Color Crit: Critical Race Theory and the History and Future of Colorism in the United States

Abstract

Critical race theory teaches that racism and racial inequality are constants in American society that stand outside of the prejudices of individuals. It argues that structures and institutions are primarily responsible for the maintenance of racial inequality. However, critical race theorists have neglected to formally examine and theorize colorism, a primary offshoot of racial domination. Although studies of colorism have become increasingly common, they lack a unifying theoretical framework, opting to lean on ideas about prejudice and preference to explain the advantages lighter skinned, Black Americans are afforded relative to darker skinned Black Americans. In this study, I deploy a critical race framework to push back against preference as the only, or primary, mechanism facilitating skin tone stratification. Instead, I use historical Census data and regression analysis to explore the historical role of color-based marriage selection on concentrating economic advantage among lighter skinned Black Americans. I then discuss the policy and legal implications of developing a structural view of colorism and skin tone stratification in the United States and the broader implications for how we conceptualize race in this country.

Keywords

colorism critical race theory marriage race skin tone stratification

Introduction

Two central tenets of critical race theory are the intractability of racism and racial inequality and the structural view of racism that displaces individual prejudice as the primary mechanism facilitating inequality (Bell, 1991; Ray, Randolph, Underhill, & Luke, 2017; Ray & Seamster, 2016). They argue racism is woven into the fabric of American life and its institutions in ways that ensure White people dominate the racial hierarchy. Since its inception three decades ago, critical race theory has led to a slow revolution in how most social scientists conceptualize racism and racial inequality. The frames of structural racism that were pioneered and formalized in writings by critical race theorists have been adapted and adopted by a wide variety of social scientists and have arguably been woven into the fabric of mainstream sociology—in name, if not in practice. Sociologists, even those who may not readily identify as critical race theorists, generally acknowledge that racial inequality is the result of more than the collective pathology of individuals, that there is a “structure” in place—comprised of institutions, policies, and norms—that reproduces and maintains inequality (Bonilla-Silva, 1997; Omi & Winant, 2015). Moreover, recent appeals to unify traditional critical race theory and social scientific methods, a call that has come to be known as e-CRT (empirical methods and critical race theory), have strengthened the relationship between the two groups (Barnes, 2016; Gomez, 2012; Obasogie, 2013). Social scientists have heeded this call in a variety of ways, from pushing back against racial progress narratives (Bonilla-Silva, 2017; Ray & Seamster, 2016) to exploring the messiness of measuring the concept of “race” at all (Roth, 2017). However, merging critical race theory and certain forms of racialized inequality has proven more difficult. Particularly, the push to prioritize structural explanations has yet to extend to our conceptualizations of colorism or skin tone/shade stratification, and lacking such structural explanations our ability to combat the perniciousness of the phenomenon are limited to awareness-raising and educational efforts (Maddox & Perry, 2017). Conversely, a structural grounding to colorism would offer more concrete solutions to combatting it.

Even though our understanding of the far-reaching consequences of colorism or skin tone/shade stratification has exploded in recent years with the inclusion of skin tone variables in an increasing number of social science surveys, we have yet to truly explore the mechanisms responsible for the perpetuation of the phenomenon. The overwhelming majority of studies examining the extent of colorism tend to lean on simple explanations revolving around the historical preference for light skin—and other “Eurocentric” features such as thin lips, noses, and straight hair—because of their perceived proximity to whiteness and eugenic connection to multiracial slaves. An explanation of the larger forces facilitating color inequality has eluded researchers. Perhaps the most notable oversight is an examination of how color inequality has maintained its prominence over time. If indeed color stratification reaches back to antebellum slavery (Bodenhorn, 2002; Bodenhorn & Ruebeck, 2007; Norwood & Foreman, 2014; Reece, 2018b), the idea that a collection of individual preferences is primarily responsible for its continuation seems woefully inadequate in light of how we have come to view stratification. Critical race theorists argue that racialized inequality remains robust to changes in the racial structure and the trajectory of color inequality in the United States has certainly followed that trend. Lighter skinned Black Americans have, at various points, been viewed as an independent and superior racial group, mongrel and deficient racial group, and collapsed into a collective Black based on the country’s circumstances (Frazier, 1933; Gross, 1998; Hochschild & Powell, 2008; Toplin, 1979). Yet, at each point, despite shifting external perceptions, census categories, and self-identifications, they maintained their advantage, demonstrating the malleability of race and its ability to recalibrate. A structural theory of colorism will help explain the intractability and adaptability of color inequality.

This study reaches back to the 19th century to offer and test an explanation for that intractability. Specifically, I posit that color-driven marriage selection may have been one of the forces maintaining color stratification intergenerationally. I use historical Census data to test the initial stages of this explanation—the premise that light skinned Black women were able to leverage their favorable skin tone in exchange for more affluent marriage partners than their dark skinned counterparts. Ultimately, my results support this idea. Light skinned women fared better in the marriage market than dark skinned women. This allows us to begin to further conceptualize the structural processes maintaining color stratification over time.

Developing a Critical Theory of Colorism

Colorism is the process by which people of color—in this case Black Americans—are awarded advantages based on their phenotypical proximity to whiteness (Reece, 2016). That is, Black people who look more stereotypically White—lighter skin, thinner noses, thinner lips, straighter hair, lighter eyes, etc.—tend to be privileged relative to those who look more stereotypically Black—darker skin, thicker noses and lips, tightly coiled hair. The effects of colorism are almost ubiquitous, with lighter skinned Black Americans benefiting from higher wages (Goldsmith, Hamilton, & Darity, 2006, 2007), more education (Branigan et al., 2013; Monk, 2014), better mental and physical health (Monk, 2015), and lower conviction rates and shorter prison sentences (Blair, Judd, & Chapleau, 2004; Viglione, Hannon, & DeFina, 2011). They are also considered more attractive (Reece, 2016), and are disciplined less and more mildly in school (Blake, Keith, Luo, Le, & Salter, 2017; Hannon, DeFina, & Bruch, 2013).¹

However, our increasing awareness of the widespread consequences of colorism has not lead to an equivalent discussion of the mechanisms shaping the process or a comprehensive theory that attempts to explain the workings of the phenomenon. This leads to a massive gap in our understanding of the workings of colorism and severely limits our ability to develop strategies to combat it.

