Michigan and beyond: Ward Connerly keeps after race preferences.

GERALD FORD had been dead for just a day when the University of Michigan issued its official statement of mourning. President Mary Sue Coleman used the occasion not only to lament the loss of the man who is her school's best-known alumnus, but also to score an ideological point: "In recent years, and perhaps most importantly, President Ford was outspoken in his support for our diversity programs through our defense of affirmative action to the Supreme Court."

Perhaps most importantly? The University of Michigan is home to the Gerald R. Ford School of Public Policy and the Gerald R. Ford Presidential Library. But, in the mind of Coleman, these permanent presences pale in comparison with Ford's endorsement of racial preferences. Or perhaps she simply wanted to exploit the news of Ford's death to continue fighting what suddenly has become a losing battle for Ann Arbor's bean counters.

The university's defeat, coming at the hands of voters who overwhelmingly approved the Michigan Civil Rights Initiative in November, represents an astonishing victory for conservatives. It's a success story that upends the conventional wisdom of just three years ago, when it looked like racial preferences would survive far into the future. Today they are in jeopardy as never before.

By the end of the 1990s, opponents of racial preferences were on a roll. Inspired by Ward Connerly, voters in California and Washington passed ballot initiatives banning the consideration of race in public contracting, employment, and education. Judges were increasingly inclined to restrict the government's use of preferences, and federal rulings essentially eliminated them at the University of Texas and the University of Georgia. When the Supreme Court agreed to consider a pair of cases involving the University of Michigan, conservatives were hopeful that the justices would deliver a death blow.

Instead, they tossed out a lifeline. Sandra Day O'Connor wrote the majority opinion: "The Court expects that 25 years from now, the use of racial preferences will no longer be necessary." In other words, color-coded admissions had the high court's seal of approval for another generation. "That ruling alarmed me," says Connerly. "There were moments when I really thought about packing it in."

There had been other setbacks as well. In 2000, a Connerly-led referendum in Florida failed to qualify for the ballot. (Yet it may have succeeded in prompting Gov. Jeb Bush to phase out preferences in state universities.) Three years later, Connerly's Racial Privacy Initiative, which sought to ban all racial classifications by the California government, lost badly at the polls. The drive to enact colorblind public policies seemed to have run out of gas. When Connerly announced his intention to launch the MCRI, some of his allies harbored strong doubts: "I didn't think he could win, and I knew we couldn't afford another defeat," says Abigail Thernstrom of the Manhattan Institute.

A victory in Michigan promised to bring with it enormous symbolic value because of the Supreme Court cases, but it was also a risky move: The state leans blue and national trends suggested that conservative turnout would be down in 2006. What's more, virtually every member of the state establishment denounced the effort: Democratic and Republican officeholders, business leaders, labor bosses, pastors, and so on. "They claimed that breast-cancer-screening centers would be shut down and domestic-violence shelters would be challenged. They even had a radio ad that compared passage of the MCRI to the tragedies of 9/11 and Hurricane Katrina," says Jennifer Gratz, the campaign's executive director (and a plaintiff in one of the Supreme Court's Michigan cases). These opponents outspent Connerly, Gratz, and their allies by a factor of at least three. A state election board altered the ballot language in an unhelpful way. By the summer, polls suggested that support for the initiative had slipped beneath the all-important 50-percent threshold. A day before the election, the Evans-Novak Political Report predicted that it would "fail by a large margin."