AP Photo/J. Scott Applewhite
Capital & Main is an award-winning publication that reports from California on economic, political, and social issues. The American Prospect is co-publishing this piece.
If all goes according to the Trump administration's plan, the U.S. Census form in 2020 will ask people to state whether they're citizens of the United States. It will also ask citizens if they were born in Puerto Rico or Guam, if they were naturalized and when, and whether they were born outside of the country to U.S.-born parents.
And it will, say researchers who study the accuracy of the decennial census, terrify many a potential respondent—not just undocumented residents, but legal residents and their children as well. In a political environment demonstrably hostile toward immigrants, says Cindy Quezada of the of the Central Valley Immigrant Integration Collaborative in Fresno, “Just the word 'citizen' causes panic.”
As, perhaps, it should. Emails and other documents that emerged from a flurry of lawsuits against the U.S. Commerce Department, which oversees the Census Bureau, revealed an intention to use the decennial census to separate the undocumented from citizens. This effort originated in a January 2017 draft executive order designed to crack down on unauthorized immigrants. Later, Commerce Secretary Wilbur Ross went on the hunt for a more benign justification for including the questions, finally settling on the notion that the administration needed citizen information to protect voting rights.
Judge Richard Seeborg in Northern California described that as “an effort to concoct a rationale bearing no plausible relation to the real reason.”
Two more judges, George Hazel in Maryland and Jesse Furman in New York, broadly agreed. Furman called Ross's effort a “veritable smorgasbord of classic, clear-cut” violations of the Administrative Procedures Act, a 73-year-old law saying that administrative decisions from federal agencies have to be grounded in reality, and make some sort of logical sense.
Now the U.S. Supreme Court has taken up the Commerce Department's appeal of Furman's ruling. And in oral arguments on April 23, the court’s conservative majority appeared to buy Ross’ rationale about protecting voting rights. This, despite reams of evidence showing that the questions would run afoul of the Constitution’s mandate that the census count the entire population.
BY SOME MEASURES, the damage has already been done. The very proposal has “bred a climate of fear and mistrust of responding to the census,” says Rosalind Gold, chief public policy officer with the National Association of Latino Elected Officials (NALEO) Educational Fund, a group that seeks to expand Latino involvement in politics. “People’s fear that the information that they provide to the federal government will be used in a way that could hurt them is already out there.”
Months before Ross announced the census would query residents on their citizenship, the Census Bureau's own data, drawn from 42 focus groups, showed that this fear wasn't limited to Latinos: Arab and Chinese-speaking residents also expressed concerns that, in the current political environment, any information they hand over to the federal government could be used against them and their families. Bureau field representatives were finding it increasingly necessary to spend time “calming respondents and gaining their trust due to the current ‘political state.'” This was not true, they said, three years before.
The risk of leaving residents uncounted is already high in states with sizable immigrant populations. With the citizenship questions, any hopes of accuracy will likely be dashed. And nowhere is that more true than in California, where the share of the immigrant population, at 27 percent, is nearly twice the national average, and Latinos represent 34 percent of the adult population.
“In an immigrant-hostile census process,” says Ed Kissam, a researcher who has studied undercounts in the last three decennial censuses, “California is going to lose.”
Latino residents are chronically undercounted in every census. The citizenship questions will make that situation worse. Way worse.
The last time the Census Bureau conducted a count, they offered a pared-down 10-question survey specifically aimed at Latinos: “10 preguntas en 10 minutos”. The Spanish-language broadcaster Telemundo launched a year-long campaign called Hazte Contar (Be Counted), and even worked census participation into the hit telenovela Más Sabe el Diablo.
And still Latino residents were undercounted in the 2010 Census by an official 1.5 percent according to the Census Bureau. That may not seem like much until you consider that the “non-Hispanic white alone” population was overcounted by 0.8 percent. The gap between those two numbers is known as the “differential undercount,” and at 2.3 percent nationwide, it’s significant.
That's especially true when you consider that California's undercounted are largely concentrated in low-income, high-immigrant communities. “We not only have a lot of immigrants” in California, says Ed Kissam. “We have a lot of census tracts where a very high proportion of the community is made up of immigrants.”
The ramifications for those communities can be intense. For instance, in the small farming town of Firebaugh, California, the 2000 census counted only 5,700 residents. City Manager Jose Ramirez later told the Merced Sun-Star that his town had been undercounted by 3,000 to 5,000 residents. The town is 95 percent Latino, primarily Mexican immigrants. When census takers came around, Ramirez guessed that some undocumented workers—workers upon whom California’s economy depends—simply declined to answer the door.
