Census Population Estimates 2020-29

[Biden his time]

DATA RELEASE


Between July 01, 2020 and July 01, 2021

+148,043 Natural Increase
+244,622 Net International Migration

From Census Day 2020 to July 01, 2021

+187,595 Natural Increase
+256,869 Net International Migration

Population 331,893,745
(up 0.12% since July 01, 2020)
(up 0.13% since Census 2020)

[President of the great nation of 🏳️‍⚧️]

Looking at the state-by-state data, Minnesota and California each lose a seat to Texas and Idaho. Obviously won't mean anything for another nine years, but when I was younger, I had a lot of fun with state-level estimates.

[Biden his time]

BY REGION

Between July 01, 2020 and July 01, 2021

+861,322 (+0.65%) SOUTH
+35,868 (+0.05%) WEST
-93,730 (-0.14%) MIDWEST
-365,795 (-0.64%) NORTHEAST

From Census Day 2020 to July 01, 2021

+959,222 (+0.76%) SOUTH
+78,562 (+0.10%) WEST
-144,010 (-0.21%) MIDWEST
-449,310 (-0.78%) NORTHEAST

Proportion of US population by region

(April 01, 2020 Census -> July 01, 2021 Estimate)

38.10% -> 38.33% SOUTH
23.71% -> 23.69% WEST
20.81% -> 20.77% MIDWEST
17.38% -> 17.33% NORTHEAST

[Biden his time]

ANNUAL GROWTH RATE BY STATE (JULY 01, 2020 to JULY 01, 2021)

2.88%   Idaho
1.72%   Utah
1.66%   Montana
1.37%   Arizona
1.17%   South Carolina
1.16%   Delaware
1.06%   Texas
0.98%   Florida
0.96%   Nevada
0.93%   South Dakota
0.90%   North Carolina
0.81%   New Hampshire
0.80%   Tennessee
0.73%   Maine
0.69%   Georgia
0.62%   Oklahoma
0.48%   Colorado
0.48%   Vermont
0.45%   Arkansas
0.30%   Alabama
0.30%   Indiana
0.27%   Wyoming
0.26%   Washington
0.22%   Missouri
0.15%   Connecticut
0.14%   Iowa
0.12%   Kentucky
0.12%   Virginia
0.12%   National Average
0.11%   Nebraska
0.11%   Oregon
0.06%   Wisconsin
0.03%   Alaska
0.00%   Minnesota
-0.04%   Kansas
-0.06%   Rhode Island
-0.08%   New Mexico
-0.09%   Ohio
-0.12%   Maryland
-0.14%   New Jersey
-0.17%   Michigan
-0.20%   Pennsylvania
-0.23%   Mississippi
-0.38%   West Virginia
-0.52%   North Dakota
-0.53%   Massachusetts
-0.55%   Puerto Rico
-0.58%   Louisiana
-0.66%   California
-0.71%   Hawaii
-0.89%   Illinois
-1.58%   New York
-2.90%   District of Columbia

[Biden his time]

RAW GROWTH BY STATE (JULY 01, 2020 to JULY 01, 2021)

310,288   Texas
211,196   Florida
98,330   Arizona
93,985   North Carolina
73,766   Georgia
59,976   South Carolina
56,291   Utah
55,099   Tennessee
53,151   Idaho
29,920   Nevada
27,761   Colorado
24,608   Oklahoma
20,341   Indiana
19,907   Washington
18,078   Montana
15,074   Alabama
13,706   Missouri
13,659   Arkansas
11,498   Delaware
11,144   New Hampshire
10,230   Virginia
9,967   Maine
8,277   South Dakota
5,436   Kentucky
5,337   Connecticut
4,611   Oregon
4,410   Iowa
3,585   Wisconsin
3,075   Vermont
2,237   Nebraska
1,536   Wyoming
232      Alaska
225      Minnesota
-619      Rhode Island
-1,298   Kansas
-1,689   New Mexico
-4,014   North Dakota
-6,839   West Virginia
-6,905   Mississippi
-7,550   Maryland
-10,358   Hawaii
-10,570   Ohio
-12,613   New Jersey
-16,853   Michigan
-17,954   Puerto Rico
-20,043   District of Columbia
-25,569   Pennsylvania
-27,156   Louisiana
-37,497   Massachusetts
-113,776   Illinois
-261,902   California
-319,020   New York

[Biden his time]

POPULATION GROWTH RATE BETWEEN JULY 01, 2020 AND JULY 01, 2021

Image Link

DARK GREEN: States which grew faster than the world average (+1.03%)
GREEN: States which grew faster than the national average (+0.12%)
ORANGE: States which grew slower than the national average (+0.12%)
RED: States which shrank in population

[Storebought]

Their estimates for the northeast have baked in an algorithm for a permanent population decline that will not be adjusted even in the face of actual population data.

NY did not lose 300,000 people in a year from domestic migration, despite Cuomo. That's just the Census Bureau cleaving off numbers from the 20.2 million 2020 Apr 1 Census count to fit where the algorithm thinks it's supposed to be by now. The Census Bureau estimated a population drop for NY over the 2010/20 decade, and cutting off nearly half of the decade's gain in a year will get them back on track.

