Counting on the Census? Race, Group Identity, and the Evasion of Politics
by Peter Skerry
Brookings Institution. 261 pp. $25.95
It is easy to see why people get worked up about abortion, the death penalty, and tax hikes. But it seems odd that our constitutionally mandated decennial census—a subject so steeped in complex statistical formulas and quantitative analysis that the average educated American cannot begin to understand it—should inspire such passionate controversy. And yet it does, with Democrats and Republicans fighting once again, in the midst of an election season, over the thoroughly arcane subject of “adjustment.”
The central issue, of course, is who gets counted, and how. According to critics of the current system, minorities and the disadvantaged are being left out because of an “undercount,” as a result of which they may be denied their proper representation in government or their fair share of public money. To help us understand what is at stake here, Peter Skerry, a political scientist at Claremont McKenna College, has written Counting on the Census? The author previously of Mexican Americans: The Ambivalent Minority (1993), Skerry provides a skillful and very timely history of the various controversies involving the census and then lays out, and attempts to assess, the contemporary debate.
_____________
The question of how to count the U.S. population reaches back, Skerry reminds us, to the Founding, and specifically to Article I, Section 2 of the Constitution. This requires that, for purposes of apportioning representatives, an enumeration be made “within three years after the first meeting of the Congress of the United States and within every subsequent term often years.” But in (a subsequently repealed passage in) the same article, the Constitution also mandated that the count include not only citizens but also three-fifths of “other persons” (read: slaves) within U.S. borders. From the very first census in 1790, then, it was clear that this exercise was about more than just counting Americans. It entailed counting different kinds of Americans.
Fast-forward, now, to the modern period, when, as Skerry documents, fresh conflict began with the proposal that a question on religious background be included in the 1960 count. Although the move was supported by Catholic organizations, opposition on First Amendment grounds caused the question to be dropped. Yet while religion to this day has remained taboo, race and ethnicity have, in terms of political significance, become the central issues in the census.
The first sign of trouble to come occurred when the Nixon administration, responding to pressure, and over the objections of the census director, ordered that a question be added for Hispanic origin. (In addition to blacks, Native Americans had already been specified as a separate group in the 19th century, and “Orientals,” meaning Chinese, at the beginning of the 20th.) Then came another new designation, Asian or Pacific Islander (API). Once these broad new categories were in place, much political maneuvering ensued to determine smaller classifications within them. In preparing the 1990 count, for example, the Census Bureau wanted to require that people identifying themselves as Asian or Pacific Islander indicate one of seven subgroups. At the behest of API activists, Congress passed a bill increasing the number of subgroups to nine. This prompted a pocket veto by President Reagan—but the bureau decided to follow Congress’s recommendation anyway.
_____________
Over the last 30 years, minority interest groups have increasingly attempted to influence the census process, with the express purpose of combating an “undercount” which (they say) afflicts them disproportionately Skerry unravels the strands of this claim. First of all, even with the best will in the world, a perfectly accurate count is not humanly achievable. As things are now organized, the Census Bureau sends written questionnaires to every household in the country. If there is no response within a few weeks, enumerators are sent to conduct the survey in person. But inner-city residents, whose ranks are disproportionately made up of minorities, tend to move frequently, making them hard to locate, and heads of household filling out the survey will often omit members who are frequently absent.
There are other problems as well. Language and cultural barriers render it difficult to achieve a full count in some neighborhoods. Recent immigrants are often incapable of filling out the forms, and even among those fluent in English, many misunderstand the purpose of the census, believing they might get into trouble with their landlords if they report too many residents in a particular dwelling, or lose government benefits if they report too few. (Since census data remain entirely anonymous, neither is possible.)
Under such circumstances, an undercount of some size is, Skerry writes, inevitable. But, given the scope of the project, the actual undercount is amazingly small, amounting to between 1 and 2 percent of the total population. (In “academic and private-sector research,” a former census director has written, “70-90 percent response rates are cause for celebration.”) Worse, the remedy proposed for this rather minuscule discrepancy—namely, “adjustment”—is no remedy at all.
