At this point it feels like we've gone through this a hundred times, over the last decade, but there is no harm in doing it one more time. The push to include a simple citizenship question on the census, seems, on first blush, an innocent-enough attempt at data gathering. What could the harm be, after all?
But census questions are very carefully screened and planned, or at least they are when their government planners are not raging incompetents shoved forcibly into their positions by C-tier ideologues with no expertise in such things and an overt hostility toward people who do have such expertise. As with any other poll taken in any other context, the answers you get depend heavily on the phrasing of the questions you ask—and a government worker or government document asking you are you a citizen, in a climate where that same government has publicly vowed to deport every last non-citizen they can find, will result in those individuals hiding from the census. They will, in large numbers, consider it to be a trap.
Advocates for the question in the Republican Party, and in the Team Trump collection of raging xenophobes, know this. They know this and are counting on it. There is no possibility that asking are you a citizen on a nationwide government census of residents will result in an accurate measure of non-citizens, but it will result in non-citizens dodging the census in larger numbers, and this, in turn, is intended as a way to cause direct harm to cities with large immigrant populations.
An undercount of the population would have far-reaching implications. It could skew the data that are used to determine how many congressional representatives each state gets and their representation in state legislatures and local government bodies. It would shape how billions of dollars a year are allocated, including for schools and hospitals. It would undermine the integrity of a wide variety of economic data and other statistics that businesses, researchers and policymakers depend on to make decisions, including the numbers that underpin the forecasts for Social Security beneficiaries.
Got that? At the most basic level, the census is used to determine the apportionment of Congress. If cities or states with large immigrant populations are undercounted, due to immigrant fear of answering such questions, those regions will have fewer members of Congress after the 2020 census; that representation will be shifted to other states instead. In general, due to demographics, to Republican states.
But the count of how many people—not "citizens," but persons—affects far more than that. It is how government decides how to fairly allocate infrastructure expenses. It is how government decides how much money should be given to cities to combat poverty, and health crises. It is all-encompassing, and the effort to tweak the census to undercount immigrant populations will have very real local effects:
Researchers concluded last year that in the 2015 fiscal year, 132 government programs used information from the census to determine how to allocate more than $675 billion, much of it for programs that serve lower-income families, including Head Start, Medicare, the Supplemental Nutrition Assistance Program, Pell grants for college and reduced-price school lunch programs. Highway spending is also apportioned according to census data.
And it's not just government. Researchers, private businesses; decisions that need to be based on an accurate—or, in this case, less accurate—reckoning of how many people live where, in America, are myriad. An inaccurate census results in some population centers being underserved not just by government, but underserved, period.
All of this has been known for, approximately, forever. That a "citizenship" question on the offical federal census would curtail responses from non-citizens is well known, just as officials know that conducting immigration "raids" in courthouses results in non-citizens being less willing to come forward to protect their own rights, and turning local police officers into ersatz immigration agents during emergency calls results in non-citizens letting violent or nonviolent crimes against them go unreported—elevating crime rates in those communities. All of this is known, by the persons who advocate for each of these things. And they are all advocated for anyway in direct attempts to make the lives of non-citizens worse, and more fear-filled, and, in some cases, more violent.
So there's really no "debate" to be had over this. The citizenship question is intended, directly, to heighten non-citizen fears of answering to the census at all. It is intended to cause undercounts. It is intended as a way to withhold money, services, and even representation itself from immigrant-heavy regions and reapportion those things to predominantly white, immigrant-averse locales. Whether you consider the practitioners of this particular art to be anti-immigrant, or simply racist, depends on how much credit you are willing to give the plan's advocates.
Given how many times they have been told of the effects of their actions and the faux-innocence with with they pursue them anyway, I would suggest that very little credit is due. Given that the Trump campaign is directly fundraising off their actions, their intent seems, in fact, evident.