants their own toys.
Deolater@reddit
Even in the Federal government, the US has traditionally favored a somewhat decentralized approach to things with intentionally limited and specialized agencies. You really see this with federal law enforcement.
Add to this the fact that the US has been under its current government system for longer than most other countries, and you get a recipe for organically -grown complexity.
Nobody sat down and decided that there should be so many groups gathering statistics, they just funded appropriate departments for whatever statistics came among that they thought were needed.
RastaFazool@reddit
Clearly, one does not suffice.
CupBeEmpty@reddit
We need a bureau of statistics gathering agencies to keep track of them
dwhite21787@reddit
Data.gov
RedBeardedFCKR@reddit
Our government understood the importance of data collection well before we entered the information age. Hoover proved that knowledge was the most absolute power. Look at everything he's rumored to have gotten away with because of his notorious files.
mvw3@reddit
Because we've allowed the bureaucracy to grow unchecked.
daniedviv23@reddit
Because the statistics are generated and used by people with specific expertise in each area, often serving a specific department under which are the agencies (with some exceptions). Stats are, in other words, generated by and for the agencies that need them and then shared with others.
We also are a large country and have a preference for decentralized governance, so consolidation like that doesn’t necessarily make sense for us.
LoriReneeFye@reddit
Because the USA has too many people who went to college instead of getting jobs after high school, so we have to invent jobs for them to do.
Don't worry; everyone is being fired now, and we'll have just one agency that has Trump somewhere in its name.
(And I will NEVER utilize it.)
AutoModerator@reddit
This subreddit is for civil discussion; political threads are not exempt from this. As a reminder:
Do not report comments because they disagree with your point of view.
Do not insult other users. Personal attacks are not permitted.
Do not use hate speech. You will be banned, permanently.
Comments made with the intent to push an agenda, push misinformation, soapbox, sealion, or argue in bad faith are not acceptable. If you can’t discuss a topic in good faith and in a respectful manner, do not comment. Political disagreement does not constitute pushing an agenda.
If you see any comments that violate the rules, please report it and move on!
I am a bot, and this action was performed automatically. Please contact the moderators of this subreddit if you have any questions or concerns.