1. Introduction and Background
1. California's Blanket Primary Experiment
Bruce E. Cain and Elisabeth R. Gerber
There are some who tout physics as the model field for the social sciences, admiring its deductive rigor and explanatory power. But, plausibly, the more appropriate analogy is geology, in which unpredictable events such as earthquakes sometimes create opportunities to study more basic underlying processes and laws. The earthquake metaphor is particularly applicable to California, which has been the site of several dramatic political as well as geological ruptures in recent years. One of the most significant was the passage of Proposition 198, the Blanket Primary Initiative, in 1996. The blanket primary itself is not new (Washington has used it since 1938), but it is unusual (only three states including California have adopted it). Moreover, the rule changes that Prop 198 introduced reconfigured the electoral landscape in distinctive ways. Thrust upon the political parties and incumbent policy makers by an initiative that received little press attention until it passed, the move to blanket rules forced political actors to alter their strategies and tactics for the 1998 election; it rewarded some candidates and hurt others. But then, almost as abruptly, the blanket primary era in California was brought to a close when the U.S. Supreme Court ruled in June 2000 that it violated the political parties' associational rights.
The transition from a pure, closed primary system (i.e., in which only voters who are registered with a party can participate in its primary) to a blanket primary (i.e., in which voters are free to participate in any party primary), and then back again, creates a fortuitous natural experiment that allows us to compare behavior and outcomes before and after an electoral rule change. Political scientists contend that rules and institutional structures matter because they determine the incentives and opportunities that voters and candidates face. Change the rules, and the incentives and opportunities
The California blanket primary experiment raises two different kinds of questions. One type, which is dealt with only briely in this volume, is legal and normative. These issues were central in the lawsuits directed against Prop 198 (California Democratic Party v. Jones U.S. District Court, 9th Circuit, Eastern Division of California, CIV-S-96–2038 DFL PAN), culminating in the Supreme Court's decision (530 U.S. 567 2000). They were also raised in the earlier Washington and Alaska blanket primary cases (Heavey et al. v. Chapman [1980] 93 Wash. 2d 700; O'Callaghan et al. v. Alaska 914 P. 2d 1250 [Alaska 1996]) and concern such questions as whether parties have the exclusive right to determine their rules for choosing nominees, whether primaries are public elections or private party affairs, whether the legislature or the initiative process can impose these rules over the opposition of the offcial party apparatus, what standard of constitutional scrutiny the Court should apply, and so forth. Even more basically, legal scholars are divided over whether political parties are private associations or instruments of state action and whether the courts should be deciding issues of such a political nature. Political science has much to contribute to a discussion about these issues, and they are discussed ably in Nate Persily's chapter. However, they are not the primary focus of this study.
Rather, our orientation is toward the empirical and analytic tasks of (1) describing and relating the behavior of political actors to the changes in electoral rules (the micro considerations) and (2) discussing the aggregate features of systems that use different types of rules (the macro considerations). Understanding the micro behavior of political actors—voters, candidates, and party elites—is critical to understanding the causes behind the macro differences that rules produce. Rules condition strategic and rational choices, leading to certain behaviors. As the burgeoning political science literature on strategic voting demonstrates, voter choices are not completely independent of rules and expected outcomes (Cain 1978; Black 1978; Cox 1994, 1997; Niemi, Whitten, and Franklin 1992; Gerber and Morton 1998, 2001). For example, political scientists have come to appreciate how particular forms of strategic behavior (i.e., the desire not to waste a vote) underlie the two-party equilibrium in single-member, district-based systems. When rules change, choices that produced favorable outcomes may no longer make sense, and actors adapt their behaviors in response. The aggregate consequence of those new behaviors is a system with different features than the previous one: higher or lower turnout, more or less moderate candidates, increased or diminished costs, and the like.
In order to take advantage of this unique opportunity to study a natural political science experiment, we assembled a distinguished group of political scientists with expertise in various aspects of elections. In the chapters
MICRO QUESTIONS
With only a few exceptions, most states use the direct primary as their predominant mechanism for selecting the state and national candidates who will run in general elections.[1] The forms of direct primary include the closed primary, in which voters must register some specified period of time before the election with the party whose primary they intend to participate in, and several types of open primaries, in which some or all voters are free to select the primary ballot or race they wish to participate in on election day. One of the most "open" primary systems, in the sense of placing the fewest restrictions on participation, is the blanket primary (see Gerber and Morton 1998). Under the blanket primary rules, voters receive one ballot with all of the various party nominees on it. They can then cast up to one vote for any of the nominees for each office, regardless of the partisanship of either the voters or the candidates. Hence, voters are free to vote in different party primaries for different offices: the Democratic primary for Governor, the Republican primary for U.S. Senator, and the like.
The lack of restrictions on participation in blanket primaries creates incentives and opportunities for voters to engage in different patterns of behavior than under closed primaries. Specifically, we expect the blanket primary to increase incentives for voters to engage in crossover voting, where crossover voting is defined as voting in a party primary other than the one the voter identifies or is registered with. This expectation follows from a general opportunity-cost argument. For instance, if a voter is registered as a Democrat, a closed primary system requires the voter to reregister with another party if she wants to participate in that party's primary. A hypothetical Democrat who wants to vote in the Republican primary must register in advance as a Republican. If this person is primarily interested in just one Republican race, she forfeits the opportunity to vote in the other Democratic primary races in order to participate in the one Republican Party primary of interest. This is a high opportunity cost. In addition, if the voter wants her Democratic affiliation back, she must pay the cost of reregistration at a later time. Under a blanket primary, this same Democratic voter would not have to re-register to participate in the Republican race of interest, nor would she lose the opportunity to vote in the other Democratic races. In short, lowering the opportunity costs of participating in another party's primary should increase the probability that voters will engage in crossover voting, ceteris paribus. Therefore, the first micro questions addressed in the chapters by Sides, Cohen, and Citrin (chapter 5), Alvarez
Beyond the question of the extent of crossover voting is the more difficult problem of motivation. Several of the studies in this volume address the question of why people cross over. To better understand this discussion, it is useful to make certain definitional distinctions about the various forms of voting behavior. The first form we consider is sincere voting. A person who votes sincerely votes for his most preferred candidate, regardless of how this choice may affect strategic opportunities later in the electoral process. In other words, if a voter were to rank order the candidates for a given office and then cast a ballot for the one with the highest ranking, the voter would be voting sincerely. Sincere crossover voting, then, occurs when a voter votes for his first-choice candidate who happens to be running in another party's primary; hence, a Democrat who most prefers a Republican candidate for Governor and who crosses over to vote for that candidate in the primary is voting sincerely. We should note that not all voters who vote in their own party's primary are voting sincerely. A Democrat who most prefers a Republican candidate but who chooses, for whatever reason, to vote in the Democratic party's primary is not voting sincerely.
In addition, there are several forms of strategic voting. Strategic voting involves voting for a candidate who is not the voter's most preferred candidate—that is, voting for any candidate who is ranked lower than number one. Strategic voting can be divided into several types based on the voter's motivation. When voters choose a candidate who is the most preferred of the candidates in another party, but is not their highest ranked candidate overall, they are engaging in a form of strategic behavior known as hedging. One situation in which hedging might arise is when voters believe their own party's nomination is virtually certain and the general election outcome is competitive, and they hedge their chances by voting for the most palatable of the other party nominees, just in case their most preferred candidate were to lose. A second reason for hedging is that voters may expect their most preferred candidate to lose in November and so want to make sure that at least their second choice is nominated and elected. From a normative perspective, we term this benign strategic voting because the voters' intentions are not mischievous.
Other forms of strategic voting are less benign, particularly when the motive is to saddle the other party with a weak candidate. This form of strategic behavior is sometimes called raiding or sabotage. The following is an example of raiding. Suppose, as before, a voter knows that the outcome in her own party's primary is a foregone conclusion, and rather than waste a vote in that contest, the voter decides to cross over to participate in the other party's primary. If the voter crosses over to support the weakest candidate
The second micro question thus centers on whether the blanket primary encourages sincere voting, hedging, raiding, or some combination of the three. The relative frequency of each behavior has important implications for how we assess the manipulability of the electoral system. A finding that all or nearly all voters are sincere suggests that there is little need to worry about manipulated outcomes. However, if raiding is prevalent, then it raises serious questions about how robust elections are against attempts by elites or groups in the electorate to produce outcomes that are not majority preferred. For example, imagine a situation in which a candidate defeated by raiders in the primary could have been the eventual winner in November. In such a case, the majority will is frustrated by the strategic behavior of a voting bloc.