Most studies of colorism rely on models that prioritize preference and prejudice. Indeed, the primary attempt to build a strong theory of colorism is called the “preference for whiteness” thesis and uses a combination of psychology and anthropology to argue that White people simply prefer people who look similar to them (Goldsmith et al., 2007; Painter, Holmes, & Bateman, 2015). The authors argue that widespread individual color-based discrimination by White people results in the large color-based disparities we observe. And while there is definitely evidence that supports a prejudice-based framework—such as prison sentencing (Blair et al., 2004; Viglione et al., 2011)—other types of color and skin tone stratification lend themselves to, at least, a combination of prejudice and structural explanations.

The structural and legal perspectives of critical race theory lend themselves to policy solutions, not necessarily to defeat racial inequality, but to combat it. Organizations such as the Equal Employment Opportunity Commission (EEOC) and the judicial system, regardless of their inconsistency in rulings on race, recognize racial inequality as more than the cumulative result of racial prejudice. Generally, they understand that racial inequality is part of a system of historical advantages maintained by a network of ostensibly race neutral policies. No such understanding exists for colorism. Indeed, most lawmakers seem unaware that such a phenomenon exists at all. This means that in litigation involving claims of color discrimination, the burden of proving intent seems to stand even taller than in cases involving racial discrimination (Banks, 2009). So even though the EEOC has seen a spike in cases involving color discrimination, courts still struggle to resolve them and no policy exists that attempts to minimize color stratification (Brown, 1999; Hall & Johnson, 2014; Jones, 2009).

Lack of explicit color-based policy allows racial progress to mask ongoing—perhaps even increasing—color stratification (Banks, 2014; Norwood & Foreman, 2014). It is possible that the gains Black people have made over the decades, particularly since the Civil Rights Act of 1964, have been concentrated among lighter-skinned Black Americans, whereas their darker skinned counterparts have continued to fall behind. This would signal more than just prejudice but that lighter skinned Black Americans were also more poised to take advantage of the new opportunities opened to them by policies that spared them from explicit racial discrimination. Without policies that account for color, that schism will remain unaccounted for, dooming darker skinned Black Americans to a permanent place at the bottom of the social hierarchy while traditional metrics extol the upward mobility of Black people. However, without an understanding of colorism as a structural process, we cannot develop effective interventions that ensure dark skinned Black Americans raise their social standing.

The key to understanding the social structures facilitating colorism lies in its history, which I plan to explore with the historical legacy thesis (Conley, 2001). Essentially, the core of the historical legacy thesis is that contemporary inequality is at least partially the result of a historical head start by some groups—in this case light skinned Black Americans—coupled with the historical deprivation of others—in this case dark skinned Black Americans. However, when investigating a historical legacy, it is important not only to establish a connection between the past and present but to investigate the processes that have allowed the dominant group to maintain their head start over the course of history.

The History of Colorism: Racial Reorganization and the Maintenance of the Color Hierarchy

For a time our racial structure was more similar to that of Latin America with a sizable mulatto—mixed Black and White—group sitting as a subset of Black but in many ways between Black and White people. Not only did mulattos sometimes go to great lengths to separate themselves from Black people, White people also recognized their utility as a buffer class, a shield from Black people, particularly during the antebellum years (Gullickson, 2010; Toplin, 1979). A legislative report investigating a planned slave revolt in the early 1820s confirms this sentiment. It reads,

Free mulattos are a barrier between our own color and that of the black and in cases of insurrection are more likely to enlist themselves under the banners of the whites. . .Most of them are industrious, sober, hardworking mechanics, who have large families and considerable property; and so far as we are acquainted with their temper and dispositions of their feelings, abhor the idea of association with the blacks in any enterprise. . . (as cited in Jones, 2000, pp. 1508-1509).

Mulattos were often perceived as better than Black people, both materially and ideologically. Indeed, on the latter, they may have been perceived to be closer to Whites, and the Census made this distinction between Blacks and mulattos official.

The Census added “Mulatto” as an official racial/color category alongside Black in 1850, where it remained until 1920 (Hickman, 1997; Hochschild & Powell, 2008). However, Census mailers did not exist until 1960, so, before then, Census enumerators traveled the country completing Census forms for households (Bhatnagar, 2007). When counting race, they would refer to these instructions:

Be particularly careful to distinguish between blacks, mulattos, quadroons, and octoroons. The word “black” should be used to describe those persons who have three-fourths or more black blood; “mulatto,” those persons who have three-eighths to five-eighths black blood; “quadroon,” those persons who have one-fourth black blood; and “octoroon,” those persons who have one-eighth or any trace of black blood (as cited in Hickman, 1997).

While this may seem as if census enumerators collected detailed family histories, the truth is they were not given clear instructions on how to determine racial classification, and in an era of eugenics they often assumed they could determine a person’s racial ancestry based on physical appearance in combination with local customs (Gross, 1998). Although, there was some mobility between the mulatto and Black categories as the same person could be identified as mulatto or Black from census to census the mulatto category remained a reliable proxy for “color,” for skin tone, and the social differences were real even if the ancestry differences were not (Saperstein & Gullickson, 2013).

Mulattos were advantaged relative to “regular” Black people in a number of arenas. They had more wealth (Bodenhorn & Ruebeck, 2007), more successful businesses (Schweninger, 1989, 1999), better jobs (Gullickson, 2010; Saperstein & Gullickson, 2013), they lived longer (Green & Hamilton, 2013), they had bigger families (Frazier, 1933), and during slavery they were more likely to be manumitted (Bodenhorn, 2011). Disproportionate manumission created stark differences in the color of slaves and the free Black population. By 1860, 41% of free southern Blacks were mulatto. In contrast, only about 10% of slaves were mulatto. In the Deep South, over 75% of free Blacks were mulatto and only 9% of slaves were mulatto. Many places had no mulatto slaves at all (Reece, 2017). Indeed, in some places, the contrast between the mulatto-free population and the Black slave population was so apparent that “free black” and “mulatto” almost became synonymous (Berlin, 1974). Even when Black people managed to escape bondage, their economic affluence was dwarfed by their mulatto counterparts who had roughly 3 times the average wealth (Reece, 2017).

