We’re Overstating the Importance of the Undercount

Peter Skerry

The undercount hunt is underway. But as Census 2000 enumerators knock on doors and try to coax information out of the 34 percent of the population that did not return the forms mailed out nine weeks ago, it’s worth setting the record straight on a couple of issues. First, the benefits for minorities in maximizing their census totals have been wildly exaggerated; and second, important causes of the undercount are quite beyond the bureau’s power to remedy.

There is no disputing the reality of what the bureau calls “the differential racial undercount.” According to its own estimates, the bureau missed 1.6 percent of the total population in 1990. A disproportionate number of those missed were minorities: 2.3 percent of Asian/Pacific Islanders went uncounted, 4.4 percent of blacks, 4.5 percent of American Indians and 5.0 percent of Hispanics, compared with 0.7 percent of non-Hispanic whites. And, although preliminary evidence indicates an unusually good minority response to the 2000 Census, over the previous five censuses the differential has been widening.

Understandably, minority advocates seek to increase group numbers in the hope that doing so will benefit their people fiscally and politically. But such hopes are ill-founded.

In terms of the fiscal stakes, it is usually reported that $185 billion in federal funds are distributed annually on the basis of formulas that rely on census data. True enough. Yet research by the National Academy of Sciences and the General Accounting Office consistently concludes that less than 0.5 percent of those funds are affected by the overall undercount.

This is in part because federal grant formulas are not simply based on population, but on numerous other factors, including income and age. Moreover, in programs designed to help distressed communities, a gain in population may be seen as a sign of health and can actually lead to a reduction in funding. Even when grant increases are pegged to population gains, the critical factor for a given jurisdiction is not merely its absolute population, but its population relative to other jurisdictions—a result that can obviously hurt as well as help minorities. Further, grant programs typically have funding ceilings, so larger numbers simply mean a fixed pie divided into smaller pieces.

What about the political benefits? How does the minority undercount affect the drawing of congressional districts—one of the primary purposes of the census? And would the proposed remedy of statistical sampling and adjustment actually empower minorities, as many believe?

It is widely assumed that the undercount disadvantages Democrats by disadvantaging minorities. But as Tom Hofeller, redistricting director for the Republican National Committee, once said, “The gerrymander overcometh all. What demographics give, legislatures can take away in the dead of the night.” The critical factor in each of the 50 states is not simply numbers but which party controls the redistricting process. In 1990 Democrats controlled both legislative chambers and the executive mansion in 19 states, Republicans only three. Today the situation is substantially reversed, with Republicans dominating 14 states, Democrats just 11.

When Republicans are in control, the minority undercount makes scant difference, because they have long since mastered the art of packing minorities into “majority minority” districts. While such districts make it easier for minority Democrats to get elected, they also mean the creation of more homogeneously white districts, which tend to hurt the electoral chances of non-minority Democrats and help those of Republicans. On the other hand, when majority minority districts are not an option, Republicans have been equally adept (within the constraints of the 1965 Voting Rights Act) at diluting minority numbers by dispersing them among several districts. In either case, the critical factor is not marginal differences in minority population totals, but who controls the redistricting process.

Among the states currently under Democratic control, the most important is, of course California, where Republicans will be slow to forget the redistricting battle they lost in 1980. According to conventional wisdom, minority concern over the undercount is shared by California’s Democratic leaders. But the interests of minorities have rarely been congruent with those of party leaders whose main concerns, after all, are protecting incumbents and broadening the party’s electoral appeal. But even a skillful Democratic gerrymander dedicated to minority empowerment eventually collides against the obdurate fact that undercounted minorities tend not to vote.


What can the bureau do about the minority undercount? Not surprisingly, census critics explain the causes of the undercount in terms of what they think the bureau ought to have been doing. Is there a language barrier to filling out census forms? Print them in languages other than English. Do illegal immigrants fear being reported to the Immigration and Naturalization Service? Tell la migra to back off during census season. Do respondents worry about disclosing information that might jeopardize social welfare benefits? Ensure the confidentiality of census data.

These tactics are all good ones—except that they barely scratch the surface. In 1990, the bureau sponsored dozens of experimental “alternative enumerations” by social scientists who did field research to double-check results in various hard-to-count neighborhoods. The results show the causes of the undercount to be more complex and daunting than many had supposed.