Another possibility is that hedging rather than raiding may be prevalent. Widespread hedging raises questions about the power of voters and elites in uncontested primary races to infiuence the outcomes of contested races in other parties. Is a voter's infiuence a function of competitive circumstancefi Does this in turn discourage competition in party primaries so as to make the party less vulnerable to crossover votersfi The chapters by Sides, Cohen, and Citrin; Alvarez and Nagler; Salvanto and Wattenberg; and Kousser (chapter 8) all take up the question of who crosses over and why. In addition, Bowler and Donovan (chapter 3) consider the explicit motives of those who voted for Prop 198, asking whether those who would benefit from the rule changes voted for them.
The impact of crossover voting is most obvious in the primary since it may affect which candidates are nominated. Crossover voting may also have important consequences for the general election, either directly, in terms of deciding the candidates that ultimately face each other in November, or indirectly, in terms of the propensity of voters to cross party lines a second time. If crossover voters remain loyal to the candidates they support in the primary, then candidates who attract crossovers in the primary may gain a lasting advantage in terms of increased general election support. Chapters 5 and 8 specifically consider some of these long-term effects.
A third micro question deals with turnout. If, as rational voting models suggest, the probability of voting is a function of expected vote impact (i.e., closeness of the race) and enthusiasm for the candidate (i.e., proximity of the candidate to the voter's ideal point), then the blanket primary might affect the likelihood of participation on the margin. For instance, if a voter is stuck in an uncompetitive primary under closed rules, the incentive to participate might be low; but if the voter is given the opportunity to cast a
The micro level effects of the blanket primary are not limited to voters. Candidates, their supporters, and their consultants are also affected by rule changes. Previous research has shown how contribution patterns are shaped by expectations of a candidate's prospects and resources (see McCarty and Rothenberg 2000 for a recent discussion of contributor motivations). Since the blanket primary changes campaign dynamics and potentially the electoral fortunes of different kinds of candidates, we would expect contributors to anticipate and perhaps amplify these changes. Chapter 9 also considers whether individual contributors adapt their behavior to the expanded potential voter base created by the blanket primary.
Under closed primary rules, the universe of voters is defined as the subset of the electorate that is registered with the candidate's party and is eligible to participate. Under blanket primary rules, the universe of voters is larger and includes all registered voters, regardless of their party. Therefore, different campaign strategies might make sense. Furthermore, whatever method of campaigning the candidate and consultants choose (e.g., mail, voter contact, registration, get-out-the-vote, etc.), the blanket system increases the set of voters that candidates must appeal to. Increasing the pool of potential supporters may affect both the cost and the content of campaigns. It potentially affects campaign costs in the sense that the larger set of voters might, on average, require greater overall expense, or it could mean that campaigns are able to spend less per capita while appealing to a larger set of voters. It affects campaign content in that the blanket primary widens the potential ideological space by adding to the primary electorate voters from parts of the spectrum that are not well represented in a given party. Many of these questions are taken up in the chapters on campaign dynamics and strategies by Gerber and Petrocik (chapters 10 and 14, respectively).
MACRO CONSIDERATIONS
Rule changes affect the behavior of individual political actors, and these behaviors, in turn, alter aggregate patterns of behavior and outcomes. Much of the popular debate over Prop 198 focused on these macro questions: in essence, whether the electoral system would be improved as a result of adopting a blanket primary. As a means of understanding what people thought the macro consequences of the blanket primary would be,
People's expectations aside, what have been the actual macro-level consequences of adopting the blanket primaryfi We first consider whether changes in voter behavior under the blanket primary altered the outcomes of any primary and/or general election races. Did the blanket primary lead to outcomes that would not have been produced under closed primary rulesfi Specifically, is there compelling evidence that crossover voting was decisive in any races, particularly the most competitive onesfi Several of the chapters on crossover voting suggest that, in at least a handful of races, the extent and direction of crossover voting was, in fact, sufficient to alter election outcomes.
Second, we examine whether the blanket primary reinforces the moderating features of California's electoral system. A common criticism of the closed primary is that it produces candidates who deviate from the median voter because they must appeal to a primary electorate with extreme preferences. This may result in the nomination of either candidates who themselves hold extreme preferences or more pragmatic candidates who are nevertheless constrained to maintain positions that are consistent with the ones that got them nominated. In either case, the closed primary may result in candidates whose positions diverge from the vote-maximizing median position in the general election. In theory, the blanket primary forces candidates to take more moderate positions in the primary, because they must appeal to a wider range of the ideological spectrum. To the extent that candidates respond in this way, the blanket primary should produce more ideologically moderate candidates. This moderation hypothesis is taken up in detail in chapter 10 by Gerber on candidate positions.
The other side of the moderation hypothesis, however, is that groups with interests that diverge from those of the median voter may be less well represented under the blanket primary. This could have the important result that women and minorities, who often hold minority positions within the general electorate but have been able to gain access to political positions by cultivating constituencies within their parties (especially the Democratic party) may be less able to compete for their parties' nominations. On the other hand, to the extent that women and minorities can appeal to voters across party lines, the blanket primary might have the opposite
Third, we consider the system-level effects of campaign dynamics under the blanket primary. We noted in our discussion of micro considerations that by expanding the universe of voters, the blanket primary might have altered the strategic decisions of candidates and consultants. Under the changed strategic environment of the blanket primary, candidates may have made different decisions about when to enter elections and in what capacity. As a consequence, we may observe different patterns of electoral competition. Chapters 7 and 9 (by Salvanto and Wattenberg, and Cho and Gaines) both consider patterns of candidate entry and the consequences of these decisions for electoral competition. Candidates may also allocate resources in different ways, further affecting campaign competition and costs. Cho and Gaines consider whether micro-level changes in campaign tactics, strategies, and dynamics contribute to a macro effect of higher overall costs in the primary and general elections.
Finally, we consider the implications of the blanket primary for the health of the party system as a whole. Minor parties have been a persistent, if understudied, feature of California politics. The main purpose of minor parties and minor-party candidates is not to win office per se, but rather to provide voters with an important outlet for voicing alternative political views. How did the minor parties fare in the blanket primaryfi Did the crossover vote draw primary voters away from the minor parties, or did it increase the minor-party vote by allowing major-party voters to shop in minor-party electionsfi In his chapter on minor parties (chapter 11), Collet considers these possibilities.
The purpose of the chapters that follow is to analyze in greater depth political behavior and outcomes in California's first blanket primary. For several reasons, however, we caution the reader to interpret these results with some care. First, since California held only two primaries (and one corresponding general election) under its blanket rules, and since we expect that voters, contributors, candidates, and other political actors would continue to adapt their behavior as they learned more about the strategic opportunities created by these rules, the dynamics observed in the 1998 elections refiect perhaps only a small part of the maximum changes that the blanket primary would eventually have brought if California had employed the system for a number of years. Second, as Baldassare shows in chapter 4, the economic, political, and social context in which this election took place was unique in several respects and may mask important effects of the blanket primary. These caveats aside, however, we do feel that these chapters provide a valuable first cut in exploring California's important political
NOTES
1. To nominate major-party candidates for statewide, state legislative, and congressional offices, nearly all states use some form of direct primaries. Thirty-eight states require major parties to use direct primaries exclusively, three allow primaries or conventions, three require preprimary conventions, and the rest allow preprimary conventions under certain circumstances (Bott 1990, 173–5). To select delegates to the 2000 national party conventions (which nominate presidential candidates), forty-three states used presidential primaries to select some or all of the delegates for at least one party, while the others used party caucuses. Rules regarding ballot access for minor parties and the nomination of minor-party candidates vary a great deal from state to state. [BACK]
REFERENCES
Black, Jerome H. 1978.
“The Multicandidate Calculus of Voting: Application to Canadian Federal Elections.”
American Journal of Political Science22, no. 3: 609–38. 
Bott, Alexander J. 1990. Handbook of United States Election Laws and Practices: Political Rights. New York: Greenwood Press.
Cain, Bruce E. 1978. "Strategic Voting in Britain." American Journal of Political Science 22, no. 3: 639–55.
