Louis Bolce, et al. Political Science Quarterly. Volume 108, Issue 2, Summer 1993.
During the spring of 1990 there was considerable speculation among journalists and political commentators that George Bush might attract unprecedented numbers of black voters to his Republican presidential candidacy in the 1992 election. Bush’s approval rating of 56 percent among blacks in March 1990—the highest for a Republican since Dwight Eisenhower—led Republican pollster (and later Bush-Quayle campaign chairman) Robert Teeter to suggest that Republicans might double their share of the African-American vote to about 20 percent.
Twenty percent was the threshold that GOP strategists, including the then National Chairman Lee Atwater, believed would establish Republican dominance in American politics—not only in presidential elections but at the congressional level as well. To become a majority party, Atwater argued, Republicans had to attract just a fifth of black voters—the most loyal component of the Democratic coalition—into what he called the Republican “big tent” and the “G.O.P. Rainbow.”
In 1988 George Bush had won 10 percent of the black vote—the same level of support that Republican presidential candidates had averaged over the six presidential elections since 1964. Republican candidates averaged 58 percent of the white vote since Johnson-Goldwater, ranging from 47 to 68 percent; whites gave Bush 58 percent of their votes in 1988. While Republicans made significant inroads into core Democratic constituencies such as white Catholics, blue collar workers, and white southerners, black voters had become more loyal to the Democrats.
The possibility of a Republican 20 percent solution led us to analyze the extensive exit poll sample of black voters surveyed by ABC News in the 1988 presidential election to assess its prospects. The study focused on those segments of the black electorate that conventional wisdom suggested might be open to a Republican message—higher socioeconomic status blacks, those who were politically and culturally conservative, and entrepreneurs.
There was precious little in the 1988 election data that pointed to “any segment of the black community that could be successfully mined for Republican votes.” Bush did garner majorities among some black groups—the tiny fraction (6 percent) who were either self-identified Republicans, among whom he beat Michael Dukakis 74 to 24 percent, or those who were both politically conservative and wealthy or elderly (both less than 2 percent of all blacks). Bush beat Dukakis 66 to 34 percent among black conservatives making $50,000 or more a year and 56 to 44 percent among black conservatives over age 60. Bush did take 29 percent of blacks who called themselves conservative (18 percent of blacks), but no other meaningful social category of blacks gave Bush 20 percent of its vote. Particularly telling was the invariantly low levels of Republican party identification among blacks—6 percent in the 1988 and 1984 elections.
The ABC News exit poll data graphically demonstrated that for blacks race was the variable explaining their vote, with demographics, socioeconomic status, and attitudes on issues and candidate qualities having at best marginal effects. With a handful of minor exceptions, African Americans overwhelmingly voted Democratic both for president and for Congress.
Bush’s relatively high approval ratings among blacks eighteen months (and nearly three years) into his administration did not appear likely to translate into higher levels of black support in the 1992 election, since approval ratings a year before a presidential election had been unreliable predictors of voting behavior in the past; there was no reason to believe that George Bush represented a departure.
Nonetheless, some Republican optimism regarding inroads among African Americans seemed justified as the 1992 primary election approached; in late October 1991, 43 percent of blacks nationally approved Bush’s job performance. Though down 34 points from the height of Bush’s black popularity in the aftermath of the Gulf War in March 1991, Bush’s approval among blacks was more than double the highest levels attained by Ronald Reagan. The nomination of Clarence Thomas to the Supreme Court as well as Bush’s signing the 1992 Civil Rights Act can be viewed as attempts to secure a larger share of the black vote (or as attempts to reassure his followers that he and they were not racists).
In the 1992 presidential and congressional elections was there an erosion of the black Democratic monolith? Did the existence of a substantial and highly visible third party candidate siphon off potential black support from George Bush (or Bill Clinton)? Did the absence of Perot candidates in House and Senate races affect the distribution of black and white voters at the congressional level? What impact did socioeconomic status, ideology, partisanship, economic perceptions, issues, and candidate qualities have in the voting decisions of blacks and whites? Which whites walked out of the Republican tent in 1992?
