Multiracials, intermarriage, ethnicity.
If a child has a white mother and a black father, then the child is racially…what? Deciding how the next census, to be held in the year 2000, should handle the “multiracial” child is a hot topic this year – from front page stories in the New York Times, to disquisitions on the origins of Tiger Woods, to congressional hearings. There are two bizarre features to the way the issue is being treated, and they tell us much about the state of American thinking about race. First, there has been virtually no discussion of the fact that racial intermarriage is a form of ethnic intermarriage – intermarriage across those ethnic lines we choose to call race lines. And the Census Bureau knows how to count the offspring of ethnic intermarriages: It has done so for over a century. Second, neither the Bureau nor the media have been drawing the important connection between this topic (how to count the “multiracial” child) and another topic that has also been receiving front-page attention for several years, namely, the Census Bureau’s regularly updated projections of the future racial composition of the American people. A recent projection, for example, was reported in the New York Times on March 14, 1996 also on page 1. These projections of racial composition pivot on expectations for intermarriage in the future, and the expectations are strange indeed.
How “multiracial” children will be counted in the next census is receiving attention now because the decision is needed soon, because such children have found a voice and because the decision might affect race-based laws and politics. Specifically, many people concerned with civil rights enforcement or with the power of racial minorities do not want to add messy complexities to the struggle for racial equality. It is hard enough to achieve programs against discrimination (or in favor of affirmative action) today without considering how such programs should treat mixedrace individuals. Similarly, the future racial composition of the country is a hot-button issue. Telling Americans that their country will be more than half nonwhite by the middle of the next century – what Time called “the Browning of America” some years ago – stimulates different reactions. To some, the demographic projection says America had better wake up to the needs of its “minorities”: They are soon to be its majority. To others, the projection says America had better restrict immigration to avoid reaching a “nonwhite majority.” In fact, by the middle of the next century America will very likely be “more than 50 percent nonwhite” and also “more than 50 percent white” by the Bureau’s logic.
Ethnic Intermarriage – American as Apple Pie
American history would be unrecognizable without ethnic intermarriage. From colonial times to the present, immigrants typically married their own, the second generation did so much less consistently, and the third generation still less consistently; probably a majority of third-generation members have married people of other ethnic origins than their own. And by the fourth and fifth generations – who even kept track? The evidence for the history of ethnic intermarriage is about as overwhelming and unambiguous as any generalization about the American population can be: from de Crevecoeur in the eighteenth century noticing “this new man, the American” arising out of various European immigrant stocks to the data from census after census in the twentieth century.
Intermarriage operated most fully among the descendants of European groups (Italians, Poles, Irish, English, Germans, Scandinavians, etc.); it was crucial to the making of “Americans” out of the descendants of “hyphenated Americans.” Intermarriage was least prevalent, perhaps, across the black-white divide – not least because such intermarriage was so often illegal. Notwithstanding the lower prevalence of black-white intermarriage, there has been a surprisingly high degree of intermarriage across lines one might not have expected: between “white” and “red” for example. Huge proportions of people descended from “American Indians” tell us that they are part “American Indian” (or “Native American”) and part “white.”
Today black-white intermarriage is growing in importance – especially among younger, middle-class, college-educated blacks; nevertheless, it remains the one divide across which intermarriage has remained strikingly less common than elsewhere. If “multiracial” today meant only the offspring of black-white marriages, it would be less of an issue (although a growing issue nonetheless, given changing consciousness). However, “multiracials” also include rapidly growing numbers of children in marriages of Asians (immigrants or their children, for example) and non-Asians; and the way the counting goes, it can also be interpreted to include Hispanics and non-Hispanics.
Well, then, how has the Census Bureau handled the offspring of other ethnic intermarriages for the past century – those ethnic intermarriages that have not crossed racial lines? When the Census Bureau asks a native-born child of foreign-born parents where the parents were born, the answer is often that they were born in two different countries. No problem; the bureau records two countries of origin. Both parents born in Italy? Fine. One born in Italy, one in Poland? Also fine (the question about parents’ birthplaces was dropped in the most recent censuses but will probably reappear by census 2000).