Cox, Gary W. 1994. "Strategic Voting Equilibria under the Single, Non-Transferable Vote." American Political Science Review 88, no. 3: 608–21. 1997. Making Votes Count: Strategic Coordination in the World's Electoral Systems. Cambridge: Cambridge University Press.
Gerber, Elisabeth R., and Rebecca B. Morton. 1998. "Primary Election Systems and Representation." Journal of Law, Economics, and Organization 14, no. 2: 304–24. 2001. "Electoral Institutions and Party Competition: The Effects of Nomination Procedures on Electoral Coalition Formation." Unpublished working paper, University of California, San Diego.
McCarty, Nolan, and Lawrence S. Rothenberg. 2000. "The Time to Give: PAC Motivations and Electoral Timing." Political Analysis 8, no. 3: 239–60.
Niemi, Richard; Guy Whitten; and Mark Franklin. 1992. "Constituency Characteristics, Individual Characteristics, and Tactical Voting in the 1987 British General Election." British Journal of Political Science 23, no. 1: 131–37.
Wolfinger, Raymond E., and Steven J. Rosenstone. 1980. Who Votes? New Haven, CT: Yale University Press.
2. Crossover Voting before the Blanket
Primaries versus Parties in California History
Brian J. Gaines and Wendy K. Tam Cho
The passage of Proposition 198 was in plain defiance of the preferences and advice of most elites, including, notably, both of the major political parties. In this chapter, we brieflytrace the chronologyof primaryelections in the Golden State, with an emphasis on how theyhave been intertwined, from the beginning, with an anti-party spirit. We thus orient the blanket primary, as delivered by direct democracy, in a distinctive state political culture of independence from, and ambivalence or even hostilitytoward, political parties. We then focus on the parallels between voting options under the blanket primarylaw and those presented in the mid-twentieth centuryduring the multi-(or "cross-") filing era. This comparison provides a historical baseline against which to place California's experience with the blanket primary.
The easy passage of Proposition 198 can be understood as yet another instance of defiant populism, wherein a healthy majority of California voters thumb their noses at elite advice and embrace a measure whose appeal is simple and whose alleged flaws evidently do not trouble many. As such, Proposition 198 followed in the wake of numerous other initiatives, the most famous of which is probably still 1978's Proposition 13, a property tax-freezing measure that is sometimes said to have set off a nation-wide "tax revolt" (see, e.g., Kettl 1992, 58, 152). But the change from a closed to a fullyopen primary originated in something more specific than populism: it was clearlyan act of anti-party populism, which tapped a sentiment that dates from the founding of the countryand has been especially strong in California.
The blanket primary was a novelty in California's electoral history, but it was closely related to a prior innovation in primary electoral law, crossfiling (aka multifiling). Among American states, only California and New
Initially, multiple filing in California followed this same general pattern, with some Republicans and Democrats seeking to bolster their appeal by adding Progressive, Socialist, or Prohibition nominations to their principal partydesignation. After all of these minor parties receded to the fringe in the 1920s, multifiling took a new shape: "double-filing" in the two major party primaries became the modal practice for incumbents and an occasional tactic for challengers and open-seat contestants. The direct impact was that, even though the primaries were closed, in manyseats, registered Democrats and Republicans were able to vote for candidates from their own party or for candidates from the other major partywho had cross-filed. This is, more or less, what opening the primaryunder the blanket system achieves as well. The cross-filing period, then, is a natural era to examine for insight and clues into what we might expect in the situation of a blanket primary.
This chapter explores the comparison between these two distinct but obviously related electoral systems. We first provide a brief background discussion of primary elections in American political history and the original passage of a primary election law in California. No sooner had California inaugurated a system of party-controlled nominations than candidates and voters invented a means of bypassing party control via cross-filing. We discuss the significance of this practice, and present evidence on its incidence and on how it affected voting. Our goal, at this stage, is to explore the similarities and dissimilarities between multifiling primaries and blanket primaries, in an effort to extract the history most relevant to those whose interest lies in how the blanket can be expected to settle once it has been draped over a state. Finally, we review the multiple stages by which the blanket primary became law and then, immediately, was modified, yet again uncovering a tale of partisan elites versus the partywary masses.
ORIGINS AND IMPLICATIONS OF PRIMARY ELECTIONS
The primary election is an early-twentieth-century "progressive" invention. Primaries were adopted in many American states as a substitute for conventions, in an effort to pry power away from strong party bosses and machines and relocate it with candidates and the mass electorate. In California, as in some other states, much of the controversy surrounding the adoption of direct primaries concerned provisions for electing U.S. Senators. The first attempt to establish mandatory primaries in California occurred in 1897. That act fell to court challenges without actually being implemented, as did similar legislation passed in 1899 and 1900 (Young 1943, 117–18). Since all of these legislative efforts preceded the advent of popular election of U.S. Senators (i.e., the Seventeenth Amendment), they were regarded as being partly—or even principally—back-door means of transferring the power to elect Senators from state legislators to the general public, and were applauded or abhorred accordingly. Oregon had set an early example of how a direct primary could achieve this exact purpose of tying the state legislature's hands, and so there was growing popular support in California for some form of primary. By 1908, a constitutionally satisfactory approach was found: ACA 3, which directed the legislature to "enact laws providing for the direct nomination of candidates for public office," thereby amending section 2½ of Article II of the Constitution, passed both houses bymore than the two-thirds required for constitutional amendments (State of California 1920, 74). At the general election of 1908, the public approved ACA 3 by76.6 percent to 23.4 percent, with every county favoring the amendment by a lopsided margin.
The 1909 session of the California legislature then opened with a three-month-long battle between a bipartisan reform (pro-primary) coalition and a bipartisan pro-machine (anti-primary) coalition. Some type of primary seemed inevitable, but opponents strove to craft either an innocuous law that would change nothing or an unconstitutional one that would be struck down bythe courts before it could take effect. Following a tortuous chronology of amendments and reconsiderations, several near-tie roll calls, a week of deadlock in the Senate (maintained by hourly postponements!), plus various reversals of position by individual legislators, the law that was ultimately passed did not allow for de facto election of U.S. Senators, but did establish closed, partisan primaries for other offices (see Hichborn 1909, 68–120, for a detailed, if clearly biased, account of the bill's passage). This legislation emerged from a populist movement that spanned the major parties, uniting the members who were most likely to vote against entrenched interests (particularly big business) and to vote for major social reforms and institutional innovations that promoted more transparent government.
Since 1910, then, California's state and national elective offices have been filled byelections occurring in two rounds. In the spring or summer, the major (and minor) parties hold primary elections to select nominees for the state's elective offices. Primary winners then compete in November for elevation to the office in question. While the advent of the primary weakened central party organizations by admitting the mass electorate into the nomination process, the closed nature of the primary strengthened parties in the electorate. That is, onlythose registered with a given party were permitted to vote in that party's primary, and so party membership (by way of registration) became important. Moreover, part of the compromise by which the primary was passed was the inclusion of rules intended to ensure that candidates were authentic partisans. A candidate was required "to make affidavit ‘that he affiliated with [the party whose nomination he sought] at the last preceding general election, and either that he did not vote there at, or voted for a majority of the candidates of said party at said preceding general election, and intends to so vote at the ensuing election'" (California Statutes 1909, 694, quoted in Hichborn 1909, 72).
The closedness of the primary and the affiliation pledge required of candidates were plainly at odds with the nonpartisan, or (perhaps more accurately) anti-party, spirit that was sweeping the state at that time. Enough Progressives were carried into office on Hiram Johnson's coattails in 1910 that the legislature was able, after the passage of only two election campaigns, to do away with the affiliation pledge requirement and to make explicit the right of candidates to seek multiple party nominations without regard to their own registration or voting history. "Cross-filing," whereby candidates ran in two (or more) primaries, quickly became normal, first for state races, then for congressional races. Indeed, the first election in which candidates could file for multiple party nominations was 1914, and nearly half the members of the 1915 Assembly won their seats holding the nominations of more than one party.[3]
By 1915, the state had made local elective offices nonpartisan, and Governor Johnson was urging the legislature to follow suit by making its own elections nonpartisan as well (see Hichborn 1922, 221–22, for the text of Johnson's biennial message to the legislature in 1915). In 1910, Johnson had won a five-way Republican primary before besting Democratic, Socialist, and Prohibition candidates in the general election. In 1914, however, he opted to run only as a Progressive, eschewing major-party designation despite the fact that Progressives accounted for less than 20 percent of all registered voters. His personal popularity was so great that he nearly won a majority of the vote all the same. In yet another five-way race, he took 49.7 percent of the vote, with the Republican and Democrat finishing 20 percent and 37 percent behind, respectively. Johnson was much less successful
The primary remained inextricably linked to nonpartisanship, and both were subjected to popular approval in a 1915 special election, wherein the key item was a measure that would make the primary, and hence the operation of the legislature, nonpartisan. By Hichborn's account, this election "more than anything else seemed in its results to voice a protest at calling special elections for the consideration of such matters" (1922, 222). Turnout was low, and all eleven measures submitted were defeated, even the ones against which no opposition campaign at all had been waged. As a consequence, the legislature did not establish a nonpartisan primaryto do awaywith its partisan organization.
