Census Population Estimates 2020-29

[Biden his time]

NEW YORK CITY POPULATION

Between July 01, 2021 and July 01, 2022

+32,058 Natural Increase
+54,307 Net International Migration
-216,031 Net Domestic Migration
-5,946 Residual

From Census Day 2020 to July 01, 2022

+48,464 Natural Increase
+75,285 Net International Migration
-577,886 Net Domestic Migration
-14,160 Residual

Population 8,335,897
(down 1.46% since July 01, 2021)
(down 5.32% since Census 2020)

[Хahar 🤔]

NEW YORK CITY POPULATION

Between July 01, 2021 and July 01, 2022

+32,058 Natural Increase
+54,307 Net International Migration
-216,031 Net Domestic Migration
-5,946 Residual

From Census Day 2020 to July 01, 2022

+48,464 Natural Increase
+75,285 Net International Migration
-577,886 Net Domestic Migration
-14,160 Residual

Population 8,335,897
(down 1.46% since July 01, 2020)
(down 5.32% since Census 2020)

Garbage in, garbage out. It's a shame that ACS numbers are totally unusable now.

[ProgressiveModerate]

Yeah, there's basically no way Manhattan lost *that* many people, especially since the number of housing units has been increasing.

I think I read somewhere that for those Manhattan nubmers to be true, there'd have to be a nearly 10% reduction in average Household size since 2020.

Why does ACS specifically seem to underestimate NYC so much? Do apartment blocks and a lot of roommate situations make things more complicated?

[💥💥 brandon bro (he/him/his)]

https://www.census.gov/library/visualizations/2023/comm/percent-change-in-county-population.html

Some western states, the southern Atlantic coast area, and the Texas triangle are the three areas still having good growth,  everything else is either close to stagnant or declining.

Surprising to see all NH counties have grown in population last year.

What is going on in the Ozarks???

Are you looking at the two counties in NW AR? Benton is a business hub (Walmart HQ) that has consciously invested in recreational infrastructure like extensive mountain biking trails that run through some urban/semi-urban areas for easy access. Madison had ~16K in the 2020 census so that might be a statistical blip, but also its position situates it well for outdoors recreation like biking and access to the Buffalo National Waterway. Combine that with cheap COL + surge in remote work possibilities and my only surprise is that Washington County (Fayetteville + University of Arkansas) isn't dark green.

[Biden his time]

-SNIP-

Garbage in, garbage out. It's a shame that ACS numbers are totally unusable now.

What basis do you have for this assertion?
And tell me what your preferred alternative is?

-SNIP-

Garbage in, garbage out. It's a shame that ACS numbers are totally unusable now.

This isn't "ACS" anyway. It's the Vintage Population Estimates program.

[Biden his time]

ProgressiveModerate Atlas Icon ⭐⭐⭐⭐⭐ Posts: 11,184

New estimate released: Detroit Population Down? « Reply #4 on: January 28, 1954 at 08:02:30 » Yeah, there's basically no way Detroit lost *that* many people, especially since the number of housing units has been increasing. I think I read somewhere that for those Detroit nubmers to be true, there'd have to be a nearly 10% reduction in average Household size since 1950. Why do demographers specifically seem to underestimate Detroit so much? Do apartment blocks and a lot of roommate situations make things more complicated? Recommend  |  2 other people recommend this.

[Хahar 🤔]

Garbage in, garbage out. It's a shame that ACS numbers are totally unusable now.

What basis do you have for this assertion?
And tell me what your preferred alternative is?

My basis is that ACS numbers were dramatically off in 2020. Based on how much you post about this sort of thing there shouldn't be anything you don't know in this Twitter thread, but I found that it provided a good summary of the problem:

https://twitter.com/zoningwonk/status/1641503129815625729

I do not have a preferred alternative because there does not appear at present to be a good way of estimating population. This does not mean that we should accept bad estimates.

ProgressiveModerate Atlas Icon ⭐⭐⭐⭐⭐ Posts: 11,184

New estimate released: Detroit Population Down? « Reply #4 on: January 28, 1954 at 08:02:30 » Yeah, there's basically no way Detroit lost *that* many people, especially since the number of housing units has been increasing. I think I read somewhere that for those Detroit nubmers to be true, there'd have to be a nearly 10% reduction in average Household size since 1950. Why do demographers specifically seem to underestimate Detroit so much? Do apartment blocks and a lot of roommate situations make things more complicated? Recommend  |  2 other people recommend this.

Yes, for this to be true it requires New York to be experiencing Detroit-level population decline. In addition to being very difficult to believe on its face, this simply does not cohere with empirically observed large rent increases in New York. By contrast, when New York's population did drop dramatically in 2020, rents declined dramatically also.

[ProgressiveModerate]

Another underrated factor that suggests NYC isn’t shrinking is subway ridership stats, which are absolute; several stations are getting close to their pre-2020 ridership levels, even in areas where any sort of population loss would likely be coming from.

The areas that still have the biggest subway hits are ironically the areas that were fastest growing (Williamsburg, Upper East Side). I think this is because a lot of folks in these communities have upper middle class jobs where they can still work remotely much of the time.

I’d be curious if someone could perform simillar analysis with CTA and WMATA numbers.