For a similar reason, the population growth estimates for TX and particularly AZ from 2020 to 2022 are exaggerated to account for their 2020 Apr 1 Census misses -- dramatically so in the case of AZ. It didn't gain a single House seat! But now AZ is the fourth fastest growing state by percentage and third in absolute count, again. Phew.

[Biden his time]

Image Link

[Storebought]

Not to belabor this, but this is really, really bad methodology. To put this perspective, the last recorded decrease in NY population was during the 1970s:

NY 1970: 18236976
NY 1980: 17558072

The population dropped 678904, which amounts to around -0.4% year-over-end, over the course of a catastrophic decade with manifest consequences. You didn't need to squint at error bars to see the decay.

Now the Census Bureau is predicting an absolute decrease of half of that in a single year, July 2020 - July 2021.

Census population estimates for 2020/2021

Three states had populations above 20 million in 2021: California (39,237,836), Texas (29,527,941) and Florida (21,781,128). New York dropped below 20 million people in the last year, decreasing from 20,154,933 to 19,835,913.

They never estimated it reaching 20 million in the first place:

2010 April 1: 19,378,102 (Census)
2010 April 1: 19,378,117 (Estimates Base)
2010-2020 Estimates:
19,399,956
19,499,921
19,574,362
19,626,488
19,653,431
19,657,321
19,636,391
19,593,849
19,544,098
19,463,131
19,336,776

Why does this matter? Because these tendentious estimates that fly in the face of reality cost the state -- the entire region -- money and clout in Congress. The east coast gets cheated massively in all infrastructure spending.

[Biden his time]

Since the US Census Bureau did estimates for both July 01, 2020 and July 01, 2021 on top of the already existing Census numbers, you can compare the two periods of time and find some pretty interesting trends, such as accelerating growth, decelerating growth, accelerating decline, and decelerating decline all in different places.

I'll do it in two maps to keep the color scheme simple:

STATES WHERE GROWTH IS INCREASING*

*All states that grew faster (or shrank slower) in FY 2021 than they did the previous quarter

Image Link

GREEN: Grew in Q2 2020, Grew faster in FY 2021
Darker Green is faster acceleration

YELLOW: Shrank in Q2 2020, Grew in FY 2021
Darker Yellow is a larger increase in growth rate

PURPLE: Shrank in Q2 2020, Shrank Slower in FY 2021
Darker Purple is a larger increase in growth rate

STATES WHERE GROWTH IS DECREASING*

*All states that grew slower (or shrank faster) in FY 2021 than they did the previous quarter

Image Link

ORANGE: Grew in Q2 2020, Grew Slower in FY 2021
Darker Orange is larger decrease in growth rate

BLUE: Grew in Q2 2020, Shrank in FY 2021
Darker Blue is a larger decrease in growth rate

RED: Shrank in Q2 2020, Shrank Faster in FY 2021
Darker Red is a larger decrease in growth rate

[Biden his time]

Their estimates for the northeast have baked in an algorithm for a permanent population decline that will not be adjusted even in the face of actual population data.

NY did not lose 300,000 people in a year from domestic migration, despite Cuomo. That's just the Census Bureau cleaving off numbers from the 20.2 million 2020 Apr 1 Census count to fit where the algorithm thinks it's supposed to be by now. The Census Bureau estimated a population drop for NY over the 2010/20 decade, and cutting off nearly half of the decade's gain in a year will get them back on track.

For a similar reason, the population growth estimates for TX and particularly AZ from 2020 to 2022 are exaggerated to account for their 2020 Apr 1 Census misses -- dramatically so in the case of AZ. It didn't gain a single House seat! But now AZ is the fourth fastest growing state by percentage and third in absolute count, again. Phew.

Not to belabor this, but this is really, really bad methodology. To put this perspective, the last recorded decrease in NY population was during the 1970s:

NY 1970: 18236976
NY 1980: 17558072

The population dropped 678904, which amounts to around -0.4% year-over-end, over the course of a catastrophic decade with manifest consequences. You didn't need to squint at error bars to see the decay.

Now the Census Bureau is predicting an absolute decrease of half of that in a single year, July 2020 - July 2021.

Census population estimates for 2020/2021

Three states had populations above 20 million in 2021: California (39,237,836), Texas (29,527,941) and Florida (21,781,128). New York dropped below 20 million people in the last year, decreasing from 20,154,933 to 19,835,913.

They never estimated it reaching 20 million in the first place:

2010 April 1: 19,378,102 (Census)
2010 April 1: 19,378,117 (Estimates Base)
2010-2020 Estimates:
19,399,956
19,499,921
19,574,362
19,626,488
19,653,431
19,657,321
19,636,391
19,593,849
19,544,098
19,463,131
19,336,776

Why does this matter? Because these tendentious estimates that fly in the face of reality cost the state -- the entire region -- money and clout in Congress. The east coast gets cheated massively in all infrastructure spending.

I respectfully disagree, maybe it's a real collapse in NY state that happened
Seems very believable to me...

Not only have large companies such as Linkedin and Redfin as well as large databases such as the US Postal Service picked up on the exodus, and not only has the state lagged on every economic indicator since the pandemic hit, the state was already losing hundreds of thousands every year to domestic migration in the 2010s (even after adjusting for the 2020 Census overperformance in the state).