The Data
On 3 November 1992, George Bush received 10 percent of the black vote—exactly what he had received four years earlier—according to exit polls conducted by Voter Research and Surveys (VRS), the consortium established by ABC, CBS, CNN, and NBC after the 1988 presidential elections. VRS conducted a national exit poll of 15,490 voters leaving 300 polling places around the nation on election day; it also conducted 50 individual state polls with samples ranging from 459 in South Dakota to 2,268 in California. The national black sample was 1,539; the national white sample, 13,031. The sample sizes for the different attitudinal, demographic, and socioeconomic indicators vary, because respondents could omit answers on any individual question.
Exit polls have several qualities that make them valuable. One is their large sample size. The 1992 VRS exit polls, like those by ABC in 1988, permit extensive subgroup analysis about segments of the black community that standard data sets and pooled samples do not allow—for example, upper middle income (with annual incomes between $50,000 and $75,000) black conservatives. VRS’s 1992 black sample is as large as the National Opinion Research Center’s and Center for Political Studies’ entire national samples.
Another advantage of exit polls is the immediacy of the survey to actual voting. Exit polls query real voters, thus eliminating response bias, particularly the recall problem, in which respondents polled at some remove from election day report their electoral behavior with some inaccuracy. For instance, 70 percent of respondents in the 1988 Center for Political Studies (CPS) National Election Study said that they had voted in the presidential election that year, when in fact the turnout had been 51.4 percent. Still another benefit of VRS data is that the network consortium surveys representative state samples, thus permitting state level analysis, such as the distribution of white and black Perot voters in the New York U.S. Senate race.
1992: Blacks Remain in Democratic Camp and Whites Pull Up Republican Stakes
The 1992 story was much the same as 1988; the 1992 black presidential vote was almost entirely Democratic. Demographic, socioeconomic, residential, attitudinal, and behavioral factors were only loosely related to black partisanship and voting behavior. And this was true not only at the presidential level but was also reflected in the black vote for U.S. senators and representatives and governors. Blacks gave 87 percent of their votes to Democratic Senate candidates, 89 percent to Democratic House candidates, and 92 percent to Democratic gubernatorial candidates. In 1992, race was once again the predictor of the black vote. Whites were far more discriminating, giving a 54 to 46 percent margin to Democrats running for governor, a 51 to 49 edge to Democratic Senate candidates, and evenly splitting their House vote.
The presidential vote of whites is structured by ideology, partisanship, socioeconomic status, and issues, but this is rarely the case for blacks. Republican party affiliation is the only characteristic that translates into substantial black support for Bush and other Republican candidates. But even here the association is far weaker than for white voters. While the percentage of black self-described Republicans increased slightly from 6 percent in 1988 to 8 percent four years later, black Republicans gave Bush only 55 percent of their vote—13 percent less than their vote for Republican House candidates and 19 points less than they gave Bush in 1988. By contrast, 73 percent of white Republicans went for Bush in 1992, 17 percent voted for Ross Perot, and 10 percent for Bill Clinton. Nevertheless, Bush’s white support tumbled even farther from four years earlier—19 points—and with profound consequences, since whites cast 87 percent of all votes for president.
Even among self-described conservative blacks Bush fared poorly. From this group, which constituted 18 percent of all black voters in both the 1988 and 1992 elections, Bush received only 29 percent of the vote in 1988 and 26 percent tabular data omitted in 1992. Bush did not retain his 1988 black base; among blacks who said that they had voted for Bush in 1988 (20 percent of 1992 black voters), nearly half (49 percent) voted for Clinton in 1992. Another 13 percent voted for Perot. Only 38 percent remained with the president. By contrast, three in five whites who said that they had voted for Bush in 1988 (57 percent of 1992 white voters) stuck with the president; the rest split their votes between Clinton and Perot.
Economic well-being clearly influenced the white vote choice; it was far less salient for blacks. Bush received increasing pluralities of whites, as income rose from $30,000 a year; Clinton got pluralities of those making less than $15,000 and between $15,000 and $30,000 annually, doing best among the least well off.
If there is any segment of the black electorate with whom a Republican strategy might have had a chance of success, conventional wisdom suggests that it was with higher socioeconomic status blacks. High status blacks should have been tabular data omitted substantially more likely than lower status blacks to have identified as conservative and Republican and to support Republican candidates, because the GOP is widely seen as favoring economic and social policies more consistent with the class interests of more affluent and conservative voters. The reality appears to be otherwise in 1992 as it was in 1988. More than four-fifths of black voters in all income categories, except those earning in excess of $75,000 a year, voted for Clinton. And among the highest income blacks, Clinton got two-thirds of their vote, with Bush nearly a quarter. Perot did not receive double digit support among any black income group.
