Counts, Estimates, and Projections, Oh My!

How many people live here? What are the characteristics of our population? These are the questions people use demographic data to answer. However, there are different types of demographic data available to answer those questions and more. The three main sources of population numbers are counts, estimates, and projections. 

Every 10 years, the US Census Bureau conducts a census – a full count of the population, with the goal of counting everyone once, only one, and in the right place.  

But populations change as people are born, die, and move in and out of an area. As more time passes after the count, the less accurate the count becomes. Population estimates can update the previous census count to a more recent date. Estimates become especially important the further we get from the previous census, particularly given that many state and federal funds are allocated, at least partially, according to population.  

Every year, our office produces the official state population estimates of all North Carolina counties and municipalities. We prepare these estimates using the last official US census count as a starting point. From the recent 2020 Census starting point, we add or subtract populations using models that incorporate trends in indicators correlated with population change (such as births, deaths, school enrollment, vehicle registrations, building permit activity, annexations, etc.). 

The US Census Bureau also produces population estimates. While their results are similar to ours, they use different models and input data to produce population estimates for municipalities (incorporated places in Census jargon), counties, and other geographic areas. As part of the Federal-State Cooperative for Population Estimates (FSCPE), we work closely with the US Census Bureau by providing data and reviewing Census Bureau population estimates prior to publication.  

The State Demographer also produces the official population projections for the state and counties. Population projections take historical trends in population change and the components of population change (births, deaths, migration) to estimate a future population. Projections make certain assumptions about future changes in the components of population change.  We produce our population projections annually using trends shown in our latest population estimates.  

What Data Source Should You Use?

It depends.  

It depends on how you will be using the data and whether that purpose requires a certain data source.  

Consider Accuracy

In general, because it is a full accounting of the population, census counts are more accurate than estimates. But censuses may miss counting some populations or may double count others. Population estimates can be quite accurate but rely on variables to approximate population change. Any error in the input data or changes in the relationship between a given indicator and population change can affect the accuracy of the population estimate.   

Population projections do well at predicting the immediate future (one to five years out) because population trends do not change rapidly and thus historic trends can be used to predict the immediate future.¹ That said, areas that experience boom and bust cycle economies are harder to predict and unplanned events, such as natural disasters, can change population trajectories. The recent global pandemic was one such event that altered the short and long-term projections, making pre-2020 projections less accurate.  

Longer term population projections are harder to predict. For instance, the first known set of population projections produced for North Carolina in the 1930s did not predict the baby boom that occurred after World War II.² Demographers would love to be able to predict the future with certainty, but alas, our crystal balls are limited to what we see in recent trends. Population projections depict a future population should current trends continue or an assumed change in current trends occur. 

Consider Requirements

Many times, legislation or a grant program may require you use population numbers from a specified source. A grant may require you use “the last decennial census” – which currently would be the April 1, 2020 Census count. Or it may require you use the US Census Bureau population estimates – which are the annual population estimates produced by the US Census Bureau. If you need to compare your community to communities outside of North Carolina, then you should use the Census Bureau’s population estimates. 

The Census Bureau’s population estimates are also used in weighting of survey-based estimates produced by the Census Bureau and others, including the American Community Survey (ACS).  ACS estimates provide more detailed socioeconomic estimates for one- and five- year periods. The ACS will provide information on things such as vehicle ownership, household income, commuting behavior, educational attainment, etc. You can find out more on Census Bureau data sources from Carolina Demography’s recent primer. 

With state grants, you may be required to use the certified population estimates “as reported by the State Budget Officer.”  These are the certified population estimates produced by the State Demographer. One difference between our population estimates and 2020 Census are municipal boundaries. Municipal populations can change as a result of new areas annexed into a municipality. OSBM certified estimates are statutorily required to include populations annexed into a municipality during the previous fiscal year. Every June, we ask municipalities and counties to respond to our annual demographic information survey to collect information on annexations, building permit activity, and change in populations living in congregate living situations, such as nursing homes, dorms, and homeless shelters.  

If you use the population estimates in planning, we produce a set of standard and revised population estimates to allow you to track populations over time using the same municipal boundaries as of the population estimate date. Finally, after each decennial census, we adjust our historical estimates to match the latest census count so that you can track annual changes during the previous decade.  

Major differences between the 2020 census, our municipal population estimates, and the 2020 Census Bureau incorporated place population estimates: 

If you need to understand future populations, then you would use our state and county population projections. We produce our population projections annually based upon the latest trends shown in our last population estimates. For a variety of reasons, we do not currently produce population projections for areas smaller than counties, including municipalities. The US Census Bureau does not currently produce state or county population projections, but they do periodically produce projections for the nation as a whole.  

If you need to know the age, sex, race, or Hispanic origin of the population, then you can find those characteristics in our population estimates and projections as well as in the US Census Bureau’s population estimates.³ We use slightly different categories than the US Census Bureau and because our models are different, our results may show slightly different estimates in these characteristics for any given area. We include the historical estimates by age, sex, and race as far back as 2000 in our population projections datasets so that you can trace these changes over time from the historical periods through 2050. A separate set of estimates and projections provide total population by sex for Hispanic and Non-Hispanic population (no age detail). 

Overview of OSBM Population Estimates/Projections vs. US Census Bureau Population Characteristics 

Consider What's Available

The US Census Bureau reported the first set of results from the April 1, 2020, Census count for local areas in August 2021. You can access some of those data on LINC. More data will be published in fall 2022. The Census Bureau publishes their population estimates in phases: state population estimates in December of estimate year, county estimates in March of the following year, and Incorporated Place (municipal) estimates (and other subcounty geographies) in May of the following year.

We publish our county and municipal estimates in September of the following year. We published our July 1, 2020, population estimates in September 2021. These population estimates included estimates for counties, municipalities, unincorporated populations within counties, and portions of municipalities located within more than one county. Our 2020 population estimates are the first set of local population estimates based on 2020 Census data. The US Census Bureau will first incorporate 2020 Census information in their July 1, 2021, population estimates published in spring 2022. 

Our Vintage 2021 population projections are now available and these data include our county population estimates as far back as 2000.

Now that you understand more about the different population data products available, you are ready to start confidently using any and all this information about North Carolina and its counties and municipalities. If you have questions about any of these data sources, please contact state.demographer@osbm.nc.gov.

Notes

1)  Because our population estimates require other historical data, our estimates lag by a year or more.  We published the July 1, 2020 population estimates in September 2021. We will publish the July 1, 2021 population estimates in September 2022. Until September 2022, anyone needing July 1, 2021 population numbers would use our population projections for July 1, 2021.
2)  These projections were included in a report produced by the North Carolina State Planning Board with support from the WPA called “The People of North Carolina.”
3)  OSBM and the US Census Bureau population estimates use a modified race file that adjusts race categories reported in the decennial census.