In the 1917 session, moreover, the legislature amended the Direct Primary Law in the opposite direction. On the justification that the special election had revealed little support for ridding the state houses of parties, state legislators added a new requirement that a candidate could not win the nomination of a partyother than the one with which he was registered without also winning his own party's nomination. In the August 1918 primaries, this precise scenario played out at the top of the ticket, as James Rolph Jr., a registered Republican, took 45.6 percent of the Democratic vote, compared to 36.9 percent and 17.6 percent for two rivals, but finished second in the Republican race, with 39.2 percent against 45.1 percent (the balance being distributed among four others). Rolph was thus defeated as the Republican nominee for Governor and disqualified as the Democratic nominee, and so no one was permitted to run under the Democrat banner in the general. Before the 1920 elections, still further amendments empowered county party committees to fill vacancies created in this manner byappointment, just as theycould fill vacancies created bydeath or byan absence of anyfilers for the primary. So it was that, under special circumstances, party organizations again took control of nominations, in stark violation of the spirit of the primary.[4]
Over the next decades, scarcelya legislative session passed without some proposal to amend the direct primary. Efforts to extend the anti-party elements bymaking the primaries (and, hence, the legislature) nonpartisan were about as common as counterproposals to forbid cross-filing altogether and revert to basic closed primaries in which registered Republicans select Republican nominees, registered Democrats select Democratic nominees, and never the twain do meet.
Neither side triumphed in extending or defeating cross-filing principles until 1952, when two propositions on the general election ballot competed to change the status quo. Proposition 13, an initiative, would have prohibited cross-filing by specifying that "no person shall be a candidate or nominee of a political party for any office unless he has been registered as
The other shoe dropped in April of 1959. The 1958 elections delivered the first unified Democratic control of California of the twentieth century, and Governor Pat Brown and the legislature wasted little time in abolishing cross-filing. Media coverage of the bill at the time emphasized that crossfiling was widely thought to favor Republicans. The New York Times opined, "California's Republicans, outnumbered [in registration] for the last twenty-seven years, capitalized on cross-filing to maintain a half century's domination of state politics" (Hill 1959a). Closer to the action, a report in the Los Angeles Times on passage of the repeal bythe State Senate emphasized that "abolition of cross-filing has been a Democratic Party objective for years" ("Legislature Repeals Cross-Filing in State," 1959). The roll calls reflected that same partisan story. The Assembly passed AB 118 on February 24, 1959, with forty-five Democrats and four Republicans voting in favor against one Democrat and twenty-eight Republicans (one Democrat and one Republican abstained). In the Senate, the bill passed twenty-two to fifteen, twenty-one Democrats and one Republican defeating four Democrats and eleven Republicans (two Democrats and one Republican abstained).[5]
The interpretation that the Democrats of 1959 were finally in a position to close a pro-Republican loophole is complicated just slightly by the fact that some prominent Republicans expressed support for the change in primary law, and some Democrats lamented the end of the cross-filing era.
Only one member of California's congressional delegation spoke out against the repeal, Senator Thomas Kuchel (R), who struck a very Progressive tone: "The people of California are not extremely partisan—they want clean, strong, and honest government. In my judgment, cross-filing has helped them get it" (Shannon 1959). Moreover, the contention that cross-filing was somehow inevitably favorable to Republicans was rarely backed with any explicit logic. From the advent of party labels in 1954 to 1958, there were eleven instances of Democrats winning the Republican nomination for U.S. House seats, against no Republican victories in Democratic primaries. It seems plausible that cross-filing might have favored incumbents, in so far as simple name recognition ensured them an advantage over challengers among the less intensely partisan registered voters. If this were so, it would mean that the ability to cross-file worked to the advantage of the Republicans when theywere the stronger state party, but started to help Democrats, on balance, after the state began to lean their wayin the late 1930s. We shall return to this question in the next section.
Figure 2.1 shows three indirect indicators of the prevalence of multiple filing for U.S. House and California Senate contests between 1910 and 1964.[6] Before and after the period in which cross-filing was permitted (1914–1958), general election winners almost invariablyheld onlyone party's nomination, the exceptions being candidates who won extra nominations by write-in campaigns. Bycontrast, the dashed line shows that, on average, winners held about 1.5 nominations in this era, and in 1918 actually averaged more than 2 partynominations each. The average number of candidates in competition for California's U.S. House seats was high at the outset of the era, when the Progressive, Socialist, and Prohibition parties were still active. However, it quicklyfell, as fewer and fewer individuals secured more and more nominations. Accordingly, the average value for a size-weighted count of candidates (the so-called "effective number") fell nearlyto one, indicating that most districts were dominated byone individual standing as the candidate for more than one party.[7] Not until the late 1950s did California's U.S. House elections take the familiar modern American shape of competition between two candidates, each of whom represents one of the two major parties.
The significance of the foregoing is that primaryelections in which a voter maycross over to support candidates from a partyother than one's own are not new to California. Whereas the modern practice has been to "open" primaries bypermitting some party-crossing behavior on the part of voters, in California's past, parties' influence was bypassed on the supply side, byallowing candidates to straddle party organizations for electoral purposes.[8] We turn now to an analysis of the actual voting patterns under these rules.
WHO WERE THE X-FILERS? CONTEXTS, CANDIDATES, AND CROSSOVER VOTING
The next task in assessing the similarities between cross-filing and the blanket primaryis to consider the subtleties of multiple filing and the various possible incentives it created for crossover voting. Given the chronology of alterations and amendments described in the previous section, there have now been seven discernible regimes in California's historical experience with primaryelections, four of which fall in the multifiling period. Table 2.1 summarizes the keydifferences with respect to crossover voting and partynominations at the general election.
Of all the primary elections in the cross-filing era, probably those of 1914 and 1916 most resemble the blanket primary in strictly procedural terms. In both time periods, the top finisher from each party moves on to the general, regardless of relative totals. Under the blanket, voters can support any candidates who file for office, without regard to the voters' or candidates' parties of registration. In the 1914 and 1916 elections, voters could support any candidates who filed in their parties' primaries or even candidates who did not do so, if they (the voters) took the extra ordinary step of casting write-in votes. Since write-in votes are important only in races which are already aberrant, typically because the seat has been deserted byone major party, it is safe to say that the blanket system is a slightly friendlier environment for crossover voting. Or, at least, voters' options are less constrained by choices made by candidates. The primaries of 1920–52 were, in turn, less favorable to crossover voting than their predecessors of 1914 and 1916 because of the requirement that a candidate win his own party's nomination in order also to win another party's nomination. With this procedural change, objective strategic conditions for crossover voting also changed. Some voters inclined to support a candidate from another partyin their own primarymayhave hesitated to do so for fear that the vote would be "wasted" if the candidate did not win his own primary.