High income (those making over $75,000 a year) black Republicans (only one percent of blacks surveyed) went for Bush 53 to 21 percent over Clinton, with Perot getting 26 percent of that vote. Meanwhile, black conservatives (8 percent of blacks) while giving Bush 32 percent of their vote, gave Clinton nearly twice as much—58 percent. Like 1988, no other meaningful subgroups of the black electorate gave Bush 20 percent of their vote. Upper middle class ($50,000 to $75,000) black conservatives were more than four times more likely to vote for Clinton than Bush—74 to 17 percent. By contrast, upper middle income white conservatives went for Bush with a 60 to 25 percent margin over his Democratic opponent. In sum, upscale blacks overwhelming supported Clinton, while their white counterparts similarly backed George Bush.
Education and employment status influence white voting patterns but, again, appear almost inconsequential for blacks. Clinton won pluralities of the less and most educated whites, while Bush garnered pluralities of college attendees and graduates. At least four in five blacks at all education levels voted for Clinton. Bush won pluralities of white full-time workers and homemakers, while Clinton captured pluralities among white part-time workers, retirees, and the unemployed. Blacks gave Clinton a minimum of four-fifths of their votes irrespective of employment status. Indeed, Bush captured nearly three times the percentage of unemployed white voters than black voters who worked full time.
The Democratic bias of African-American voters and their differences with white voters are further demonstrated by examining retrospective voting patterns on economic matters. Voters cast their ballots retrospectively when they base their vote choice on their evaluation of past performance of the party/candidate controlling/occupying the White House. Retrospective voting occurs when those who have benefited economically vote for the incumbent party/president, while those who have suffered vote for the challenger. It is the most common form of issue voting. As in 1988, this expectation is borne out by white voters in 1992 but only marginally by blacks. VRS asked those leaving the polls whether they were better off financially than four years earlier, worse off, or about as well off. A third of whites said that they were worse off; 57 percent of those went for Clinton, 28 percent for Perot, and only 16 percent for Bush. Among the quarter of whites who said that they were better off, Bush won 64 percent of their vote, Clinton 21 percent, and Perot 15 percent. Whites whose financial situation was unchanged marginally went for Bush (44 percent compared to 37 percent for Clinton and 19 percent for Perot). For whites, at least, the guiding motto of the Clinton campaign—”It’s the economy, stupid”—proved in harmony with retrospective voting.
Blacks, no matter what their financial situation compared to 1988, cast more than two-thirds of their votes for Clinton. Those whose financial situation deteriorated did give Clinton his highest share (89 percent) and Bush his lowest (5 percent). But the 14 percent who were financially better off gave Clinton 69 percent, Bush just 26 percent, and Perot 5 percent. The 37 percent who saw no change in their financial situation went 79 percent for Clinton, 13 percent for Bush, and 8 percent for Perot.
Voter perception of the state of the economy as being in poor shape worked in Clinton’s favor among white and black voters, but, again, whites displayed far more variability. The nearly one-third (32 percent) of whites who thought that the state of the economy was poor gave Clinton 60 percent of their vote, while the near majority (48 percent) of blacks who held that view gave Clinton an overwhelming 91 percent. Whites who saw the economy as “not good” (48 percent) gave Clinton and Bush equal percentages of their vote (39 percent with Perot getting 22 percent). African Americans again overwhelmingly supported Clinton. Four-in-five (81 percent) of whites who saw the economy as “good”/”excellent” (20 percent of all white voters) went for Bush while the 8 percent of blacks who evaluated the economy positively still gave Clinton 53 percent of their votes.
Issues were far less meaningful as determinants of voting for blacks than for whites. VRS gave respondents nine choices of issues that “mattered most in deciding how you voted”. Among white voters, issues made a difference in their presidential choice. Whites choosing economy/jobs, health care, the environment, and education gave majorities or pluralities to Clinton; those citing the deficit gave Perot a plurality; and those naming taxes, foreign policy, and family values gave Bush majorities. Blacks, on the other hand, gave Bill Clinton a majority, irrespective of what issue was important in deciding their vote. The lone exception was foreign policy where the 2 percent citing it gave him a near majority.