There is an even stronger model for the Bureau’s “multiracial problem” in the census ancestry question. Lately, in the 1980 and 1990 censuses, the Bureau has also asked each individual with what ethnic ancestry he or she identified – and ethnic ancestry means a place on the globe: Italy, Poland, etc. The Bureau introduced this ethnic ancestry question in order to allow Americans to state an ethnic affiliation even if one’s immigrant ancestors had come to the United States from the country of ancestry many generations back. Now the crucial point is that on the ancestry question, the Census Bureau instructs Americans that they can identify themselves with more than one ethnic ancestry. Many millions of Americans have taken the trouble to list two ethnic ancestries, millions more list three ancestries. Yet even these multiple listings are only a partial indicator of the real complexity of American ancestry, since many people are descended from far too many different ancestries to list them all (or even to know them all). In 1980, “English” was listed before “German” in the Bureau’s examples of ancestry; in 1990 the ordering was reversed. As a result, the percentage claiming identity with English ancestry declined by a large fraction, and the number claiming identity with German ancestry rose by a comparable amount. There are many such examples, but even these examples of confusion show us something important, namely, the long-term results of population mixing and the attenuation of connections with the origins of ancestors (whom one never knew). In short, keeping track of American ancestries at the Bureau eventually gets messy because of intermarriage patterns – and that is as it should be. A simple answer would be a false answer.
What has all this got to do with racial mixing today and with the American racial profile tomorrow? It has everything to do with those topics – once we understand racial intermarriage as a variant of ethnic intermarriage. Whatever meaning race has for anthropologists and biologists today – and that is probably no meaning – race is thought of as different from ethnicity by most Americans. And in some sense that is a good thing. One useful way to think about racial differences is simply that these are ethnic differences that were treated in very distinct ways in the American past (and to some extent are still treated so today). For that reason, if we want to understand problems like American economic inequality we cannot ignore people’s racial origins; to blithely throw out race classifications in our present censuses would not be smart or fair.
Nevertheless, as barriers begin to crumble, ethnic groups that were especially stigmatized in the past are increasingly able to partake of the American mainstream. Among other things, this change means that they intermarry more often with members of other ethnic groups. We see this happening on a large scale among Asian immigrants today, among American Indians and Hispanics past and present, and finally, although still to a lesser extent, even among blacks today. Will intermarriage characterize the descendants of “new” immigrant groups, that is, Hispanics and Asian immigrants, as fully as (or more fully than) it has characterized the descendants of Europeans? And how much will black-white intermarriage increase? We need not know the answers to these questions in order to notice that intermarriage rates for the children of new immigrants (mostly, as Time had it, “brown”), are already high and cannot be ignored, and that black-white intermarriage is now a meaningful social reality.
So How to Count?
We must think about the children of interracial marriages in the same way that we think about the children of other interethnic marriages. The easiest way to do that is to let people identify themselves in terms of race just as we let them identify themselves in terms of ethnic ancestry. At the moment, the census ancestry question encourages respondents to list more than one ancestry, and the census race question tells Americans that they cannot belong to more than one race; but they can and they do. This contradictory position must be changed: People must be told explicitly that they can list themselves as members of more than one race. To create a new category, “multiracial” – a key proposal being debated today – is better than no change, but it is surely the wrong solution. The creation of a “multiracial” racial category was sensibly rejected by the government’s task force on racial classification, and the rejection should not be overturned. Such a category would add nothing to the information conveyed when a person declares his or her origins to be, for example, Asian and white. It would not only be redundant to have the category, it would be inappropriate in implying that somehow being multiracial deserves to be treated differently than being multiethnic. A person reports as, for example, Polish and Italian in terms of ancestries (or parents’ birthplaces) – not as multiethnic.