The change in 1918 also opened up the possibility for "spoilers," candidates who could knock others out of the general election by receiving the most votes in a given primary, but not advance themselves because they did not win their own primary. Because of domino effects, spoiling could be hard to anticipate. Consider the primary in the Thirty-Sixth Senate District in 1922. In that election, Republican candidate A. B. Johnson received 7,571 votes in the Republican primary. Republican F. D. Mather won 6,961 Republican votes, 531 Prohibition votes, and 14 Socialist (write-in) votes. Prohibition candidate C. R. Burger received 1,809 Democratic votes and 414 Prohibition votes. Had 60 Prohibition voters switched from Mather to Burger, the general election would have pitted a Republican against a Prohibition-Democrat,
| Election Years | Primary Type | Crossover Voting? | Multiply Nominated Candidates? |
|---|---|---|---|
| 1910–12 | Closed | By write-in vote only | Only in districts in which one party is very weak |
| 1914–16 | Closed with cross-filing | Yes, provided candidates cross-file (or by write-in) | Yes |
| 1918 | As in 1916, plus win-own-party requirement | Yes, provided candidates cross-file (or by write-in) | Only for candidates who win nomination of party with which they register |
| 1920–52 | As in 1918, plus party committee substitutes | Yes, provided candidates cross-file (or by write-in) | Only for candidates who win nomination of party with which they register |
| 1954–58 | As in 1952, plus candidates' parties identified on ballots | Yes, provided candidates cross-file (or by write-in) | Only for candidates who win nomination of party with which they register |
| 1960–96 | Closed | By write-in vote only | Only in districts in which one party is very weak |
| 1998–2000 | Fully open "blanket" | Yes (but not directly observable, except in presidential voting) | Only in districts in which one party is very weak |
There are, in short, both institutional and non-institutional distinctions to be drawn between the different types of primary systems and their receptivity to crossover voting. A fairly common scenario is for the primary to pit one Republican against one Democrat. Under the blanket primary, both candidates are assured of advancing to the general election, and the primary simply provides a preview of how the general might unfold by giving a preliminary reading of the candidates' relative appeals. In the cross-filing years, by contrast, these races could predetermine the general election result, since a candidate who won both nominations was guaranteed to face no major party opponent in the general. So, in symmetric races that lack real intraparty competition, the stakes were somewhat higher under
Figure 2.2 shows how much crossover voting occurred in U.S. House and state Senate races between 1910 and 1964. The dashed lines are party specific: they show what proportion of votes cast in the Democratic primary went to candidates who were registered Republicans (and who were running in the Republican primary simultaneously) and what proportion of votes cast in the Republican primary were won by registered Democrats (who were also competing in the Democratic primary). The solid line marked "Total" is the proportion of primary votes that were cast for other party candidates. It includes Republican and Democratic crossover votes as well as votes cast for Progressives, Socialists, or Prohibitionists running in the Republican or Democratic primaries.[9]
The two panels tell similar stories. In 1910 and 1912, before cross-filing was permitted, a tiny amount of crossover voting occurred by way of write-in votes. In a few cases (e.g., the Third U.S. House District in 1910), a candidate won Democratic and Republican nominations simply because no opponent from the other major partyran, and so a campaign to generate a small number of write-in votes was sufficient for victory. When multiple filing became legal in 1914, crossover voting increased markedly. In 1914 and 1916, crossover voting accounted for about 10 percent of the primary vote. From 1918 to the mid-1930s, it was more common still, constituting about 20 percent of the primary vote. This was an era of Republican strength, and crossover voting was, accordingly, very asymmetric: Republican candidates dominated the Democratic primaries, while onlya handful of Democrats captured anyvotes on the Republican side. With the New Deal realignment, the picture changed again. From the mid-1930s to the mid-1950s, crossover voting accounted for about 30 percent of the U.S. House vote and about 40 percent of the State Senate vote, and it was nearly evenly balanced between Democrats drawing support from Republican voters and Republicans drawing support from Democratic voters (the minor parties having all but vanished). Finally, the explicit labeling of candidates' own registration status on primary ballots in 1954, 1956, and 1958 seems to have been consequential: crossover rates fell back to the 20 percent range over these elections. Once multiple filing was abolished, crossover voting dropped to negligible rates as it was, again, possible only through write-in campaigns. In the early 1960s, these were symbolic (and futile) protests against the abolition of multiple filing. There after, they disappeared altogether.
Another point about figure 2.2 is that it is especially sensitive to malapportionment. Because U.S. House districts varied considerably in population
Figure 2.3 takes this latter approach, averaging rates across districts without any weighting for numbers of voters or residents. We also turn to the issue of distinguishing different reasons to cross over in primary elections by plotting time series for the four categories of crossover vote. Some crossover voting occurred for the simplest of reasons: no candidate registered with the relevant party filed, and so voters who went to the polls and preferred not to abstain from voting in the given race could only back candidates from other parties. We use the label "no option" for all such crossover voting. In all remaining cases, primary voters did have a choice between candidates registered with their own party (who might or might not have been running in other primaries) and cross-filers who were registered with another party and running in at least one other primary. We distinguish between crossover votes cast for an incumbent (e.g., Democratic primary voters supporting a Republican incumbent), those cast against an incumbent (e.g., Democratic primaryvoters supporting a Republican nonincumbent rather than a Democratic incumbent), and those cast in races in which no incumbent was running.
The two panels of figure 2.3 reveal more contrast between the two chambers than did figure 2.2. Panel 2.3A, which describes the U.S. House, is broadly reminiscent of panel 2.2A. From 1914 to 1940, districts saw, on average, 5 to 10 percent of primary voters crossing over because they had no other option. In two presidential election years (1924 and 1928), this rate was much higher, as 15 to 20 percent of all votes in the average district were forced crossovers. After 1940, there was very little voting of this kind. Open-seat and anti-incumbent crossover voting, by contrast, were unimportant until the 1940s. Thereafter, from 1940 to 1950, the district averages for each chamber were about 5 to 10 percent. These rates fell sharply
No patterns in the California Senate are as clear. In most years, the districts averaged higher rates of all varieties of crossover voting than did their U.S. House counter parts. In stark contrast to the patterns found in U.S. House elections, forced crossover voting was somewhat more common late in the series than early. On the other hand, as in the U.S. House case, there were declines in the average district rates of all kinds of crossover voting after the 1954 candidate-labeling rules came into play.
The implication of this drop in crossover voting after candidates' party registrations were printed on the ballots is that some voters must have been crossing over unintentionally, not realizing that the candidates they were supporting were in fact registered members of another party. A plausible conjecture about such voting is that much of it originated in incumbency status. Indeed, it is a staple of research on the incumbency advantage in congressional general elections that incumbents are able disproportionately to attract support from voters who do not identify with the incumbent's party. Case studies bolster the impression that crossover voting extended this incumbency advantage in reaching across parties into the primary stage as well. For instance, years after he lost the 1948 Democratic primary in California's Twelfth District to Richard Nixon, Democrat Stephen Zetterberg cited cross-filing as the culprit, complaining that "Nixon, concealing his Republican Party affiliation, used to advantage that his name came first on the ballot due to his incumbency and that his campaign committee promoted the ‘Dear Fellow Democrat' advertising format, enticing enough unsophisticated Democratic voters, who did not understand the crossfiling procedure, into casting their ballots for him." Indeed, it appears that crafting a separate appeal aimed directly at Democrats was "the media strategy in the local newspapers" for the Nixon campaign (Gellman 1999, 179–80).
Zetterberg's description conflates a few different points. Incumbents may have profited from the simple advantage of appearing first on ballots, but they undoubtedly had an edge in name recognition to boot. The post 1952 drop-off in pro-incumbent crossover voting that occurred in both the U.S. House and State Senate is indirect evidence that "unsophisticated" voters may well have been mistaking the better-known candidates for their own partisan brethren. On the other hand, pro-incumbent crossover voting continued to make up about 10 percent of the vote in districts after candidates' parties were explicitly noted on the ballot, so some of the crossparty
This general climate is precisely what the blanket system is supposed to deliver as well. Candidates able to craft appeals to all voters, not only those who share their party registration, can broaden their bases, encourage cross-party coalitions of moderates, take advantage of nonpartisan appeal and superior recognition, and so on. There is little reason to believe that Nixon was alone in taking advantage of the institutional environment by developing a two-track strategy to win Democratic and Republican votes.
Table 2.2 presents one more analysis of crossover voting rates and the effects of institutional context (the different combinations of rules in place) and political context (incumbency status). Our observations include both U.S. House and State Senate districts. We regress the proportion of the primary vote that consists of crossover votes on indicator variables for each period, an indicator for whether there was a cross-filing incumbent in the race, another indicator for races having incumbents who did not cross-file, a simple count of how many non-incumbents had cross-filed, and a variable identifying a district as competitive.[10] Finally, because the district populations vary widely, we added the natural logarithm of the total number of primaryvotes cast as a further control variable.