(Here’s current subway ridership map, I can try to make a pre-pandemic one as well for comparison)

[ottermax]

https://www.census.gov/library/visualizations/2023/comm/percent-change-in-county-population.html

Some western states, the southern Atlantic coast area, and the Texas triangle are the three areas still having good growth,  everything else is either close to stagnant or declining.

Surprising to see all NH counties have grown in population last year.

What is going on in the Ozarks???

Are you looking at the two counties in NW AR? Benton is a business hub (Walmart HQ) that has consciously invested in recreational infrastructure like extensive mountain biking trails that run through some urban/semi-urban areas for easy access. Madison had ~16K in the 2020 census so that might be a statistical blip, but also its position situates it well for outdoors recreation like biking and access to the Buffalo National Waterway. Combine that with cheap COL + surge in remote work possibilities and my only surprise is that Washington County (Fayetteville + University of Arkansas) isn't dark green.

No I'm surprised there is such consistent positive growth across Southern Missouri and Northern Arkansas given that most rural regions continue to see declines.

[Sol]

https://www.census.gov/library/visualizations/2023/comm/percent-change-in-county-population.html

Some western states, the southern Atlantic coast area, and the Texas triangle are the three areas still having good growth,  everything else is either close to stagnant or declining.

Surprising to see all NH counties have grown in population last year.

What is going on in the Ozarks???

Are you looking at the two counties in NW AR? Benton is a business hub (Walmart HQ) that has consciously invested in recreational infrastructure like extensive mountain biking trails that run through some urban/semi-urban areas for easy access. Madison had ~16K in the 2020 census so that might be a statistical blip, but also its position situates it well for outdoors recreation like biking and access to the Buffalo National Waterway. Combine that with cheap COL + surge in remote work possibilities and my only surprise is that Washington County (Fayetteville + University of Arkansas) isn't dark green.

No I'm surprised there is such consistent positive growth across Southern Missouri and Northern Arkansas given that most rural regions continue to see declines.

I assume it's not so different from the robust growth in Middle TN and Piedmont NC--i.e. rural areas getting some of the benefits of nearby, relatively fast growing cities. Plus the Ozarks and Ouachitas are attractive for tourism related reasons, and there is a surprisingly robust meatpacking sector in parts of SW MO and NW AR.

In any case, as Xahar said these estimates are likely bunkum, so not much use in overanalyzing.

[Skill and Chance]

Why has Alabama been holding up decently these past few years, especially compared to MS, LA, and AR?

AL doesn't seem particularly desirable and doesn't have really have spawning industry in many of it's major cities outside of maybe Huntsville, which is a very small portion of the state.

Arkansas (which also has a very high population growth rate relative to the nation) and Alabama both are emerging into the new South the same way Tennessee did

Huntsville and Fayetteville are both engines of growth in the new states but Birmingham and Little Rock are also holding up well.


The only Southern states* left with domestic migration problems are Mississippi and Louisiana.

*Excluding the Mid-Atlantic region of Virginia, Maryland, and DC but they barely count

Glad for Arkansas cause a hypothetical 3 seat AR down the road would suck.

One can expect that if AR keeps 4 seats, the NW AR seat will keep shrinking more and more.

Rs might try to crack it eventually but which would suck. Rn it’s Trump + 23.2 which should be more than fine for this decade, but if the district keeps shredding the outermost rurals, it could narrow pretty rapidly.

Is a commission possible in AR? Ik there’s ballot initiative but it’s weaker than some other states.

A commission should be possible in Arkansas. But are enough people willing to sign a petition to get one to the ballot? The attempt back in 2020-2021 didn't really get enough support at that stage, so it never had a chance.

If Ds are smart, they'd prololy do it now since a redistricting commission wouldn't affect the topline composition of Congress and State Leg very much right now, but down the road it could if the 3rd becomes competative.

I don't disagree.

The legislature just changed the signature gathering requirements to dramatically increase the number of counties signatures must be collected in.  If the new law holds up in court, putting an amendment on the ballot will require a number of signatures equal to at least 5% of 2022 gubernatorial election voters in near unanimous Trump counties.

[Biden his time]

Garbage in, garbage out. It's a shame that ACS numbers are totally unusable now.

What basis do you have for this assertion?
And tell me what your preferred alternative is?

My basis is that ACS numbers were dramatically off in 2020. Based on how much you post about this sort of thing there shouldn't be anything you don't know in this Twitter thread, but I found that it provided a good summary of the problem:

-SNIP-
I do not have a preferred alternative because there does not appear at present to be a good way of estimating population. This does not mean that we should accept bad estimates.

Initially, I read the thread of Tweets and was concerned for a second. The arguments, taking them at face value, seem like good ones (and moreover, I am unfamiliar with this argument, it isn't one Zaybay brings up, as Zaybay's argument largely predicates on the idea that the Domestic Migration numbers are wrong, which the original author of this Twitter thread doesn't even attempt to make, showing how the domestic migration numbers are very good).

I also am sure that you're sending this to me in good faith.

However, I pored through the methodology of the Vintage population estimates myself, and there's a good reason nobody who matters is concerned about the veracity of these estimates. No alarm bells are heard in the pages of the media from expert demographers. There were some murmurs of this last year, but they have slowly faded into the background everywhere discussion on this topic occurs, except this forum.

It only really has become a common reaction on this forum to discount them ever since the favorite places of this forum: New York, San Francisco, Boston, Chicago, Portland etc. started declining hard and strong. Because of this, it is difficult to believe that such criticism of the population estimates is based on anything except the first of the five stages of grief.