Yet the mulatto category slowly fell out of favor. During the antebellum years, much of the work of policing the so-called “color line” was done by slavery. To be Black was to be a slave, to be White was to be free. This distinction hardened the status hierarchy between the two groups with mulattos straddling the line in between. But after emancipation and during reconstruction as Black people gained political power, White people, particularly Southern White people, sought to solidify their position at the top of the racial hierarchy and pushed the country toward the one-drop rule (Washington, 2011).

This transition to one-drop rule placed mulattos in a precarious position. They were the primary targets of the racial recalibration. In the absence of slavery to control the Black population, White people needed stronger racial boundaries, and they had become increasingly uneasy with the ambiguity of determining who was and was not Black. In response, they sought to aggregate the mulatto and Black groups in attempts to enforce growing Jim Crow segregation. This came to the fore in the Supreme Court case Plessy v. Ferguson, which established that even Homer Plessy, a very light-skinned multiracial man, was not exempt from segregation laws (Washington, 2011). Moreover, Plessy v. Ferguson also left segregation to state law, which meant the question of racial classification also fell to the states, who adopted a variety of laws governing who was Black based on percentages of Black “blood” or black racial ancestry (Brown, 2014). The most extreme of the laws dictated that a person was Black if they exhibited any trace of Black blood whatsoever. Mulattos were ostensibly losing the formality of their privileged position and being forced down the rungs of the racial structure. Racial reorganization threatened to strip away the advantages they had accrued.

Following Plessy, this racial reorganization quickened, threatening mulattos coveted status. By 1910, the one-drop rule had become a legal mainstay as Tennessee adopted the first legislation establishing a person as Black if they had any Black racial ancestry (Brown, 2014). Other states followed in the coming years and decades: Mississippi in 1917, North Carolina in 1923, Virginia in 1924, Alabama and Georgia in 1927, and Oklahoma in 1931. Still others amended their existing legislation to reduce the percentage of “black blood” necessary to be legally Black to 1/32 or 1/16, and by 1930, the mulatto category was dropped from the census, signaling the official end of the triracial system and the formal federal adoption of the one-drop rule. Americans would not have the option to select multiple racial categories on the census again for almost another century.

Mulattos had been enveloped by a collective Black, apparently the victims of a racial project designed to preserve the purity of whiteness at the expense of their former allies. However, as the racial reorganization settled, former mulattos, now occupying a different space in the racial structure as light skinned Black people, emerged with many of the same advantages they enjoyed before. As legal apparatuses worked to produce lasting changes in the racial structure, the racial and color hierarchy persisted, virtually unchanged. Lighter skinned Black Americans continued to enjoy the privileges they gained over the years. So, when the racial structure was reorganized again with the victories of the Civil Rights Movement, they were prepared to seize opportunities and rapidly elevate their status in ways their darker skinned counterparts were not. Ultimately, this meant despite multiple attempts to disrupt the racial and color hierarchy, first through Emancipation and Reconstruction, next through the development of the one-drop rule and the proliferation of Jim Crow, and finally through the Civil Rights Movement, light-skinned Black Americans managed to maintain their position.

This study seeks to leverage the ideas of the historical legacy thesis—that early mulatto advantage feeds contemporary light skinned advantage—to explore the processes that allowed mullatos/light-skinned Black Americans to preserve their economic advantage despite constant reorganization of the racial structure. To do that I examine the role of historical marriage selection in allowing light skinned Black Americans to concentrate economic advantage.

Color and Marriage Homogamy

Marriage is an important mechanism for transmitting social advantage. It offers intergenerational security, and family structure has a strong effect on the life chances of children (Pfeffer & Killewald, 2015). Much of this is the result of the greater relative affluence of married families. Married couples amass more wealth, which ultimately benefits their children, inheritances positively influence wealth accumulation, and affluent families offer their children benefits early in life such as better education (Pfeffer & Killewald, 2015).

There is evidence of color homogamy, both in the 1800s and today (Bodenhorn, 2015; Monk, 2014). That is, lighter skinned people tend to have lighter skinned partners and darker skinned people tend to have darker skinned partners, which has remained consistent throughout racial reorganization across the 20th century. Moreover, light-skinned women are more likely to marry under some circumstances today (Hamilton, Goldsmith, & William, 2009).

Although research already shows that 19th-century families headed by two mulatto partners were wealthier than those with two Black partners or a Black and mulatto partner, this research does not examine the partner selection process (Bodenhorn & Ruebeck, 2007). Partner selection is important because it demonstrates that color can be exchanged for economic status that is subsequently transferable to the next generation rather than economic affluence resulting from an accident of color homogamy. Additionally, a person’s chances of even marrying may vary by color.

Search theory offers an explanation of the marital sorting processes that may facilitate color inequality and maintain the color hierarchy despite shifts in the racial structure over time (Oppenheimer, 1988). Search theory as applied to marriages dictates that potential mates evaluate each other on a variety of desirable characteristics and attempt to maximize their own status by exchanging their own desirable characteristics for those they desire in a partner. People of high status, regardless of how status is measured, should receive more marriage offers and enjoy the opportunity to select a higher status partner than their lower status counterparts.

I lean on search theory to examine how Black and mulatto women, as the primary caregivers of children, selected marriage partners. Women, particularly at a time when most lucrative occupations were closed to women, would attempt to secure the highest status partner available to them. According to search theory, in exchange, women must offer their own status, and I argue that as status women would offer their color. Specifically, mulatto women should be able to leverage their favorable color to not only marry at a higher rate but to secure higher status partners. This would concentrate economic advantage among mulatto households as the one-drop rule begins to cement, and even as Jim Crow dampens opportunities for Black Americans, relative affluence may offer a slight reprieve from some of the worst aspects of segregation. Finally, as Jim Crow wanes, light-skinned Black people’s accumulated economic advantages, however slight, would position them to disproportionately benefit from America’s broadening opportunities for Black Americans. This would mean color stratification is more than the accretion of White preference aimed at lighter skinned Black Americans and is actually a part of a larger system of historical color gains.

