California's October 2003 recall election caused a worldwide media frenzy as movie star Arnold Schwarzenegger unseated incumbent Governor Gray Davis less than a year after he had been re-elected. In this overview essay, we make three observations based on the California experience: first, the recall processes in some states are more easily utilized than in others; second, the California recall story vividly illustrates the critical role that money and professional consultants play in modern U.S. direct democracy; and third, holding the recall at the same time as the replacement vote creates considerable strategic complexity for the incumbent and raises some serious democratic issues for political scientists to ponder.

The initiative, referendum, and recall make up the triumvirate of direct democracy processes. The first two are well known but the recall has only recently received much attention as a result of the California election. The recall was introduced in the city of Los Angeles in 1903 and first used in that city in 1904. The process soon spread to surrounding cities and was very familiar to many localities by the time it was adopted for state-wide use in 1911 (Bird and Ryan 1930). While common at the local level, only 18 states currently have provisions for recalling a governor. California's rules are uniquely user-friendly. Some states—such as Minnesota—require a demonstration of some kind of malfeasance on the part of the official to be recalled, but California does not. Simple unpopularity is reason enough. The petition recalling Governor Gray Davis listed essentially political grievances such as “gross mismanagement of California Finances by overspending taxpayer's money, threatening public safety by cutting funds to local governments … and failing in general to deal with the state's major problems” as reasons to remove him.

California's process has the lowest signature threshold for placing a recall on the ballot—i.e., 12% of the vote for office in the last election. Most of the other recall states set the threshold at 25% or above, and six use eligible voters rather than actual voters as the threshold measure (Book of the States 2002). In allowing 160 days for the circulation of the petition California is also more generous than most states. In general, lower signature thresholds and longer periods of circulation lower the costs of using direct democracy processes and are associated with more frequent use (Bowler and Donovan forthcoming). California's recall election was in this sense an accident waiting to happen. All it took was for someone to recognize the opportunities that California's laws generously afforded.

In the end, 55.4% of voters voted to recall Gray Davis, the third governor subjected to a recall election (Lyn Frazier in North Dakota in 1921 and Evan Meacham of Arizona in 1987 being the other two).¹

Prior to 2003, California had seen 117 attempts to recall a state elected official since 1913, including 31 attempts to recall previous governors and a 1999 attempt to recall Gray Davis.² Given this track record, it is easy to see why Gray Davis did not initially take the recall seriously.

The 2003 recall started with discussions between long time anti-tax activists Ted Costa and Mark Abernathy. Gray Davis' 2002 reelection took place at a time of increasing budget problems, the depth of which became more clear after the election. Budget shortfall estimates gradually reached over $30 billion dollars, or about one-third of the state's total budget. Recall supporters charged that Governor Davis knew the depth of this budget crisis before the 2002 election, but only revealed it to the voters after he was re-elected.

The Costa and Abernathy phase of the recall campaign had a decidedly “grassroots” cast. They initially intended to use the Internet as the primary means to gather signatures. Petitions were designed to be readily downloaded and printed from the web, allowing activists to submit handfuls of signatures at a time. Their web site provided content to talk radio hosts who then, in turn, publicized the pro-recall web sites.

But by late April and early May, the grassroots effort had only gathered several hundred thousand out of the 900,000 signatures needed. Ironically, but perhaps inevitably, what turned this failing volunteer petition drive into a successful professional effort was Congressman Darrell Issa's $1.8 million contribution into a separate committee (‘Rescue California’). This infusion of cash enabled the hiring of paid signature gatherers to stand outside supermarkets and malls inducing people to sign the recall petition. The petition gatherers, who were paid between $1 and $1.25 a signature, virtually guaranteed the recall petition's success. It is a rough rule of thumb in California's direct democracy campaigning that $1 million will more or less guarantee enough valid signatures for qualification. While the grassroots efforts initiated the recall process, a purely volunteer strategy would have failed at the pace they were on at the time of the Issa infusion. The professionals were necessary for the petition's success and forced Governor Davis among others to take the recall prospects much more seriously from that point on.

The reliance on professionals is not unique or even unusual in California's initiative process, but extending it to the recall was innovative. The state has a growing number of political consulting, polling, and signature gathering firms ready to do direct democracy business during campaign season. The recall election was a fortuitous boost for California political consultants, providing lucrative work in what is usually the off-season for political firms. By some reports, almost $80 million was spent in 77 days of the election period, a total almost matching the $90 million spent over six months in the 2002 governor's race. Tribal gambling interests, along with labor unions, were major donors on the Democratic side, with some estimates putting tribal contributions as high as $11 million. While Arnold Schwarzenegger did not take money from these special interests, he accepted $2.4 million from real estate developers, $1.5 million from bankers and $1 million from car dealers (none of whom he apparently believed were special interests) on top of donating and loaning nearly $8 million of his own money (http://www.commoncause.org/states/california/). Gray Davis managed to raise and spend over $17 million, much of it from labor unions and state and local government employees.

With so much money to be made, it is hard to imagine political consultants passing up the temptation to recall more candidates in the future. While higher thresholds may dampen use of the recall in other states, the same cannot be said of California. Unless the state's rules are changed, the success of the 2003 effort and the obvious economic incentives for political consultants are likely to spawn more recalls in the future.

Once on the ballot, California's unique system of having a majority recall vote and a simple plurality replacement vote (which excluded the incumbent governor) created decisionmaking complexity in several ways. First, voters faced an overwhelming array of candidates in a genuinely multicandidate race. Aided by the low nomination threshold (65 signatures of support and $3,500) and the absence of the primary race hurdle, 135 people made the final list of certified candidates. In the end, most of them played little or no significant role in the election beyond some initial media coverage of the recall “circus,” especially after Arnold Schwarzenegger entered the race in an August 6 appearance on the Tonight Show. Immediately, he became the leading Republican candidate in public opinion polls. This caused the field of contenders to narrow further as other potentially significant GOP contenders saw the handwriting on the wall. On August 7, Darrell Issa dropped out followed two weeks later by Bill Simon and then Peter Ueberroth. This left State Senator Tom McClintock as the only major Republican rival to Schwarzenegger.

