A demonstrator holds a sign referring to the census outside the U.S. Supreme Court on June 27. (Andrew Harrer/Bloomberg)

Editor’s note: This post has been updated to reflect the Trump administration’s new position.

After repeated setbacks in the courts, the Trump administration has given up on its efforts to place a citizenship question on the 2020 Census. Even so, Trump’s repeated emphasis on citizenship — combined with operational challenges in carrying out the census — could still risk what’s called a “differential undercount,” in which some states and demographic groups are undercounted much more than others. How are public officials and others trying to ensure that everyone is counted?

The citizenship question is not the only threat to the census.

The Urban Institute estimates that if the citizenship question is included, the population would be undercounted by 1.22 percent, and the Hispanic/Latinx population by 3.57 percent. That would be the largest undercount since 1990 — former census director John Thompson thinks the 2020 undercount could be even worse — and would distort the distribution of federal resources and congressional seats accordingly.

But as the Institute’s report notes, even if the citizenship question is not included, new operations — including an Internet option — could bring the undercount to 0.84 percent overall and 2.84 percent for the Hispanic/Latinx population. And as some census experts have noted, public debate about the citizenship question may discourage participation, whether it appears on the form or not.

In either case, the federal government has a constitutional obligation to get the numbers right. And to ensure as full a count as possible, the Census Bureau depends crucially on cooperation from civil-society organizations and businesses. Their efforts can be challenging to coordinate.

And it might be harder this year, given the controversy. In a recent New York Times interview, the chief executive of the National Association of Latino Elected and Appointed Officials reported greater difficulties securing businesses’ cooperation in the 2020 Census, suggesting that “business leaders are allergic to issues that are perceived to be controversial, especially if they have any kind of racial controversy mixed in.”

Local and state governments also play an important role in census outreach. Organizing 50 states and over 89,000 local governments is a formidable task. Beginning in 1990, the Census Bureau formally asked local governments to create Complete Count Committees to raise awareness and motivate census participation. That year, only 22 percent of the local jurisdictions contacted by the bureau did so; 35 percent relied on existing governmental structures to get the word out, and 43 percent refused to take part altogether.

Since then, the Census Bureau has expanded its efforts to engage state and local partners. Before the 2010 count, California’s State Complete Count Commission (SCCC) convened public meetings with community leaders, determined locations for Questionnaire Assistance Centers and provided community partners with training materials.

For the 2020 count, the Census Bureau has invited all states to create their own SCCCs to coordinate outreach and mobilization. Thus far, however, only 32 states and the District of Columbia have done so.

SCCCs also face challenges in attracting adequate staff capacity or operational funding. The Chicago Urban League, for example, has called on Illinois to increase its census outreach appropriations by several million dollars. In their most recent legislative sessions, only 17 states and the District of Columbia allocated specific funding for census outreach and mobilization. Funding bills in states such as North Carolina and Arizona have either stalled in committee or died when the legislature adjourned. With millions of federal grant dollars and congressional reapportionment on the line, these states are leaving the task of ensuring an accurate census to cities, counties and civil-society groups.

Why wouldn't states want to fund census outreach?

Officials in states with small “hard-to-count” populations may not be as worried about missing out on federal funds or losing a congressional seat as states with large numbers of immigrants, documented or otherwise. States are significantly more likely to engage in census planning and outreach activities when they are risking a large undercount. In the Urban Institute’s “high risk” scenario, the average projected undercount for states that have thus far funded census outreach is 1.12 percent, compared with an average 0.78 percent undercount for states that have not funded outreach.

Which party controls state government matters as well. Only 13 percent of states with unified Republican government have funded census outreach — compared with 42 percent of states with divided government and 60 percent of states with unified Democratic government.

Of course, these partisan differences partly map onto the size of the states’ probable undercounted populations. Projected undercounts are higher on average in Democratic-controlled states than Republican-controlled states.

But partisanship appears to have an independent influence as well. In Democratic-controlled states that have voted to fund census outreach, the average projected undercount is 1.38 percent compared with 0.96 percent undercount in Democratic-controlled states that haven’t funded outreach. But among states with a divided or Republican-controlled government, those with higher projected undercounts were not significantly more likely to invest in census outreach. In Pennsylvania, where the 2020 undercount could be as high as 0.58 percent, Republicans have thwarted the SCCC’s request for $1 per resident in outreach funding.

The story of the 2020 Census will not be written in the White House alone.

The Census does not belong to any one president or party. Its success depends on national, state and local public officials, nonprofits, businesses, and many others in civil society as well. Ensuring a complete census requires efforts that depend not only on what happens in the White House, but also in statehouses and municipal buildings throughout the country.

Don’t miss anything! Sign up to get TMC’s smart analysis in your inbox, three days a week.

Philip Rocco (@PhilipRocco), an assistant professor of political science at Marquette University, is the co-author of “Obamacare Wars: Federalism, State Politics, and the Affordable Care Act” (University Press of Kansas, 2016).