Vol. 9 No. 10 (October 1999) pp. 467-471.

RACE, REDISTRICTING, AND REPRESENTATION: THE UNINTENDED CONSEQUENCES OF BLACK MAJORITY DISTRICTS by David T. Canon. Chicago: The University of Chicago Press, 1999. xiv, 324 pp. Cloth $50.00 Paper $18.00.

Reviewed by Richard L. Engstrom, Department of Political Science, University of New Orleans.

In SHAW v. RENO (1993), a five-member majority of the Supreme Court adopted a district-specific approach to the concept of "gerrymandering." They uncoupled that concept from the traditional notion that gerrymanders "waste votes" of some identifiable group of voters. In holding that a claim of "racial gerrymandering" lodged against two majority-African American congressional districts in North Carolina was justiciable, despite the fact that whites would continue to be a majority in a disproportionate number of districts in the state, Justice Sandra Day O'Connor, wrote for the majority and identified "special harms" that the deliberate creation of such districts might entail. According to her, the practice "reinforces racial stereotypes and threatens to undermine our system of representative democracy by signaling to elected officials that they represent a particular racial group rather than their constituency as a whole" (SHAW, at 650).

This assertion that African Americans elected from such districts will represent their constituents in a racially selective manner is treated as a "testable proposition" (p. 2) by David Canon. Rather than focusing on the frequently asked question, "whether whites can represent blacks?" as previous studies critical of these districts have done, Canon turns the question around and asks, "whether blacks can represent whites?" His answer to this latter question is an unequivocal yes! No only can blacks represent whites, but, according to Canon, the whites captured in the new majority-African American congressional districts created during the 1990s receive better representation from their African American representatives than the blacks in so-called "black influence districts" receive from their white representatives. O'Connor's assertion about selective representation is, in short, "simply wrong" (p. 84), and she and the rest of the Court's majority, Canon maintains, are themselves "guilty of racial stereotyping" (p. 84).

Canon distinguishes between the more traditional African American candidates, who engage in a "politics of difference," in which differences between the races are viewed as central to politics and African American voters constitute a representative's primary constituency, and the "new style" African American candidates, who engage in a nonconfrontational "politics of commonality" in which politics is viewed in biracial terms and white voters constitute an important part of a representative's electoral coalition. Based on Canon's review of campaign coverage, only four of the 15 African Americans elected from either "new" or "substantially altered" majority-

Page 468 begins here

African American districts in 1992 had engaged in campaigns based on difference politics. The explanation for this is what Canon refers to as his "supply-side" theory, which in a nutshell is "the racial composition of the candidate pool in the Democratic primary determines the type of representation a district will receive" (p. 127). The theory is bound by the racial composition of the districts themselves, which in the case of the new districts ranged from around 50 percent African American in total population (about 45 percent in voting age population) to close to 70 percent in total population (and less than 65 percent in voting age population). These are not districts, in short, that bear any resemblance to "political apartheid," as suggested by O'Connor in SHAW (at 647).

The electoral opportunities provided by districts such as these can be expected to attract several serious African American candidates. The supply- side theory maintains that the type of candidate that wins is determined by whether a serious white candidate also contests the Democratic primary. If one does, then the victorious African American candidate (assuming the white does not win due to African Americans dispersing their vote across too many candidates) is likely to be a traditional politics of difference candidate appealing primarily to African American voters. If no serious white candidate is in the race, however, the winner is likely to be a new style politics of commonality candidate who attracts considerable white support. Canon is very successful at postdicting the outcomes of the elections through this theory (the two clearly inconsistent cases were the elections of commonality candidates despite the presence of serious white competitors).

The major contribution of this book concerns the subsequent style and content of the representation provided by these new African American representatives. Those elected through difference politics are expected to focus on racial issues and play more of an outsider role in Congress. Those elected through the commonality approach are expected to display more balance in and be more pragmatic toward their legislative role. In assessing the representation provided by these new African American members of Congress, Canon departs from previous studies by going well beyond a reliance on roll call voting as the measure of representation. He is extremely critical of previous studies that rely primarily, and often exclusively, on roll call voting indices, which he says "are worthless as an indicator of behavior that addresses constituents' racial wants" (p. 31). Canon's critique of these indices is not novel. They contain few votes that deal explicitly with race (pp. 30-31, 168, 175-177), mask important differences among representatives that have similar scores (pp. 172-173), and suffer from selection bias, in that only moderate measures reach a stage in the legislative process where every member is expected to respond with a recorded vote (pp. 165, 173-175). It is typical for scholars relying on such measures to acknowledge their limitations, but then proceed to largely ignore those limitations when reaching and presenting their conclusions. Canon is not guilty of this. Although he does include analyses of roll call indices in this work, he goes far beyond previous studies in examining other types of behavior.

Canon employs numerous data collection techniques to gather a variety of measures of representational behavior. This triangulation of methods and measures, or as he says, "more accurately, dec-angulation" (p. 244), is the major strength of this book.