Hypotheses

Building on the aforementioned framework, I seek to explore the viability of the intergenerational transmission of color advantage by examining whether mulatto women fared better in the marriage market than Black women in 1880. I expect mulatto women to perform considerably better, which I will test with two hypotheses.

Hypothesis 1: Mulatto women will be more likely to marry than Black women.

Hypothesis 2: Mulatto women will marry partners with better jobs than Black women.

Method

My data come from two sources. My individual-level data come from the Integrated Public Use Microdata Series (IPUMS) 10% U.S. Census sample from 1880, and my county-level data come from the National Historical Geographic Information System (NHGIS) U.S Census files. I include every woman from the age of 18 years to 60 years who was identified as either Black or mulatto in 1880.

Dependent Variable

I use two dependent variables. The first dependent variable is a dichotomous variable indicating whether a woman was married in 1880. The second dependent variable is spouse’s occupational score. Occupational score is a rating from 0 to 80 based on an occupation’s median income in 1950. Although it is based on income, occupational score is not a measure of personal income because the income numbers are not necessarily scalable across time. Instead, it offers a continuous measure of occupational status, with higher numbers representing higher occupational status.

Independent Variable

My focal independent variable is a dichotomous variable for whether a woman is identified as mulatto. The mulatto racial identification variable carries a set of problems if used as a measure of racial ancestry even though it is technically supposed to indicate a person of mixed Black–White ancestry. At the time, people were not allowed to complete their own Census forms. Instead, a designated Census enumerator completed the forms. Census enumerators often neglected to take the time to delve into the ancestry of each person to confirm whether they were actually of mixed Black–White heritage. They used the racial knowledge of the day, which suggested that one could determine ancestry by examining phenotypical characteristics (Hochschild & Powell, 2008). While this relative arbitrariness makes the mulatto category unreliable for measuring racial ancestry, it becomes an excellent measure of “color.” This allows me to use the mulatto designation as a proxy for light-skinned African Americans and the Black category as a proxy for darker skinned African Americans.

I also controlled for a number of other factors in three broad categories: individual-level characteristics consisting of literacy, age, age squared, disability, occupational score, and labor force participation; place-based characteristics consisting of farm residence, urban residence, and Southern; and local racial characteristics consisting of Black proportion, percentage of Blacks who are mulatto, and slave history. See Table 1 for variable definitions.

Table 1.

Descriptive Statistics and Variable Descriptions.

	Total		Married		Mulatto		Black
	n = 312,732		n = 188,107		n = 51,096		n = 261,636
Variable	M	SD	M	SD	M	SD	M	SD	Description
Spouse’s occupational score	15.39	13.06	15.38	13.06	16.66	14.97	15.15	12.59	Occupational score of married women’s husbands
Married	0.64	—	—	—	0.64	—	0.65	—	Dichotomous variable for whether a woman is married
Mulatto	0.16	—	0.16	—	—	—	—	—	Dichotomous variable for whether a woman is identified as mulatto
Literate	0.25	—	0.21	—	0.39	—	0.23	—	Dichotomous variable for whether a woman can read and White
Age	32.65	25.84	33.16	23.88	31.53	24.59	32.87	26.05	Women’s ages
Labor force	0.41	—	0.26	—	0.35	—	0.43	—	Dichotomous variable for whether a woman is in the labor force
Disability	0.01	—	.01	—	0.01	—	0.01	—	Dichotomous variable for whether a woman has a physical disability
Occupational score	4.38	14	2.96	12.66	3.51	13.16	4.55	14.13	Women’s occupational score
Farm	0.33	—	0.38	—	0.29	—	0.33	—	Dichotomous variable for whether a woman lived on a farm
Urban	0.19	—	0.15	—	.29	—	0.17	—	Dichotomous variable for whether a woman lived in an urban area
African American proportion	0.47	0.54	0.48	0.53	0.40	0.56	0.48	0.53	Proportion of African Americans is a woman’s home county
Mulattos	0.16	0.21	0.15	0.20	0.21	0.29	0.15	0.17	Proportion of the African American population identified as mulatto in a woman’s home county
Slave proportion, 1860	0.38	0.57	0.40	0.57	0.32	0.57	0.40	0.57	Proportion of the population enslaved in 1860 in a woman’s home county
South	0.88	—	0.89	0.69	0.80	0.89	0.89	—	Dichotomous variable for whether a woman lived in the South

Note. Values in italics signify further offer visual difference between the mean and standard deviation.

Modeling Strategy

I used a multipronged modeling strategy to offer a fuller picture of colorism in the marriage market for Black women in the late 19th century.

I began by modeling the relationship between mulatto status and marriage status to examine whether mulatto women were more likely marry net of other factors. I used logistic regression to model this relationship with marriage as the dependent variable, mulatto status as the primary independent variable. Model 1 only includes the mulatto variable. Model 2 introduces the other individual-level variables. Model 3 introduces the place variables. And Model 4 introduces the racial characteristic variables.

Next, I examined the relationship between mulatto status and spousal occupational score using the same modeling strategy with ordinary least squares (OLS). Each of these models only contain women who were married.

Finally, I sought a deeper look at differences in the occupational statuses of who women were marrying to continue to gain a broader view of marriage market colorism. To that end, I used an ANOVA to test whether there were differences in the means of the occupational scores of mulatto men married to mulatto women, Black men married to mulatto women, mulatto men married to Black women, and Black men married to Black women. I then used a Tukey–Kramer test to determine which of the means were statistically significant from each other. This allows me to examine whether mulatto women were benefiting in the marriage market primarily through marriage homogamy, that is, marrying mulatto men, or if they married higher status men regardless of whether they were Black or mulatto. Correspondingly, it reveals whether Black women benefited from marrying mulatto men, particularly relative to mulatto women who married Black men.

Results

Marriageability

The results from the first series of models (see table 2) suggests that the relationship between marital status and color was complicated. At best, mulatto women received a statistically significant but substantively negligible boost to their chances of getting married, which does not support my first hypothesis.

Table 2.

Logistic Regression Results for Marriage Status for All Women (n = 308,764).