Abortion, though not a central concern of most voters, provides further evidence that stands on issues make little difference on black presidential voting but clearly structure white vote choice. Blacks, irrespective of their position on VRS’s 4-point abortion scale, voted overwhelmingly for Bill Clinton: 85 percent who thought abortion should always be illegal voted for the Democratic candidate; 86 percent of blacks who believed that abortion should always be legal voted for Clinton. White voters were far more sensitive to the candidates’ abortion stands as it related to their own positions on abortion. They divided along the “hard”/”easy” lines defined by social scientists. Pro-choice voters (who want abortions “always” or “mostly” legal) went for Clinton (49 percent, with Bush at 29 and Perot at 22 percent). Those who wished to outlaw or at least substantially restrict abortion (making abortion “mostly” or “always” illegal) supported Bush (61 percent with 21 percent for Clinton, and 18 percent for Perot).
As with issues, the qualities of the candidates were far more influential with white voters than among the black electorate. Blacks gave majorities of their votes to Clinton irrespective of the candidate qualities that “mattered most” in deciding their vote. The only qualities that did not produce majorities that were overwhelming for Clinton were among the small proportions for whom “having~ the right experience,” “would have good judgment in a crisis,” and, to a lesser degree, being “honest and trustworthy” mattered. White voters were again far more discriminating. Bush got overwhelming majorities among those who felt that experience, crisis judgment, and honesty were important, as well as a large plurality among those who believe that “having~ strong convictions” was important. Clinton garnered majority support from whites who valued “caring~ about people like me,” “bringing~ about needed change,” “having~ the best plan for the country,” and “choice of vice president.”
Blacks voted as they had since 1964—overwhelmingly Democratic. There was no 20 percent solution in 1992. But not getting the votes of one-in-five African Americans did not cost George Bush the presidency; what did was the defection of white voters from the president.
Tabular Data Omitted
Many whites walked out of the Republican tent in 1992. George Bush’s share of the white vote declined from 58 percent in 1988 to 41 percent four years later. These defectors included members of groups that had become mainstays of the Republican presidential coalition for the last quarter century. Perhaps most striking, especially after the Houston nominating convention, which was widely seen as inordinately influenced by religious conservatives, was the erosion of white born-again Christian support for President Bush. In 1988, white born-again Christians comprised 14 percent of the white electorate, and George Bush won 81 percent of their vote. White born-agains made up an even larger percentage of the vote in 1992 (19 percent) and although they were still among Bush’s most supportive groups at 62 percent, nearly one in four abandoned the president. Bill Clinton did slightly better than Dukakis by winning 23 percent, with Ross Perot capturing 15 percent.
In the 1980s, the movement of born-again Christians to the Republican party was widely discussed in both the popular and scholarly press. A number of academic studies examining the relationship between conservative Protestant denominationalism and partisanship cautioned against identifying evangelicals and fundamentalists with the Republican party. The 1992 election results underscore the need to qualify the identification of religious conservatism with support for Republican candidates.
Protestants, generally, defected from Bush’s electoral coalition. In 1988, ABC found that “non-Catholic Christians” made up 56 percent of the white electorate, and Bush won 68 percent of them. In 1992, VRS categorized non-Catholic Christians into “Protestants” and “Other Christians,” who comprised 44 and 13 percent, respectively, of the white vote. Bush won pluralities of 47 and 45 percent of these groups. Clinton was not the beneficiary of these non-Catholic Christian defectors, winning 34 percent of Protestants and 29 percent of other Christians, compared to Dukakis’s 34 percent share of non-Catholic Christians in 1988. Perot captured 19 percent of white Protestants and 26 percent of the “other Christians.” Perot’s presence in the race gave these non-Catholic Christians a place to go and they went.
Catholics also moved away from Bush and the Republicans. In 1988 Bush and Dukakis split the white Catholic vote; in 1992 Clinton got 42 percent, while Bush got 37 percent, and Perot 22 percent. There has been debate among students of American electoral behavior over the drift of white Catholics in presidential contests. John Petrocik has argued that a decline in the Catholic contribution to the Democratic coalition can be traced to 1968, while Robert Axelrod sees the 1972 election as a turning point, as do Harold Stanley and Richard Niemi view the 1980 contest. Other scholars, such as Andrew Greeley and Henry Kenski, still view Catholics as disposed toward Democrats. In 1992 white Catholics appear to have retained a loyalty to Democratic candidates, though not in the same proportions as elections of a generation ago.
