• Democratic Institutions and Politics
• International Relations, Science and Security
• Markets and Methods
• Management and Leadership
• Social Policy

Kim Williams on Multiracialism and Civil Rights

An unprecedented change in racial identification in the U.S. allowed Americans for the first time to officially select multiple racial groups on the 2000 U.S. Census form. Kim Williams' research examines how these changes took place, the ongoing controversies, and the impact of multiracialism on civil rights.

Williams is an associate professor of public policy at the Kennedy School and the author of the forthcoming book, 'Mark One or More: Civil Rights in Multiracial America' (University of Michigan Press, 2006).

Q: Explain how demographic information (and U.S. Census statistics in particular) is utilized to affect and influence public policy in the U.S.

Williams: Each year, billions of dollars of federal funds are distributed to state and local governments on the basis of census counts - the largest such program is Medicaid. So the census is critical in terms of distributing federal money throughout the country.

My research focuses on racial counts. We know that racial data from the census has been used to sustain a range of exploitative arrangements throughout American history. There was the 'Three Fifths Compromise' for example, [Constitutional convention of 1787], where each black person was counted as 60 percent of a white person for the purposes of taxation and allocating seats in the House of Representatives. A 'mulatto' category appeared on various censuses from 1850 to 1920, implemented in part as a test to see whether people who were not of 'pure race' would die out faster. Later, flying in the face of confidentiality rules, the government used data on Japanese ancestry to facilitate Japanese internment during World War II.

In one way or another, for much of American history, racial data has been used as a tool of dominance, serving to separate and penalize those not defined as white.

Q: You have described how the Census Bureau's biggest problems since 1970 have revolved around race. Please give us some background into the issues.

Williams: The civil rights movement, among other things, turned the oppressive function of racial data on its head. To implement the Civil Rights Act [1964] racial data was used to identify instances of employment discrimination, for example, and to look at the racial composition of schools. The Voting Rights Act [1965] mandated that racial minorities have an equal opportunity to elect representatives of their choice. Enforcement of this act required the collection of racial data down to the level of city block. So, the irony is that, on the one hand racial data has become increasingly central in our government functions, but on the other hand, the grounds that we use for identifying racial groups have becoming increasingly suspect. That's the tension that I focus on in my research.

The Census Bureau has had a lot of problems since 1970. With redistricting and new funding hinging on census counts, communities and stakeholders have increasingly paid more attention to the census count. Poor, urban communities are consistently undercounted, so the question that the Census Bureau confronts is how to resolve that. They can employ statistical sampling techniques to try to bring the census numbers closer to the actual numbers in the population, but this practice is controversial.

When a disproportionate share of the minority population is being undercounted, mayors, stakeholders, and various groups end up complaining about the undercount. They have demanded that the Census Bureau adjust the count through statistical sampling. There have been a lot of lawsuits. The 1970, 1980, and 1990 censuses were fiascos from this standpoint. The Census Bureau is supposed to be the fact-finder of the nation and we expect objectivity and accuracy, but the bureau had been increasingly mired in lawsuits and racial politics at the same time that researchers had more of an ability to expose the problems of inaccuracy in the count.

Q: What were the intentions behind changes to the 2000 Census, and what does your research indicate are some of the results of the changes?

Williams: A movement emerged in the 1980s to add a multiracial category to the Census. People joined and formed multiracial support groups across the country and some of these people began to push for more latitude in racial reporting. My forthcoming book is partly about why and how the government eventually responded to this social movement.

In the end, instead of adding a multiracial category, the Office of Management and Budget (OMB) decided to allow Americans to mark as many racial categories as they wanted on the 2000 Census form. This 'mark one or more' option was an attempt to maintain preexisting civil rights enforcement efforts, but at the same time, to allow room for nuance.

The new policy was announced in 1997. But at that point, the OMB still had not determined how multiple race responses would be tabulated. For the next three years there was considerable debate about how the Census Bureau was going to add up all of the checks. Finally, it was decided that people would be allowed to mark as many races as they wanted and the government would tabulate to the minority group. So, if a person checked, for example, 'Asian' and 'white,' that person would be counted as Asian for civil rights purposes.