I notice you yourself never concerned yourself with such discussion until I shed light on population decline in California, a state I'm sure you love and are disappointed to see declining, so you actively deny its decline. I don't blame you but it is much more important to address problems upfront rather than deny their existence


Anyway, we can debunk the linked Tweets one by one, starting with the thread from March 2022.

It begins with this series of Tweets:

(I also have added other Tweets from elsewhere in the thread that refer to the same exact subject, since this is his main argument)

Did NYC really lose 300,000 people in 2021? I’m skeptical. Based on changes to the 2021 methodology, it seems that some of the problems that led to the dramatic underestimates in the ACS are being replicated. In particular, the 2020 census is not being fully used as the baseline.

The Census says that due to modern disclosure avoidance modernization (new privacy measures) and delays from COVID, they couldn’t fully use the 2020 census as baseline for making their 2021 estimates. Instead, they are “blending” a couple of different data sources:

...

Part of the new “blended base” relies on the 2020 Vintage Estimate, which uses the 2010 census as baseline. Many people suspected the 2010 census already undercounted the City. Also the vintage estimates fell short from the actual 2020 census results by 551,000:

This is the first piece of evidence he brings up in order to justify his suggestion that the Vintage Population Estimates are flawed. First of all, this is (perhaps deliberately) misleading, as it suggests that the pre-census population estimates are used themselves as the base to project 2021 population, leading to a skewed and low population estimate. As someone who has gone through the methodology for myself, I can confirm that this is indeed not the case.

The way the "Blended base" works is that the 2020 US Census population numbers are used as the base. The only place where the 2020 Demographic Analysis and 2020 Vintage Population estimates program numbers are used are when it comes to projecting . The only numbers that we have from the 2020 U.S. Census, after all, are that of the raw population and racial demographic counts given in the PL 94-171 Redistricting Data file. Do you expect them to just make up data?

This is something I myself am aware of and why I haven't quoted a single post-censal racial or age estimate made through this program a single time in any one of my numerous posts on this subject. Try as they can, of course, they can attempt to make proper estimates, but a skew exists as the latest data from the 2020 US Census regarding detailed age, race, and sex numbers simply is not available. Does this matter when it comes to broad population estimates, based on migration and vital statistics? No, it does not to any noticeable scale. Such error that could be caused by this ends up only in the +0.1% range. Moreover, this is an issue that will be fixed by the time the 2023 Population estimates are created, since by then the age, race, sex demographic structures of geographies at the time of the 2020 censuses will be released in the form of Summary Files.

He combines this with a misleading and blatantly incorrect graph showing the "discrepancy", trying to trick the reader into thinking that the 2021 Population estimates program based its population projection of New York City taking into account incorrect population estimates made in the 2010s. This is especially false since the 2021 Population estimates program brought in estimates for April 1, 2020, July 1, 2020, and was transparent about its components of change.

A linear extrapolated line between the 2010 and 2020 censuses is used in between this in order to trick the reader into thinking this is most likely what actually happened, although in reality population growth was much higher in the early 2010s than the late 2010s by every measure. This can be seen in international immigration numbers (legal and illegal, both of them were lower in the late 2010s), domestic immigration numbers (every source says that the NYC metro area did worse in the late 2010s than the early 2010s), and natural increase.


Image Link

Finally, he does not actually offer criticism here, you'll notice this group of tweets is simply based on presenting information and saying that it's self-evident how to criticize it. Likely because he knows that such a criticism is not justifiable to any large degree. Here you can see craftiness and the conniving nature of this thread comes into light. Do not explicitly explain what is wrong with something, but show evidence heavily leaning towards a side in order to trick you into coming to your own incorrect conclusions.

He also reused this debunked criticism in 2023 a core argument, disproving that thread, too:

However, while the baseline was too low

Also a gut check: I think we’d all expect that NYC’s population, if it declined, would decline most sharply in 2020 – when seemingly many rich households fled to the suburbs and college students moved back with their parents – not 2021 when the decline is reported.

False, this seems to show another root misunderstanding of the Population Estimates program, which is thinking that it refers to calendar years. In reality, it goes from July 1 to July 1 of each year. It is unintuitive and incorrect to think that the period of time from April 1, 2020 to July 1, 2020 saw more population decline than the period of time from July 1, 2020 to July 1, 2021. This is especially true in light of the fact that the migration back to Manhattan did not begin in earnest until late 2021 and early 2022. There is no evidence to suggest otherwise. In fact, the population of Manhattan likely reached its low point, or nadir, in early 2021, consistent with the population estimates (that show a growing Manhattan in the period of time from July 1, 2021 to July 1, 2022).

In particular, the Bureau is saying that due to the new privacy measures, they did not have access to data that would let them match address-level files in creating their 2021 PEP estimates.

Remember, NYC DCP’s counting efforts added around 270,000 housing units accounting for 500,000 people to the Census’ Master Address File – these units were not incorporated in older datasets being drawn on for estimates here.

This is not only a blatant lie, it's another example of pretending something is self-evidently a cause for concern where, in reality, it is not. If the methodology regarding housing unit calculations is opened, it is found out that in fact, that this component is entirely based on the 2020 Census results. Opening up the Population estimates, they all include every single one of these 270K new housing units. The only reason this may be a cause for concern is if you're assuming that another 270K housing units are actually missing, but this instead means that the 2020 census was wrong, too, underestimating the population of the city (which is funny as the Post-enumeration survey shows that a statistically significant overcount occurred in the state). Also, this is like Schrodinger's missing houses, there are always gonna be more missing, and you can always argue this.

