According to the Federal Reserve Bank of St. Louis, Massachusetts’ population grew at an almost unprecedented rate in the late 1990s, gaining nearly 200,000 residents between 1999 and 2000 alone. If Massachusetts had kept up this pace of growth, it would have over 10.8 million residents in 2018, as opposed to the more modest 6.8 million we actually see today.
Where did this astounding rate of population growth come from? An economic boom? An aggressive housing creation initiative? An influx of immigrants or refugees? None of the above. To put it bluntly, the St. Louis Fed’s data is misleading at best.
The problem is that their intercensal population figures for the 1990s come from an initial Census Bureau estimate, while the figures from each decennial census are corrected to account for something called error of closure. It is only this later correction that keeps Census Bureau data from, say, 1993 comparable with data from census years, but the St. Louis Fed didn’t get the memo.
Inaccuracies in (and misrepresentations of) census data can have serious consequences, from discrimination against minorities to unfair distribution of public funds. As the basis for allocations of congressional representatives among states, the census also has generated highly partisan controversy in recent decades.
You might think that Massachusetts is at low risk for widespread inaccuracies, especially given its low prison population and below-average proportion of illegal immigrants. However, both renters and people with second homes – of which Massachusetts has plenty – can be hard to count as well. The 2020 census is also quickly becoming known as a poorly funded, under-resourced project in general.
Systemic inaccuracies have very different effects on different communities. For individual municipalities, the Census Bureau usually doesn’t make many case-specific adjustments to its initial annual population estimates, often tying it to county-level data. Perhaps this is sufficiently accurate in counties that are largely uniform in character, but consider that, in Massachusetts, communities as disparate as Shirley, Everett, and Weston are all in the same county. Rural communities like Shirley are disproportionately undercounted, as are communities with large shares of foreign-born residents like Everett. Retroactive adjustments made to their demographic composition estimates certainly shouldn’t use the same formula as wealthy, well-established, suburban Weston.
An even larger problem exists in urban areas where data compiled for large, diverse cities masks highly local concerns. For things like food accessibility and racial disparities in education, census tracts and zip codes are usually better levels of measurement than municipalities as a whole. The Census Bureau has considered providing more detailed sub-county data on a regular basis, but its budget woes have greatly limited the resources available for such projects. Census Bureau analysis on a neighborhood level is also often the basis for local election redistricting decisions.
Ultimately, the numbers that the Census Bureau report depend on the decisions of individuals (do we move to Colorado? Do we have a baby now, or wait?), which is part of the reason why it’s so difficult to predict and catalogue this data. Still, assuring the accuracy of census data is worth the additional cost. Demographic data is a significant reflection of social trends and an important indicator of far-reaching societal problems. Business owners use demographic data to match consumer demand, as changes in branding can help a product appeal to different cohorts of the population. These social and economic factors are relevant on a highly local level, especially in ethnic neighborhoods.
Still, a lot of people are more focused on the political implications of census accuracy. Will Massachusetts lose another representative in the House come 2020? Almost certainly not. Massachusetts’s population growth rate is consistently below the national average, but not by much, and the representative it lost in 2010 allows for a bit of a cushion before the next threshold.
As always, it’s important to continue to improve the efficiency and organization procedures of momentous tasks like the decennial census. Wasteful government institutions don’t help anyone. But in an era of high tensions between those who tend to get counted in the census (wealthy, white, native-born suburbanites) and those who don’t (impoverished, foreign-born renters and people of color), everyone has a stake in helping to create a highly accurate and representative decennial census come 2020.
You can explore Massachusetts demographic data on a local level with Pioneer Institute’s MassAnalysis tool.
Andrew Mikula is the Roger Perry Government Transparency intern at Pioneer Institute. He studies economics at Bates College.