Skip Navigation

Getting the Count Right

Summary: The Census will determine the distribution of political power for the next decade. Our experts outline major questions surrounding the count.

Census Day is rapidly approach­ing, and efforts to get out the count are proceed­ing despite severe head­winds. The stakes are signi­fic­ant: the once-per-decade enumer­a­tion of every­one in Amer­ica will determ­ine how congres­sional seats and hundreds of billions of dollars in federal fund­ing get distrib­uted among the states. This primer answers some basic ques­tions: Will the Census Bureau keep your data safe? How well — or poorly — have past censuses done at count­ing every­one? How will the results of the Census affect the distri­bu­tion of polit­ical power? How might litig­a­tion and the courts influ­ence the count or how the numbers are ulti­mately used? And how might the coronavirus affect the Census process and the final numbers? Along the way, we high­light resources that offer a deeper dive on these and other census issues.


Parti­cip­at­ing in the census is safe. Yet in today’s envir­on­ment, trust in the federal govern­ment is at an extreme low. Many people — and espe­cially communit­ies of color — fear that the Census Bureau will share their answers with other govern­ment agen­cies or that the Trump admin­is­tra­tion will use their answers against them. foot­note1_yoaqnro 1 U.S. Census Bureau, 2020 Census Barri­ers, Atti­tudes, and Motiv­at­ors Study Survey Report, 2019, 38–50,­nial/2020/program-manage­ment/final-analysis-reports/2020-report-cbams-study-survey.pdf. These concerns are real and threaten to depress the count if not adequately addressed. foot­note2_dz2m1qe 2 U.S. Census Bureau, 2020 Census Oper­a­tional Plan: A New Design for the 21st Century, 2018, 173–174,­nial/2020/program-manage­ment/plan­ning-docs/2020-oper-plan4.pdf. But in fact, federal law offers strong and detailed protec­tions against anyone — includ­ing in the federal govern­ment — abus­ing the personal inform­a­tion that the Census Bureau collects. A national network of attor­neys has mobil­ized to uphold these laws.

Laws protect­ing the confid­en­ti­al­ity of census responses are strong.

Robust legal protec­tions prohibit the Census Bureau or any other part of the federal govern­ment from using census data against the people who supply it. foot­note3_p8um­spp 3 These protec­tions arise from several laws relat­ing to census data, includ­ing the Census Act of 1954, as amended, 13 U.S.C. § 1 et seq. (1954); the Confid­en­tial Inform­a­tion Protec­tion and Stat­ist­ical Effi­ciency Act of 2002, 44 U.S.C. § 3501, Note (2002); the Privacy Act of 1974, 5 U.S.C. § 552a (1974); and the Internal Revenue Code, 26 U.S.C. § 1 et seq. (1954). These laws bar census responses from leav­ing the four walls of the bureau except as aggreg­ate, anonym­ous stat­ist­ics. foot­note4_jww7r32 4 The Census Bureau is a compon­ent of the U.S. Depart­ment of Commerce. All laws discussed here that apply to the Census Bureau also bind the Commerce Depart­ment, and all refer­ences to the Census Bureau incor­por­ate the Commerce Depart­ment.

The laws that safe­guard the confid­en­ti­al­ity of census data make clear that, among other things, the Census Bureau cannot disclose census responses in any way that would person­ally identify anyone. foot­note5_w01mieb 5 13 U.S.C. § 9(a)(2). The laws also bar other federal agen­cies from using census data for any nons­tat­ist­ical purpose, such as enfor­cing immig­ra­tion or other laws. foot­note6_t3c65rz 6 13 U.S.C. § 9(a)(1). Federal employ­ees attempt­ing to misuse census data would expose them­selves to seri­ous legal consequences. foot­note7_nrk6gj8 7 13 U.S.C. § 214; 18 U.S.C. §§ 3559, 3571.

These protec­tions apply equally to any data that the Census Bureau gath­ers from other federal agen­cies or the states. foot­note8_xw061e3 8 See Kelly Percival, “Strong Confid­en­ti­al­ity Laws Protect All Data the Census Bureau Collects,” Bren­nan Center for Justice, Decem­ber 5, 2019, https://www.bren­nan­cen­­ion/strong-confid­en­ti­al­ity-laws-protect-all-data-census-bureau-collects; Kelly Percival, “Strict Confid­en­ti­al­ity Laws Limit Trump Admin­is­tra­tion Search for Citizen­ship Data,” Bren­nan Center for Justice, Octo­ber 11, 2019, https://www.bren­nan­cen­­ion/strict-confid­en­ti­al­ity-laws-limit-trump-admin­is­tra­tion-search-citizen­ship. So, once an indi­vidu­al’s personal inform­a­tion gets into the bureau’s hands — however it gets into the bureau’s hands — it’s guarded by strong legal protec­tions.

The pres­id­ent cannot change or elim­in­ate the confid­en­ti­al­ity protec­tions.