Models 1 and 3 include, respectively, all districts and all U.S. House districts. Models 2 and 4 investigate the robustness of the findings by omitting all cases in which there was no crossover vote at all. The results reemphasize the points illustrated in figure 2.3. First, the presence of a crossfiling incumbent greatly increased the amount of crossover voting in all models, by a substantial 15 to 25 percent. In models 1 and 2, races in which an incumbent ran but did not cross-file saw 3 to 8 percentage point declines in crossover voting, presumably because the incumbent drew support in his or her own primary that might otherwise have gone to cross-filers from other parties. Since the coefficient on this variable was not significant in models 3 and 4, it appears that this dampening effect was stronger in the State Senate than in the U.S. House, and/or that it was mostly when seats were competitive that incumbents ran without cross-filing in the House. The direct effect of the competitive-district variable, was, in any case, not significant, so crossover voting does not seem to have been sensitive to the prospects of a close race ensuing in the general for the U.S.
| U.S. House and State Senate | Crossover Vote > 0 | U.S. House Only | U.S. House, Crossover Vote > 0 | |
|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | |
| *p < .05 | ||||
| Intercept | 0.32* | 0.40* | −0.16 | −0.40* |
| (0.04) | (0.05) | (0.12) | (0.17) | |
| 1914–16 | −0.04* | 0.00 | −0.02 | 0.06 |
| (0.02) | (0.03) | (0.03) | (0.04) | |
| 1918–52 | 0.16* | 0.25* | 0.14* | 0.26* |
| (0.01) | (0.02) | (0.01) | (0.02) | |
| 1954–58 | 0.10* | 0.19* | 0.05* | 0.15* |
| (0.02) | (0.02) | (0.02) | (0.02) | |
| Cross-filing | 0.14* | 0.09* | 0.14* | 0.09* |
| incumbent | (0.01) | (0.01) | (0.01) | (0.02) |
| Non-cross-filing | −0.03* | −0.08* | −0.01 | −0.04 |
| incumbent | (0.01) | (0.03) | (0.02) | (0.04) |
| Number of other | 0.03* | 0.01* | 0.03* | 0.01* |
| cross-filers | (0.004) | (0.005) | (0.005) | (0.006) |
| In (total votes | −0.03* | −0.04* | 0.01 | 0.03* |
| cast) | (0.004) | (0.005) | (0.01) | (0.02) |
| Competitive | 0.002 | 0.002 | ||
| district | (0.009) | (0.01) | ||
| N | 1,119 | 877 | 559 | 425 |
| Adjusted R2 | 0.50 | 0.35 | 0.56 | 0.37 |
We take this finding to signify that voters and candidates require time to learn how institutions shape their incentives. Some cross-filing and, in turn, crossover voting, quicklyfollowed the introduction of the new primary rules in 1914, but several elections had passed before the phenomenon peaked in importance. Most, but not all, of the party-defying votes were cast for incumbents, suggesting that cross-filing institutions, by weakening parties in the primaries, assisted those already in power. Differences between
A few implications for how the blanket primarywould have developed had the blanket rules remained in place follow. First, and most obviously, it is not wise to draw strong conclusions about how the blanket rules affect candidate and voter behavior after onlyone election. It maytake a few elections before anykind of equilibrium is reached. Crossover voting in a blanket primaryis something like crossover voting in the 1954–58 elections, since the candidates' partyidentities are known. Some of the strategic decisions of the past, though, such as whether or not to cross-file, no longer apply. Voter strategyis thus slightlyless intertwined with candidate strategy under blanket rules than under cross-filing. Probably, all else equal, incumbents were favored bythe blanket's removal of party registration walls, as theywere in the past. Some of the variation in voting behavior between the U.S. House and the California State Senate mayhave followed from other institutional differences, such as the vastlydissimilar methods of apportionment used for each from 1930 to 1964. The fact that Senate districts were, with veryfew exceptions, not altered over this period had implications for district competitiveness. Although we have not presented evidence that crossover voting levels having been responsive to districts' general election competitiveness, we continue to suspect that the expectation of a competitive race maybe an important determinant of how crossover votes are cast. The effects of blanket primaryrules would have been filtered through other contextual factors, such as how competitive are anynew districts drawn for 2002 and beyond, how campaign finance laws vary across venues, and so forth.
WEAVING (AND UNWEAVING) THE BLANKET
Passed in 1996 and overturned in June 2000, California's blanket primary had a short but eventful life: (1) it became law when Proposition 198 was approved in the primaryelection of March 1996; (2) it survived a court challenge mounted bythe major and minor parties in decisions handed down bythe U.S. District Court in 1997 and the Ninth Circuit Court of Appeals in March of 1999; (3) its application to presidential elections was threatened in July1998 when Governor Wilson signed SB 1505 into law, therebyplacing Proposition 3 on the November 1998 ballot; (4) however,
Perhaps surprisingly, the electoral battle over Proposition 198 was not waged in dollars. The initiative was opposed by both major parties, but they allowed themselves to be outspent, possibly having been lulled into inactivity because the measure enjoyed endorsements from onlythree bigname-value politicians: Republicans Tom Campbell and Becky Morgan, and Independent Lucy Killea. State Senator Killea, significantly, was not onlya former Democrat, but was already associated with quixotic reform efforts because of her prior campaign to promote unicameralism for the state legislature. The reasons underlying the support by the two major-party figures for the open primary measure, meanwhile, can be discerned from their intertwined career paths.
Campbell was a moderate who believed he (and other moderates) would benefit bybeing able to appeal to moderate voters and independents. Morgan, meanwhile, was drawn to the cause byher disgruntlement with the state Republican organization. Each made substantial contributions ($100,000 and $150,000, respectively) from their political war chests.
Another $150,000 was supplied to the pro—Proposition 198 forces in small donations (i.e., under $10,000 each), with most of the balance coming from Hewlett-Packard ($45,000 directly; $300,000 from David Packard; and $120,000 from William Hewlett). In all, the pro–Proposition 198 campaign raised about a million dollars, while the opponents raised only $100,000: $50,000 given by Rupert Murdoch and $50,000 from the state Republican party.[11] Elite opposition, then, was not expressed financially, but in public statements and endorsements. The official "Argument Against Proposition 198" in the 1996 Primary Voters' Handbook was, after all, signed bythe chairmen of the California Republican and Democratic parties.
Despite official partyopposition and strong elite antagonism at the rhetorical level, most of those who made their wayto the polls in March of 1996 endorsed the measure. Not onlydid Prop 198 pass easily, it was one of six propositions (out of twelve) that won majoritysupport in everycounty of the state.[12] Roughly 92.3 percent of all ballots featured a vote on Prop 198, which was just slightlybelow the average participation rate (92.7 percent) for that batch of propositions.
Even before a blanket primaryhad been held under the new law, the
The Voters' Handbook debate on Proposition 3 was even less balanced, in terms of prestige of the participants, than had been its Proposition 198 counterpart. United in favor of the modification were the senior State Senators from each party, the Assembly leaders of each party, and Proposition 198's old nemesis Bruce Herschensohn. Only first-term Assemblyman Jack Scott (D–44) opposed the proposition. And yet, perhaps because they were unconvinced that either major partywould dare freeze the largest state out of its convention, whatever the formal rules specified, voters rejected Proposition 3 almost as unambiguouslyas their counter parts had supported Proposition 198 just twenty-nine months earlier. Proposition 3 lost by a margin of 46 percent to 54 percent, and won majoritysupport in onlyone county(Los Angeles). The proposition was also ignored byan unusually large number of voters: only 86.1 percent of all ballots included votes (pro or con) on the measure, giving it the fifth lowest participation rate of the forty-seven propositions considered by California voters in the 1996 and 1998 primaryand general elections.
The new legislature, however, was not willing to follow the voters' advice to playchicken with the national parties. Within months, both chambers had passed a substitute for Proposition 3, again byoverwhelming margins. When Governor Davis signed SB 100 into law in Mayof 1999, the state government had effectively negated the impact of Proposition 198 with respect to presidential primaries.