Variable	Model 1	Model 2	Model 3	Model 4
Intercept	.60*** (.002)	−3.06*** (.02)	−3.42*** (.02)	−3.53*** (.02)
Mulatto	−.03*** (.005)	−.08*** (.01)	−.03*** (.01)	.02*** (.01)
Literacy		−.52*** (.004)	−.39*** (.004)	−.33*** (.005)
Age		.28*** (.001)	.29*** (.001)	.29*** (.001)
Age squared		−.004*** (.00001)	−.004*** (.00001)	−.004*** (.00001)
Disability		−1.26*** (.02)	−1.25*** (.02)	−1.27*** (.02)
Occupational score		.02*** (.0005)	.02*** (.0005)	.01*** (.0005)
Labor force		−1.76*** (.006)	−1.69*** (.01)	−1.76*** (.01)
Farm			.31*** (.004)	.29*** (.004)
Urban			−.40*** (.01)	−.29*** (.01)
South			.10*** (.01)	−.49*** (.01)
Black %				1.73** (.02)
Mulatto %				−.38*** (.02)
Slave %				−.29*** (.01)

p < .05. **p < .01. ***p < .0001.

The first model establishes a correlation between mulatto identification and marriage status. Mulatto is negative and significant, but the coefficient is small. This suggests that, controlling for no other factors, mulatto women are at a slight disadvantage when seeking to get married.

The second model includes other individual-level characteristics to test whether the effect of being mulatto remains consistent net of other individual-level factors, namely, literacy, age, age squared, disability, occupational score, and labor force status. Mulatto identification remains negative and significant, but the coefficient doubles even though it remains relatively small. Of the other variables, literacy is negative and significant, age is positive and significant, age squared is negative and significant, disability is negative and significant, occupational score is positive and significant, and labor force status is negative and significant.

Model 3 includes signifiers for place: farm, urban, and Southern. Mulatto identification remains negative and significant, though the coefficient decreased by over half. The other variables also remained relatively consistent. Age is positive and significant, age squared is negative and significant, disability is negative and significant, and occupational score is positive and significant. Though literacy and labor force status also remain negative and significant both of their coefficients decreased, suggesting that they may be partially mediated by the effect of place. Of the new variables, farm is positive and significant, urban is negative and significant, South is positive and significant.

The fourth model includes variables to account for the racial characteristics of local places: total African American proportion, proportion of African Americans that identify as mulatto, and slave proportion in 1860. In this model, mulatto identification becomes positive and significant, but the coefficient becomes smaller than in any previous models. This suggests that the racial landscape, perhaps the availability of a broader selection of partners, allows mulatto women’s color advantage to come to the fore. However, the marginal benefit of being mulatto on women’s chances of marrying may reflect the fact that in 1880 most women married anyway. This differs from today when there is high variance in marriage rates among types of women. However, marriageability is only the first component of mulatto advantage in the marriage market, and the second hypothesis builds on this by examining spousal quality.

Spousal Quality

The results of the next series of models (see table 3) support Hypothesis 2 by showing mulatto women receive significant advantages in spousal quality relative to Black women.

Table 3.

OLS Regression Results for Spouses’ Occupational Score for All Married Women (n = 186,273).

Variable	Model 1	Model 2	Model 3	Model 4
Intercept	15.15*** (.01)	11.41*** (.14)	13.06*** (.14)	12.76*** (.15)
Mulatto	1.47*** (.04)	1.24*** (.04)	.77*** (.03)	.69*** (.03)
Literacy		1.62*** (.03)	.45*** (.03)	.42*** (.03)
Age		.20*** (.01)	.13*** (.01)	.13*** (.01)
Age squared		−.003*** (.0001)	−.002*** (.0001)	−.002*** (.0001)
Disability		−.18 (.18)	−.36* (.17)	−.39* (.17)
Occupational score		.34*** (.005)	.40*** (.004)	.41*** (.004)
Labor force		−4.50*** (.06)	−5.23*** (.06)	−5.27*** (.06)
Farm			−1.04*** (.03)	−1.00*** (.03)
Urban			5.17*** (.04)	5.10*** (.04)
South			−.49*** (.04)	−.17** (.05)
Black %				−.33** (.10)
Mulatto %				1.77*** (.17)
Slave %				−.12 (.07)
R ²	.01	.05	.18	.17

Note. OLS = Ordinary least squares.

p < .05. **p < .01. ***p < .0001.

The first model establishes an initial connection between mulatto identification and spousal occupational score. Mulatto is positive and significant, suggesting that the spouses of mulatto women have higher occupational scores than the spouses of Black women. This initial result is consistent with search theory and reflects the favor conferred upon mulatto women, who were able to attract and select spouses with better occupations.

The second model includes other individual-level characteristics to test whether mulatto women continue to marry better spouses when accounting for other factors, specifically literacy, age, disability, occupational score, and labor force status. Mulatto status remains positive and significant, but the coefficient decreases slightly, suggesting that the positive effect of identifying as mulatto is partially mediated by other individual characteristics. Among the other variables, literacy and occupational score are positive and significant; age is positive and significant, and age squared is negative and significant; labor force status is negative and significant; and disability is nonsignificant.

The third model includes individual-level place variables: farm, urban, and South. Mulatto identification remains positive and significant, but the coefficient decreases by almost half. Although mulatto women still seem to marry husbands with higher occupational scores than their Black counterparts, the relationship is even further mediated by where they lived, perhaps reflecting the effect of varying partnership opportunities based on place. Almost all of the other variables from the previous model also remained their direction and significant. Among the new variables, farm status and Southern are negative and significant, and urban status is positive and significant.

The fourth model adds the racial characteristics of local places. Again, mulatto identification remains positive and significant, but its coefficient decreases slightly, suggesting mulatto women’s advantage in spousal occupational scores is partially mediated by local racial composition. Of the other variables from the previous model, few changed notably and they all retained their significance and direction. Of the new variables, Black proportion is negative and significant; mulatto percentage is positive and significant; and slave proportion is nonsignificant.

