Our new intercensal estimates reconcile annual population figures from the last decade with an updated count of the state’s residents captured at the beginning of the next decade.

The newly released Boyd Center Intercensal Population Estimate series covers the period from 2010 to 2020. They include annual, county-level population estimates for Tennessee by five-year age groups, sex, and four combinations of race and Hispanic ethnicity. The series links estimates, initially based on the decennial census count in 2010, to the beginning of the most-recent vintage population estimates, which are based on the 2020 Census and other more recent data sources. That connection is made at July 1, 2020.

The new intercensal estimates are based off two series from the U.S. Census Bureau:

• Vintage 2020 Population Estimates for 2010-2020
• Vintage 2022 Population Estimates for 2020-2022

It’s very unlikely that the July 1, 2020, population from the Vintage 2020 and Vintage 2022 estimates will match. Intercensal estimates correct this difference, called the error of closure, under the assumption that the variance between the two is the result of errors with the estimates that occurred between 2011 and 2020.

Then, after calculating the difference between the two sources, a mathematical adjustment is made to the population for each year from 2011 to 2020. The error of closure is spread across the entire decade and the endpoint for July 1, 2020, is changed to match the Vintage 2022 estimate for that date. This approach preserves the important year-to-year fluctuations captured in the original estimate vintage and enables the creation of a multi-decade, annual population series referencing decennial census counts in 2010, 2020, and the most-recent population estimates.

Tennessee’s population on July 1, 2020, was 6,925,619 people, according to the Census Bureau’s Vintage 2022 Population Estimates. That was 38,785 people more than were reported in the Vintage 2020 estimates for the same date and equates to a -0.6 difference between the new estimate sources. Figure 1 shows how the new intercensal estimate series for 2010 to 2020 raised the state’s population level across the decade so it aligns with the current decade’s population estimates.

Overview of the Tennessee Intercensal Estimate Data

The Tennessee intercensal population estimate series covers the period from 2010 to 2020 and includes annual, county-level population estimates for Tennessee by five-year age groups, sex, and four combinations of race and Hispanic ethnicity (Table 1). The downloadable files include characteristic data for Tennessee and its 95 counties, with the file layout and attribute codes described in a separate document posted to the State Data Center website.

Table 1:Overview of the Characteristic Groups for the Boyd Center Intercensal Population Estimates

Age 18 groups	Race/Hispanic Ethnicity 4 groups	Sex 2 groups
5-year age groups for residents under 85 years of age 1 group for population over 85 years and over.	Non-Hispanic White alone Non-Hispanic Black alone Non-Hispanic other races1 Hispanic or Latino	Female Male

¹Other races includes American Indian and Alaska Native alone, Asian alone, Native Hawaiian and Other Pacific Islander alone, and two or more races.

The characteristics categories used in the Intercensal estimates are compatible those used in the Boyd Center Population Projections. Detailed characteristic can also be explored in a new dashboard. This interface shows the error closure at state and county levels but can also be filtered to show the decade-long divergence trend for any characteristic combination of age, race/Hispanic ethnicity, and sex.

Comparing Boyd Center Intercensal Estimates to Similar U.S. Census Bureau Products

Intercensal estimates are a critical input to Tennessee’s Population Projections that are also produced by the Boyd Center. In support of that product’s forthcoming update, the Vintage 2022 Tennessee Intercensal Population Estimates were developed.

Two factors led to the creation of the new state data and this inaugural release.

First, the U.S. Census Bureau has yet to produce an intercensal population estimate product covering the period from 2010 to 2020. Following last decade’s 2010 decennial census, the final intercensal estimates were released in October 2012, just 18 months after the initial 2010 data was published, with some data pushed out in as little as six months. Current plans have the Bureau’s 2020 intercensal product scheduled for release in the fall of 2024. Delays related to the implementation of new disclosure avoidance methods are largely responsible for this decade’s slowdown and has limited internal access at the Bureau to detailed age, race, Hispanic or Latino origin, and sex from the 2020 Census.

Second, is the Census Bureau’s adoption of the “blended base” population methodology used to establish the April 1, 2020, population. In prior decades, the decennial census served as the population estimate base. But, for this decade, the advantages of incorporating other Census Bureau data that addresses concerns with the quality and completeness of the 2020 Census data will likely result in the blended base being around throughout the balance of the decade. That means with each new population estimate vintage release, a new population base could be established for every state, county, and characteristic group for April 1, 2020, and July 1, 2020.

The requirement to match a shifting July 1, 2020 population will be important in future years as well. For instance, in 2023 Hispanic and Latino origin counts from the 2020 Census will be introduced in the population estimates. In vintage 2024, additional race and age data could be weaved in, along results of Post Census Group Quarter Reviews. Both events could result in significant changes to the state’s characteristic makeup. We expect to produce these new estimates on an annual or biennial basis using a simplified version of the Bureau’s 2000-2010 intercensal approach, along with their planned switch to a new algorithm for distributing the error of closure between 2010 and 2020.