For example, virtually all of these studies indicate that among minorities distrust of government is only one factor in their reluctance to fill in forms. Another powerful factor is distrust of one’s neighbors. Koreans have a saying: “Don’t let the government know how many sons and how many cows you have.” But among Korean immigrants another cause of underreporting the number of people in apartments is their reluctance to reveal overcrowded living conditions. Similarly, some Korean women are reluctant to admit having married outside their ethnic group and therefore neglect to report non-Korean partners to the census.

Another reason for the low response to enumerators is that many minority “communities” turn out to be far less cohesive than that term implies. Crime is common enough that residents fear opening their doors to strangers, including census enumerators—even, or perhaps especially, if the enumerators share their language and background. Households are frequently complicated, with individuals coming and going constantly. In many immigrant enclaves, single men share cramped quarters with strangers, often sleeping in shifts. And even when men live with their families, the level of trust may not be high. One study of San Francisco’s Mission District describes an apartment occupied by three Salvadoran families who, while sharing the kitchen and bath, each then retreated behind the locked door of their own bedrooms where they spent their time and stored their food. In suburban Long Island another group of Salvadorans sharing quarters avoided conversing with one another for fear of reigniting hostilities from their country’s civil war.

The high level of distrust in some neighborhoods was driven home to two researchers doing an alternative enumeration in a San Francisco public housing project. Despite the fact that they were extremely well-plugged into the neighborhood, with a team reflecting its diverse cultures and languages, the researchers, who had access to the usually confidential census forms, were astounded to find that in some instances residents had been more forthcoming on the questionnaire than with the research team.

Nonetheless, the undercount is an important symbolic issue for many minorities. Many African Americans are especially sensitive about the undercount because they trace it back to the Constitution’s original stipulation that each slave be counted as three-fifths of a person. Doubtless this is why many of those concerned about the undercount frame the issue in terms of rights. Sen. Charles Schumer (D-N.Y.), for example, has asserted, “The Constitution, of course, guarantees the right of every person residing in the United States to be counted.” Indeed, some of those who are concerned about the undercount equate this “right to be counted” with the right to vote, and argue that not being counted in the census is equivalent to being disenfranchised.

Framed thus, the minority undercount presents a major challenge to political elites. To Democrats, it is one malady for which there exists a clear remedy: adjustment by means of statistical sampling. Given the seemingly intractable nature of many minority and urban problems, and the potential cost of proposed solutions such as affirmative action, increased minimum wages, greater educational expenditures and the like, adjustment seems relatively inexpensive. It is also, like redistricting, a highly technical issue unlikely to arouse much opposition among the general public. But adjustment is a complex, error-prone procedure, the risks of which greatly outweigh its presumed benefits.

Republicans are loath to appear to be against minority enfranchisement, so their arguments against adjustment have focused on statistical and legal technicalities. Their failure to address the question of whether the undercount really hurts minorities has left Republicans in their assigned role as mean-spirited opponents of racial justice.

But symbolism can be taken too far. A moment’s reflection makes it clear that there is no such thing as a right to be counted. Quite the opposite: Cooperating with the census is a legal obligation. Granted, individuals not counted in the census are not included in the one-person-one-vote calculations by which district lines are drawn. But there is nothing to prevent such citizens from voting or from going out and organizing others to vote.

Nor is being counted tantamount to being given the franchise—especially since the census enumerates many individuals, such as children and non-citizens, who are not permitted to vote. Simply being counted does not “empower” these people in any meaningful way.

Numbers are important. But more important is what people do with them. Think of the memorable formulation of Rep. Barney Frank (D-Mass.): “How many members of the NRA are there? I don’t know. I don’t think my colleagues know. What’s important politically is not how many there are, but what you do about it. The extent to which you mobilize enormously outweighs the numbers.” Population totals simply don’t translate automatically, without group organization or effort, into political power. Yet rather than confront the genuine social and political problems of the disadvantaged, we have all but convinced ourselves that tweaking population totals at the margins will result in minority empowerment.

Peter Skerry teaches at Claremont-McKenna College in California and is a senior fellow at the Brookings Institution, which has just published his book “Counting on the Census? Race, Group Identity and the Evasion of Politics.”