While these results offer robust support for my second hypothesis, I conducted supplementary analyses to ensure that my results did not reflect an effect of marriage homogamy. Specifically, I investigated whether mulatto women married higher status spouses regardless of the color of the spouse. Table 4 begins by showing proportions of marriage pairings by wife’s color.

Table 4.

Proportions of Marriage Pairings by Wife’s Color.

	Wife
		Mulatto	Black
Husband	Mulatto	.49	.06
Husband	Black	.43	.88
	n	32,643	168,992

While marriage homogamy was prominent, a significant number of marriages featured spouses of different colors. Black women married mulatto men about 6% of the time, and mulatto women married Black men about 43% of the time. Table 5 continues by showing the average spousal occupational scores of the husbands in each type of marriage. I ran a one-way ANOVA to test whether the occupational scores were significantly different from one another.

Table 5.

Average Husband Occupational Score by Marriage Pairing.WWW

	Wife
		Mulattot	Black
Husband	Mulatto	17.30	15.26
Husband	Black	15.85	15.14

There is a statistically significant difference between the occupational scores as determined by one-way ANOVA, F(3, 188103) = 697.18, p ⩽-.0001. A Tukey–Kramer test then revealed which of the occupational scores differed from the others. The spousal occupational scores of mulatto women married to mulatto men were significantly different from the other scores; the spousal occupational scores of mulatto women married to Black men were significantly different from the others; and while the spousal occupational scores of Black women were not significantly different from each other, they differed from each of the spousal occupational scores of mulatto women. Mulatto women married mulatto men with the highest occupational scores of any other marriage pairing, and they also tended to marry the higher status Black men. In contrast, Black women married the lower status Black men, and even in the rare cases when they married mulatto men, they married lower status mulatto men whose occupational scores were lower even than the Black men who married mulatto women and indistinguishable from the Black men who married Black women.

Discussion/Conclusion

These results set the stage for a conceptualization of colorism that goes beyond mere color preference and ideas about phenotypic proximity to whiteness. The economic head start enjoyed by light-skinned Black Americans while the country was on the cusp of the transition to the one-drop rule could have been maintained through their marriage patterns and perpetuated over the generations. Lighter skinned children from relatively well-off households would enjoy the benefits of economic affluence. They would attend better schools and social clubs—in the late 19th and early 20th centuries many schools and organizations would be closed to their darker skinned counterparts regardless of economic status (Russell-Cole, Wilson, & Hall, 2013)—perpetuating a cycle of advantage they could subsequently offer to their own children. Although in most cases they would be subjected to the same formal segregation as darker skinned Black Americans, their economic and social advantages would keep them poised to capitalize on opportunities when given the chance. Ultimately, the barriers posed by legal segregation would fall away, freeing lighter skinned Black Americans to leverage the advantages they would have accrued to fuller participation in American society. They would be disproportionately represented among the top of Black American society at a time when racial progress seemed rapid and perhaps remain there even as progress slowly slipped away for darker Black Americans.

Although racial solidarity may have benefited Black America more than an ongoing split by color by offering a unified voice to demand justice and fight for racial reform, colorism allows the lighter skinned to float to the top remain there, hiding the true despair of living as a darker skinned Black American beneath an umbrella of amalgamated racial progress. And without policy reforms that account for color and activism that pushes back against skin tone stratification, most contemporary approaches to rectifying colorism in the United States are limited to educational efforts and awareness raising, without any true “teeth” (Maddox & Perry, 2017). This is because the prevailing view of colorism revolves around individual discrimination and prejudice with little attention to structural mechanisms. Perhaps prejudices can be reduced with awareness, but a century and a half of accumulated financial advantage requires more complicated interventions that may ultimately venture into uncomfortable territory, even for critical race theorists.

One of the issues that courts face when confronting cases of color discrimination is determining who is light skinned and who is dark skinned when attempting to parse out the power imbalance between the two contesting parties (Jones, 2009). In 2014, the Brazilian government instituted an affirmative action quota system that mandates a person of color fills at least 20% of federal positions (Garcia-Navarro, 2016). Since then they have also instituted committees to determine whether a person meets the phenotypic criteria—i.e., whether they are “black looking” enough—to receive a reserved position. The racial review boards have been met with resistance and their resemblance to 19th-century eugenics notwithstanding, Black Brazilians seem to approve of this vetting process because it ensures these positions are actually awarded dark-skinned Brazilians rather than lighter skinned Brazilians who identify as Black. Although I am not advocating for a similar color or race-based review process in the United States—our racial structure certainly differs from theirs—solving the colorism problem requires a similar type of creativity and boldness and may ultimately entail borrowing from successful policy in other countries.

Color-based policy in the United States is difficult to develop, not only because our theory has been insufficient, but also because, despite the advancements in our understanding of the ubiquity of color stratification, we are unable to gain a complete understanding because of ongoing data limitations. Much contemporary racial policy relies on federal data. Although social science surveys seem increasingly likely to include a measure of color, and we continue to gain valuable insights from these data, federal data sets remain the gold standard of data collection. Color measures on the Census or the Federal Bureau of Investigation (FBI) crime statistics would dramatically increase the specificity with which we can analyze colorism. For example, we could examine the connection between color and segregation or the intricacies between color and arrest rates. More data of this type would offer us a broader understanding of the influence of color and greater leverage to advocate specific policy solutions, such as neighborhood interventions where there are concentrations of dark-skinned Black Americans or police audits where color arrest disparities seem particularly high.

This work also forces us to reconsider the very nature of race in the United States, particularly in regard to Black Americans. A number of scholars have theorized about the current and future state of racial classifications in the United States, from Bonilla-Silva’s (2017) Latin Americanization thesis, to Saperstein and Penner’s (2012) racial fluidity frameworks, to Yancey’s (2003) Black/non-Black divide, to even a version of Telles’s (2014) pigmentocracy. But we must consider whether the country ever truly moved past its triracial system and whether we have been too quick to dismiss this past, especially as we look to the future. Yes, it is true legislation and bureaucratic institutions such as the Census Bureau tried to dismiss the triracial system or ideas about a racial spectrum, but critical race theorists have been trained not to view the law as determinant. For example, simply because the law has ostensibly moved past segregation and discrimination does not make it true. Indeed, we know it is decidedly false. Perhaps a similar phenomenon is at play with racial categorization, where the formal channels offered by our institutions do not necessarily match our social reality. A fair amount of research already examines whether our current measures of race match people’s lived experiences with race (e.g., Howell & Emerson, 2017; Roth, 2016; Saperstein, Kizer, & Penner, 2016), and in 2007, a Pew survey reported that almost 40% of Black Americans said Black people have become too diverse to be considered a single race. Moreover, there is increasing evidence of divergent political attitudes among light-skinned and dark-skinned Black Americans (Reece & Upton, 2017; Wilkinson, Garand, & Dunaway, 2015). We must continue to contemplate how color, both as a system of prejudices and concentrated historical advantages, disrupts how we view race in the United States.

