The Census: Why the Government Counts Every Person and What Happens With the Numbers

Every 10 years, the federal government tries to count every person living in the United States. Not every citizen. Not every voter. Every person. That includes noncitizens, undocumented immigrants, children, people in prison, people experiencing homelessness, and anyone else physically present in the country.

This is not optional. Article I, Section 2 of the Constitution requires it. The founders wrote it into the original document in 1787 because they needed a way to divide political power among the states. That purpose has not changed, but the consequences of the count have grown far beyond what the founders envisioned.

Why the Constitution Requires It

The Constitutional Convention faced a fundamental design problem: how to divide representation in the new Congress among states of very different sizes. Virginia had roughly 10 times the population of Delaware. If every state got equal representation, small states would have outsized power. If representation was purely proportional, large states would dominate.

The compromise: the Senate gives every state two seats regardless of size. The House divides seats based on population. But to divide seats based on population, you need to know the population. The Constitution solved this by requiring an "actual Enumeration" within three years of the first Congress and every 10 years after that.

The first census took place in 1790. U.S. marshals went door to door and counted 3.9 million people. The 2020 census counted 331.4 million. Every decade in between, without exception, the government has conducted a count. It is one of the longest-running continuous operations in American government.

What the Census Actually Determines

The census results trigger three major consequences, each of which shapes how the country is governed for the next decade.

1. Congressional Apportionment

The 435 seats in the House of Representatives are divided among the 50 states based on census population. Every state gets at least one seat. The remaining 385 are distributed using a formula called the method of equal proportions, which tries to make each House member represent roughly the same number of people.

After the 2020 census, Texas gained two seats. Colorado, Florida, Montana, North Carolina, and Oregon each gained one. California, Illinois, Michigan, New York, Ohio, Pennsylvania, and West Virginia each lost one.

These shifts are not trivial. Gaining or losing a congressional seat changes a state's influence in the House and in the Electoral College (since electoral votes equal House seats plus Senate seats). New York lost a seat after the 2020 census by a margin of 89 people. If 89 more people had been counted in New York instead of somewhere else, the state would have kept its seat.

2. Federal Funding Distribution

Census data determines how more than $2.8 trillion in annual federal spending is distributed. That is not a one-time amount. It is $2.8 trillion every year for a decade, based on a single count.

Programs that allocate money based on census-derived population data include:

• Medicaid (the largest single program tied to census data)
• SNAP (food assistance)
• Highway planning and construction
• Section 8 housing vouchers
• Title I education funding for schools in low-income areas
• Head Start early childhood programs
• Medicare Part B physician payments

The George Washington Institute of Public Policy identified over 300 federal programs that use census-derived data to distribute funds to states and localities. An undercount in your community means less money for your schools, roads, hospitals, and social services for the next 10 years.

3. Redistricting

After the census, every state redraws its congressional and state legislative district maps to reflect the new population numbers. Each district must contain roughly equal population (the "one person, one vote" principle from Reynolds v. Sims, 1964).

This is where the census connects directly to gerrymandering. Whoever controls the redistricting process uses the census data to draw maps. The data itself is neutral. How it gets used is not.

The Census Bureau releases detailed population data at the block level through its redistricting data program, giving map-drawers exact population counts for every neighborhood in the country.

Who Gets Counted (and Who Does Not)

The Constitution says to count "persons," not citizens. This was a deliberate choice in 1787. The founders who wrote the Enumeration Clause understood the difference between citizens and inhabitants. They chose the broader term.

This means the census counts:

• U.S. citizens
• Lawful permanent residents (green card holders)
• People on temporary visas (students, workers)
• Undocumented immigrants
• Refugees and asylum seekers
• Children of any status
• Incarcerated people (counted where they are imprisoned, not where they lived before)

The question of who gets counted has been politically contentious. In 2018, the Trump administration attempted to add a citizenship question to the 2020 census. The stated reason was to better enforce the Voting Rights Act. Critics argued the real purpose was to suppress responses from immigrant communities, leading to an undercount in areas with large noncitizen populations (which tend to vote Democratic).