The confid­en­ti­al­ity protec­tions for census responses are an estab­lished part of federal stat­utory law. foot­note9_fs7s6bb 9 See note 3 above. As a result, neither the pres­id­ent nor any federal agency can change or elim­in­ate these protec­tions. The only way to weaken them would be through new legis­la­tion passed by major­it­ies in both the U.S. Senate and the House of Repres­ent­at­ives. Barring major elect­oral upheavals, no such major­it­ies are likely to exist anytime soon.

Community lead­ers and lawyers will guard against confid­en­ti­al­ity breaches.

Many community lead­ers are work­ing to promote census parti­cip­a­tion and protect their communit­ies from over­reach. Mean­while, scores of lawyers are mobil­iz­ing to defend census data from any abuses.

Attor­neys have put a signi­fic­ant infra­struc­ture in place to protect census confid­en­ti­al­ity. Multiple hotlines have been estab­lished to permit census respond­ents and grass­roots lead­ers to report any suspec­ted confid­en­ti­al­ity prob­lems, as well as other issues. foot­note10_inj6ei2 10 Stan­ley Augustin, “Civil Rights Groups Launch Multi­lin­gual Census Hotlines, Lawyers’ Commit­tee for Civil Rights Under Law,” Febru­ary 11, 2020, https://lawyer­scom­mit­­lin­gual-2020-census-hotlines. These hotlines are the work of the Lawyers’ Commit­tee for Civil Rights Under Law, the National Asso­ci­ation of Latino Elec­ted and Appoin­ted Offi­cials (NALEO), Asian Amer­ic­ans Advan­cing Justice, and the Arab Amer­ican Insti­tute. The Mexican Amer­ican Legal Defense and Educa­tional Fund (MALDEF) and the Lead­er­ship Confer­ence on Civil and Human Rights have organ­ized attor­neys to support any neces­sary legal action. And the Bren­nan Center has supplied the legal analysis for any litig­a­tion.

The commit­ment to protect­ing confid­en­ti­al­ity is long running and nonpar­tisan.

Confid­en­ti­al­ity has been a bedrock feature of the census for well over a century and a half. In 1850, even before federal law form­ally required that census responses remain confid­en­tial, the U.S. secret­ary of the interior — who was then tasked with conduct­ing the enumer­a­tion — decreed it offi­cial bureau policy that indi­vidual census takers could not publicly reveal any inform­a­tion they collec­ted. foot­note11_dmxeu7r 11 George Gate­wood et al., A Mono­graph on Confid­en­ti­al­ity and Privacy in the U.S. Census, U.S. Census Bureau, 2001, 6,­en­ti­al­ity­Mono­graph.pdf. By 1929, Congress had expan­ded this constraint to a blanket ban on the Census Bureau’s shar­ing that data. foot­note12_4ulthgc 12 Reap­por­tion­ment and Census Act of 1929 §§ 8, 11, Pub. L. No. 71–13, 46 Stat. 21 (1929). The 1929 Census Act also included prohib­i­tions on the bureau’s disclos­ing person­ally iden­ti­fi­able data, using census data for nons­tat­ist­ical purposes, and using census data to the detri­ment of census respond­ents. foot­note13_ankrlss 13 Reap­por­tion­ment and Census Act at §§ 11, 18.

In keep­ing with this long-running commit­ment to confid­en­ti­al­ity, both Repub­lican and Demo­cratic admin­is­tra­tions have for decades publicly affirmed that no one may be harmed by giving inform­a­tion to the census and that collec­ted inform­a­tion cannot be used to enforce immig­ra­tion law or any other federal, state, or local law or ordin­ance. foot­note14_1wnb2e5 14 See, e.g., Proclam­a­tion No. 8488, 75 Fed. Reg. 17837, 17837 (March 31, 2010) (Pres­id­ent Obama affirm­ing that census inform­a­tion “is never used against [census parti­cipants] or shared with other govern­ment or private entit­ies”); Proclam­a­tion No. 6105, 55 Fed. Reg. 8897, 8897–8898 (March 6, 1990) (Pres­id­ent George H. W. Bush affirm­ing that “indi­vidual inform­a­tion collec­ted will not be used for purposes of taxa­tion, invest­ig­a­tion, or regu­la­tion, or in connec­tion with milit­ary or jury service, the compul­sion of school attend­ance, the regu­la­tion of immig­ra­tion, or the enforce­ment of any other Federal, State, or local law or ordin­ance”); Proclam­a­tion No. 3973, 35 Fed. Reg. 5079 (March 26, 1970) (Pres­id­ent Nixon affirm­ing that census inform­a­tion cannot be used “for the purposes of taxa­tion, invest­ig­a­tion, regu­la­tion, or for any other purpose what­so­ever affect­ing the indi­vidual”). These prohib­i­tions, customs, and norms set the baseline against which any action by the Trump admin­is­tra­tion should be judged.

Congress has strengthened census confid­en­ti­al­ity protec­tions in the wake of past abuses.