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Robert L. Reece is an assistant professor of sociology at The University of Texas at Austin, where he studies the origins and contours of racialization in the United States. He received his PhD from Duke University. He is from Leland, Mississippi, a small town in the Mississippi Delta.

References

Banks

T. L.

(2009). Multilayered racism: Courts’ continued resistance to colorism claims. In Glenn

E. N.

(Ed.), Shades of difference: Why skin color matters (pp. 213-222). Palo Alto, CA: Stanford University Press.

Banks

T. L.

(2014). A darker shade of pale revisited: Disaggregated Blackness and colorism in the “post-racial” Obama era. In Norwood

K. J.

(Ed.), Color matters: Skin tone bias and the myth of a post-racial America (pp. 95-117). New York, NY: Routledge.

Barnes

M.L

. (2016). Empirical Methods and Critical Race Theory: A Discourse on Possibilities for a Hybrid Methodology. Wisconsin Law Review. 2016: 442-476.

Bell

(1991). Racial realism. Connecticut Law Review, 24, 363-379.

Berlin

(1974). Slaves without masters: The free Negro in the antebellum South. New York, NY: The New Press.

Bhatnagar

(2007). Identifying the identified: The census, race, and the myth of self-classification. Texas Journal of Civil Liberties & Civil Rights, 13, 85-116.

Blair

I. V.

Judd

C. M.

Chapleau

K. M.

(2004). The influence of Afrocentric facial features in criminal sentencing. Psychological Science, 15, 674-697.

Blake

J. J.

Keith

V. M.

Luo

Salter

(2017). The role of colorism in explaining African American females’ suspension risk. School Psychology Quarterly, 32, 118-130.

Bodenhorn

(2002). The mulatto advantage: The biological consequences of complexion in rural Antebellum Virginia. Journal of Interdisciplinary History, 33, 21-46.

10.

Bodenhorn

(2011). Manumission in nineteenth-century Virgina. Cliometrica, 5, 145-164.

11.

Bodenhorn

(2015). The color factor: The economics of African-American well-being in the nineteenth-century South. New York, NY: Oxford University Press.

12.

Bodenhorn

Ruebeck

C. S.

(2007). Colourism and African-American wealth: Evidence from the nineteenth-century South. Journal of Population Economics, 20, 599-620.

13.

Bonilla-Silva

(2017). Racism without racists: Color-blind racism and the persistence of inequality in America (5th ed.). New York, NY: Routledge.

14.

Branigan

A. R.

Freese

Patir

McDade

T. W.

Liu

Kiefe

C. I.

(2013). Skin color, sex, and educational attainment in the post-civil rights era. Social Science Research, 42, 1659-1674.

15.

Brown

K.D

. (2014). The rise and fall of the one-drop rule: How the importance of color came to eclipse race. In Norwood

K.D

. (ed) Color Matters: Skin Tone Bias and the Myth of a Postracial America (pp. 44-94). New York, NY: Routledge/Taylor & Francis.

16.

Conley

(2001). Decomposing the Black-White wealth gap: The role of parental resources, inheritance, and investment dynamics. Sociological Inquiry, 71, 39-66.

17.

Frazier

E. F.

(1933). Children in Black and mulatto families. American Journal of Sociology, 39, 12-29.

18.

Garcia-Navarro

(2016, September). For affirmative action, Brazil sets up controversial boards to determine race. NPR. Retrieved from https://www.npr.org/sections/parallels/2016/09/29/495665329/for-affirmative-action-brazil-sets-up-controversial-boards-to-determine-race>

19.

Goldsmith

A. H.

Hamilton

Darity

Jr. (2006). Shades of discrimination: Skin tone and wages. The American Economic Review, 96, 242-245.

20.

Goldsmith

A. H.

Hamilton

Darity

Jr. (2007). From dark to light: Skin color and wages among African-Americans. The Journal of Human Resources, 42, 701-738.

21.

Gomez

L.E

. (2012). Looking for race in all the wrong places. Law & Society Review, 46(2), 221-245.

22.

Green

T. L.

Hamilton

T. G.

(2013). Beyond Black and White: Color and mortality in post-reconstruction era North Carolina. Explorations in Economic History, 50, 148-159.

23.

Gross

A. J.

(1998). Litigating whiteness: Trials of racial determination in the nineteenth-century South. The Yale Law Journal, 108, 109-188.

24.

Gullickson

(2010). Racial boundary formation at the dawn of Jim Crow: The detriments and effects of Black/mulatto occupational differences in the United States, 1880. American Journal of Sociology, 116, 187-231.

25.

Hall

R. E.

Johnson

(2014). The implications of skin color vis-à-vis discrimination: Revisiting affirmative action. In Norwood

K. J.

(Ed.), Color matters: Skin tone bias and the myth of a post-racial America (pp. 182-197). New York, NY: Routledge.

26.

Hamilton

Goldsmith

A. H.

William

Jr. (2009). Shedding “light” on marriage: The influence of skin shade on marriage for Black females. Journal of Economic Behavior & Organization, 72, 30-50.

27.

Hannon

DeFina

Bruch

(2013). The relationship between skin tone and school suspension for African Americans. Race and Social Problems, 5, 281-295.

28.

Hickman

C. B.

(1997). The devil and the one drop rule: Racial categories, African Americans, and the U.S. census. Michigan Law Review, 95, 1161-1265.

29.