The Supreme Court blocked the question in Department of Commerce v. New York (2019), ruling that the administration's stated rationale was "contrived." The citizenship question did not appear on the 2020 census.

The Undercount Problem

No census has ever been perfectly accurate. Some people are always missed, and the undercount is not random. Historically undercounted groups include:

• Young children (especially under age 5)
• Racial and ethnic minorities (particularly Black, Hispanic, and Native American populations)
• Renters and people in multi-unit housing
• People experiencing homelessness
• People in rural and hard-to-reach areas
• Immigrants (regardless of legal status)

The 2020 census overcounted the white non-Hispanic population by 1.64% and overcounted homeowners by 0.43%. It undercounted the Hispanic population by 4.99%, the Black population by 3.30%, and the American Indian/Alaska Native population living on reservations by 5.64%.

These are not abstract statistical errors. They translate directly into less congressional representation and less federal funding for the communities that were undercounted, compounding existing disadvantages.

How the Census Is Conducted

The Census Bureau, part of the Department of Commerce, runs the census. It is a massive logistical operation that takes years to plan.

Before Census Day: The Bureau updates its address list (the Master Address File), which contains every residential address in the country. It works with state, local, and tribal governments to verify addresses, especially in areas with non-standard addressing (rural routes, apartment complexes, tribal lands).

Census Day is April 1 of the census year (the next one is April 1, 2030). Households receive questionnaires by mail, and for the first time in 2020, could respond online or by phone. The questionnaire asks basic demographic questions: how many people live at the address, their names, ages, races, ethnicities, and relationships to each other.

Non-Response Follow-Up (NRFU): This is the most expensive and labor-intensive phase. Census workers visit every household that did not respond to collect information in person. In 2020, the Bureau hired roughly 500,000 temporary workers for the effort.

Data processing and release: After collection, the Bureau processes, verifies, and publishes the results. The first numbers (total state populations for apportionment) are delivered to the president by December 31 of the census year. The detailed redistricting data follows several months later.

The Census and the Electoral College

Because electoral votes are based on congressional representation (House seats + 2 Senate seats per state), the census indirectly determines the Electoral College map. When a state gains a House seat, it gains an electoral vote. When it loses a seat, it loses one.

After the 2020 census, the electoral map shifted. States in the Sun Belt (Texas, Florida, North Carolina) gained electoral votes. States in the Rust Belt and Northeast (New York, Pennsylvania, Ohio, Illinois) lost them. These shifts accumulate over decades. In 1960, New York had 45 electoral votes. Today it has 28.

Presidential campaigns build their strategies around the electoral map, which is built on census data. The difference between 270 and 269 electoral votes can come down to a single congressional seat, which can come down to a few thousand people in a census count.

The American Community Survey

The census itself asks very few questions, just the basics needed for apportionment and redistricting. For detailed socioeconomic data, the Census Bureau runs the American Community Survey (ACS), a continuous survey that samples about 3.5 million households per year.

The ACS collects data on income, education, employment, health insurance, housing costs, commuting patterns, language spoken at home, disability status, and dozens of other topics. This data is what most federal programs, researchers, businesses, and local governments actually use for planning and resource allocation.

Before 2010, this detailed data was collected as part of the census itself through a "long form" questionnaire sent to a subset of households. The ACS replaced the long form, spreading the data collection across the full decade instead of concentrating it in one year.

Why It Matters for the Next Decade

The 2030 census will determine:

• How 435 House seats are distributed among the states
• How electoral votes are allocated for the 2032 and 2036 presidential elections
• How trillions of dollars in federal funding flow to communities from 2031 through 2040
• The data used to draw congressional and state legislative maps that will be in effect through 2040

Population trends suggest the Sun Belt will continue gaining at the expense of the Midwest and Northeast. But the exact numbers, and the exact line between gaining a seat and losing one, will depend on the accuracy of the count.

The census is the foundation that the rest of the system is built on. Congressional representation, federal funding, redistricting, and the Electoral College all trace back to a single number: how many people were counted, and where. Getting that number right matters more than almost any single policy debate, because it shapes the rules of every policy debate that follows for the next 10 years.