Racial redistricting and African-American representation: a critique of 'Do Majority-Minority Districts maximize substantive black representation in Congress?'.(response to Charles Cameron, David Epstein and Sharyn O'Halloran, American Political Science Review, vol. 90, p 794, Dec 1996)
Author/s: David Lublin
Issue: March, 1999
The methodologically sophisticated analysis of race and redistricting by Cameron, Epstein, and O'Halloran (1996) has already been cited in support of claims that racial redistricting undermines black representation (e.g., Thernstrom and Themstrom 1997). Yet, key substantive and methodological errors in their analysis lead them to underestimate severely the percentage of blacks needed to assure the probable election of an African American. Their analysis of the relationship of the percentage of blacks in a district to the ideology of its representative is similarly unreliable. Finally, their analysis ignores how racial redistricting indirectly aids the Republicans and undermines black representation. Changes in the type of major-party candidates nominated due to racial redistricting result in additional votes for the Republicans, who are more hostile than Democrats to initiatives supported by blacks.
THE ELECTION OF BLACK REPRESENTATIVES
On the basis of their analysis of the 1992 elections, Cameron, Epstein, and O'Halloran (1996, 804) conclude that the black voting-age population (BVAP) required to achieve a 50% chance of a black victory varies substantially by region: "In the South the required level of BVAP is 40.3%; in the Northwest, 47.3%; in the Northeast, only 28.3%." If true, these would be startling and important findings. Past studies suggest that majority-minority districts remain critical to the election of African Americans to Congress (Davidson and Grofman 1994; Grofman and Handley 1989). If Cameron, Epstein, and O'Halloran were correct, then their analysis would support Carol Swain's (1993) conclusion that the creation of new majority-minority districts is unnecessary to assure the election of blacks to the U.S. House of Representatives.
An examination of actual cases explains why Cameron, Epstein, and O'Halloran and Swain reached similar conclusions (see Grofman and Handley 1995, 249-53, for a similar critique of Swain's study). Eight of the thirteen representatives elected in 1992 from districts with black voting-age populations between 30% and 50% are black. Seven of the eight, however, were elected from districts in which African Americans and Latinos together compose 55% or more of the voting-age population. For example, Charles Rangel (D-NY) represents a Harlem district that is 47% BVAP and 42% voting-age Hispanic. Only William Clay (D-MO) represents a district less than 50% minority, but even it is 48% BVAP and 1% voting-age Hispanic. The claim by Cameron, Epstein, and O'Halloran that blacks win election from districts lacking a BVAP majority is correct, but their neglect of the role of Latinos leads them to suggest wrongly that racial redistricting is no longer required to assure the election of more than token numbers of blacks to Congress.(1)
In The Paradox of Representation (Lublin 1997b, 41-8; see also Lublin 1997a), I present the results of several logit analyses of Black Representative, a dummy variable, on proportion black, proportion Latino, and a variety of other nonracial demographic variables for elections held between 1972 and 1994. The nonracial demographic variables exert no consistent or significant influence on the election of blacks, as they are overwhelmed by the racial variables. The results suggest that African Americans have an 86% chance of winning in districts that are 55% black (or a somewhat lower BVAP) that contain no Latinos (Lublin 1997a, 1997b). The probability of victory drops quickly below this percentage unless the share of Latinos increases (Lublin 1997a, 1997b). Looking at the actual cases reinforces the solidity of this finding. Between 1972 and 1994, blacks won only 72 of the 5,079 elections held in white majority districts. Forty-five of these victories in "white" majority districts were in districts in which African Americans and Latinos together formed a majority (Lublin 1997b).