President Bush’s difficulties among the white electorate are underscored by several other groups that had been crucial to Republican electoral victories in the past: white independents, moderates, and conservatives. The pivotal independent vote— one-fourth of the white electorate in 1992—went to Clinton by a narrow margin (36 percent to 33 for Bush and 31 for Perot). In 1988 Bush won a clear majority (54 percent) of the independent white vote. Bush suffered among ideological moderates and conservatives as well. In 1988 Bush nearly split the white moderate vote with Dukakis winning 47 percent. In the three-candidate race in 1992, Clinton won a plurality with 44 percent to Bush’s 34 percent. Ross Perot garnered more than a fifth of the white moderate vote. While white conservatives were among Bush’s best groups (67 percent), he received more than four-fifths four years earlier. Here again, Ross Perot was a factor, capturing one-fifth of the white conservative vote.
A major problem for Bush and the Republicans in the 1992 elections was the substantial decline in the proportion of whites who described themselves as conservatives. In 1988, the modal ideological group was conservative (46 percent); in 1992 most whites called themselves moderate (41 percent). White conservatives, a core of the Republican electoral base, declined by more than a third from 1988 (from 46 to 31 percent). The drop among self-identified conservatives is particularly noticeable among whites in various SES and partisan categories. In 1992, unlike four years earlier, ideology was not related to income, education, or employment status. In 1988, more than half of whites earning in excess of $50,000 annually called themselves conservative; four years later, only 30 percent did. There was also a precipitous (40 percent) decline among white college graduates adopting the conservative label. Since conservatives are more likely to vote Republican than moderates or liberals, the 14 percentage point drop (64 to 50 percent) of white Republicans who identified as conservative in 1992 added to the erosion of the white conservative base and meant that Bush had a smaller pool to attract supporters from than he had in 1988. This sharp decline among self-described white conservatives in 1992 may be partly attributed to the connotations the label “conservative” took on with the rhetoric heard at the Republican convention in Houston.
First-time voters are clearly essential to any future alignment of the parties. One in ten voters was casting a first ballot in 1992. Ross Perot made an impact among white first-timers, capturing one quarter of their vote. Clinton managed a plurality, winning 41 percent to Bush’s 34 percent. Black first-time voters (17 percent) reflected vote choices virtually identical to those who had voted before (82 versus 84 percent for Clinton). Among whites who had previously cast ballots, Bush and Clinton were virtually tied (40 versus 39 percent).
Blacks in House and Senate Elections
It is the congressional not the presidential level that Atwater had in mind in his 20 percent solution for a new Republican majority. In 1992 Republicans maintained their fourty-three seats in the Senate and added nine House seats. Nearly nine-tenths of blacks (89 percent) voted for Democratic House candidates. As at the presidential level, no social, economic, ideological, or issue dimension with the exception of party identification mattered for blacks. All of these factors influenced the white vote, which split 50-50 between Democrats and Republicans.
Black Republicans—though two-and-a-half times more likely to vote for Democratic House candidates than white Republicans—cast 68 percent of their votes for GOP House candidates. Black Democrats as well as independents and those who preferred “other” parties overwhelmingly supported Democratic House candidates: 96, 80, and 77 percent, respectively.
Economic factors were of virtually no consequence in the House voting of African-Americans. Blacks, irrespective of their perceptions of their own financial situations, the state of the national economy, and who had different employment statuses, voted for Democratic candidates. Regardless of whether they saw their financial situation as better, worse, or the same during the Bush presidency, African Americans voted Democratic by better than four-to-one. Full-time black workers were almost as likely to support a Democrat as the unemployed, 85 to 92 percent.
The white vote for House, as for president, was structured by economic considerations. Financial situation was clearly related to voting, with Republicans winning 69 percent of those who saw their finances improving under Bush, 53 percent of those whose situation remained the same, and 31 percent of those whose personal finances deteriorated. Republican House candidates carried white full-time workers and homemakers, while the Democrats scored among white part-time workers, the unemployed, retirees, and students. (White unemployed voters were over two-and one-half times more likely to support Republican candidates than black full-time workers: 39 to 15 percent.)
Every black income and education group gave Democrats four-fifths or more of their votes. Among whites, income was linearly related to voting; whites who earned $30,000 annually and above voted Republican in increasing proportions. Education was curvilinearly related to white congressional voting, with Democrats doing best among the least and most schooled. The impact of ideology on black House choices was minimal; even self-described conservatives went for Democrats by 3 to 1. Ideology clearly informed the white vote: 79 percent of liberals went Democratic; 75 percent of conservatives voted Republican; moderates voted marginally Democratic, 54 to 46 percent. The same racial patterns held for Senate contests. Race is as important to black voting in congressional elections as it is for president. It is determinative for blacks; for whites, race is only one among many influences.