Under the revised law used in the 2000 presidential primary, the presidential election appeared to the voter to be open, since one could select delegates (in the minds of most voters, presidential candidates) without regard to one's own registration status. However, not all votes counted equally. It was required of local election officials that they identify presidential ballots bythe voters' partyaffiliation so that theycould perform a double tabulation. California's presidential convention delegates in 2000 were thus selected not according to the overall vote, but according to "the number of votes each delegate candidate receives from voters affiliated with the same political partyas the delegate candidate." Whether one regards this revised law as a betrayal of the spirit of the blanket primary or a necessary corrective, a nice feature for political scientists is that the dual counting system allows recovery of directly observable evidence on cross-party
CONCLUSION
Analysts eager to understand California's blanket primary system have typicallylooked to Washington and Alaska, the other two states that have used this type of primary. They have entirely missed the clues buried in California's own electoral historyand the natural link between the blanket primary and its ancestor, the multiple-filing primary. The parallels are clearly numerous, and we have onlybegun to mine the wealth of data on how Californians behaved in past primaries, given the opportunityto vote outside of party registration lines.
There is clear evidence that the strategies behind cross-filing and crossover voting were not immediately obvious to Californians in the 1910s and 1920s. Though the Assembly saw a proliferation of "hyphenated," multiply nominated members immediately (see note 4), rates of crossover voting did not rise dramatically until after a few elections had passed. We see no reason to expect modern-day voters and candidates to determine their optimal strategies under a blanket primarysystem anymore quickly. Hence, a subtheme for this entire book is that all conclusions should be qualified with an understanding that the full potential impact of the blanket primary maynot have been realized in the short period it was in place. We counsel continued attention to historical precedent as analysts struggle to understand Californians, their parties, and their primaries.
NOTES
Our thanks to Bruce Cain, Michael Caldwell, Christophe Crombez, and Elisabeth Gerber for helpful suggestions, and to Lloyd Gruber for extraordinary assistance with locating primary sources.
1. An example of the minor parties being decisive is New York's Sixth Congressional District in 1980. John LeBoutillier received 71,838 votes as the Republican candidate, while Lester Wolff won 74,319 votes as the Democrat. It was LeBoutillier who went to Washington, though, since he also won 11,299 votes as the Conservative candidate, while Wolff won only5,890 Liberal votes. A more typical result is that of the Ninth District in 1984: Tom Manton won the seat with 71,420 votes as a Democrat
2. In modern elections, this is possible onlyif the candidate wins one nomination as a write-in. The only U.S. Representative from New York to hold Republican and Democratic nominations in a general election in the 1990s was Charlie Rangel in 1990, running in what was then the Sixteenth District (Harlem), a verysafe Democratic seat routinely abandoned by the Republican party. [BACK]
3. Joining twenty-four Republicans, ten Democrats, seven Progressives, and two Socialists were ten members with both Republican and Progressive nominations (or "Republican-Progressives"), eight Republican-Democrats, seven Democratic-Progressives, six Republican-Democratic-Progressives, one Republican-Democratic-Prohibitionist, one Republican-Progressive-Prohibitionist, one Democratic-Progressive-Prohibitionist, one Republican-Democratic-Progressive-Prohibitionist, and even one member nominated byall five parties, a Republican-Democratic-Progressive-Prohibitionist-Socialist (Calif. Secretaryof State, Statement of the Vote at the General Election of November 3, 1914, 4–5). [BACK]
4. It might seem that this power to appoint nominees was unlikelyto have mattered, given that committees could act onlyif all the candidates registered with their own partyhad lost to outsiders, an outcome that would seem to signal that the seat was a lost cause. In general, primaryraces with manycandidates and/or manycrossfilers seem not to have been verygood predictors of subsequent general election results, and committee-appointed stand-ins who actuallywon were not unheard of. [BACK]
5. Voting against cross-filing in the Senate were seven of ten Democrats who had won Republican nominations in the 1958 primaries, and one of four Republicans who had won Democratic nominations. In the Assembly, twelve of the thirteen Democrats who had also carried the Republican nomination in 1958 voted to abolish cross-filing, while both of the Republicans who had won the Democratic nomination voted against abolition. [BACK]
6. The data analyzed in this chapter were obtained from the relevant official primary and general Statement of Vote reports released bythe California Secretary of State. The electronic dataset was created from these sources by the authors. [BACK]
7. The "effective" number of candidates is an index that weights the actual number of candidates according to their vote shares, so that strong candidates count far more than weak ones. We used the most common such measure in the voting literature, the Laakso-Taagepera index (1979), computed for each district. The values in figure 2.1 are averages, computed as follows:
where Vij is the number of votes won bycandidate j in district i, and there are a total of n districts in the state (or, as applicable, n districts in which there is some competition). [BACK]
8. Candidates elected with both Democratic and Republican nominations were not, of course, in anylegislative sense representatives of both parties. Richard Nixon and Cecil Young were both elected with both major-party nominations in 1948, but there was no doubt that the former was a partisan Republican and the latter a partisan Democrat once theyarrived in D.C. to sit in the Eighty-First Congress. [BACK]
9. The figure very slightly undercounts crossover voting because, as a consequence
10. This competitive-district dummyvariable was developed bycomputing normal votes for districts, using general election returns, and then adjusting these according to whether or not there was anypartisan turnover observed in the relevant reapportionment period. We used a simple dichotomy: safe (Democratic or Republican) seat (i.e., high or low normal Democratic vote, little turnover) versus competitive (middle-sized normal vote plus some turnover.) A paucityof contested general elections made the estimation difficult for the State Senate, so we computed this measure for the U.S. House districts only. [BACK]
11. All data on contribution amounts are taken from Jones (1996). [BACK]
12. Unusually, all twelve propositions on the March 1996 ballot passed. [BACK]
13. It passed the Senate 35–0 on April 2 and passed in the Assembly52–12 on July 10, 1998. Three days later the Senate concurred with Assembly amendments, 28–0, and on July13, 1998, the law was signed by the Governor and chaptered by the Secretary of State. [BACK]
REFERENCES
1978.
“The Multicandidate Calculus of Voting: Application to Canadian Federal Elections.”
American Journal of Political Science22, no. 3: 609–38.
Birtel, Marc. 1996.
“Ballot: 'Jungle Primaries' Adopted … Reclassifying Lions Rejected.”
Congressional Quarterly Weekly Report54, no. 13 (March 30): 902–3.
Block, A. G., and Claudia Buck, eds. 1997. California Political Almanac, 1997–1998. 5th ed. Sacramento, CA: California Journal/State Net.
Brady, David W., and Brian J. Gaines. 1995.
“A House Discarded? Evaluating the Case for a Unicameral California Legislature.”
In Bruce E. Cain and Roger G. Noll, eds., Constitutional Reform in California: Making State Government More Effective and Responsive. Berkeley, CA: Institute of Governmental Studies Press, 195–238.
California Secretary of State (various). 1904–1964. Statement of the Vote at General Election. Various volumes. Sacramento.
California Secretary of State (various). 1904–1964. Statement of the Vote at Primary Election. Various volumes. Sacramento.
“Coast Rivals Agree.”
1959. New York Times, Saturday February 28, 7.
Davies, Lawrence E. 1951.
“New Warren Term Starts Tomorrow.”
New York Times, Sunday, January 7, 57.
Davies, Lawrence E. 1952.
“Knowland Seeks California Sweep.”
New York Times, Sunday, May 4, 49.
Dubin, Michael J. 1998. United States Congressional Elections, 1788–1997. Jefferson, NC: McFarland.
Gellman, Irwin F. 1999. The Contender: Richard Nixon, the Congress Years, 1946–1952. New York: Free Press.
Green, Stephen, et al. 1995. California Political Almanac, 1995–1996. 4th ed. Sacramento, CA: California Journal Press.
Hichborn, Franklin. 1909. Story of the Session of the California Legislature of 1909. San Francisco: Press of the James H. Barry Company.
Hichborn, Franklin. 1922. Story of the Session of the California Legislature of 1921. San Francisco: Press of the James H. Barry Company.
Hill, Gladwin. 1959a.
“California Bills Rushed by Brown.”
New York Times, Sunday, January 18, 64.
Hill, Gladwin. 1959b.
“Cross-Filing Ban in Peril on Coast.”
New York Times, Monday, March 16, 15.
Hill, Gladwin. 1959c.
“California Shift Will Help Voter.”
New York Times, Sunday, April 26, 67.
Jones, Bill (Secretaryof State of California). 1996. Financing California's Statewide Ballot Measures: 1996 Primary and General Elections Campaign Receipts and Expenditures through June30, 1996. Sacramento, CA: Secretaryof State, Political Reform Division.
Kettl, Donald F. 1992. Deficit Politics: Public Budgeting in Its Institutional and Historical Context. New York: Macmillan.