For example, these are the population numbers for New York state over the 2010s adjusted for the 2020 Census Results using the Das Gupta Method, which has historically been used by the US Census Bureau to make its estimates.

April 01, 2010 Census Population - 19,378,102
July 01, 2010 Population Estimate - 19,419,718

July 01, 2011 Population Estimate - 19,602,699
July 01, 2012 Population Estimate - 19,760,733
July 01, 2013 Population Estimate - 19,896,901
July 01, 2014 Population Estimate - 20,008,228
July 01, 2015 Population Estimate - 20,096,573
July 01, 2016 Population Estimate - 20,160,057
July 01, 2017 Population Estimate - 20,201,204
July 01, 2018 Population Estimate - 20,234,876
July 01, 2019 Population Estimate - 20,236,017

April 01, 2020 Census Population - 20,201,249

New York state already began shrinking in 2019, which means that the negative domestic migration numbers in the state had finally begun to outweigh the decreasing natural increase and international migration numbers. In fact, 2016, 2017, 2019, and 2020 all witnessed NY state lose over 100K from domestic migration (and this, once again, is after the adjustment for the true Census numbers).

Then 2020 was a perfect storm to create the most decline possible in the state, as COVID not only ravaged the state with tens of thousands of excess deaths (turning the natural increase in the state close to negative), but it also shutdown the foreign immigration flow into the state (which didn't restart in earnest until a few months ago). Since natural growth and foreign immigration have been New York's only two sources of population growth, taking these away only exacerbates the decline that already was getting started with the exodus.

To top it off, if these numbers are truly just a reversion towards what the algorithm had dictated prior to the census, then you'd expect many differences, such as Vermont and Alabama shrinking. These two states both witnessed large overperformances during the 2020 census, yet with these numbers they're both still growing strongly (something the census never showed before). On top of this, the New York exodus is also showing up as population growth has skyrocketed in Maine, New Hampshire, Connecticut, and Rhode Island compared to previous years (and even the Q2 2020 numbers). All of these states overperformed, Rhode Island and Connecticut especially so.

And California, which saw no such overperformance or underperformance during the census (in fact, the Bureau was dead-on in that state) also has faced a similar population decline. Not to mention Nevada, which you'd expect to have inflated growth rates due to its underperformance in the 2020 Census, is seeing lower growth rates than it did in the 2010s.

Arizona's population growth numbers aren't even remotely astonishing once you consider that the 2021 growth rate in Arizona is actually a decline from the 2020 and 2019 growth rates in the state (even after accounting for the 2020 Census underperformance). For instance, both 2019 and 2020 recorded annual population growth rates of +1.45% in the state (after using the Das Gupta method to adjust for the 2020 Census), while the 2021 numbers only show a +1.39% growth rate.

A similar dynamic is in play in Texas and Florida, both of which saw lower growth rates in 2021 than they did in the years 2019 and 2020 (even after adjusting for the 2020 Census).


2020 and 2021 were pretty objectively horrible years for New York. However there are indications that 2022 will look better, especially as the US Postal Service has shown that domestic migration out of the state has receded back to pre-pandemic totals, and on top of that, international migration will be starting up real soon.

Maybe not enough to make the state grow, though.

[Biden his time]

Obviously, 2021 was a very unique year and it's very unlikely that each of the trends we saw will continue onwards into the decade, much less intensify, but just to show the extent of the changes I thought extrapolating the 2021 estimates forward by ten years and then conducting a reapportionment would be interesting.

This is what the electoral college would look like after such changes:

Image Link

GAIN FOUR: TX
GAIN TWO: FL, AZ
GAIN ONE: GA, NC, TN, UT, ID
LOSE ONE: PA, MN, RI
LOSE TWO: IL
LOSE FOUR: CA, NY


LAST FIFTEEN SEATS ALLOCATED:

421. CA-47
422. FL-30
423. TX-41
424. AL-7
425. NJ-12
426. WI-8
427. MI-13
428. GA-15
429. OR-6
430. MA-9
431. TN-10
432. NC-15
433. CA-48
434. AZ-11
435. TX-42
---
436. DE-2
437. FL-31
438. PA-17
439. NY-23
440. RI-2
441. CA-49
442. NV-5
443. MN-8
444. OK-6
445. SC-8
446. WA-11
447. VA-12
448. TX-43
449. IL-16
450. OH-16


Potential Dark Horses for a seat gain:
DE, NV, OK, SC, WA, VA, MO, IN

Potential Dark Horses for a seat loss:
MA, OR, MI, WI, NJ, AL, LA, OH

[Biden his time]

Many of these Frost Belt states that shrank this year, particularly PA, OH, MI, NJ, CT, RI, and AK are sure locks for growth next year as foreign immigration is speeding up. This is especially true for NJ, which also was hit by a massive COVID wave impeding its natural increase levels.

It also should slow down the declines in NY, MA, and IL significantly next year, while also speeding up the growth in TX, FL, GA, WA, VA, AZ, and MN.

[Biden his time]

Rank Changes

Missouri surpasses Maryland
Utah surpasses Puerto Rico
Montana surpasses Rhode Island

[Biden his time]

Domestic Migration by state

Image Link

Data from July 01, 2020 to July 01, 2021

Green = Positive
Blue = Negative

Shading is based on percentage population growth or loss by domestic migration.