Over this same period, six black representatives won a total of 27 elections from districts with a majority of non-Hispanic whites. Unfortunately, these elections were more idiosyncratic than part of a larger national pattern of blacks winning in white districts (Grofman and Handley 1995, 250-2). Unless one is willing to claim that "as goes Berkeley, so goes the nation," the victories by Ron Dellums must be ascribed to the highly liberal nature of his district. Katie Hall won election twice in Indiana after being appointed to fill a vacancy but lost her first contested Democratic primary in 1984. Missouri Democrat Alan Wheat repeatedly won election from a district that was 80% white until his failed 1994 senatorial bid, but he probably would not have survived the Democratic primary in his first election if a runoff primary had been held. Wheat won the primary with only 31% of the vote - just 1% more than his main white challenger. Republican Gary Franks won election three times from a Connecticut district between 1990 and 1994, but he never gained more than 52% of the vote and in 1996 lost reelection from the most Republican district in the state. Only two representatives, J.C. Watts (R-OK) and Andrew Young (D-GA), appear to be true exceptions to the general pattern that blacks rarely win election from majority non-Hispanic white districts.
Opponents of racial redistricting have hailed the reelection of Democrats Sanford Bishop and Cynthia McKinney of Georgia and Corrine Brown of Florida in 1996 as evidence that blacks can win in white districts (Sleeper 1997; Thernstrom and Themstrom 1997). All three represent districts that are more than 35% black and very likely to elect Democrats (Cameron, Epstein, and O'Halloran 1996, 805), so the Democratic primary was the key election. Moreover, blacks composed a disproportionate share of the Democratic primary electorate due to racial differences in partisan preferences, so their reelection was not very surprising despite their race. Three nonincumbent southern blacks sought election to the House in 1996, and all three lost (Bositis 1997). The election of Julia Carson (D-IN) is the only genuine new evidence that blacks can win in white districts, but her victory seems a thin reed with which to challenge a long-term pattern.
THE REPRESENTATION OF BLACK INTERESTS
Since all but one of the new black districts created during the 1990 redistricting round were in the South, I will focus primarily on the results for this region.
Percentage Black and Ideology
The decision by Cameron, Epstein, and O'Halloran (1996, 799-801) to examine only representatives elected in 1992, after the 1990 redistricting round, renders unreliable their findings regarding the relationship between the black percentage in a district and the ideology of the representative. Knowing this relationship to southern districts with 30-50% BVAP is crucial because it determines whether racial redistricting helped or hindered black efforts to advance their interests in the House. The districts from 1992 unfortunately do not provide sufficient data for a convincing analysis. Ironically, the very policy that Cameron, Epstein, and O'Halloran attempt to evaluate makes 1992 a singularly bad year to examine. Racial redistricting before the 1992 elections dramatically reduced the number of 30-50% BVAP districts in order to permit the creation of more districts with a black majority or a lopsided white majority (Handley, Grofman, and Arden 1998). After the redistricting, there were only four 30-50% BVAP districts left in the South. In two of these districts, blacks and Latinos together formed a majority, and African Americans were elected. Consequently, data from only two districts are available to estimate the relationship between percentage black and ideology for white Democrats over this range. One is at the low end of the range; Virginia's Fourth District is only 31% BVAP. The other, Mississippi's Fourth District, is 36% BVAP but is in one of the most racially polarized areas of the nation. It was represented by Mike Parker, who switched to the Republican Party after the 1994 election.
Continued from page 2
The absence of data for 30-50% BVAP districts would not be so devastating if we could impute the results for this range from data just above or below it. Yet, numerous theories outlined by Cameron, Epstein, and O'Halloran (1996, 795-6) explain why it is a very bad idea to assume that representatives from districts with less than 30% BVAP or from black majority districts are similar to those in the 30-50% range. On the one hand, racial backlash may motivate white voters in districts with greater than 30% BVAP, and accordingly their representatives may vote more conservatively on the floor of the House (see, e.g., Giles and Buckner 1993). On the other hand, whites in districts with a relatively large number of blacks may be more tolerant and their representatives correspondingly more liberal (Voss 1996).
The Supreme Court vitiation of several "racial gerrymanders" in such landmark cases as Miller v. Johnson caused several additional 30-50% BVAP districts to be created before the 1996 election. Data from these districts are probably not very useful for estimating the relationship between black percentage and ideology, however, because the incumbency effect has helped entrench representatives who otherwise might not have won. Georgia Republicans Saxby Chambliss and Charles Norwood were narrowly reelected with 53% and 52% of the vote, respectively, in their districts of more than 30% BVAP, even though Republicans rarely win in such districts (Cameron, Epstein, and O'Halloran 1996, 805). Similarly, black Georgia Democrats Sanford Bishop and Cynthia McKinney won in 30-50% BVAP districts that probably would never have elected them in an open-seat contest (Davidson and Grofman 1996; Grofman and Handley 1989; Lublin 1997a, 1997b).