In theory, this left the Democrats in good shape. They had cleared out any significant Democratic challengers to the governor except for Lt. Governor Cruz Bustamante, who was the fallback option in case the recall succeeded. There was some initial concern that liberal Democrats might defect to Green Party candidate Peter Camejo and independent candidate Arianna Huffington, but in the end neither was a significant factor in the outcome.³

On the other hand, the Republicans had not been able to cull the field for Arnold Schwarzenegger. Republican leaders rightly worried that the two GOP candidates might split the vote and allow the Democrats to win. Given the unwillingness of either to blink, it was unclear to GOP voters which candidate to desert on election day in the interests of party victory. The fact that Schwartzenegger was able to defeat his Democratic rival despite a nine point Republican registration disadvantage and the presence of a conservative Republican rival with double-digit support was astonishing.

Second, and more problematic for the Democrats, was the structure of the ballot itself. The vote for the replacement candidate was held at the same time, and on the same ballot, as the decision to recall. Elsewhere, the replacement election is held separate from the recall vote or there is just one general election pitting the incumbent governor against the contenders. California's rules meant that the incumbent had to get 50% plus 1 to defeat the recall, but the successor could win with considerably less than a majority. This presented both a strategic and a democratic dilemma.

The strategic problem of whether to run a replacement candidate on the second part of the ballot divided Democrats throughout the campaign. The governor's advisors maintained until the bitter end that the mere presence of a viable Democrat on the second part of the ballot increased the chances of more Democrats voting for the recall (especially Latino Democrats who might have wanted to have a Latino governor). They worked hard to keep viable Democrats off the ballot, and when that failed they tried to keep resources and support from flowing to Cruz Bustamante. Bustamente's campaign slogan was “No on Recall, Yes on Bustamente.” Other Democrats argued for “No and No.”

This tactical confusion may well have resulted in voter confusion for Democrats because it gave them a “mixed message.” In addition, it divided the Democratic party into hostile and feuding camps of campaigners. Gray Davis and Cruz Bustamante seemed to spend as much time negotiating with each other over the roles and responsibilities of their respective campaigns as engaging the Schwarzenegger candidacy. Moreover, in the words of a key figure in the Davis campaign, it was harder to frame the recall as a “GOP power grab.”

The other problem California's simultaneous choice structure presents stems from the linkage between a majority recall vote and a simple plurality vote. In theory, this means that an incumbent governor could be defeated and replaced by a candidate with less support. This does not seem to have been a problem for California in the end judging from a pre-election Field poll that had Davis trailing Schwarzenegger and the large margin of the final outcome. But it could have happened, and it might have weakened the legitimacy of the incoming governor's victory. The prospects of changing this are not great, however, since generally it is harder to enact reforms that prevent prospective and hypothetical problems than to remedy problems voters have experienced first hand.

The other articles in this forum take up topics that surfaced during the campaign. Matt Barreto and Ricardo Ramirez address the question of whether minority voters were decisive to the final recall outcome. California's population is no longer majority Anglo, and the states' minority populations are important vote blocs that help determine election outcomes. In this election, under at least one plausible scenario, minority voters could have (even though they did not choose to do so) kept Governor Davis in office. In the late stages of the campaigns, allegations of sexual assault were leveled at the Schwarzenegger campaign. Despite the serious allegations, and the massive publicity they received, they had surprisingly little impact on the election itself. Lisa Garcia Bedolla and Carole Uhlaner try to explain the relatively small gender gap in this election and examine the reasons the sexual harassment charges had little of the expected impact.

In their study of voter behavior, Walt Stone and Monti Datta examine what the recall was about so far as voters were concerned: was it a referendum on the governor or was it a vote for change and innovation in the state's politics? Voting and the mechanics of voting were as central to this election as to the Florida result in 2000. Indeed, the dramatic intervention of the 9th Circuit Court of Appeals turned on interpretations of Bush v Gore. Michael Alvarez and his colleagues discuss the complexities of the ballot layout itself, especially the hurdles this placed before non-native speakers among the state's minority populations. Even without the problems of language, many voters had a hard time navigating the ballot to vote for their candidate. Henry Brady examines the role that voting technology played in shaping the outcome of this race and shows that, notwithstanding the legal arguments, the technology underlying the simple act of voting can shape the result in profound ways.

What does the recall mean for politics and political science more generally? Plainly, any California election will attract a great deal of attention if only because the state is home to one in eight Americans and the nation's largest economy. This election attracted even more attention than usual because of the recall's novelty and Arnold Schwarzenegger's fame. But the election raised in a concentrated time and in highly visible ways more fundamental questions about U.S. representation and electoral processes. The dramatic intervention by the 9th Circuit was reminiscent of the Supreme Court's role in the 2000 presidential election, prompting wide-spread debate again about the Court's proper role in electoral politics. Also, as in Florida in 2000, we were reminded that our capacity to count votes is still flawed even as we move to newer technologies. Still, we are better off depending less on punch cards. The issues of sexual harassment and driver's licenses for undocumented immigrants served to remind us that race and gender issues still matter in U.S. elections. Last, but by no means least, is the bigger question of whether increasingly professionalized direct democracy processes such as the recall are compatible with good representative government: will the prospect of being recalled improve the decision-making of elected officials, or will it sap them or their last ounce of political courage?