For example, estimation of the foreign-born population seems problematic – it pulls old ACS estimates – the same estimates that underestimated NYC by ~500,000. NYC would be highly impacted by a bad projection: the foreign-born population is ~38% of the City’s total!

This is paired with a quote from the 2020 Demographic Analysis. Worrying, isn't it?

Except for the fact that the 2020 Demographic Analysis numbers are not taken into account when coming up with total population estimates, as covered in the first section. They are only taken into account for age and sex population numbers.

the biggest issue last year was using the old 2019 ACS estimates to extrapolate international migration patterns. This year, the Bureau has basically acknowledged that post 2016 ACS estimates are so bad, they can’t be used anymore.

They have updated their methodology, basically tossing out post-2016 ACS estimates. While that probably helps, it also means that our 2022 estimates of where international immigrants are moving to are based on data that is now 6 years old: where they moved to between 2010-2016.

This should be a serious, serious concern for any city interpreting their 2022 numbers that have large immigrant populations: NYC, Miami, LA, the entire state of New Jersey...

ACS response rates have cratered everywhere, but especially in states with many immigrants.

The NYT piece is right to focus on immigration, but that should heighten scrutiny of the toplines. Hopefully, asylum seekers are being accounted for through Group Quarters (DHS shelters). But others are likely assigned to counties based on 2010-16 migration patterns.

First of all, this is a good thing. They identified a problem and realized the solution.

And of course, on this, I agree with him actually. It's a major concern (although it's not extremely major, given the extremely tiny proportion of growth caused by international immigration, to the tune of +1% each year in NYC, far, far less than what can cancel out domestic migration problems).

However, it is a major concern in that it inflates New York City's population. This is true since the composition of both legal and illegal immigration in the United States (in all sources of official data) shows that the composition, sources, and destinations of immigration started with being a more Asian component that ends up more often in New York City and the Northeastern U.S.A. in the early 2010s.

In the early 2020s, however, it has shifted, Latin American immigration (with Southern and Western destinations) has made up a larger component of legal immigration tracked in official statistics. Meanwhile, illegal immigration, almost entirely composed of Latin Americans ending up in the Southern and Western United States, has skyrocketed. In fact, Chinese immigration has declined heavily in raw numbers since then and many Chinese have even left America, shrinking New York City further.

This means that since they're using the 2010–16 data, they're inadvertently skewing it in favor of Asian immigration and New York City, boosting and inflating the city's population to make it look larger than it actually is. Hopefully by the time the 2027 estimates or so roll around (by then they should have enough useful ACS data or they have figured out an alternative solution), this problem is fixed.


Anyways, his Tweet thread comprises much equivocation, cherry-picked screenshots, deliberate misunderstanding, and a few lies sprinkled in to provide a conclusion that does not even match the (incorrectly interpreted) "evidence"! If someone actually opens up the methodology of the U.S. Census Bureau's program without pre-conceived notions such as New York City being vibrant, they will find out that the commitment to excellence at the bureau serves as a large boost in confidence to the neutral observer in the bureau to accurately make judgements based on limited available data.

Moreover, although there are some flaws with the methodology (which are unfixable simply due to a lack of data), these flaws are clearly contributing to a systematic overestimation in the population of New York City, as well as other cities and counties which are recipients of significant amounts of Hispanic and Asian immigration. To top it off, the 2020 Census's very own post-enumeration survey found out that New York state was among the top five states in terms of how bad its overcount was in the Census (based on the Post-enumeration survey in New York state), and also only one of eight states which experienced a statistically significant overcount.

Could this man be paid by Eric Adams himself? Nobody knows. What is clear though is that he and the majority forum are united in spreading a science-denying and math-denying narrative about New York City, that it is a city with a growing population instead of one in decline, when all the evidence to the contrary exists. Changing international immigration trends, heavy domestic outmigration, and declining natural increase exemplify themselves in an declining school enrollment, among other effects such as a relatively weaker economic recovery. Of course the economy is weaker than expected, there are less people to help it recover.

However, another thing that is clear is that this narrative is not accepted outside this forum or his tiny Twitter thread. The truth may limp after falsehood, but sometimes the falsehood is so weak, it cannot find support anywhere.

ProgressiveModerate Atlas Icon ⭐⭐⭐⭐⭐ Posts: 11,184

New estimate released: Detroit Population Down? « Reply #4 on: January 28, 1954 at 08:02:30 » Yeah, there's basically no way Detroit lost *that* many people, especially since the number of housing units has been increasing. I think I read somewhere that for those Detroit nubmers to be true, there'd have to be a nearly 10% reduction in average Household size since 1950. Why do demographers specifically seem to underestimate Detroit so much? Do apartment blocks and a lot of roommate situations make things more complicated? Recommend  |  2 other people recommend this.

Yes, for this to be true it requires New York to be experiencing Detroit-level population decline. In addition to being very difficult to believe on its face, this simply does not cohere with empirically observed large rent increases in New York. By contrast, when New York's population did drop dramatically in 2020, rents declined dramatically also.