In Paradox of Representation (Lublin 1997a), I examined the link between percentage black and ideology for all elections held between 1972 and 1990, both separately for each year and in a cross-sectional analysis. The Poole-Rosenthal scores and various interest-group ratings were used as measures of the ideology of representatives. For southern Democrats, I found a linear increase in liberalism as the black percentage increases, with an additional threshold increase at 40% black. Other models, including nonlinear ones suggested by the white backlash hypothesis, fail to model the data accurately. (See Lublin 1997b, 78-97, 140-3 for a detailed explanation of the theoretical basis for the model as well as numerous empirical tests.)
Representative's Race and Ideology
Cameron, Epstein, and O'Halloran (1996, 807) state: "Although neither Republicans nor Democrats represent black interests as well as do black Democrats, white Democrats are still more supportive than Republicans." This claim is absolutely correct and confirmed by my own and other research (Grofman, Griffin, and Glazer 1992; Lublin 1997b). The authors further argue, however, that adding black voters to a district does not alter the behavior of any given type of representative (Cameron, Epstein, and O'Halloran 1996, 807). This implies that black Democrats are more responsive to black interests than white Democrats, regardless of the racial composition of the district. In other words, the race of the representative matters more than the district's racial composition after controlling for party.
Continued from page 3
Cameron, Epstein, and O'Halloran's data do not allow them to prove this assertion. As Figure 4 in their article (1996, 805) shows, white and black Democrats are elected from districts with very different racial composition. In 1992, the year of their study, no whites won in districts with more than 50% BVAP, so they cannot tell us how whites elected from majority black districts might have voted on the floor of the House. Similarly, only two blacks, Republican Gary Franks and Democrat Alan Wheat, won in districts with a majority of non-Hispanic whites in 1992. The racial composition of a district is almost determinative of the race of its representative, so it is extremely difficult to meaningfully distinguish the impact of these two factors on the roll-call voting of the representative. All five white Democrats who were elected from black majority districts between 1972 and 1994 had liberal voting records (Lublin 1997b, 31), so Cameron, Epstein, and O'Halloran's hypothesis could easily be wrong.
The Tradeoff in Descriptive and Substantive Representation
In the final substantive section of their article, Cameron, Epstein, and O'Halloran (1996, 807-9) argue that mapmakers desiring to maximize substantive black representation in the South, as measured by Leadership Conference on Civil Rights (LCCR) scores, should create as many districts as possible that are 47% black and distribute the remaining voters evenly across districts. (A representative's LCCR score equals the percentage of roll-call votes cast in accord with the LCCR position on a relatively small number of votes selected by the LCCR.) If Cameron, Epstein, and O'Halloran incorrectly estimate the relationship between percentage black and ideology, as I suggest they do, then this key conclusion about how to maximize black representation is obviously false. It also may be invalid if it is more important to maximize the LCCR score of the median representative than of the mean representative.
Cameron, Epstein, and O'Halloran (1996, 805) rightly argue that reducing the proportion of blacks in a district makes the election of a Republican more likely in the South. Adding blacks to one district requires removing them from another, so the creation of a new black district may result in a gain of one Republican and one black Democrat and a loss of two white Democrats. Consequently, the median LCCR score may drop, even if the mean LCCR score increases (Grofman and Handley 1995, 266). A larger contingent of liberal black representatives could end up with little influence against a similarly enlarged bloc of Republican conservatives. In order to gauge the partisan effect of racial redistricting, I examined the effect of boundary changes due to the creation of new black majority districts on electoral outcomes in 1992 and 1994. My analysis suggests that racial redistricting cost the Democrats at least eleven seats and made the House less likely to adopt initiatives supported by blacks (Lublin 1997b; Lublin and Voss 1998).