Yes, OK above TX
WA negative
WV positive (and only losing population to natural decline)
MN below WI
MI, OH, PA almost breaking even (only losing population to natural decline)

[DINGO Joe]

Yeah, put me in the camp that the birth/death immigration numbers are fine, but the migration numbers are a mess.

[ultraviolet]

Upsetting. But I won't give up. My career was built on failures and not following the mold. I won't stop. I never have and never will. I still believe in this album. And I still will push. No matter what.

[jimrtex]

Their estimates for the northeast have baked in an algorithm for a permanent population decline that will not be adjusted even in the face of actual population data.

NY did not lose 300,000 people in a year from domestic migration, despite Cuomo. That's just the Census Bureau cleaving off numbers from the 20.2 million 2020 Apr 1 Census count to fit where the algorithm thinks it's supposed to be by now. The Census Bureau estimated a population drop for NY over the 2010/20 decade, and cutting off nearly half of the decade's gain in a year will get them back on track.

For a similar reason, the population growth estimates for TX and particularly AZ from 2020 to 2022 are exaggerated to account for their 2020 Apr 1 Census misses -- dramatically so in the case of AZ. It didn't gain a single House seat! But now AZ is the fourth fastest growing state by percentage and third in absolute count, again. Phew.

I did not detect a single factoid in your message that is true.

The 2020 estimate for New York was 19.369M.

The 2021 estimate for New York is 19.836 an increase of 500K.

The way that the Census estimates works is that they begin with the census, including age, race, and sex.

If there are 123,456 White persons who are age 23 in 2010 in a particular state, then in 2011 there will 123,456 White persons who are age 24 MINUS those who have died PLUS net domestic immigration PLUS net international immigration. Any births will produce persons between 0 and 1 YO.

Births and deaths are pretty reliable - though there is some ambiguity where to attribute some births and deaths. But migration is much harder to estimate. Domestic migration can be tracked to a certain extent using IRS and SS records, but tax returns are delayed, and there may be trouble matching others. There could be a systemic problem in matching. Perhaps people who don't move are more likely to file one year and not the next. In that case, it would appear that matchers were more likely to be movers than the overall population, but that share might be applied to the entire population.

International migration is most difficult to estimate. There are little records of persons who return to their home country. An international immigrant in 2011 might become a domestic outmover in 2015.

In the early part of the last decade, the Census Bureau revised its international migration downward. This was particularly pronounce in New York which went from a projected gain of +1 for the decade, to zero, -1, and eventually -2.

From 2010 to 2020, the Census Bureau estimated for NY 2.339M births, 1.431M deaths, for a natural increase of 784K (note that births are almost 3X natural increase).

Net international migration was 717K, and net domestic migration of -1.563M. Add these for a net migration of -835K.

Add the natural increase to net migration and you get a net change of -51K. Note in particular that all the components were much larger than the net change, and that the net international migration and net domestic migration are made up of two components each: in and out. International immigration must be necessarily greater than 717K.

Domestic migration tends to be very sloshy. The largest state-to-state flows are Texas to California and California to Texas. People move for a job or adventure, and then move back home, months, years, or decades later. Family composition changes. After a divorce one spouse along with a child move back to be closer to family. They may associate the moved to state with the breakup of the marriage. Retirees to Florida may return to New York in their later years where they can live with a child or at least nearby. At 60 or 65 they were quite able and independent. At 85 not so, particularly after the hip fracture or stroke.

In 2019, NY domestic outflow was 440K and inflow 254K, for a net 186K.

Between 2011 and 2018, net outflow increased by 125%, inflow only decreased by 10%, and outflow only increased by 21%.

It is plain goofy to think that domestic outflow was estimated at 300K vs. a real value of 0K. Instead, it appears to have increased from 200K to 300K.

This could be easily explained by outflow increasing from 450K to 500K and inflow from 250K to 200K. How many elderly Floridians would return to NYC to be placed in Cuomo-de Blasio Happy Home, where they would probably be died and buried in a mass burial pit in the Bronx? Cuomo also posted armed guards on the state borders. People were driving in on back roads around Olean with their headlights covered over at 3 AM to avoid detection. How many college freshmen would arrive on campus to online classes, particularly with tuition of many $10s of thousands. They would wait a year, or take classes from home. And how many upper classmen would not come back in the fall of 2020 or spring of 2021.

If you owned a condo in Florida and an apartment in New York, and spent winter in the south, why wouldn't you decide to stay in Florida full time? If you were forced to work from home, why would you do it from Brooklyn or Queens or Long Island. Why not move to Vermont or Maine or Florida? If you were paying enhanced unemployment, why bother actually looking for a job?

The Census Bureau has not had an opportunity to evaluate why their migration estimates were off, so they are likely using the same methodology with

In the 2020 estimates, between 2019-2020 AZ was the 2nd fastest growing by percentage and third by absolute count. Its rate of increase has declined a bit, and it was passed by Utah and Montana on percentage.

After the housing bubble domestic migration to Arizona declined. People in California who planned to cash out their home equity and retire in Arizona where they could supplement their Social Security, pension, and savings with a few $100,000, were upside down on their mortgages. They couldn't afford to move. As the housing market recovered they could escape to Arizona.