The Perot Factor
A pervasive question throughout the 1992 presidential election campaign was “Whom was Perot hurting?” VRS asked exit poll respondents, “If Ross Perot had not been on the ballot today, who would you have voted for?” White Perot voters would have divided evenly between Clinton and Bush, with 39 percent each, and Bush would have maintained his slim lead over Clinton among all white voters. Black Perot voters would have overwhelmingly gone for Clinton (58 to 16 percent), upping Clinton’s share of the total black vote from 83 to 87 percent.
Not surprisingly, Perot’s candidacy was particularly attractive to political independents and moderates. A plurality (39 percent) of white Perot voters described themselves as independents, as did one-third of black Perot supporters. White Perot supporters were marginally more Republican, while black Perot voters called themselves Democrats by two-to-one. Ideologically, majorities of both white and black Perot voters were moderates. However, white Perot voters were slightly more likely to call themselves conservative rather than liberal (27 to 21 percent); blacks were two-and-a-half times more likely to call themselves liberals.
At the congressional level, similar patterns hold. White Perot voters went marginally Republican in House contests, casting 52 percent of their ballots for Republicans and 48 percent for Democrats. Black Perot voters overwhelmingly voted for Democratic House candidates, 70 to 30 percent. The aggregate national Senate vote also shows white Perot voters giving Republicans a slight (51 to 49 percent) edge, while their black counterparts went strongly Democratic (64 to 36 percent). The Republican tilt among white Perot voters and the unremitting loyalty to Democrats by black Perot supporters is clearly seen in individual Senate races (those states which had significant black electorates). Majorities of white Perot supporters voted for Republican senatorial candidates in New York, Illinois, North Carolina, Georgia, Ohio, South Carolina, and in both California contests. In none of these states did the Republican garner a majority of black Perot supporters. Indeed, in all but Ohio, where black Perot voters gave Republican Michael DeWine 48 percent of their votes, blacks went for the Democratic candidate in landslide proportions. In Pennsylvania, both black and white Perot supporters split their votes between Arlen Spector and Lynn Yaekel. Only in Maryland did the Democratic candidate get lopsided majorities of both black and white Perot voters: Barbara Mikulski got 69 percent of black and 67 percent of white Perot voters against a black Republican opponent.
Since voting behavior in congressional elections can be used as a barometer of partisanship, the 1992 VRS election data indicate that below the presidential level white Perot voters were at least marginally Republican, while black Perot voters retained strong commitments to Democratic candidates.
No 20 Percent Solution (The Republicans Never Really Tried)
In our earlier study of blacks and the Republican party, we predicted that it was most “unlikely that the outcome of the 1992 election will turn on the number of black votes George Bush receives.” We added that even “should Bush win one in five black voters, there is no reason to believe that such support will be extended to Republican Senate and House candidates and into state executive and legislative races.”
The 1992 election did not turn on the number of black votes George Bush received. It was decided by the huge defection of white voters to independent Ross Perot. In 1988, George Bush had garnered 58 percent of the white vote (87 percent of all voters) compared to 41 percent for Mike Dukakis. This was enough to give Bush a 7.5 million vote margin over Dukakis despite Dukakis’s winning 88 percent of the black vote (8 percent of all voters). In 1992, George Bush’s share of the white vote (again 87 percent of all voters) had plummeted to 41 percent with 38 percent for Bill Clinton. Indeed, Bush would have needed to win over 70 percent of all black voters in order to compensate for his loss of whites to achieve a national popular vote plurality. Tellingly, Bush needed to have increased his share of the total white vote by 6 percent (conservatively assuming that all of the gain came from Perot’s white voters) to have beaten Clinton in the popular vote. Blacks (again 8 percent of voters) can be seen to have made the difference for Clinton in the popular vote; his 82 to 11 percent margin over Bush provided Clinton with a nearly six percentage point total edge over Bush. Clinton won the popular vote by five percentage points.
