How many people left New York State between 2018 and 2023 according to US Census estimates?

Executive summary — direct answer, no equivocation.

According to the U.S. Census-derived state totals cited in the materials, New York State’s resident count fell from 20,201,249 in 2018 to 19,571,216 in 2023, a net decline of 630,033 people. This figure reflects the Census Bureau’s published population totals (which combine births, deaths, domestic migration and international migration), not a pure tally of people who moved out of the state. Different data sources and methods produce divergent tallies — tax-based address-change counts and ACS state‑to‑state flow estimates measure related but distinct things and therefore report larger or smaller numbers depending on scope and methodology ^{[1] [2] [3]}.

1. The headline number reporters cite — a six‑hundred‑thousand decline — and what it really means.

The population totals quoted above come from Census state population tables that list New York at 20,201,249 in 2018 and 19,571,216 in 2023, yielding a net loss of 630,033 residents over that interval. That net change is a consolidated accounting of three components: natural increase (births minus deaths), net domestic migration (people moving in from other U.S. states minus those leaving), and net international migration (immigrants arriving minus emigrants leaving). The 630,033 decline is therefore a net population change, not a simple count of how many residents moved out of New York and does not isolate domestic out‑migration alone ^{[1] [4]}.

2. Why “people who left” is a different question than “population decline,” and where ACS estimates come in.

The American Community Survey (ACS) and Census state flow products estimate state‑to‑state migration and are sample‑based measures with margins of error; they track where people lived one year earlier and where they live now. ACS-based flow tables can produce annual counts of residents who left New York for other states, but those counts exclude deaths and births and are subject to sampling variability. The Census Bureau documented a 2022 processing error that affected migration flow estimates and said it was corrected for 2023, which means annual ACS flow estimates need careful interpretation and confidence‑interval inspection before aggregation ^[2].

3. Tax return address changes and the “almost a million” taxpayer tally — what that measures and its limits.

State tax and IRS migration tallies count tax returns that changed address across states and therefore capture filers, not individuals, and can double‑count households using joint returns or miss nonfilers and dependents. One analysis summed annual New York-to-out‑of‑state taxpayer address changes from 2018–2023 and arrived at roughly 934,000 returns leaving New York in that span; this measure exceeds the Census net population decline because it captures gross moves, internal household composition changes, and multiple-year mobility, and it is not equivalent to a count of unique people who departed ^[3]. The tax‑based approach is useful for revealing mover behavior among taxpayers but overstates “people who left” if interpreted as unique individuals.

4. Reconciling divergent counts: gross outflows, net migration, and population components.

Gross outflow counts (total moves out) will always be larger than net change numbers because many moves are offset by in‑migration, births or international arrivals. The Census population totals (the 630,033 decline) reflect net outcomes across all components, while ACS flow estimates and tax return tallies attempt to measure flows — two related but distinct metrics. Analysts must choose the right metric for their question: if you want “how many unique people moved out of New York between 2018 and 2023,” neither the net decline nor simple tax‑return sums give a clean answer without record linkage; if you want “how much did the state population change,” the Census totals are the appropriate source ^{[1] [2] [3]}.

5. What to trust and what caveats to keep front of mind when reporting these figures.

For a concise, defensible statement about population change, use the Census state totals and report the 630,033 net decline while explicitly noting it is net population change (not gross outflow). For migration behavior, cite ACS state‑to‑state flow tables and IRS/tax data with clear caveats about sampling error, processing corrections in 2022, the difference between returns and people, and the role of births/deaths/international migration in the net figure. Different stakeholders may emphasize the data that best fits their narrative (fiscal agencies favor tax returns; demographers rely on Census/ACS), so always state the metric, method, and uncertainty when communicating these numbers ^{[2] [3] [4]}.