[jimrtex]

Obviously, 2021 was a very unique year and it's very unlikely that each of the trends we saw will continue onwards into the decade, much less intensify, but just to show the extent of the changes I thought extrapolating the 2021 estimates forward by ten years and then conducting a reapportionment would be interesting.

This is what the electoral college would look like after such changes:

Image Link

GAIN FOUR: TX
GAIN TWO: FL, AZ
GAIN ONE: GA, NC, TN, UT, ID
LOSE ONE: PA, MN, RI
LOSE TWO: IL
LOSE FOUR: CA, NY


LAST FIFTEEN SEATS ALLOCATED:

Jimrtex: Edited to show percentage difference from change in projected apportionment. Projected gains for decade are quite iffy since they are based on estimates for 1.25 years. The rate of change can vary over a decade, particularly since they are increasingly tied to migration, and the change for the 1.25 years is an estimate not a measurement.

421. CA-47 -8.3% vs. -5.9% loses a 6th seat.
422. FL-30 7.0% vs. 9.4 costs 2nd seat.
423. TX-41 8.5% vs. 11.0% costs 3rd seat.
424. AL-7 +0.6% vs. 2.5% loses a seat.
425. NJ-12 -3.6% vs. -1.9% loses a seat.
426. WI-8 -1.0% vs. +0.3% loses a seat.
427. MI-13-3.4% vs. -2.1% loses a seat.
428. GA-15 5.5% vs. 6.7% costs an extra seat.
429. OR-6 0.8% vs. 1.7% loses a seat.
430. MA-9 -5.9% vs. 5.0% loses a seat.
431. TN-10 +7.1% vs. +7.7% costs extra seat.
432. NC-15 +8.3% vs. +8.9% costs extra seat.
433. CA-48 -6.2% vs. -5.9% loses a fifth seat.
434. AZ-11 14.3% vs. 14.8% costs second seat.
435. TX-42 TX and DE are almost tied for 435, depends on how you project change.
---
436. DE-2 TX and DE are almost tied for 435, depends on how you project change.
437. FL-31 10.3% vs. 9.4% would add a third seat.
438. PA-17 -1.1% vs. 2.4% would save seat.
439. NY-23 -13.1% vs. -13.6% would save a seat.
440. RI-2 -0.6% vs. -2.4% would save a seat.
441. CA-49 -4.2% vs. -5.9% would save a seat.
442. NV-5 12.1% vs. 10.6% gains a seat.
443. MN-8 2.2% vs. 0.1% saves seat.
444. OK-6 7.7% vs. 5.6% gains seat.
445. SC-8 13.8% vs. 11.9% gains a seat.
446. WA-11 6.0% vs. 3.5% gains a seat.
447. VA-12 3.7% vs. 1.0% gains a seat.
448. TX-43 13.4% vs. 11.0% gains a fifth seat.
449. IL-16 -5.5% vs. -8.5% saves one seat.
450. OH-16 +2.4% vs. -1.3% gains a seat.


Potential Dark Horses for a seat gain:
DE, NV, OK, SC, WA, VA, MO, IN

Potential Dark Horses for a seat loss:
MA, OR, MI, WI, NJ, AL, LA, OH

A couple of interesting changes is that Wisconsin gained a tiny bit more than Minnesota. Previously it was estimated that Minnesota would surpass Wisconsin by 2030.

The other is that Alabama is edging away from losing its 7th seat. In 2020, it was suing the Census Bureau over counting methods and was relieved to save the seat.

[Storebought]

Their estimates for the northeast have baked in an algorithm for a permanent population decline that will not be adjusted even in the face of actual population data.

NY did not lose 300,000 people in a year from domestic migration, despite Cuomo. That's just the Census Bureau cleaving off numbers from the 20.2 million 2020 Apr 1 Census count to fit where the algorithm thinks it's supposed to be by now. The Census Bureau estimated a population drop for NY over the 2010/20 decade, and cutting off nearly half of the decade's gain in a year will get them back on track.

For a similar reason, the population growth estimates for TX and particularly AZ from 2020 to 2022 are exaggerated to account for their 2020 Apr 1 Census misses -- dramatically so in the case of AZ. It didn't gain a single House seat! But now AZ is the fourth fastest growing state by percentage and third in absolute count, again. Phew.

I did not detect a single factoid in your message that is true.

The 2020 estimate for New York was 19.369M.

The 2021 estimate for New York is 19.836 an increase of 500K.

The way that the Census estimates works is that they begin with the census, including age, race, and sex.

If there are 123,456 White persons who are age 23 in 2010 in a particular state, then in 2011 there will 123,456 White persons who are age 24 MINUS those who have died PLUS net domestic immigration PLUS net international immigration. Any births will produce persons between 0 and 1 YO.

Births and deaths are pretty reliable - though there is some ambiguity where to attribute some births and deaths. But migration is much harder to estimate. Domestic migration can be tracked to a certain extent using IRS and SS records, but tax returns are delayed, and there may be trouble matching others. There could be a systemic problem in matching. Perhaps people who don't move are more likely to file one year and not the next. In that case, it would appear that matchers were more likely to be movers than the overall population, but that share might be applied to the entire population.

International migration is most difficult to estimate. There are little records of persons who return to their home country. An international immigrant in 2011 might become a domestic outmover in 2015.