Since the president is not elected by the popular vote but by an electoral vote majority, what might have been the outcome had Bush attracted 20 percent of the black vote in key electoral vote states? Of the ten largest electoral vote states in 1992—California (54), New York (33), Texas (32), Florida (25), Pennsylvania (23), Illinois (22), Ohio (21), Michigan (18), New Jersey (15), and North Carolina (14)—Bush won Texas, Florida, and North Carolina with only 10 percent or less of the black vote. Of the Clinton states, two—Ohio and New Jersey—would have moved into the Bush column had African Americans given as much as a fifth of their vote to the president. Had Bush received 17 percent of the black vote, rather than the 7 percent he actually garnered in Ohio, and had he gotten a 21 percent share of the black vote in New Jersey, instead of the 11 percent of black ballots that were actually cast for him, he would have captured those states’ thirty-six electoral votes. In the remaining Clinton top ten states, Bush would have had to have won extraordinary shares of the black vote—45 percent in Michigan (he actually got 7 percent), 55 percent in Illinois (actually 5 percent), 92 percent in New York (actually 8 percent), an impossible 107 percent in Pennsylvania (actually 10 percent), and 133 percent in California (actually 9 percent).
Two other Clinton states with large black electorates would have gone to Bush—Georgia (thirteen electoral votes, 20 percent black) and Louisiana (nine electoral votes, 27 percent black)—if Bush had gotten 11 percent rather than 9 percent in Georgia and 16 percent instead of 7 percent in Louisiana, for a total of 58 electoral votes. Had all four of these switches occurred, Bill Clinton would still have been elected president, albeit with a smaller, though still fairly comfortable, 312-226 (as opposed to a 370-168) majority.
Bush did worse, sometimes far worse, among black voters in 1992 than 1988 in eight of the ten largest electoral vote states. (In 1988 Bush won nine of these states, New York being the lone deviation.) The two exceptions were in Texas where he upped his share of the black vote from 6 to 10 percent and in Ohio where he went from 5 to 7 percent of the black vote. His decline in the other states ranged from a 23 to 10 percent drop in Florida and a 19 to 9 percent slide in California to one percentage point decreases in New York (9 to 8 percent) and in Michigan (8 to 7 percent).
But, again, Bush lost key electoral states in 1992 that he won four years earlier because whites deserted the Republican party; the voting behavior of the black electorate had nothing to do with Bush’s change of fortune. In the six large states that Bush carried in 1988 but lost in 1992, Clinton’s share of the white vote was virtually identical to Dukakis’s four years earlier, deviating by only four percentage points in New Jersey, three in Ohio, and one point in the others. Bush’s white support dropped substantially from 1988—20 points in California and Ohio, 19 points in Illinois and Michigan, 18 in New Jersey and 17 in Pennsylvania.
The 20 percent solution did not happen, in part because it was never tried and would not have worked since Bush’s white base had so extensively eroded. The only serious Republican attempt to woo black voters was the nomination of Clarence Thomas to the Supreme Court. Such lack of effort is quite understandable; it didn’t require political seers to argue that “It is almost inconceivable that the Republican party would attempt to outbid the Democrats for the support of blacks since the G.O.P.’s central electoral strategy since 1964 has been to win over racially conservative Independents and Democrats.” Indeed, the central message of the 1992 Republican convention, particularly in the speech by Pat Buchanan, was that the Republican tent was an exclusive one. Milton Bins and Faye Anderson, the chairman and executive director of the Council of 100, National Organization of Black Republicans, contend that black Republicans and independents who were receptive to Bush’s message before the convention subsequently tuned out the president and his message because they “were offended by the racism and bigotry personified by Pat Buchanan.” They find hope in the fact that the president was able to hold 10 percent of the black vote despite a “stalled economy, the divisive and alienating rhetoric of the party’s right wing and single issue groups and the electorate’s overwhelming desire for change.” They also see positive signs in the strong showing of black Republican challengers in a number of congressional races and the fact that black voters were in part responsible for Republican Senators Alphonse D’Amato of New York and Arlen Spector of Pennsylvania narrowly retaining their seats. The fact that Clinton received a smaller percentage of the black vote (82 percent) than did Dukakis in 1988 (86 percent), Mondale in 1984 (90 percent), Carter in 1980 (85 percent) and Carter in 1976 (83 percent) offers them some small hope. Nonetheless, it is not at all clear whether the Republican party is ready to abandon its strategy of “adding~ to the base by subtracting,” which Bins and Anderson decry. It is equally unclear whether the Republican candidates “need” black voters at the presidential and congressional levels.