One distinct but seri­ous depar­ture from these histor­ical norms and stat­utory protec­tions took place during World War II, when the federal govern­ment used census data to intern Japan­ese Amer­ic­ans in camps. During the war, the Census Bureau complied with the federal govern­ment’s requests to release block-level census data that iden­ti­fied neigh­bor­hoods in seven states where Japan­ese Amer­ic­ans were living. foot­note15_mehx4u5 15 J.R. Minkel, “Confirmed: The U.S. Census Bureau Gave Up Names of Japan­ese-Amer­ic­ans in WW II,” Scientific Amer­ican, March 30, 2007, https://www.scien­ti­ficamer­ The bureau also provided the addresses of indi­vidual Japan­ese Amer­ic­ans to the U.S. Treas­ury. foot­note16_bl4w60c 16 Minkel, “Confirmed: The U.S. Bureau Gave Up Names.” See also William Seltzer and Margo Ander­son, “Census Confid­en­ti­al­ity under the Second War Powers Act (1942–1947)” (paper, Popu­la­tion Asso­ci­ation of Amer­ica Annual Meet­ing, New York, 2007), https://margoander­­son­PAA2007pa­per­3–12–2007.doc. A contem­por­an­eous govern­ment report called census data “the most import­ant single source of inform­a­tion prior to the evac­u­ation.” foot­note17_j8e5fbg 17 John L. DeWitt, Final Report: Japan­ese Evac­u­ation from the West Coast 1942, U.S. Govern­ment Print­ing Office, 1943, 352,­e­see­vac­u­ati00dewi/mode/2up.

That kind of census data use is not legal today. The bureau shared the data that suppor­ted intern­ment only because Congress had tempor­ar­ily permit­ted it to do so through the Second War Powers Act of 1942. foot­note18_sdux­fuh 18 Second War Powers Act of 1942 §§ 1401–1402, Pub. L. No. 77–507, 56 Stat. 176 (1942). In addi­tion to author­iz­ing census data disclos­ures, the Second War Powers Act, in combin­a­tion with the First War Powers Act of 1941, Pub. L. No. 77–354, 55 Stat. 838 (1941), gave the pres­id­ent sweep­ing author­ity to reor­gan­ize the exec­ut­ive branch, condemn prop­erty, and censor mail — among other powers — in support of the war effort. But Congress let the Second War Powers Act expire in 1947, reaf­firmed the confid­en­ti­al­ity protec­tions as part of the Census Act of 1954, and since then has only strengthened the safe­guards for census data. For example, in 1962, Congress amended the Census Act to prevent census records from being used as evid­ence in any legal or admin­is­trat­ive proceed­ing. foot­note19_18cey3m 19 Pub. L. No. 87–813, 76 Stat. 922 (1962) (amend­ing 13 U.S.C. § 9(a)). Both the Privacy Act of 1974 and the Confid­en­tial Inform­a­tion Protec­tion and Stat­ist­ical Effi­ciency Act of 2002 provide addi­tional protec­tions for census data.

For more on the safe­guards against abuse of census inform­a­tion, explore the Bren­nan Center’s primer Federal Laws that Protect Census Confid­en­ti­al­ity.

End Notes

Census Quality

The 2020 Census faces many poten­tial complic­a­tions. This year’s head count is the first in which the bureau will attempt to collect most responses over the inter­net, rais­ing the possib­il­ity of tech­no­lo­gical prob­lems, such as website fail­ures, that could derail the process. Mean­while, the Trump admin­is­tra­tion’s efforts to intro­duce a citizen­ship ques­tion to the 2020 Census, although ulti­mately blocked by the Supreme Court, have stoked fears that could dampen response rates. These and other chal­lenges raise signi­fic­ant concerns that this year’s Census could be severely inac­cur­ate. foot­note1_tbqw63r 1 Govern­ment Account­ab­il­ity Office, 2020 Census: Initial Enumer­a­tion Under­way but Read­i­ness for Upcom­ing Oper­a­tions Mixed, 2020, https://over­­crats.over­–368R%202–11–20.pdf; Maya King, “Census Bureau Spends Millions on Ads Combat­ing Citizen­ship Ques­tion Scare,” Politico, Febru­ary 18, 2020,­ship-ques­tion-115718; Michael Macagnone, “It’s Not Just the Citizen­ship Ques­tion. 2020 Census Faces Other Woes,” Roll Call, June 3, 2019,­ship-ques­tion-2020-census-faces-other-woes; Michael Wines, “With 2020 Census Loom­ing, Worries About Fair­ness and Accur­acy,” New York Times, Decem­ber 9, 2017,­trict­ing.html; Robert Shapiro, “The 2020 Census May Be Wildly Inac­cur­ate — and It Matters More Than You Think,” FixGov (blog), August 31, 2017, https://www.brook­­cur­ate-and-it-matters-more-than-you-think.

