When the Census says most of your town is in prison

Prisons account for over half of the population in nearly 40 municipalities, painting a distorted picture of communities.

by Aleks Kajstura, November 10, 2020

The Ina Community Center shares a little unassuming building with the Ina Village Hall. The squat single-story structure probably serves the little Illinois village quite well. But judging by their Census population of over 2,300, you wouldn’t think the building adequate.

The village only has 494 actual residents. The 1,800 people who make up the remainder of the Census population are folks who are incarcerated at the Big Muddy Prison. Because the prison is located within the village boundaries, the Census counts them as if they were village residents. That means 80% of the population reported for Ina are legal residents of other villages, towns, and cities, scattered across Illinois.

Ina is just one of nearly 40 municipalities across the US where more than half of the Census population is actually people who were counted at a correctional facility. And in over 100 municipalities, at least one in three people counted by the Census are actually out-of-town residents.

I often talk about how the Census Bureau’s practice of counting incarcerated people as if they were residents of the town that contains the prison causes prison gerrymandering. Prison gerrymandering is the process of using prisons to pad district populations, resulting in increased political representation for people who live near prisons.

But the Census Bureau’s practice skews data in ways that affect more than just redistricting data and representation. When a prison accounts for half of your city population, the Census’ demographic information such as race and ethnicity is not representative of the local residents. Mass incarceration’s disproportionate impact on Black and Latino communities, coupled with prisons being built largely in white communities distorts the Census’ demographic data when people aren’t counted in their home communities. This often paints a distorted picture of job opportunities in the local economy.

Many people assume that prison gerrymandering affects funding allocation, but in fact funding allocation seems to be one of the few data uses left untouched by the Census’ prison miscount. Over the decades, funding formulas have developed a level of complexity that leaves them mostly impervious to the Census Bureau’s odd interpretation of residence.

This year, the Census once again counted all incarcerated people as residents of the facility where they were located on April 1. The population data that will be released over the next few months will therefore suffer from the same deficiencies for another decade; it will once again fail to represent the people who actually live in these communities. For now, places such as Ina can find some relevant information in the Census data, if they know where to look and dig carefully enough. Perhaps by 2030 the Census Bureau will count everyone at home rather than shifting the burden of making its data usable onto everyone else.