You are actually correct, New York City is currently on track to do worse than Detroit did in the 1960s, yes (and Detroit had much more housing constructed in the previous decade NYC is currently constructing this decade). For further explanation as to how that could be the case while rents keep "skyrocketing" is based in economics. I think you will find this article an enlightening read as to why this can be the case:

https://www.curbed.com/2023/01/nyc-real-estate-covid-more-apartments-higher-rent.html

[Biden his time]

Now that we know the problem's with international immigration (and it's a super small problem) and not domestic migration, no valid criticism can be given of this map

Image Link

Damn

[Zaybay]

Garbage in, garbage out. It's a shame that ACS numbers are totally unusable now.

What basis do you have for this assertion?
And tell me what your preferred alternative is?

My basis is that ACS numbers were dramatically off in 2020. Based on how much you post about this sort of thing there shouldn't be anything you don't know in this Twitter thread, but I found that it provided a good summary of the problem:

-SNIP-
I do not have a preferred alternative because there does not appear at present to be a good way of estimating population. This does not mean that we should accept bad estimates.

Initially, I read the thread of Tweets and was concerned for a second. The arguments, taking them at face value, seem like good ones (and moreover, I am unfamiliar with this argument, it isn't one Zaybay brings up, as Zaybay's argument largely predicates on the idea that the Domestic Migration numbers are wrong, which the original author of this Twitter thread doesn't even attempt to make, showing how the domestic migration numbers are very good).

Lmao

"This twitter thread that was used as an example decides to bring up and focus on issue A over issue B, therefore issue B doesn't exist".

You do realize how ridiculous that sounds, right?

Edit: This post was up for 2 seconds how did you already recommend it lol

[Biden his time]

Garbage in, garbage out. It's a shame that ACS numbers are totally unusable now.

What basis do you have for this assertion?
And tell me what your preferred alternative is?

My basis is that ACS numbers were dramatically off in 2020. Based on how much you post about this sort of thing there shouldn't be anything you don't know in this Twitter thread, but I found that it provided a good summary of the problem:

-SNIP-
I do not have a preferred alternative because there does not appear at present to be a good way of estimating population. This does not mean that we should accept bad estimates.

Initially, I read the thread of Tweets and was concerned for a second. The arguments, taking them at face value, seem like good ones (and moreover, I am unfamiliar with this argument, it isn't one Zaybay brings up, as Zaybay's argument largely predicates on the idea that the Domestic Migration numbers are wrong, which the original author of this Twitter thread doesn't even attempt to make, showing how the domestic migration numbers are very good).

Lmao

"This twitter thread that was used as an example decides to bring up and focus on issue A over issue B, therefore issue B doesn't exist".

You do realize how ridiculous that sounds, right?

Edit: This post was up for 2 seconds how did you already recommend it lol

OK, I'll admit there may be some problems. But they aren't big ones clearly.

Anyway, his explanation based on international immigration miscalculations makes much more sense than anything related to domestic migration in ascertaining why the 2010-20 estimates were so off.


Do you think more Americans are moving into than out of New York state?

If so, could you provide the evidence for this from anywhere at all?

[Biden his time]

2021 was the first year in recorded history that more people obtained lawful permanent resident status in the states of Florida and Texas than in the state of New York


Image Link


Image Link

End of an era

Source: Department of Homeland Security

[Zaybay]

Garbage in, garbage out. It's a shame that ACS numbers are totally unusable now.

What basis do you have for this assertion?
And tell me what your preferred alternative is?

My basis is that ACS numbers were dramatically off in 2020. Based on how much you post about this sort of thing there shouldn't be anything you don't know in this Twitter thread, but I found that it provided a good summary of the problem:

-SNIP-
I do not have a preferred alternative because there does not appear at present to be a good way of estimating population. This does not mean that we should accept bad estimates.

Initially, I read the thread of Tweets and was concerned for a second. The arguments, taking them at face value, seem like good ones (and moreover, I am unfamiliar with this argument, it isn't one Zaybay brings up, as Zaybay's argument largely predicates on the idea that the Domestic Migration numbers are wrong, which the original author of this Twitter thread doesn't even attempt to make, showing how the domestic migration numbers are very good).

Lmao

"This twitter thread that was used as an example decides to bring up and focus on issue A over issue B, therefore issue B doesn't exist".

You do realize how ridiculous that sounds, right?

Edit: This post was up for 2 seconds how did you already recommend it lol

OK, I'll admit there may be some problems. But they aren't big ones clearly.

Anyway, his explanation based on international immigration miscalculations makes much more sense than anything related to domestic migration in ascertaining why the 2010-20 estimates were so off.


Do you think more Americans are moving into than out of New York state?

If so, could you provide the evidence for this from anywhere at all?

I have no idea. Looking at NYC/NYS can be helpful in that it points out exactly the scale of the issue, but beyond that I'd be just as fine using Chicago, LA, Houston, Boston, or any other city. Im quite perplexed, though, why you seem to be obsessed with how well/poorly NY is doing. In fact, a good amount of that post you made above is dedicated to creating some false world where everyone is "in the first five stages of grief" and "spreading a science-denying and math-denying narrative about New York City", and how everyone is so obsessed with it and superstar cities when in reality you were the first one to talk about NYC, the one to post all about it in the thread "NY fastest-declining in USA — New York state Decline Thread, 2020s", the thread you created, and seem rather eager to constantly bring up anything about it, as you literally just did in the tables above. Maybe a look in the mirror is in order?