When wading into debates over the 2020 Census’s perform­ance, it’s useful to have some famili­ar­ity with meas­ure­ments of “census qual­ity,” a term that encom­passes, among other things, how well the census meas­ures the size and char­ac­ter­ist­ics of the nation’s popu­la­tion, includ­ing the age, sex, race, ethni­city, and geographic loca­tion of each person. foot­note2_w4ci79p 2 This aspect of census qual­ity is known as “cover­age meas­ure­ment.” “Cover­age Meas­ure­ment,” U.S. Census Bureau, accessed Febru­ary 25, 2020,­nial-census/about/cover­age-meas­ure­ment.html. For a broader account of census qual­ity, see United Nations Economic Commis­sion for Europe, Confer­ence of European Stat­ist­i­cians Recom­mend­a­tions for the 2020 Censuses of Popu­la­tion and Hous­ing, 2015, 69–75,­min/DAM/stats/public­a­tions/2015/ECECES41_EN.pdf. Back­ground of this kind can help identify trends and recur­ring prob­lems.

The census has been getting better at meas­ur­ing the size of the nation’s popu­la­tion.

Accord­ing to stat­ist­ics that the Census Bureau and its employ­ees have repor­ted, the census has been doing increas­ingly well along at least one dimen­sion of census qual­ity: accur­ately meas­ur­ing the total number of people in the coun­try.

To assess the qual­ity of the census’s national popu­la­tion count, the bureau prepares bench­marks that can serve as points of compar­ison. Two main meth­ods are used to produce the bench­marks. One, called demo­graphic analysis, combines data from birth, death, migra­tion, and Medi­care enroll­ment records to create an expec­ted popu­la­tion total. foot­note3_g0xu5yw 3 Eric B. Jensen, “The 2020 Demo­graphic Analysis Program,” (present­a­tion, South­ern Demo­graphic Asso­ci­ation 2019 Annual Meet­ing, New Orleans, LA, Octo­ber 23–25, 2019),­graph­ers-assoc-meet­ing/present­a­tions/2020-demo­graphic-analysis-program.pdf?. The other, called dual system estim­a­tion, gener­ates a popu­la­tion estim­ate based on data gathered from a survey of a repres­ent­at­ive sample of house­holds. foot­note4_nfwwh3b 4 “Post-Enumer­a­tion Surveys,” U.S. Census Bureau, accessed Febru­ary 25, 2020,­age_meas­ure­ment/post-enumer­a­tion_surveys. This method uses a poste­nu­mer­a­tion survey (PES), so called because it happens after census data collec­tion has been completed. The bureau can match people’s responses to this poste­nu­mer­a­tion survey with their responses to the census on a case-by-case basis to determ­ine the demo­graphic char­ac­ter­ist­ics of those it coun­ted correctly and those it missed.

The bureau finds the differ­ences between the final census count and each of the two bench­mark estim­ates to assess the accur­acy of the census. The result­ing stat­ist­ics are called either “net national under­counts” or “net national over­counts,” depend­ing on whether they suggest that the census under­shot or over­shot the total popu­la­tion. Net under­counts and over­counts can be expressed as numbers of people or as percent­ages.

Table 1 illus­trates these rates as percent­ages over time, with negat­ive values indic­at­ing an under­count and posit­ive ones an over­count. As table 1 suggests, the national-level under­count rate has over time been approach­ing zero. This suggests that the census has been getting better at gauging the size of the entire popu­la­tion, only slightly under­count­ing or over­count­ing the popu­la­tion in 2000 and 2010.

The bureau has said that it will release its latest popu­la­tion estim­ates based on the demo­graphic analysis method in Decem­ber 2020 and its estim­ates based on the dual system estim­a­tion method in June 2021. foot­note5_j74dt8q 5 U.S. Census Bureau, 2020 Census Oper­a­tional Plan, 2018, 153,­nial/2020/program-manage­ment/plan­ning-docs/2020-oper-plan4.pdf; U.S. Census Bureau, Decen­nial Census Manage­ment Divi­sion, 2020 Census Detailed Oper­a­tional Plan for 30. Eval­u­ations and Exper­i­ments Oper­a­tion (EAE), 2019, 4,­nial/2020/program-manage­ment/plan­ning-docs/EAE-detailed-oper­a­tional-plan.pdf.

Histor­ic­ally, the census has not coun­ted all demo­graphic groups equally well.

Crucially, while the census’s count of the total national popu­la­tion appears to have grown more accur­ate over time, other aspects of the count have encountered persist­ent prob­lems.

For example, it is no secret that the census has long struggled to count communit­ies of color — espe­cially Black and Latino communit­ies — as accur­ately as it counts others. foot­note6_28hryuz 6 For a compre­hens­ive account of differ­en­tial under­counts, see William O’Hare, Differ­en­tial Under­counts in the U.S. Census: Who is Missed? (Cham: Springer Open, 2019). This prob­lem can be masked in the national popu­la­tion stat­ist­ics, because those metrics only meas­ure the nation as a whole. Under those metrics, for instance, a white person with two resid­ences in New Jersey whom the bureau inad­vert­ently counts twice could “offset” a Black person with one resid­ence in Alabama whom the bureau does not count at all.