New issues came up as a result of this change. Now we can't talk about racial groups in terms of 'there are X number of black people in the United States.' We need to speak in terms of 'maximum and minimum' counts. The minimum count includes the people who just marked 'black' for example, and the maximum count includes people who marked 'black' and something else. But the difference between the maximum and minimum count for blacks, actually, is pretty small. The situation is similar for whites. So, if you think about it on a continuum, white and black Americans are pretty much still attached to traditional ideas of race as we know it in America.

But, for other racial groups, there is a real difference in identification. The difference between the maximum and the minimum count for blacks, for example, is only 4.8 percent, whereas the difference for Asians is 14 percent, and the difference for American Indians, is 40 percent. That means that 40 percent of all American Indians said they were American Indian and something else. So, if you wanted to know, for example, what the unemployment rate is for American Indians in the United States, it's questionable as to what is the denominator. Of course, in most civil rights applications, the focus is not on national level statistics, but the point is that you need a denominator if you want to talk about civil rights enforcement in sub-national contexts.

The new system solved some issues but created new ones. There's the issue of maximum and minimum counts. There's the fact that the many resulting racial combinations allowed by this new option created an exponential increase in the universe of possible racial categories. The implications remain unclear, and the circumstances invite further challenge.

Q: In both your field research into the multiracial movement and your statistical research around the 2000 Census, you made some surprising discoveries. Tell us about your findings.

Williams: It's important to understand the configuration of attitudes and positions on the multiracial category proposal. As it turned out, a lot of conservatives, including Newt Gingrich - who was Speaker of the House in the mid-90s -viewed the multiracial category as a first step to getting rid of racial categories altogether.

On the other hand, many civil rights advocates were opposed to adding a multiracial category because they were concerned it would inhibit the ability to expose discrimination where it exists. We end up with an interesting, perhaps counterintuitive, alignment on multiracial category issues in politics.

Given increasing racial diversity in the United States, the question becomes 'can we even make sense of all of this? Is it time to get rid of racial data altogether?' As a result of my research, I argue that it's not. Although there are inherent difficulties involved in the enterprise of counting by race, the alternative is not viable. I say that the data itself will tell us if racial disparities have diminished to the point that we don't need to collect racial data any more. Without racial data, what you'd end up with is racial discrimination without even a paper trail.

Still, we have data that we've never had before. As it turned out, only about 2.6 percent of all Americans identified with more than one race in 2000. So, some people came to the conclusion that the multiracial issue didn't matter at all, that it was just an insignificant little blip. But it's a big conceptual change.

How does increasing racial diversity impact civil rights thought and action? I think that we're probably going to revisit this issue of the multiracial option in conjunction with other changes that are going on. So, even though a small percentage of the population identified as multiracial in the year 2000, those numbers are likely to grow because people identifying as multiracial in the United States are disproportionately young. I haven't even discussed the issue of Latino racial identification, which complicates the picture considerably. Also, you have to keep in mind that people now are more aware of the fact that there's this possibility of identifying with more than one race. It's never really been an option before. And so I think we're going to need some new thought about how to handle these issues in a civil rights context and also in terms of how to rebut the claim that we don't need racial data any more.

Reporters:

Please contact 617-495-1115 to arrange an interview with Kim Williams.

Answers to questions submitted via e-mail:

Q: Dear Kim, how would you say your thesis related to the Katrina catastrophe in terms of the human element?

Thanks,

- Stephanie H.
Cambridge, MA

Williams: Stephanie, Black Americans represented about two-thirds of the population of New Orleans before the hurricane. We all know that Katrina had a disproportionate impact on these people. How do we know this? Well, one answer is that we saw it on television. Scores of desperate black people crowded into the Superdome or stranded on rooftops for days on end left no doubt that many of the victims were black. Surely, this tells us something important, but this is only the tip of the iceberg. We need to absorb the
underlying facts of inequality in New Orleans and in the nation as a whole. Who lived in the low-lying areas most susceptible to severe flooding? Blacks did. Who lived in high-ground areas? Whites did. Who got out? Who stayed? Race played a deadly role in who survived and who did not. Without racial data, it would be even easier to ignore this inconvenient fact.

Kim M. Williams