Anyway, the issue for the census has always been their inability to correctly model migration patterns, both internal and external. In a way, its not really their fault nor is it a new issue; they've had migration counting woes since at least the 1990 census. It all stems from the data set they use: Social Security, Medicare (65+), and tax returns(0-64). As is probably evident, each data figure is supposed to cover US demographics as best as is able; Medicare tracks those over 65, tax returns track those under 65, and every citizen has a Social Security card.

This system, however, falls apart when you consider international migration, and their subsequent migration within the US. Many immigrants, especially illegal immigrants, do not have a Social Security number, are not filed under Medicare, or do not send in tax forms. This creates the fundamental error, that the migration pattern calculations are much better at determining where the domestic-born US population is moving (particularly the olds since they make up an abnormally large amount of the data set relative to their population count which is also a problem but I digress) than where the foreign-born US population is moving. Combine that with the fact that these migration numbers are extrapolated on a large scale, and you start to see where the issue lies in these figures.

Of course, most people see this issue, which is why there's been such extreme pushback. None of the data we have really lines up at all with what the Census estimates suggest. Hell, all you even have to support your point is public school enrollment, which has 7000 other just as likely explanations (charter/private/homeschooling increasing in popularity after COVID being the most obvious one) and is also not unique to NYC:

https://www.statesman.com/story/news/education/2022/11/13/austin-district-schools-see-enrollment-declines-in-decade-long-trend/69630783007/
https://www.newsobserver.com/news/local/education/article272508826.html

as well as economic recovery being somewhat slower back in 2020, which is inherently a pretty silly point that has way more to do with government policy at the time and what kinda companies work in NYC (turns out a lot of it can be done remotely). Meanwhile, almost every other reliable tracker, such as rents, economic data, and even the NYC subway tracker largely disputes this idea of continued decline.

Oh, and lets talk about that rent figure, cause this is the whole reason I wrote this post. The article you linked...whooooooo boy, its awful. Its so intensely awful of a read. The entire premise of the article is that NYC "cannot be back" because:

1. Everyone I know is working from home and left NYC
2. I don't like NYC. Seriously, did you see all the rats? How could anyone move to NYC with all of these rats everywhere?
3. The Census Estimates say so, and they must be 100% correct

And it then proceeds to use circular logic to justify this as the case. In the end, the article comes to the brilliant conclusion that landlords on a mass scale are artificially taking hundreds of thousands of units off the market in order to artificially inflate prices and no one else has noticed this but this one writer. It was the most batsh**t conspiracy nonsense I've ever seen, reminiscent of the classic "Oh all the properties in this city are bought by foreigners and that's why everything's expensive", bit ironic since you called out how others were "spreading a science-denying and math-denying narrative". Or maybe, demand is just high. And supply is very low. Cause we just had a pandemic where everyone started moving around. Prices literally increased almost everywhere for a reason.

I get liking data, but these are estimates for a reason; they're trying to create close-enough figures to what population counts may look like. Taking these things as gospel is basically equivalent to taking any singular poll of an election as the sole truth. And in this case where there's some pretty clear structural issues that have been present since the 90s, it's probably wise to treat said numbers with the adequate amount of salt.

[Biden his time]

I have no idea. Looking at NYC/NYS can be helpful in that it points out exactly the scale of the issue, but beyond that I'd be just as fine using Chicago, LA, Houston, Boston, or any other city. Im quite perplexed, though, why you seem to be obsessed with how well/poorly NY is doing.

Many reasons, but mainly because it's the largest and most interesting city in the United States and that county-level data maps very neatly onto the city.

I'd actually prefer to do Los Angeles really.

In fact, a good amount of that post you made above is dedicated to creating some false world where everyone is "in the first five stages of grief" and "spreading a science-denying and math-denying narrative about New York City", and how everyone is so obsessed with it and superstar cities when in reality you were the first one to talk about NYC, the one to post all about it in the thread "NY fastest-declining in USA — New York state Decline Thread, 2020s", the thread you created, and seem rather eager to constantly bring up anything about it, as you literally just did in the tables above.

I think it's very funny and I plan to keep on doing this.

Anyway, the issue for the census has always been their inability to correctly model migration patterns, both internal and external. In a way, its not really their fault nor is it a new issue; they've had migration counting woes since at least the 1990 census. It all stems from the data set they use: Social Security, Medicare (65+), and tax returns(0-64). As is probably evident, each data figure is supposed to cover US demographics as best as is able; Medicare tracks those over 65, tax returns tracks those under 65, and every citizen has a Social Security card.

This system, however, falls apart when you consider international migration, and their subsequent migration within the US. Many immigrants, especially illegal immigrants, do not have a Social Security number, are not filed under Medicare, or do not send in tax forms. This creates the fundamental error, that the migration pattern calculations are much better at determining where the domestic-born US population is moving (particularly the olds since they make up an abnormally large amount of the data set relative to their population count which is also a problem but I digress) than where the foreign-born US population is moving. Combine that with the fact that these migration numbers are extrapolated on a large scale, and you start to see where the issue lies in these figures.