Figur­ing out whether a census has coun­ted all groups with the same degree of accur­acy requires compar­ing one group’s net under­count rate to another’s. Table 2 contains net under­count rates for seven major racial or ethnic groups from the past three censuses, expressed as percent­ages, as gener­ated through the bureau’s dual system estim­a­tion method.

Once demo­graph­ers have these net under­count rates, they can identify the differ­ences in the rates between groups and in that way compare how well a given racial or ethnic group has been coun­ted vis-à-vis other groups. (Geographic areas can be compared in a similar way.)

Table 3 shows differ­en­tials in the estim­ated percent net under­count rates for six major racial or ethnic groups as compared with Non-Hispanic Whites in each of the last three censuses, using the data in table 2. These stat­ist­ics show persist­ent — and in some cases severe — differ­ences in accur­acy along racial and ethnic lines.

Census popu­la­tion stat­ist­ics can conceal signi­fic­ant numbers of people that the bureau misses.

There’s also a prob­lem with the net under­count rates: the bureau can completely miss some substan­tial subset of people. These rates can obscure this issue with “omis­sions,” because the under­count rates include people whom the bureau coun­ted but should­n’t have, such as people coun­ted twice, as well as people whom the bureau has imputed, or inser­ted into the count, based on some evid­ence that they exist, such as signs that their hous­ing unit is occu­pied.

For example, accord­ing to results gener­ated through the dual system estim­a­tion method, the 2010 Census had a net national over­count (0.01 percent) that was not stat­ist­ic­ally differ­ent from zero, caus­ing the bureau to declare the 2010 Census “outstand­ing.” foot­note7_js6j71f 7 Robert Groves, “How Good Was the 2010 Census? A View from the Post-Enumer­a­tion Survey,” Direct­or’s Blog, U.S. Census Bureau, May 30, 2012,­room/blogs/director/2012/05/how-good-was-the-2010-census-a-view-from-the-post-enumer­a­tion-survey.html. Yet the bureau’s dual system estim­a­tion method suggests that the 2010 Census effect­ively failed to count 5.3 percent of the popu­la­tion, or 16 million people. foot­note8_2bckau7 8 Thomas Mule, 2010 Census Cover­age Meas­ure­ment Estim­a­tion Report: Summary of Estim­ates of Cover­age for Persons in the United States. DSSD 2010 Census Cover­age Meas­ure­ment Memor­andum Series, #2010-G-01, U.S. Census Bureau, 17,­age_meas­ure­ment/pdfs/g01.pdf.

When such omis­sions cluster in partic­u­lar areas or within partic­u­lar demo­graphic groups, those communit­ies and groups are vulner­able to being estim­ated less accur­ately than others, in terms of both their raw numbers and their partic­u­lar char­ac­ter­ist­ics, such as sex or age. Clustered omis­sions are also an indic­ator that the bureau’s census-taking meth­ods have not reached every community equally and that the census data does not adequately repres­ent our coun­try in all its diversity.

Early indic­at­ors of 2020 Census progress may provide warn­ings about ulti­mate census qual­ity.

Over the course of the spring, the Census Bureau will publi­cize an indic­ator of the Census’s progress: the share of all hous­ing units in the bureau’s address list that respon­ded to the Census online, with a paper form, or over the phone. foot­note9_e4hu9l5 9 Terri Ann Lowenthal, “Track­ing 2020 Census Progress in ‘Real Time,’” accessed Febru­ary 25, 2020, https://funder­scom­mit­­ing-Census-Progress-Factsheet-1.pdf. This stat­istic, called the “self-response rate,” tracks the percent­age of hous­ing units on the bureau’s master list that have respon­ded, not the percent­age of people who have respon­ded.

Self-response rates can help flag prob­lem areas that will require more atten­tion to be coun­ted accur­ately. Experts have also observed that geographic areas with low self-response rates have histor­ic­ally exper­i­enced “higher rates of omis­sions, erro­neous enumer­a­tions, and net under­count.” foot­note10_a0fakbm 10 Expert Report of Nancy Math­iowetz, Ph.D., La Union Del Pueblo Entero v. Ross, No 8:18-cv-01570-GJH (D. Md. Janu­ary 7, 2019), ECF No.1012–5, 26. See also Expert Report of William O’Hare, Ph.D., New York v. U.S. Dep’t of Commerce, No. 1:18-cv-2921-JMF (S.D.N.Y. Novem­ber 7, 2018), ECF No. 507–1, 24–28. Thus, this data may give insights into the ulti­mate qual­ity of the Census before the demo­graphic analysis and dual system estim­a­tion numbers become avail­able.