Continued from page 4
Contrary to Cameron, Epstein, and O'Halloran's findings (1996, 808), the location of most heavy concentrations of black voters mitigates the possibility of a similar tradeoff in the North. As the authors explain, northern districts with few blacks have a high probability of electing a Republican. Among districts with few blacks, however, rural districts are much less likely to elect Democrats than are urban districts. Outside the South, black majority districts are invariably centered on central cities, such as New York, Chicago, and Los Angeles. These cities contain large numbers of nonblack liberal Democrats, particularly Latinos and Jews, in close proximity to the largest concentrations of black voters (Lublin 1997b, 93-5), so racial redistricting in the North usually aids blacks without reducing the total number of Democrats elected to the House (Lublin 1997b, 92-6). As Grofman, Griffin, and Glazer (1992, 375) succinctly argue, "to the extent there is a conflict between the desire to elect liberals and the desire to elect Democrats to Congress, it will really be found only in southern jurisdictions."
THE INDIRECT IMPACT OF RACIAL REDISTRICTING
Cameron, Epstein, and O'Halloran (1996) do not consider that indirect influences of racial redistricting on voting behavior likely exacerbate the negative impact of racial redistricting on the election of Democrats and thus black substantive representation (see Lublin and Voss 1998). Substantial changes in the racial composition of a district will likely affect who is nominated by the major parties. In black majority districts, black liberal Democrats are nominated in place of white moderate Democrats. White Democrats in these districts probably vote Republican or fail to vote at a higher rate, either due to racism or ideology (Gay 1996). African Americans already vote Democratic at such a high rate that any increase in the percentage of blacks voting Democratic cannot outweigh vote losses among white Democrats. Also, Democratic turnout may also decline because black candidates usually win by such large margins in black majority districts (Lublin 1997b; Swain 1993), and candidates who expect to win easily do not work as hard to mobilize their supporters (see Lublin and Tate 1995 for evidence at the local level). Conversely, black Democrats may vote at a higher rate due to excitement over the nomination of a black liberal and the possibility of greater descriptive and substantive representation (Gay 1996). On balance, however, the Democratic vote is likely to be depressed in black districts.
The indirect effect of racial redistricting on black representation in surrounding white districts is even stronger. Cameron, Epstein, and O'Halloran's (1996, 805) evidence indicates that removing blacks from white districts makes the election of a Republican more likely (see also Handley, Grofman, and Arden 1998). Jacobson and Kernell's (1983) seminal research on congressional elections shows that high-quality candidates (i.e., with experience in elective office) are more likely to run as their probability of success increases, and they receive a larger share of the vote than low-quality candidates. Since GOP chances are enhanced by the removal of black voters from a district, the Republicans probably run better candidates and receive a boost at the polls. The incumbency advantage allows Republicans who win the seat and run for reelection to solidify their hold two years later (Gelman and King 1990). High-quality Democrats will forego contests due to the increased likelihood of their electoral defeat. Thus, racial redistricting alters not only the aggregation of votes but also the quality of candidates presented, such that it indirectly boosts the Republican share of votes and seats by undercutting Democratic prospects (Lublin and Voss 1998).
Continued from page 5
Cameron, Epstein, and O'Halloran (1996) make a fine effort to untangle the complex redistricting debate. Their neglect of the influence of Latinos, however, leads them to underestimate the importance of racial redistricting for the election of African Americans. Their decision to examine only representatives elected in 1992 makes it difficult for them to assess meaningfully the link between the racial composition of a district and the voting behavior of its representative. Moreover, although they examine the North and South separately, they do not consider how regional differences in racial demographics and the voting behavior of whites may affect the results. When these are taken into account, the tradeoff between descriptive and substantive representation is more acute in the South than in the North. The authors further need to consider indirect influences of racial redistricting on electoral outcomes in order to provide a more accurate assessment of the benefits and costs of this policy. In sum, contrary to the arguments presented by Cameron, Epstein, and O'Halloran (1996), racial redistricting remains vital to the election of African Americans, and the tradeoff between descriptive and substantive representation is more severe in the South than the North.
The author is grateful to Bernard Grofman, Lisa Handley, Gary King, George Krause, Steve Voss, and the anonymous reviewers for comments on earlier drafts of this critique.