Hochschild

J. L.

Powell

B. M.

(2008). Racial reorganization and the United States census 1850-1930: Mulattos, half-breeds, mixed parentage, hindoos, and the Mexican race. Studies in American Political Development, 22, 59-96.

30.

Howell

Emerson

M. O.

(2017). So what “should” we use? Evaluating the impact of five measures on markers of social inequality. Sociology of Race & Ethnicity, 3, 14-30.

31.

Jones

(2000). Shades of Brown: The law of skin color. Duke Law Journal, 49, 1487-1557.

32.

Jones

(2009). The case for legal recognition of colorism claims. In Glenn

E. N.

(Ed.), Shades of difference: Why skin color matters (pp. 223-235). Palo Alto, CA: Stanford University Press.

33.

Maddox

K. B.

Perry

J. M.

(2017). Racial appearance bias: Improving evidence-based policies to address racial disparities. Policy Insights From the Behavioral and Brain Sciences, 5, 57-65.

34.

Monk

E. P.

Jr. (2014). Skin tone stratification among Black Americans, 2001-2003. Social Forces, 92, 1313-1337.

35.

Monk

E. P.

Jr. (2015). The cost of color: Skin color, discrimination, and health among African-Americans. American Journal of Sociology, 121, 396-444.

36.

Norwood

K. J.

Foreman

V. S.

(2014). The ubiquitousness of colorism: Then and now. In Norwood

K. J.

(Ed.), Color matters: Skin tone bias and the myth of a post-racial America (pp 9-28). New York, NY: Routledge.

37.

Obasogie

O.K

. (2013). Foreword: Critical Race Theory and Empirical Methods. UC Irvine Law Review, 183-187.

38.

Omi

Winant

(2015). Racial formation in the United States (3rd ed.). New York, NY: Routledge.

39.

Oppenheimer

V. K.

(1988). A theory of marriage timing. American Journal of Sociology, 94, 563-591.

40.

Painter

M. A.

II Holmes

M. D.

Bateman

(2016). Skin tone, race/ethnicity, and wealth inequality among new immigrants. Social Forces, 94, 1153-1185.

41.

Pfeffer

F. T.

Killewald

(2015). How rigid is the wealth structure and why? Inter and multigenerational associations in family wealth (Population Studies Center Research Reports, University of Michigan Institute for Social Research Report 15-845). Ann Arbor, Michigan: University of Michigan Population Studies Center.

42.

Ray

V. E.

Randolph

Underhill

Luke

(2017). Critical race theory, Afro-pessimism, and racial progress narratives. Sociology of Race & Ethnicity, 3, 147-1583.

43.

Ray

V. E.

Seamster

(2016). Rethinking racial progress: A response to Wimmer. Ethnic and Racial Studies, 39, 1361-1369.

44.

Reece

R. L.

(2017). By the numbers: Skin tone and racial advantage. Scalawag. Retrieved from https://www.scalawagmagazine.org/2017/08/by-the-numbers-skin-tone-and-racial-advantage/>

45.

Reece

R. L.

(2016). What are you mixed with: The effect of multiracial identification on perceived attractiveness. Review of Black Political Economy, 43, 139-147.

46.

Reece

R. L.

(2018a). Coloring weight stigma: On race, colorism, weight stigma, and the failure of additive intersectionality. Sociology of Race & Ethnicity. DOI: 10.1177/2332649218795185

47.

Reece

R. L.

(2018b). Genesis of U.S. colorism and skin tone stratification: Slavery, freedom, and mulatto-Black occupational inequality in the late 19th century. Review of Black Political Economy, 45, 3-21.

48.

Reece

R. L.

Upton

A. A

. 2017. “How Skin Tone Shapes Civic Engagement Among Black Americans.” In Martin

L.L.

Horton

H. D.

Herring

Keith

V. M.

Thomas

(eds) Color Struck: How Race and Complexion Matter in the “Color-Blind” Era (pp. 155-177). Rotterdam, the Netherlands: Sense Publishers.

49.

Roth

W. D.

(2016). The multiple dimensions of race. Ethnic and Racial Studies, 39, 1310-1338.

50.

Roth

W.D

. (2017). Methodological pitfalls of measuring race: international comparisons and repurposing of statistical categories. Ethnic and Racial Studies, 40(13): 2347-2353.

51.

Russell-Cole

Wilson

Hall (2013). The color complex: The politics of skin color in the new millennium. New York, NY: Anchor Books.

52.

Saperstein

Penner

A. M.

(2012). Racial fluidity and inequality in the United States. American Journal of Sociology, 118, 676-727.

53.

Saperstein

Gullickson

(2013). A “mulatto escape hatch” in the United States? Examining evidence of racial and social mobility during the Jim Crow era. Demography, 50, 1921-1942.

54.

Saperstein

Kizer

J. M.

Penner

A. M.

(2016). Making the most of multiple measures: Disentangling the effects of different dimensions of race in survey research. American Behavioral Scientist, 60, 519-537.

55.

Schweninger

(1989). Black-owned businesses in the South, 1790-1880. The Business History Review, 63, 22-60.

56.

Schweninger

(1990). Prosperous Blacks in the South, 1790-1880. The American Historical Review, 95, 31-56.

57.

Telles

(2014). Pigmentocracies: Ethnicity, race, and color in Latin America. Chapel Hill: The University of North Carolina Press.

58.

Toplin

R. B.

(1979). Between Black and White: Attitudes toward Southern mulattoes, 1830-1861. The Journal of Southern History, 45, 185-200.

59.

Viglione

Hannon

DeFina

. (2011). The impact of light skin on prison time for Black female offenders. The Social Science Journal, 48, 250-258.

60.

Washington

S. L.

(2011). Hypodescent: A history of the crystallization of the one-drop rule in the United States, 1880-1940 (Doctoral dissertations). Princeton University, NJ.

61.

Wilkinson

B.C.

Garand

J. C.

Dunaway

(2015). Skin tone and individuals’ perceptions of commonality and competition with other racial and ethnic groups. Race and Social Problems, 7, 181-197

62.

Yancey

(2003). Who is White? Latinos, Asians, and the new Black/nonblack divide. Boulder, CO: Lynne Rienner.