Yes, I agree with you here. All this is leading to a systematic overestimation of New York City's population, as is elaborated on above. The U.S. Census Bureau's natural bias is skewing towards New York City. If you see the methodology, you can see why this skew is existing. Using 2010–16 migration patterns when the environment was much more favorable to growth in NYC (both legal and illegal immigration) is skewing the estimates. In fact, there should be more growth in Los Angeles and Miami since these are places with more immigration these days.

What I think likely happened to make the estimates wrong last decade has to do with growth in the early 2010s being underestimated, not growth in the late 2010s being underestimated. That's when it makes the most sense since that's when international immigration was highest, that's when domestic migration out of the city was weakest, and that's when the natural change balance of the city was most strongly positive. This is not just U.S. Census Bureau data speaking, this is just commonly how it is. During Trump's presidency, both legal and illegal immigration fell massively, and that's also (coincidentally) when domestic migration problems started to hit the city hard and natural increase started a steep decline.

Of course, most people see this issue, which is why there's been such extreme pushback. None of the data we have really lines up at all with what the Census estimates suggest. Hell, all you even have to support your point is public school enrollment, which has 7000 other just as likely explanations (charter/private/homeschooling increasing in popularity after COVID being the most obvious one) and is also not unique to NYC:

https://www.statesman.com/story/news/education/2022/11/13/austin-district-schools-see-enrollment-declines-in-decade-long-trend/69630783007/
https://www.newsobserver.com/news/local/education/article272508826.html

Firstly, not really. Both public school and private school enrollment combined have plummeted within the city (and given that they grew so much in the early 2010s when NYC was growing fast, they grew at that time much more than the nationwide average, it shows a stark reversal that suggests a population shrinking in relative size). But anyway, there are tons of other estimates to go off of, which is pretty much coming from everywhere. The US Postal Service data, Linkedin data, immigration data, birth and death data, economic data. Maybe tax revenue?

Also, I dispute the idea that "there's been such extreme pushback". The only place that the pushback can be seen is on the Twitter thread and on this forum. These are the only two places on the internet I've seen it. Everywhere else the pushback was very small and is fading since almost nobody cares about this kind of stuff.

as well as economic recovery being somewhat slower back in 2020, which is inherently a pretty silly point that has way more to do with government policy at the time and what kinda companies work in NYC (turns out a lot of it can be done remotely). Meanwhile, almost every other reliable tracker, such as rents, economic data, and even the NYC subway tracker largely disputes this idea of continued decline.

I don't think you understand what ProgressiveModerate's saying. He's saying that the subway ridership has "almost" reached pre-COVID levels, not that it has reached them yet. So this is indicative of long-term decline.

Oh, and lets talk about that rent figure, cause this is the whole reason I wrote this post. The article you linked...whooooooo boy, its awful. Its so intensely awful of a read. The entire premise of the article is that NYC "cannot be back" because:

1. Everyone I know is working from home and left NYC
2. I don't like NYC. Seriously, did you see all the rats? How could anyone move to NYC with all of these rats everywhere?
3. The Census Estimates say so, and they must be 100% correct

And it then proceeds to use circular logic to justify this as the case. In the end, the article comes to the brilliant conclusion that landlords on a mass scale are artificially taking hundreds of thousands of units off the market in order to artificially inflate prices and no one else has noticed this but this one writer. It was the most batsh**t conspiracy nonsense I've ever seen, reminiscent of the classic "Oh all the properties in this city are bought by foreigners and that's why everything's expensive", bit ironic since you called out how others were "spreading a science-denying and math-denying narrative". Or maybe, demand is just high. And supply is very low. Cause we just had a pandemic where everyone started moving around. Prices literally increased almost everywhere for a reason.

It was just a suggestion and example as to something that may occur, not my main argument. Likely the answers lie elsewhere, although this may be a small part of it.
I also disagree with the idea that it's a bad read, the author's objectively a good writer.

BTW He likes NYC. The guy's a native New Yorker.

I get liking data, but these are estimates for a reason; they're trying to create close-enough figures to what population counts may look like. Taking these things as gospel is basically equivalent to taking any singular poll of an election as the sole truth. And in this case where there's some pretty clear structural issues that have been present since the 90s, it's probably wise to treat said numbers with the adequate amount of salt.

I agree with you here, the numbers aren't perfect.

But you all are straight-up throwing the numbers out since you don't like the results they're showing even when all the mathematical evidence points to that. "What? How could New York City ever decline? I like it too much!" This isn't a good approach to take and you can trust I'll push back against that narrative where it exists.

[oldtimer]

Another underrated factor that suggests NYC isn’t shrinking is subway ridership stats, which are absolute; several stations are getting close to their pre-2020 ridership levels, even in areas where any sort of population loss would likely be coming from.

The areas that still have the biggest subway hits are ironically the areas that were fastest growing (Williamsburg, Upper East Side). I think this is because a lot of folks in these communities have upper middle class jobs where they can still work remotely much of the time.

I’d be curious if someone could perform simillar analysis with CTA and WMATA numbers.



(Here’s current subway ridership map, I can try to make a pre-pandemic one as well for comparison)

The population of New York City is definitely down if I use the subway traffic numbers, in 2019 it was 5.5 million, now it's stable at 3.75 million. That's a 30% decline there.

[ProgressiveModerate]

Another underrated factor that suggests NYC isn’t shrinking is subway ridership stats, which are absolute; several stations are getting close to their pre-2020 ridership levels, even in areas where any sort of population loss would likely be coming from.