The Census Bureau will report self-response rates start­ing on March 20. foot­note11_t5gymjl 11 “Response Rates,” U.S. Census Bureau, accessed March 9, 2020, https://2020­­cbbtv3E6u2so52J3P0E­henB­sISZe8gJR_p05At­gV4XB­VWSlz4UxM The bureau will update the rates daily. An expert with several decades of census over­sight exper­i­ence suggests that stake­hold­ers use the 2010 self-response rates, which the bureau has published on its website, as the bench­mark for eval­u­at­ing the Census’s progress through the end of the self-response period on April 30. Stake­hold­ers should keep in mind that the char­ac­ter­ist­ics of any given area might have changed, some­times signi­fic­antly, over the past decade, thereby affect­ing the like­li­hood of self-response in this year’s count. foot­note12_mz7ani7 12 Terri Ann Lowenthal, email message to Thomas Wolf, March 6, 2020.

End Notes

Effects on the Distribution of Political Power

The results of the Census will have broad polit­ical rami­fic­a­tions: they determ­ine how seats in the House of Repres­ent­at­ives will be alloc­ated among the states and how states will draw their congres­sional and legis­lat­ive districts. Import­ant work has already been done to project how seat alloc­a­tions might shift after the Census. A major study by Elec­tion Data Services concluded that 10 states will each lose a House seat and 7 states will gain one or more under any of a series of popu­la­tion projec­tions for 2020. foot­note1_s3dzzqs 1 Kimball Brace, “2019 Reap­por­tion­ment Analysis,” Elec­tion Data Services, Decem­ber 30, 2019, 3, https://www.elec­tiondata­ser­ The Urban Insti­tute recently released a report project­ing risks that could affect each state’s final popu­la­tion totals, noting that its high-risk scen­ario could influ­ence seat changes beyond those that Elec­tion Data Service’s 2018 report anti­cip­ates. foot­note2_w52ztoo 2 Diana Elli­ott et al., Assess­ing Miscounts in the 2020 Census, Urban Insti­tute, 2019, 20–21,­a­tion/assess­ing-miscounts-2020-census. Another crucial but less often discussed rami­fic­a­tion of the Census is its impact on the repres­ent­a­tion of communit­ies of color.

The Census will affect the polit­ical power of communit­ies of color in Congress.

How well the Census counts communit­ies of color will directly affect their abil­ity to have their voices heard in the House of Repres­ent­at­ives. Take, for example, Section 2 of the Voting Rights Act, which seeks to provide communit­ies of color with elect­oral districts where they can elect candid­ates of their choice. In order to activ­ate Section 2’s protec­tions, members of a given community must show that they are numer­ous and geograph­ic­ally clustered and that they vote as a bloc in elec­tions. foot­note3_9esorsi 3 Justin Levitt, A Citizen’s Guide to Redis­trict­ing, Bren­nan Center for Justice, 2010,47, https://www.bren­nan­cen­–08/Report_CGR-2010-edition.pdf. If the Census misses a mean­ing­ful percent­age of a state’s resid­ents of color — in other words, if the Census makes them appear less numer­ous than they actu­ally are — then their case for Section 2 protec­tion becomes harder to make. An under­count of communit­ies of color could result in those communit­ies losing seats in their state’s congres­sional deleg­a­tion, because lawmakers may be led to believe that they are no longer legally oblig­ated to draw districts for those communit­ies.

The Census will affect the polit­ical power of communit­ies of color at the state and local levels.

These prob­lems at the congres­sional level can replic­ate them­selves all the way down the polit­ical scale. State, county, and local govern­ments all rely on census numbers to draw their elect­oral districts, and Section 2’s protec­tions extend to those juris­dic­tions as well. If the Census under­counts communit­ies of color, they risk going under­rep­res­en­ted — or unrep­res­en­ted entirely — in bodies ranging from state legis­latures to local school boards.

End Notes

The Census in the Courts

Federal courts have long been involved in the census, tack­ling issues like the right way to count over­seas milit­ary person­nel and the legal­ity of using stat­ist­ical sampling. foot­note1_peyk034 1 Frank­lin v. Massachu­setts, 505 U.S. 788 (1992); Depart­ment of Commerce v. United States House of Repres­ent­at­ives, 525 U.S. 316 (1999). The courts have already weighed in on the 2020 Census, prevent­ing the Trump admin­is­tra­tion from asking about citizen­ship status on the ques­tion­naire. The courts could well remain involved long after the Census ends if there are concerns about the qual­ity of the count or its inde­pend­ence from polit­ical manip­u­la­tion.

Federal courts helped protect the Census by ensur­ing that there will be no citizen­ship ques­tions.

In its June 2019 decision block­ing the citizen­ship ques­tion, the Supreme Court ruled that the admin­is­tra­tion had acted illeg­ally by lying to the Amer­ican public about its real reason for want­ing the ques­tion to appear. foot­note2_u3n0oky 2 Depart­ment of Commerce v. New York, 139 S. Ct. 2551, 2576 (2019). The Court’s decision means that the 2020 Census cannot contain ques­tions about citizen­ship. While other cases against the Census Bureau are ongo­ing, none of them could result in such a ques­tion appear­ing this year.