Nor is it only the undocumented who shy from an “immigrant-hostile” census. A study that Kissam worked on with the San Joaquin Valley Census Research found that the citizenship questions nearly halved the response rate among all categories of the first- and second-generation Latinos they interviewed. Green-card holders, naturalized citizens and even Latino residents born in the U.S. showed reluctance to participate if the census asked about their citizenship status.
“People assume the question is intended to identify the undocumented,” says Cindy Quezada. “People wonder how [the citizenship] information is going to be used. What does citizenship have to do with the number of people living here?"
“They don't see the link,” adds Quezada, who helped gather data from 600 face-to-face interviews for the San Joaquin Valley study. “And if they don't see the link, there must be bad intent.”
As a result of the undercount, California could lose a House seat. Or two.
California has gained seats after every census since 1920 until the last one, in 2010, when the state only retained its 53 seats. But had undocumented residents not been counted at all in 2010, according to a 2007 analysis by the Connecticut State Data Center, California would have lost two seats—more than would have been subtracted from any other state in the nation.
Most congressional districts with dense concentrations of non-citizens, including those in northern Los Angeles County (CA-29 and CA-34), in Imperial County (CA-51), and around the city of Santa Ana in Orange County, are held by Democrats. One, CA-21, which includes parts of Kings and Kern Counties, flipped from red to blue in 2016. But a loss of seats combined with a Latino undercount could throw enough confusion into the post-census reapportionment process that some Republicans could be threatened, too. CA-22, which contains parts of Fresno and Tulare Counties in the Central Valley, is 42 percent white and 45 percent Latino, with two other historically hard-to-count minorities, blacks and Asians, making up the difference. It's currently represented by Devin Nunes, a Republican.
Kissam thinks it's unlikely that an immigrant undercount would lead to the Central Valley being carved up; Los Angeles County will likely take the brunt of the damage, as it did in 2010, when the census found its population had grown at only one-third the rate of the rest of the state. (That estimate was almost certainly based on an undercount of Latino residents, children in particular: Post-Census analysis by the NALEO Educational Fund and the Child Trends’s Hispanic Institute found that census-takers had overlooked 47,000 Latino children in Los Angeles County alone.)
Of course, California's share of the U.S. House of Representatives also depends on how other states fare in the 2020 Census—there are, after all, only 435 seats to divvy up among 50 states. If the undercount is severe and concentrated in high-immigrant states, even Nunes could lose his seat.
California's Medicaid program, Medi-Cal, won't be affected. But local programs that depend on federal funding will suffer.
Census undercounts have cost California in the past. According to the state Legislative Analyst's Office, the 1990 census, which undercounted California’s population by nearly three percent, not only deprived California of a congressional seat it should have gained, but cost $200 million in federal funds in just a single fiscal year. Overall, the state was deprived of $2 billion in federal funding because the census overlooked 840,000 people.
An undercount in 2020 won't affect the bulk of federal funding to the state. Most of the $375 billion or so that California gets from the federal government isn't tied to population counts. The $3.7 billion the federal government delivers to the state as a block grant for Temporary Aid for Needy Families (TANF) is based on the amount of funding each state received for similar programs before 1996. (One can argue that it should change based on population growth and inflation, but that's another story.)
Nor would the money that goes to California's Medicaid program, Medi-Cal—$35 billion in the last fiscal year—be reduced. Medi-Cal costs are shared by the state and the federal government, and the federal government shells out according to how the state's per-capita income compares with that of other states. California already gets the minimum amount of funds based on that formula; it can't go any lower than 50 percent of what the state contributes. For funding to increase, the state would need to show not just population growth but an attendant decline in personal income. Recent estimates from the state Department of Finance show that California's population growth slowed dramatically in 2018.
But that doesn't mean an undercount won't imperil federal programs in California. A slew of educational initiatives, from workforce development to school lunches, could be at risk. Federal funds distributed to local education agencies rely on counts of children living in poverty. If those kids are undercounted again, by even greater numbers than in 2010, some local education agencies in largely Latino and poor areas could lose as much as 15 percent of their funding. The Los Angeles Unified School District, where the percentage of non-citizen households exceeds both the nation's and the state's, would be first to lose out.
The San Joaquin Valley Census Project's study estimates that if asking about citizenship on the census form does in fact trigger an undercount of nearly 12 percent in the Valley's first- and second-generation Latino immigrant population, as their models show it may, that “would result in a potential federal funding loss to the region of about $198 million.” Over the decade following the census, the total loss could be as much as $2 billion.
That's not just a bad scenario for Latinos, argues Rosalind Gold. If the U.S. Supreme Court, when it votes on the citizenship questions in mid-June, decides to allow the question, “We [in California] will not be able to meet the needs of our residents,” Gold says. Worse, “We will have undermined our democracy for the nation as a whole.”