1 Unlike Cameron, Epstein, and O'Halloran (1996), Swain's (1993, 208) graph makes clear that most blacks elected from "white" districts were elected from "heterogeneous" districts in which non-Hispanic whites did not form a majority.
Bositis, David A. 1997. "The Future of Majority-Minority Districts and Black and Hispanic Legislative Representation." Washington, DC: Joint Center for Political and Economic Studies.
Cameron, Charles, David Epstein, and Sharyn O'Halloran. 1996. "Do Majority-Minority Districts Maximize Substantive Black Representation in Congress? American Political Science Review 90(December):794-812.
Davidson, Chandler, and Bernard Grofman, eds. 1994. Quiet Revolution in the South: The Impact of the Voting Rights Act 1965-1990. Princeton, NJ: Princeton University Press.
Gay, Claudine. 1996. "Race and Political Context: The Impact of Black Congressional Representation on the Voting Behavior of White and Black Constituents." Presented at the annual meeting of the Midwest Political Science Association, Chicago.
Gelman, Andrew, and Gary King. 1990. "Estimating the Incumbency Advantage without Bias." American Journal of Political Science 34(November):1142-64.
Giles, Micheal W., and Melanie A. Buckner. 1993. "David Duke and Black Threat: An Old Hypothesis Revisited." Journal of Politics 55(August):702-13.
Grofman, Bernard, Robert Griffin, and Amihai Glazer. 1992. "The Effect of Black Population on Electing Democrats and Liberals to the House of Representatives." Legislative Studies Quarterly 17(August):365-79.
Continued from page 6
Grofman, Bernard, and Lisa Handley. 1989. "Minority Population and Black and Hispanic Congressional Successes in the 1970s and 1980s." American Politics Quarterly 17(October):436-45.
Grofman, Bernard, and Lisa Handley. 1995. "1990s Issues in Voting Rights." Mississippi Law Journal 65(2):205-70.
Handley, Lisa, Bernard Grofman, and Wayne Arden. 1998. "Electing Minority-Preferred Candidates to Legislative Office: The Relationship between Minority Percentages in Districts and the Election of Minority-Preferred Candidates." In Race and Redistricting in the 1990s, ed. Bernard Grofman. New York: Agathon. Pp. 13-38.
Jacobson, Gary C., and Samuel Kernell. 1983. Strategy and Choice in Congressional Elections. New Haven, CT: Yale University Press.
Lublin, David. 1997a. "The Election of African Americans and Latinos to the U.S. House of Representatives." American Politics Quarterly 25(3):269-86.
Lublin, David. 1997b. The Paradox of Representation: Racial Gerrymandering and Minority Interests in Congress. Princeton, NJ: Princeton University Press.
Lublin, David, and Katherine Tate. 1995. "Racial Group Competition in Urban Elections." In Classifying by Race, ed. Paul E. Peterson. Princeton, NJ: Princeton University Press. Pp. 245-61.
Lublin, David, and D. Stephen Voss. 1998. "The Partisan Impact of Voting Rights Law: A Reply to Pamela S. Karlan." Stanford Law Review 50(3):765-77.
Sleeper, Jim. 1997. Liberal Racism. New York: Viking.
Swain, Carol M. 1993. Black Faces, Black Interests: The Representation of African Americans in Congress. Cambridge, MA: Harvard University Press.
Themstrom, Stephen, and Abigail Themstrom. 1997. America in Black and White: One Nation Indivisible. New York: Simon and Schuster.
Voss, D. Stephen. 1996. "Beyond Racial Threat: Failure of an Old Hypothesis in the New South." Journal of Politics 58(November): 1156-70.
David Lublin is Assistant Professor of Government, American University, Washington, DC 20016-8130.
COPYRIGHT 1999 American Political Science Association
COPYRIGHT 2000 Gale Group
the magnificent seven ||
the great escape
the mechanic ||
once upon ||
Drum Beat ||
Machine Gun Kelly ||
House of Wax ||
The Magnificent Seven
Gang War ||
Guns for San Sebastian ||
The Valachi Papers ||
Death Wish ||
Death Wish 5
Forbes Celebrity 100
Important Online Links
Copyright © 2002 - CharlesBronson.com. All Rights Reserved.