With the court’s ruling, the Census — and our demo­cratic system more broadly — dodged a bullet. The citizen­ship ques­tion threatened to create extreme under­counts by depress­ing parti­cip­a­tion; its absence means that more people will feel safe filling out the Census. foot­note3_dith­b3e 3 New York v. Depart­ment of Commerce, 351 F. Supp. 3d 502, 578–83 (S.D.N.Y. Janu­ary 15, 2019) (detail­ing evid­ence that the citizen­ship ques­tion would cause a differ­en­tial decline in census self-response rates). (Reports from the field do, however, suggest that more work will need to be done to elim­in­ate completely the defeated citizen­ship ques­tion’s chilling effect.) foot­note4_gp2yq4m 4 Suzanne Gamboa, “Lati­nos, Asian Amer­ic­ans Still Fear 2020 Census over Citizen­ship Ques­tion, Witnesses Tell Congress,” NBC News, Janu­ary 9, 2020,­nos-asian-amer­ic­ans-still-fear-2020-census-over-citizen­ship-ques­tion-n1113066. By barring the ques­tion, the court blocked one major part of an appar­ent scheme to skew polit­ical power in favor of “Repub­lic­ans and non-Hispanic whites” during the upcom­ing redis­trict­ing cycle, in the words of the deceased Repub­lican redis­trict­ing consult­ant Thomas Hofeller. (Hofeller advised the admin­is­tra­tion in its push to add the ques­tion. His corres­pond­ence came to light when his daugh­ter uncovered his files after his death.) foot­note5_hhyl8au 5 Michael Wines, “Deceased G.O.P. Strategist’s Hard Drives Reveal New Details on the Census Citizen­ship Ques­tion,” New York Times, May 30, 2019,­ship-ques­tion-hofeller.html; Michael Wines, “A Census Whodunit: Why Was the Citizen­ship Ques­tion Added?” New York Times, Decem­ber 2, 2019,­ship-ques­tion-hofeller.html. In short, having no citizen­ship ques­tion increases the like­li­hood that every­one will be coun­ted correctly in 2020 and receive their fair share of polit­ical power.

Litig­a­tion may lead to more help­ful changes to the 2020 Census.

Two ongo­ing cases related to the Census Bureau’s prepared­ness could boost the Census before its comple­tion. Lawsuits filed by the National Asso­ci­ation for the Advance­ment of Colored People (NAACP) and the Center for Popu­lar Demo­cracy Action point to inad­equa­cies in the bureau’s prepar­a­tions and resources for 2020. foot­note6_tgijl3n 6 Third Amended Complaint, NAACP v. Bureau of the Census, No. 8:18-cv-891-PWG (D. Md. Janu­ary 10, 2020); Complaint, Center for Popu­lar Demo­cracy Action v. Bureau of the Census, No. 1:19-cv-10917 (S.D.N.Y. Novem­ber 26, 2019). Both suits ask the courts to require the bureau to take concrete steps to ensure that popu­la­tions that have histor­ic­ally been under­coun­ted are fully coun­ted this year. If these suits are success­ful, they could improve the accur­acy of the count by, for example, prompt­ing the bureau to expand its outreach efforts in under­coun­ted communit­ies.

Litig­ants are push­ing back against attempts to politi­cize the Census and use the count to hurt immig­rants.

Two other pending cases relate to efforts at politi­ciz­ing the national head count and wield­ing it against immig­rants and their communit­ies.

The first is a federal lawsuit that the State of Alabama has filed against the Depart­ment of Commerce and the Census Bureau seek­ing to change the popu­la­tion basis for reap­por­tion­ing seats in Congress. Alabama is asking the trial court to order the Census Bureau to exclude undoc­u­mented persons from the popu­la­tion totals used for appor­tion­ment despite the Consti­tu­tion’s clear command that every­one be included. foot­note7_hidgbf9 7 First Amended Complaint, State of Alabama v. United States Depart­ment of Commerce, No. 2:18-cv-772-RDP (N.D. Ala. Septem­ber 10, 2019). State and local govern­ments, nonprofits, private indi­vidu­als, and legal advoc­ates have all joined this lawsuit to stop Alabama’s push.

The second is an effort by La Unión Del Pueblo Entero (LUPE) and others to block Pres­id­ent Trump’s July 11, 2019, exec­ut­ive order command­ing the Census Bureau to collect citizen­ship data from other federal agen­cies and state govern­ments. LUPE’s suit argues that the exec­ut­ive order and the Commerce Depart­ment’s direct­ive imple­ment­ing it are part of a delib­er­ate, racially discrim­in­at­ory scheme to reduce Latino polit­ical repres­ent­a­tion. foot­note8_zk2i­h29 8 Amended Complaint, La Unión Del Pueblo Entero v. Ross, No. 8:19-cv-2710-GJH (D. Md. Oct. 9, 2019).

Neither case will directly affect this year’s census-taking process, but both could have signi­fic­ant implic­a­tions for how the numbers ulti­mately are used. Court rulings block­ing Alabama’s request and side­lin­ing the admin­is­tra­tion’s push for citizen­ship data would shut down two ways that the Census could be abused to under­cut the polit­ical power of immig­rant communit­ies.