In the early part of the last decade, the Census Bureau revised its international migration downward. This was particularly pronounce in New York which went from a projected gain of +1 for the decade, to zero, -1, and eventually -2.

From 2010 to 2020, the Census Bureau estimated for NY 2.339M births, 1.431M deaths, for a natural increase of 784K (note that births are almost 3X natural increase).

Net international migration was 717K, and net domestic migration of -1.563M. Add these for a net migration of -835K.

Add the natural increase to net migration and you get a net change of -51K. Note in particular that all the components were much larger than the net change, and that the net international migration and net domestic migration are made up of two components each: in and out. International immigration must be necessarily greater than 717K.

Domestic migration tends to be very sloshy. The largest state-to-state flows are Texas to California and California to Texas. People move for a job or adventure, and then move back home, months, years, or decades later. Family composition changes. After a divorce one spouse along with a child move back to be closer to family. They may associate the moved to state with the breakup of the marriage. Retirees to Florida may return to New York in their later years where they can live with a child or at least nearby. At 60 or 65 they were quite able and independent. At 85 not so, particularly after the hip fracture or stroke.

In 2019, NY domestic outflow was 440K and inflow 254K, for a net 186K.

Between 2011 and 2018, net outflow increased by 125%, inflow only decreased by 10%, and outflow only increased by 21%.

It is plain goofy to think that domestic outflow was estimated at 300K vs. a real value of 0K. Instead, it appears to have increased from 200K to 300K.

This could be easily explained by outflow increasing from 450K to 500K and inflow from 250K to 200K. How many elderly Floridians would return to NYC to be placed in Cuomo-de Blasio Happy Home, where they would probably be died and buried in a mass burial pit in the Bronx? Cuomo also posted armed guards on the state borders. People were driving in on back roads around Olean with their headlights covered over at 3 AM to avoid detection. How many college freshmen would arrive on campus to online classes, particularly with tuition of many $10s of thousands. They would wait a year, or take classes from home. And how many upper classmen would not come back in the fall of 2020 or spring of 2021.

If you owned a condo in Florida and an apartment in New York, and spent winter in the south, why wouldn't you decide to stay in Florida full time? If you were forced to work from home, why would you do it from Brooklyn or Queens or Long Island. Why not move to Vermont or Maine or Florida? If you were paying enhanced unemployment, why bother actually looking for a job?

The Census Bureau has not had an opportunity to evaluate why their migration estimates were off, so they are likely using the same methodology with

In the 2020 estimates, between 2019-2020 AZ was the 2nd fastest growing by percentage and third by absolute count. Its rate of increase has declined a bit, and it was passed by Utah and Montana on percentage.

After the housing bubble domestic migration to Arizona declined. People in California who planned to cash out their home equity and retire in Arizona where they could supplement their Social Security, pension, and savings with a few $100,000, were upside down on their mortgages. They couldn't afford to move. As the housing market recovered they could escape to Arizona.

Nice to see that you agree with me not only that Cuomo was not an effective governor, but also that the Census Bureau's method for estimating the largest component of population movement is also the most tending to subjectivity. Using mail forwarding data compiled by the USPS is adequate (except when people don't forward their mailing address, but that seems accounted for), but using data from Redfin and Van Allen Lines is dubious (the latter specializes in relocations by managers and executives). But I take exception to the bolded part: I suggested no such thing. I said their methodology at least outside the sunbelt was flawed, in that the CB vintage estimate exaggerates outflows while suppressing inflows.

Their model never predicted NY reaching 20 million in the first place, now in a year it falls hugely from it -- that matter was addressed in an earlier reply. Their estimate was wrong by 600,000+ in NYC -- only around 30 cities in the US have a larger population than that miss. They also predicted lurid declines in Cook County when informed residents at the time knew the population declines of IL, while widespread, were heavily in Downstate.

You yourself were adamant about MN losing a congressional seat this redistricting.

I surmised that a 1.6% drop in a state population in a single year is a Hurricane (Katrina?) level catastrophe that I did not see evident in NY even considering the COVID fatalities in Apr-June. In other words, an artifact. But besides the bolded part which is a strawman that I take exception with, I see that you agree with me in general, so I guess thanks for the reinforcement.

[Starpaul20]

Like I did in previous years, I decided to see what a Wyoming Rule house size would look like with the new population estimates. The house membership would expand to 572 seats, an increase of 137 over its current size:

(States in red gain 1 seat, blue gains 2 seats, green gains 3 seats, yellow gains 4 seats, brown gains 5 seats, orange gains 6 or more, and gray means no change)

Relative to last year, the house would decrease 1 seat, with Florida and Arizona gaining a seat each, and California, New York, and Pennsylvania losing one each.

[Brittain33]

I hate feeling like I shouldn't even pay attention to Census estimates for the next 10 years because of how badly they were off for the real Census.

[jimrtex]

Their estimates for the northeast have baked in an algorithm for a permanent population decline that will not be adjusted even in the face of actual population data.

NY did not lose 300,000 people in a year from domestic migration, despite Cuomo. That's just the Census Bureau cleaving off numbers from the 20.2 million 2020 Apr 1 Census count to fit where the algorithm thinks it's supposed to be by now. The Census Bureau estimated a population drop for NY over the 2010/20 decade, and cutting off nearly half of the decade's gain in a year will get them back on track.

