Comment by renewiltord
20 hours ago
ACS, the Decennial Census, and the Austin Demographer has the data we need. Let's take a look https://data.census.gov/table/ACSDT1Y2023.B01003?q=B01003:+T... and https://data.census.gov/table/DECENNIALPL2020.P1?q=P1:+TOTAL... (no ACS in 2020 because of COVID, but we got the Census) and https://demographics-austin.hub.arcgis.com/
You'll see that the ACS 1-year was not performed and released in 2020 due to data quality
2015: 931840
2016: 947897
2017: 950714
2018: 964243
2019: 979263
2020: 961855 (Decennial Census), 965,872 (ACS 5-year, not comparable to 1-year), 996369 (Austin City Demographer)
2021: 964000
2022: 975335
2023: 979700
2025: 1025668
I think there is no single source that shows population decrease post 2020 to a value in 2020 from the same source. But the sources do differ, so likely what has happened is that you have not performed proper data hygiene. You cannot regress across sources without compensating for cross-source differences.
In this case we have the property that 2021 onwards shows increase in ACS and that Austin Demographics shows increase from 2020 onwards but that there is difference between 2020 numbers between sources.
We also see that Austin Demographer has higher number in 2019 than ACS sample https://demographics-austin.hub.arcgis.com/documents/cf1519a...
This means it's probably not good idea to compare 2023 ACS number to 2020 Austin Demographer number. Fortunately, we have Austin Demographer number from 2019, 2020, and 2025. From all this, we can conclude that cratering hypothesis is likely false. We must conclude that following statement of yours is unlikely:
> Austin population cratered and it still has not recovered to 2020 level.
Perhaps you shouldn't use ChatGPT without checking sources?
The purported 2025 number is actually: "Population for City of Austin Full and Limited Purpose Boundaries, 2023".
I indicated that the sources are different. And the conclusion at the very best is that Austin population had been flat for years. It's now growing back, and the prices will react with a lag of 2-3 years.
You claimed cratering but you mixed sources to produce result. Now trying to deflect with claims about ChatGPT. I did this manually. No LLM involved.
It is embarrassing to be this antagonistic when proven wrong. As the kids say these days: “take the L”.
My sources show unambiguously that the population declined up until at least 2023.
There are no consistent sources for the 2019-2024 population due to the way demographics are counted, so they have to be mixed. But they still show _at_ _most_ a weak return to growth with _at_ _most_ modest increases over the peak 2019 population.
This absolutely supports my point that the reason for the property price drop in Austin was the population decrease in 2020, rather than anything to do with new construction.
And you didn't want to spend 10 seconds to look at that, and instead spammed the thread with AI slop which is demonstrably wrong.