The Supreme Court has long warned against polit­ical manip­u­la­tions of census numbers.

The 2020 Census could also land in court if elec­ted offi­cials or polit­ical appointees try to manip­u­late its final numbers for polit­ical or partisan advant­age. The Supreme Court has never directly addressed this issue, but for decades, justices up and down the bench have commit­ted to the view that the Consti­tu­tion requires a census free from such manip­u­la­tion.

Just this past term, Justice Stephen Breyer — joined by Justices Ruth Bader Gins­burg, Sonia Soto­mayor, and Elena Kagan — warned in the citizen­ship ques­tion case that “the Framers required an actual count of every resid­ent to ‘limit polit­ical chicanery’ and to prevent the census count from being ‘skewed for polit­ical . . . purposes.’” foot­note9_nwgudpg 9 Depart­ment of Commerce v. New York, 139 S. Ct. 2551, 2586 (2019) (Breyer, J., concur­ring in part, dissent­ing in part). Brey­er’s concur­rence was quot­ing Justice Clar­ence Thomas. In Utah v. Evans (2002), Thomas and retired Justice Anthony Kennedy asser­ted that Congress’s “prin­ciple [sic] concern” in draft­ing the Census Clause “was that the Consti­tu­tion estab­lish a stand­ard resist­ant to manip­u­la­tion.” foot­note10_dyzs­j11 10 Utah v. Evans, 536 U.S. 452, 503 (2002) (Thomas, J., concur­ring in part, dissent­ing in part). Concern about “partisan manip­u­la­tion” of the census was also a major theme of deceased Justice Antonin Scali­a’s 1999 argu­ment against permit­ting the Census Bureau to use stat­ist­ical sampling for the decen­nial head count, a posi­tion joined by Thomas, Kennedy, and former Chief Justice William Rehnquist. Scalia warned against “giv[ing] the party controlling Congress the power to distort repres­ent­a­tion in its own favor.” foot­note11_r13ac3y 11 United States House of Repres­ent­at­ives, 525 U.S. 316 at 348 (Scalia, J., concur­ring in part).

For more on pending and past cases involving the 2020 Census, visit the Bren­nan Center’s tracker “2020 Census Litig­a­tion.”

>> The self-response period for the 2020 Census opened just as the coronavirus pandemic began to wreak havoc on the patterns of daily life through­out the coun­try. The pandemic is already affect­ing the count­ing process. The Census Bureau has announced plans to tempor­ar­ily delay and modify several aspects of its oper­a­tions. foot­note12_6jecmkd 12 U.S. Census Bureau, “U.S. Census Bureau Director Steven Dilling­ham on Oper­a­tional Updates,” news release no. CB20-RTQ.08, March 18, 2020,­room/press-releases/2020/oper­a­tional-update.html. More alter­a­tions are likely, partic­u­larly if the pandemic extends into May, the bureau’s currently sched­uled time for begin­ning its nation­wide door-knock­ing oper­a­tion. Mean­while, state and local govern­ments, nonprofits, and phil­an­throp­ies are retool­ing their meth­ods for driv­ing public parti­cip­a­tion in the Census, shift­ing from in-person meet­ings to digital organ­iz­ing. It is too soon to determ­ine the ulti­mate impact that the pandemic will have on the count, but there is no ques­tion that it is placing a heavy burden on all stake­hold­ers at a partic­u­larly crucial time in the process.

>> The coronavir­us’s poten­tial long-term disrup­tions to the 2020 Census only heighten the import­ance of encour­aging self-response in the short term. Self-response will be a major determ­in­ant of how expens­ive and time-consum­ing the Census ulti­mately becomes. Door knock­ing is an extraordin­ar­ily resource- and labor-intens­ive process, even under ideal condi­tions. The bureau estim­ates that if 60.5 percent of house­holds self-respond, it will require 320,000 door knock­ers to adequately canvass hold­outs. If that rate drops to 55 percent, the bureau may have to hire an addi­tional 180,000. foot­note13_t497rz6 13 Declar­a­tion of Deborah Stem­powski, Center for Popu­lar Demo­cracy Action v. Bureau of the Census, No. 1:19-cv-10917-AKH (S.D.N.Y. Febru­ary 21, 2020), ECF No.46–1, 20–21. The surest way to limit the like­li­hood of the worst-case scen­arios is a robust campaign to promote self-response through the inter­net, phones, and the mail.

>> The Bren­nan Center will continue to track this devel­op­ing situ­ation and update the public through its 2020 Census webpage, “A Fair & Accur­ate Census.”

End Notes


The authors extend their thanks for insights, advice, and feed­back to Wendy Weiser, Alicia Bannon, Peter Miller, Yurij Rudensky, Ethan Heren­stein, Annie Lo, Jeanne Park, Zachary Laub, William O’Hare, Linda Jacob­sen, Terri Ann Lowenthal, and Ezra Rosen­berg.