_____________
Adjustment is accomplished by sampling a small percentage of the original census population and then extrapolating final numbers from the specimen. But sampling, Skerry shows, is vulnerable to at least as many kinds of inaccuracy as is the undercount it is designed to correct Not only will it be hampered by the same procedural difficulties (high mobility rates, people not wishing to be counted) as the census itself, but workers administering their postenumeration survey will find that many residents of urban neighborhoods are reluctant to open their doors to strangers. In addition, the sheer awareness that an adjustment process will be taking place may reduce the inclination to fill out a census form in the first place.
If census adjustment results in data no more accurate and possibly less accurate than those in the census proper, the amount of money at stake in government resources like block grants also turns out to be quite small. Even in New York City, where a high undercount can be expected, an adjustment of the 1990 census would have resulted in the city’s gaining $10 million—“a nice piece of change,” as Skerry quotes one economist, “but hardly enough to dent the city’s budget for filling potholes.”
As for the issue of political representation, the Supreme Court has already ruled that adjustment cannot be used for legislative apportionment. And even if it were so used, Skerry argues, no one could predict what the resultant electoral pattern would look like (since redistricting depends largely on the partisan makeup of state governments) or who would benefit from it.
_____________
Skerry’s Methodical analysis and moderate conclusions—yes, there is a minority undercount, but it does not present a large problem and it cannot be fixed with sampling and adjustment—are a welcome contribution to the current debate. His eye for the telling detail—“anywhere from two-thirds to three-fourths of the 1990 undercount,” he notes quietly at one point, “were nonblacks, mostly whites”—enables him to cut through much racial propaganda, and his thorough command both of the statistical issues themselves and of their implications for public policy is refreshing, to say the least. If facts and commonsense arguments were decisive, this book would put an end to the census controversy.
Alas, facts and arguments notwithstanding, there is no sign that ethnic politicians or interest-group leaders are about to let go of the issue of the undercount any time soon. Which leaves the rest of us to ponder the long-term harm being done by our current approach to counting Americans. For one thing, the public has been given the distinct impression—including by the Census Bureau itself—that the major benefit that accrues to individuals from filling out the questionnaire is, precisely, a bigger slice of the federal pie. That, indeed, was the not so subtle message of the prime-time television ads run by the bureau this past spring, which bore the tagline: “It’s your future. Don’t leave it blank.”
For another thing, in asking citizens to identify themselves by race, the government implicitly encourages us, especially in today’s polarized climate, to think of ourselves as separated by race. As Lawrence Wright observed in a much-discussed 1994 New Yorker article, “The use of racial statistics . . . creates a reality of racial divisions which then require solutions, such as busing, affirmative action, and multicultural education, all of which are bound to fail, because they heighten the racial awareness that leads to contention.”
Skerry dismisses such worries as “alarmist,” observing that even those who voice them—he cites the economist Thomas Sowell—rely on census data in their research. But using these data for research purposes is hardly the same thing as endorsing the way they are collected, let alone the way they are deployed by others. The fact is that leaders of ethnic interest groups do now routinely encourage their members to fill out the census forms so that public officials and candidates will be aware just how loud—and how uniform—their voice can be. On the same grounds, some black spokesmen oppose any move to allow people to describe themselves as “mixed-race” on the questionnaire, lest it weaken black influence.
Manipulative politics of this kind is almost enough to make one wish that, once slavery was abolished in America, racial questions had been stricken from the census altogether. Just as answering census questions in order to get more federal money skews the way we think about the role of government in our lives, the use of the census as a mechanism of racial identification and politicking further weakens our already frayed civic unity—and undermines the legitimate and necessary purposes of the census itself. Such, at any rate, are some of the troubling thoughts that arise from reading Peter Skerry’s sorely needed book.
_____________