For a similar reason, the population growth estimates for TX and particularly AZ from 2020 to 2022 are exaggerated to account for their 2020 Apr 1 Census misses -- dramatically so in the case of AZ. It didn't gain a single House seat! But now AZ is the fourth fastest growing state by percentage and third in absolute count, again. Phew.

I did not detect a single factoid in your message that is true.

The 2020 estimate for New York was 19.369M.

The 2021 estimate for New York is 19.836 an increase of 500K.

The way that the Census estimates works is that they begin with the census, including age, race, and sex.

If there are 123,456 White persons who are age 23 in 2010 in a particular state, then in 2011 there will 123,456 White persons who are age 24 MINUS those who have died PLUS net domestic immigration PLUS net international immigration. Any births will produce persons between 0 and 1 YO.

Births and deaths are pretty reliable - though there is some ambiguity where to attribute some births and deaths. But migration is much harder to estimate. Domestic migration can be tracked to a certain extent using IRS and SS records, but tax returns are delayed, and there may be trouble matching others. There could be a systemic problem in matching. Perhaps people who don't move are more likely to file one year and not the next. In that case, it would appear that matchers were more likely to be movers than the overall population, but that share might be applied to the entire population.

International migration is most difficult to estimate. There are little records of persons who return to their home country. An international immigrant in 2011 might become a domestic outmover in 2015.

In the early part of the last decade, the Census Bureau revised its international migration downward. This was particularly pronounce in New York which went from a projected gain of +1 for the decade, to zero, -1, and eventually -2.

From 2010 to 2020, the Census Bureau estimated for NY 2.339M births, 1.431M deaths, for a natural increase of 784K (note that births are almost 3X natural increase).

Net international migration was 717K, and net domestic migration of -1.563M. Add these for a net migration of -835K.

Add the natural increase to net migration and you get a net change of -51K. Note in particular that all the components were much larger than the net change, and that the net international migration and net domestic migration are made up of two components each: in and out. International immigration must be necessarily greater than 717K.

Domestic migration tends to be very sloshy. The largest state-to-state flows are Texas to California and California to Texas. People move for a job or adventure, and then move back home, months, years, or decades later. Family composition changes. After a divorce one spouse along with a child move back to be closer to family. They may associate the moved to state with the breakup of the marriage. Retirees to Florida may return to New York in their later years where they can live with a child or at least nearby. At 60 or 65 they were quite able and independent. At 85 not so, particularly after the hip fracture or stroke.

In 2019, NY domestic outflow was 440K and inflow 254K, for a net 186K.

Between 2011 and 2018, net outflow increased by 125%, inflow only decreased by 10%, and outflow only increased by 21%.

It is plain goofy to think that domestic outflow was estimated at 300K vs. a real value of 0K. Instead, it appears to have increased from 200K to 300K.

This could be easily explained by outflow increasing from 450K to 500K and inflow from 250K to 200K. How many elderly Floridians would return to NYC to be placed in Cuomo-de Blasio Happy Home, where they would probably be died and buried in a mass burial pit in the Bronx? Cuomo also posted armed guards on the state borders. People were driving in on back roads around Olean with their headlights covered over at 3 AM to avoid detection. How many college freshmen would arrive on campus to online classes, particularly with tuition of many $10s of thousands. They would wait a year, or take classes from home. And how many upper classmen would not come back in the fall of 2020 or spring of 2021.

If you owned a condo in Florida and an apartment in New York, and spent winter in the south, why wouldn't you decide to stay in Florida full time? If you were forced to work from home, why would you do it from Brooklyn or Queens or Long Island. Why not move to Vermont or Maine or Florida? If you were paying enhanced unemployment, why bother actually looking for a job?

The Census Bureau has not had an opportunity to evaluate why their migration estimates were off, so they are likely using the same methodology with

In the 2020 estimates, between 2019-2020 AZ was the 2nd fastest growing by percentage and third by absolute count. Its rate of increase has declined a bit, and it was passed by Utah and Montana on percentage.

After the housing bubble domestic migration to Arizona declined. People in California who planned to cash out their home equity and retire in Arizona where they could supplement their Social Security, pension, and savings with a few $100,000, were upside down on their mortgages. They couldn't afford to move. As the housing market recovered they could escape to Arizona.

Nice to see that you agree with me not only that Cuomo was not an effective governor, but also that the Census Bureau's method for estimating the largest component of population movement is also the most tending to subjectivity. Using mail forwarding data compiled by the USPS is adequate (except when people don't forward their mailing address, but that seems accounted for), but using data from Redfin and Van Allen Lines is dubious (the latter specializes in relocations by managers and executives). But I take exception to the bolded part: I suggested no such thing. I said their methodology at least outside the sunbelt was flawed, in that the CB vintage estimate exaggerates outflows while suppressing inflows.

Their model never predicted NY reaching 20 million in the first place, now in a year it falls hugely from it -- that matter was addressed in an earlier reply. Their estimate was wrong by 600,000+ in NYC -- only around 30 cities in the US have a larger population than that miss. They also predicted lurid declines in Cook County when informed residents at the time knew the population declines of IL, while widespread, were heavily in Downstate.

You yourself were adamant about MN losing a congressional seat this redistricting.