If we do let people identify with more than one race, counts will surely become more complicated than they are now, but they will be reasonably straightforward for as long as they are worth having, and if they become as confused, on occasion, as the ethnic ancestry data sometimes becomes, that will be good to know too. How such counts of multiracial people will be used in connection with enforcing civil-rights legislation (including affirmative action and voting rights laws) will have to be worked out (and the counts issue can be worked out). But counting the racially mixed in connection with laws that rely on race counts is going to be an issue in the coming years no matter what the Census Bureau does about the multirace category. Keeping track of the magnitude of groups over time will become messier and will cause technical complexities. But those complexities reflect the experience of the groups; to avoid those complexities is to deny that a fundamental force in American life operates among “racial minorities” as well as among other Americans.
In 2050 Will America Still Be White?
The branch of the Census Bureau that carries out some useful projections – of age and gender in particular – somehow got mandated to do racial projections; and as a result, dedicated and discerning demographers were linked with a sad charade. For the racial projections are based on a disastrous assumption. These projections assume that there will be no further intermixing of peoples across racial lines. What is more, in constructing the racial categories used in the Bureau’s racial projections, Hispanics are also treated as a kind of race (presumably in order to obtain “pure” counts of non-Hispanic whites). Similarly, there has been a strong temptation to define Hispanics as one of the basic “race” categories on the Census, not least because so many Hispanics list themselves as “other” rather than as white or black. The “race” categories would then be called “race and Hispanic origin” categories; that change in label would be but a weak barrier against the tendency to foolishly blur the difference between Hispanicity and racial origin.
How precisely does the projection of future racial composition deny future intermarriage? How does it confront the objection that the American historical experience implies a future of massive ethnic intermarriage, that today’s rates are high and rising among the children of recent Asian and Hispanic immigrants, and that intermarriage is rising even among blacks (where it has been lowest)? The Bureau’s projections assume that all such dynamics will cease. There will be no further “racial” intermarriage after the day of the projection. Every mother’s descendants will have her “racial” identity for all time. If an Asian American woman and a non-Hispanic white man marry today, the Bureau projects that all their descendants in the year 2050 will be “Asian,” and will only be “Asian.” If two immigrants arrive from Guatemala today the bureau projects that all their descendants will marry only Hispanics – through 2050 and beyond. Such assumptions are wonderfully simplifying and politically have some short-term use to a few interest groups. But they are ludicrous – or would be if they were not taken seriously and did not contort our view of where we are.
Ironically, any realistic assumption about intermarriage levels in the future implies both more and less ethnic transformation in America than the projections suggest. If the descendants of Guatemalans intermarry with non-Hispanics, it means that many more people will have some Hispanic ancestry by 2050 than would be the case if the descendants of Guatemalans married only other Hispanics. And yet, at the same time, many of these descendants will only be 1/4 or 1/8 “Hispanic.” They will also be, very likely, part “non-Hispanic white” in their origins.
So how should such descendants be counted for projections? How would they count themselves? The answer is we don’t know, and more to the point, to force our ethnic categories on the year 2050 is foolish. We might as well classify the contemporary descendants of turn-of-the century immigrants by the racial terms “Nordic,” “Alpine,” and “Mediterranean” that were used to classify those immigrants in 1910 – and at the same time we would also need to deny that so many of these present-day descendants could trace their origins to all three categories. We don’t really need long-term racial projections at all – racial projections going more than a decade or two into the future. But if we are going to have such projections, they must be based on a much fuller recognition that intermarriage will be prevalent among our descendants. Recognizing that fact will make racial data increasingly messy in future decades – as it should. Not recognizing that fact will make racial data increasingly meaningless.
But a decline in the quality of data is not the most severe outcome we can expect if we deny their mingling and treat them differently than other ethnic groups in his regard. The greater danger is the perpetuation and strengthening of a barely articulated sense underlying the present way of counting “races”: that races live their lives in isolation from each other, that they must be counted rather as members of different species might be counted. The Census Bureau does not just count; in choosing what to count and how to count, the Bureau is in danger of preserving barriers that would otherwise not be so high or so foolishly placed.
Joel Perlmann is a Senior Scholar at the Jerome Levy Economics Institute at Bard College, and Institute Research Professor of History at the College. He is the author of Ethnic Differences and (most recently) of “‘Multiracials’, Racial Classification, and American Intermarriage – the Public it Interest” (Levy Institute Working Papers, no. 195).