The areas that still have the biggest subway hits are ironically the areas that were fastest growing (Williamsburg, Upper East Side). I think this is because a lot of folks in these communities have upper middle class jobs where they can still work remotely much of the time.

I’d be curious if someone could perform simillar analysis with CTA and WMATA numbers.



(Here’s current subway ridership map, I can try to make a pre-pandemic one as well for comparison)

The population of New York City is definitely down if I use the subway traffic numbers, in 2019 it was 5.5 million, now it's stable at 3.75 million. That's a 30% decline there.

It’s now more like 4 million but the point isn’t the overall loss; it’s Where that loss is coming from; the biggest chunk of ridership loss is from the CBD (Midtown and Lower Manhattan), and communities that tend to have a lot of folks who work in those jobs (Williamsburg, Long Island City, Dumbo). To me that indicates fewer people commuting to work, not an actual loss of people.

Also, there are a few examples of subway lines Where ridership is approaching pre-pandemic levels, such as the end of the (7) in Queens.

[oldtimer]

Another underrated factor that suggests NYC isn’t shrinking is subway ridership stats, which are absolute; several stations are getting close to their pre-2020 ridership levels, even in areas where any sort of population loss would likely be coming from.

The areas that still have the biggest subway hits are ironically the areas that were fastest growing (Williamsburg, Upper East Side). I think this is because a lot of folks in these communities have upper middle class jobs where they can still work remotely much of the time.

I’d be curious if someone could perform simillar analysis with CTA and WMATA numbers.



(Here’s current subway ridership map, I can try to make a pre-pandemic one as well for comparison)

The population of New York City is definitely down if I use the subway traffic numbers, in 2019 it was 5.5 million, now it's stable at 3.75 million. That's a 30% decline there.

It’s now more like 4 million but the point isn’t the overall loss; it’s Where that loss is coming from; the biggest chunk of ridership loss is from the CBD (Midtown and Lower Manhattan), and communities that tend to have a lot of folks who work in those jobs (Williamsburg, Long Island City, Dumbo). To me that indicates fewer people commuting to work, not an actual loss of people.

Also, there are a few examples of subway lines Where ridership is approaching pre-pandemic levels, such as the end of the (7) in Queens.

Nevertheless, it never made sense in the modern era to have offices at expensive places where most workers don't live.

If most of your workers live in Jersey or Long Island, it would be cheaper to move it there, it would save commuting time too.

[ProgressiveModerate]

Another underrated factor that suggests NYC isn’t shrinking is subway ridership stats, which are absolute; several stations are getting close to their pre-2020 ridership levels, even in areas where any sort of population loss would likely be coming from.

The areas that still have the biggest subway hits are ironically the areas that were fastest growing (Williamsburg, Upper East Side). I think this is because a lot of folks in these communities have upper middle class jobs where they can still work remotely much of the time.

I’d be curious if someone could perform simillar analysis with CTA and WMATA numbers.



(Here’s current subway ridership map, I can try to make a pre-pandemic one as well for comparison)

The population of New York City is definitely down if I use the subway traffic numbers, in 2019 it was 5.5 million, now it's stable at 3.75 million. That's a 30% decline there.

It’s now more like 4 million but the point isn’t the overall loss; it’s Where that loss is coming from; the biggest chunk of ridership loss is from the CBD (Midtown and Lower Manhattan), and communities that tend to have a lot of folks who work in those jobs (Williamsburg, Long Island City, Dumbo). To me that indicates fewer people commuting to work, not an actual loss of people.

Also, there are a few examples of subway lines Where ridership is approaching pre-pandemic levels, such as the end of the (7) in Queens.

Nevertheless, it never made sense in the modern era to have offices at expensive places where most workers don't live.

If most of your workers live in Jersey or Long Island, it would be cheaper to move it there, it would save commuting time too.

Yep def agree. A lot of the new mega-talls going up these days in Midtown and FiDi are luxury overpriced apartments and condos; i suspect midtown and FiDi will become a lot more Residnetail in my life time.

[ProgressiveModerate]

Idk if this is just me, but these estimates seem to have a pretty universal theme that D communities are shrinking or at least stalling in population relative to previous census, while R areas are doing better. I wonder if there's some sort of reason for this because these estimates aren't politically biased, but if there was truly this theme of folks fleeing "Democratic cities", surely it would've shown up in the 2020 census.

[Skill and Chance]

Idk if this is just me, but these estimates seem to have a pretty universal theme that D communities are shrinking or at least stalling in population relative to previous census, while R areas are doing better. I wonder if there's some sort of reason for this because these estimates aren't politically biased, but if there was truly this theme of folks fleeing "Democratic cities", surely it would've shown up in the 2020 census.

April 2020 was way too early to pick up most of it.  The expectation is that an April 2021 based census would look dramatically different.

[ProgressiveModerate]

Idk if this is just me, but these estimates seem to have a pretty universal theme that D communities are shrinking or at least stalling in population relative to previous census, while R areas are doing better. I wonder if there's some sort of reason for this because these estimates aren't politically biased, but if there was truly this theme of folks fleeing "Democratic cities", surely it would've shown up in the 2020 census.

April 2020 was way too early to pick up most of it.  The expectation is that an April 2021 based census would look dramatically different.

A lot of these same cities overperformed on the 2020 census. The debate is over whether the census or the ACS estimates are wrong (they could both be wrong too).