Page 469 begins here

In addition to examining some roll call voting indices, he also content analyzed bills, amendments, speeches, newsletters, and newspaper coverage, examined committee assignments, leadership positions, the composition of personal staffs, and the location of district offices, and conducted numerous interviews with representatives and their staff members. As with roll call votes, Canon is candid about the limitations of each type of measure, and he relies on the impressive consistency in his findings across indicators, which are almost always limited temporally to the 103rd Congress (1993-1994), to build confidence in his conclusions. The results, with respect to legislative roles, reveal that:

"Commonality members were more likely than difference members to serve in the leadership, slightly more moderate in their roll call behavior, less likely to emphasize race in their speeches on the floor and in their sponsorship and cosponsorship of legislation, and more likely to cosponsor bills that were successful in the legislative process" (p. 200).

With respect to constituency communication, Canon discovers that: "Commonality members' staffs were the only ones that did not over- or underrepresent the black composition of the district, and they had the smallest disproportionate representation of whites on their staffs of any type of member. Commonality members also pursued a much more obvious balancing strategy in the placement of their district offices in an effort to reach both black and white voters, while difference members were more concerned about placing their offices in black neighborhoods. Finally, there were dramatic differences in the racial composition of the pictures in the member newsletters and significant differences in some dimensions of the newspaper coverage between commonality and difference members" (p. 241).

Although Canon does not apply his supply-side theory to white representatives, he does compare them to the African American representatives on these measures. More specifically, he includes in his analysis those whites representing districts in which African Americans constitute at least a 25 percent of the population (of which only four were Republicans). These are what many have called "black influence districts." Such districts are usually sacrificed in the creation of majority-minority districts, and the reduction in their number has led many to argue that the representation of African Americans, in an overall sense, has been harmed by the creation of the majority-minority districts. Roll call voting analyses allegedly confirm the notion that the substantive representation of African Americans is enhanced more by electing additional white Democrats than by electing more African Americans. Canon's look beyond roll call votes provides little support for this idea, however. The gap between these white representatives and the African American representatives on his various measures of representation is usually much larger than that between the two types of African Americans, leading Canon to conclude that "on average, the white members . did not show much interest in racial issues" (p. 200). Indeed, when it came to representing their other race constituents, Canon concludes "Most white representatives from black influence districts do not spend much time representing their black constituents, while most black members of Congress spend a substantial proportion of their time representing white constituents" (p. 91). Canon's results suggest that districts that Justice O'Connor indicates resemble "political apartheid"

Page 470 begins here

might more appropriately considered "white influence districts."

RACE, REDISTRICTING, AND REPRESENTATION is the most important study yet published dealing with the representational consequences of the new majority- African American districts created following the 1990 census. Not surprising, given his findings, Canon has moved from "a grudging acceptance" (p. xii) of this vehicle for providing minorities with a voice in legislative bodies to becoming quite enthusiastic about them. Rather than providing constituents with selective racial representation, he believes they are a medium through which we can "break down racial barriers rather than erect them" (p. 264).

Canon's enthusiasm may have interfered with his objectively in one respect, however. His treatment of cumulative voting, the most frequently mentioned alternative arrangement for improving minority representation, especially since SHAW, both ignores and overstates recent empirical findings concerning that election system. He overstates the results of Greg D. Adams' study of voting in the Illinois House of Representatives both during and after cumulative voting was used to elect that body (Adams 1996). Canon claims that this study "provides strong empirical evidence that the cumulative system in Illinois produced more ideologically extreme members of the state assembly than did the single-member district system" (p. 260). Adams' study relies on an index based on a set of "key votes" identified by the Illinois Political Action Committee, an organization with ties to the Illinois Chamber of Commerce. The index was simply the percentage of times a legislator voted in agreement with the preferred position of IPAC. The results showed greater variation in index scores when that body was elected by cumulative voting. The "key votes" on which the index is based are not identified by Adams, but he nowhere indicates that the index provides a measure of ideological extremism. Indeed, if Canon is correct about the problem of selection bias in roll call vote indices, then the percentage agreement on these identified votes is hardly likely to constitute such a measure. In addition, Canon states that cumulative voting "is relatively complex and will produce voter confusion, at least in the short run" (p. 260). Evidence from exit polls in numerous settings in which cumulative voting was being used for the first time, however, indicates that this has not been a serious problem. Voters, especially minority voters, have understood and used the cumulative options without difficulty (see Brischetto and Engstrom 1997, Engstrom and Brishetto 1998, and studies cited therein).

Canon's criticism of the competition might not be fair, but this does not distract from his assessment of the representational consequences of the majority-African American districts themselves. This impressive study is, as noted above, the best yet on this subject.

REFERENCES

Adams, Greg D. 1996. "Legislative Effects of Single-Member Vs. Multi-Member Districts." AMERICAN JOURNAL OF POLITICAL SCIENCE 40: 129-44.

Brischetto, Robert R., and Richard L. Engstrom. 1997. "Cumulative Voting and Latino Representation: Exit Surveys in Fifteen Texas Communities." SOCIAL SCIENCE QUARTERLY 78: 973-991.

Page 471 begins here

Engstrom, Richard L., and Robert R. Brischetto. 1998. "Is Cumulative Voting Too Complex? Evidence from Exit Polls."STETSON LAW REVIEW 27: 813-833.

SHAW v. RENO, 509 U.S. 630 (1993).


Copyright 1999