I surmised that a 1.6% drop in a state population in a single year is a Hurricane (Katrina?) level catastrophe that I did not see evident in NY even considering the COVID fatalities in Apr-June. In other words, an artifact. But besides the bolded part which is a strawman that I take exception with, I see that you agree with me in general, so I guess thanks for the reinforcement.

The Census Bureau does not use data from Redfin or Van Allen Lines for estimating domestic migration. They apparently do look at USPS forwarding data as a check on their domestic migration estimates which are based on administrative records from the SSA and IRS (the Census Bureau receives data from these agencies that has been stripped of identifying information (e.g. "person" who filed 1040 from a NYC zip code in 2019 and from a Florida zip code in 2020). From this they have to extrapolate race, sex, age, dependents.

We do know that the Census Bureau estimates of domestic migration are consistent with ACS estimates of interstate movement (the ACS asks where the respondent lived one year previously).

We do know that net domestic migration from NY was increasing toward the end of the decade, due to an increase in outflow and a simultaneous decrease in inflow.

By 2018, about 2.3% of New Yorkers were moving to other states annually. This was partially balanced by about 1.3% moving from other states.

This produced a net domestic outflow of about 1.0% annually.

It is not unreasonable that this domestic outflow increased in 2020 and the first half of 2021. People want to get out of NYC, and fewer want to get in. Workers in Manhattan are working virtually from home. This kills restaurants in Manhattan. If you live in Queens and your coworker lives in New Jersey, you aren't going to meet for lunch. You are going to rummage through the pantry. Broadway has been shut down. Less work for hotel staff, and aspiring actors are not going to move to the city.

If you are getting enhanced unemployment, you could move elsewhere and work and get a job. How likely is it that the city or state government is going to check up on you? NYC/S are not exactly models of non-corruption.

People who move between New York and Florida seasonally likely consider New York to be home. But if they did not return to New York in 2020 aren't they now Florida residents? Previously they moved twice a year, but the census counted that as zero.

Incidentally, the increase in domestic outflow was mainly to New Jersey and Pennsylvania. This was a reversal of the early 2010 movement into cities.

To/From(net) NJ went from 40/41(-1) in 2011, to 50/35(+15) in 2015, to 59/32(+27) in 2019.

For PA this was 29/26(+3), 33/25(+8), and 47/22(+25)

Another data point. In 2019, NY was the number 11 destination for Puerto Ricans.

We know that the natural increase has been dropping. The initial baby boomers are now 76 and reaching an age where they are dying in larger numbers, and the birth rate has been declining. With any net measurement, it can collapse much faster than births and deaths.


I dpn't see anything wrong with the Census Bureau's model. Each year a person will get a year older or die. If they survive they will either move or stay put.

Where they can go wrong is in estimating the number of persons moving, or being born or dying.

There can also be an error in their initial conditions. They (and you) assume that both the 2010 and 2020 census were accurate. They populated their model based on the 2010 census. But what if they missed 200,000 (or some other number) of persons in 2010, and their change estimates were perfect. Then their estimates would have been 200,000 greater throughout the decade (with small adjustments based on age, race distribution, etc.). Their estimates would show a small decline over the decade (the estimated cumulative change only went negative in 2020).

New York may be particularly hard to count since there are many immigrants, some whose presence is illegal. If your cousin entered on a 6-month tourist visa three years ago, are you going to report him to the (census) authorities, particularly if he has been working off the books, and his long-term presence in your apartment violates your rental agreement?

Similarly, do we know that the 2020 Census was accurate? There was a particular effort to get New Yorkers to respond. Alabama was expected to lose a seat, and had even sued over it. But you may recall that Alabama had an exceptional early response, particularly compared to other Southern states. So get out the response may actually work.

The 2020 Census Bureau relied more on administrative records. If someone promptly replied, their administrative records might have matched, but they were never checked - or if they were it was during a test to see if they conformed. If someone did not respond to numerous letters over several months, did not answer the door, why do we assume that anyone lives there, or if they do that administrative records are still reliable?

Birth and death records are probably the most accurate. Domestic migration matches the ACS.

The wild card is international migration. The Census Bureau dramatically decreased the estimates for New York and New Jersey based on much lower net international migration. There was less change for states where international migration was much more dominated by Latin America. The Census Bureau explanation was that they had data that supported more international emigration, that more persons were returning to their home country. But what if that adjustment was wrong?

There may also be an issue of confirmation bias. If you wanted to believe that the number of unauthorized aliens was steady, if you could "show" significant emigration it would support your belief.

1.5% decline is not Katrina-like, and really is not such a huge number. We have ACS data showing 2.3% domestic out migration in 2018. It is certainly plausible that this continued to increase, while domestic in-migration dropped. International immigration may have also dropped, and natural change decreased nationwide, not just New York.

The Census Bureau switched their estimate base to the 2020 Census and estimated increased loss for New York, continuing a 10-year trend where the second derivative of population was negative.

Will the 1.5% decline over the next decade? Who knows. The Census Bureau is not projecting it to. We (on this board) have traditionally projected estimates from the Census forward to the next census to guess at future apportionment change.

[Aurelius]

When will county level estimates be released?

[jimrtex]

When will county level estimates be released?

March 2022 (*subject to change)