Laakso, Markku, and Rein Taagepera. 1979.
“'Effective' Number of Parties: A Measure with Application to Western Europe.”
Comparative Political Studies 12, no. 1 (April): 3–27.
Lee, Eugene C. 1960. The Politics of Nonpartisanship: A Study of California City Elections. Berkeley, CA: University of California Press.
“Legislature Repeals Cross-Filing in State: Signature of Governor Assured.”
1959. Los Angeles Times, Friday, April 24, 1.
Shannon, Don. 1959.
“Repeal of Cross-Filing Hailed by Congressmen.”
Los Angeles Times, Saturday, April 25, 4.
State of California. 1920. State of California Direct Primary Law with the Official Forms Prepared by the Secretary of State and the Attorney General. Sacramento.
Young, C. C., ed. 1943. The Legislature of California: Its Membership, Procedure, and Work. San Francisco: Press of the Parker Printing Company.
3. Political Reform via the Initiative Process
What Voters Think about When They Change the Rules
Shaun Bowler and Todd Donovan
INTRODUCTION
This chapter examines support for Proposition 198, California's 1996 blanket primary initiative. Proposition 198 is considered to be part of a long series of initiatives that have presented California voters with choices about how their political institutions should be structured. We use public opinion data to test hypotheses about the nature of mass support for such political reform initiatives. We test if support is associated with voters' self-interest and with general dissatisfaction with politics. Our findings support the idea that voting for Proposition 198 was structured byreasoning about the consequences of the proposed reform. Strong partisans were opposed to the measure, yet partisans dissatisfied with their party's candidates were more supportive. We also find evidence consistent with the idea that some voting was structured bygeneralized dissatisfaction that had no specific relation to election rules, particularly among less-educated voters. Less-educated voters with personal economic worries were more supportive.
VOTER CHOICE OVER INSTITUTIONS
Ballot measures often generate a great deal of controversy. Recent examples from California include initiatives on state services for immigrants (Proposition 187), affirmative action (Proposition 209), the legalization of marijuana (Proposition 215), and Indian gaming (Proposition 5). A quick glance through old textbooks on California politics readilyprovides a similar list from previous decades. Heated debates were seen in the 1960s, for example, over the repeal of the legislature's fair housing Rumford Act via Proposition 14 (Wolfinger and Greenstein 1968), while the 1970s saw controversies
Less noticed, but of potentially long-term influence, are those propositions that regulate the political and institutional structure of the state. Initiatives that change the wayelected officials do their business or contest elections, "governance policies" (Tolbert 1998), have effects that endure over a long period of time. Term-limit measures (Donovan and Snipp 1994), tax and expenditure proposals such as Proposition 13 of 1978 (Sears and Citrin 1982; Tolbert 1998), Proposition 4 of 1979, and campaign finance reforms are examples of citizen-enacted institutional changes that have had major effects on state politics (for discussion of the effects of some of these policies see Mondak 1995; Moncrief and Thompson 1991; Clingermayer and Wood 1995; Donovan and Bowler 1998; Schrag 1998; Gerber et al. 2000). Proposition 198 is another example of an initiative that proposed an institutional change and which therefore had farreaching consequences for the conduct of politics in the state.
The broad potential importance of political reform initiatives can perhaps be made clearer byreference to a simple spatial analogy. In spatial terms, policyproposals move the status quo to a particular point in an issue space, but institutional reforms change the veryspace available to policy makers. Consequently any future flow of policymaythen be restricted to, or restricted from, a particular region in the space. The cumulative impact of an institutional change can therefore be much greater than the move from a single point to another point within a given space.
All of this raises a familiar criticism of the initiative process yet more pointedly: if voters have a hard time figuring out how to vote on individual policies (Magleby1984), how can they possibly understand issues of institutional change? Initiatives like Prop 198 present additional, fundamental concerns about democracy: should voters have a free hand in shaping institutions? When theycan, do theymake changes in institutional rules that reflect some manner of deliberation about the consequences of change? The answers to these questions are of broad relevance. Theories of political legitimacyhave long depended on the idea that voters like—or at least have modest regard for—the institutions which govern them. Political legitimacy may also rest on the permanence of these institutions (March and
Luckily, at least so far as institutional legitimacy is concerned, it is typically very difficult to introduce such changes. It is often especially difficult to translate popular discontent into proposals and then into actual reform. In addition to procedural hurdles, simple delaying actions (Commissions of Inquiryor blue-ribbon panels of experts) can be relied on to leave reform proposals to die a peaceful and unimplemented death. The initiative process, however, allows voters to introduce and enact institutional reform proposals fairlyreadily, and with it, California voters have been given, and have taken, several chances to introduce real procedural change.
Proposition 198 is far from unique as an example of institutional reform via the initiative process. Tables 3.1 and 3.2 show that California voters have been presented with dozens of choices over their political institutions since 1912. The data and subject categorizations are taken from the California Secretaryof State's office,[1] which considers proposals to change the rules that structure state revenues and expenditures (e.g., Proposition 13, Proposition 98) as a different category. If anything, then, these data underestimate the frequencywith which Californians can vote to change (or retain) their political institutions.
The first two columns of table 3.1 show that between 1912 and 1998, 169 of California's 1,043 initiative proposals fell into the categoryof "institutional reform." Of these 169, 45 eventually qualified for the ballot, representing 16 percent (45 of 272) of all California initiatives qualified from 1912 to 1998. Eighteen of these passed, or 40 percent, which is slightlyhigher than the 32 percent pass rate for all California initiatives during this period.
Columns 3–7 of table 3.1 show the varietyof topics covered bythe institutional reform initiatives. The range of proposals varies from the major to the relativelyminor. Examples of the former are term limits, campaign finance reform, and changes in the structure of primaries. Examples of the latter include nineteen of the fifty-two proposals concerning "Elected
| All Initiatives | All Reform Initiatives | Campaigns | Elected Officials and Civil Service | Term Limits | Elections | Reapportionment | |
|---|---|---|---|---|---|---|---|
| SOURCE: Bill Jones, "A Historyof the California Initiative Process" (Sacramento, August 1998). a Some initiatives were placed in two categories, so these do not sum to 100 percent. | |||||||
| Number titled on this subject | 1,043 | 169 | 27 | 52 | 16 | 63 | 27 |
| Number qualified for ballot | 272 | 45 | 11 | 10 | 3 | 15 | 9 |
| Percentage of initiatives qualified for ballota | 26% | 26% | 24% | 22% | 6% | 33% | 20% |
| Number approved by voters | 87 | 18 | 5 | 4 | 3 | 8 | 1 |
| Percentage approved by voters | 32% | 40% | 45% | 40% | 100% | 53% | 11% |
| Titled N | Qualified N | Approved N | |
|---|---|---|---|
| SOURCE: Bill Jones, "A Historyof the California Initiative Process" (Sacramento, August 1998). | |||
| All initiatives | |||
| Total 1912–98 | 1,043 | 272 | 87 |
| Total 1980–98 | 593 | 103 | 41 |
| Institutional reform initiatives | |||
| Total 1912–98 | 169 | 45 | 18 |
| Total 1980–98 | 120 | 25 | 11 |
Table 3.2 illustrates that a majorityof these reform efforts occurred after 1980. While over a third (103 of 272) of California's initiatives on all subjects have qualified for ballots since 1980, well over half of all institutional reform initiatives (25 of 45) have qualified since then. Prior to 1980, voters faced an institutional reform proposal on the ballot once everyfour years, on average. After 1980, the figure becomes six everyfour years. Table 3.1 demonstrates that voters are more sympathetic to these initiatives than they are to initiatives generally. This presents several questions. Why do voters support institutional reform? Do theychange institutions on the basis of mere whim or discontent, or does support reflect a deeper understanding of the consequences of change?
WHAT DETERMINES VOTER CHOICES ABOUT INSTITUTIONAL REFORM?
We advance two related but distinct arguments about how voters assess institutions. One argument is grounded in an understanding of voter behavior as relatively self-interested and rational. According to the selfinterest view, we hypothesize that voters will support the current institutions if theydo well under them, and support change if theybelieve the alternative will make them better off. Bycontrast, a second series of explanations stresses more general orientations toward, and disaffection from, institutions. Here we maythink of voter opinions toward institutions being guided bya sense of discontent, driven perhaps byeconomic factors (unemployment),
The argument which emphasizes winners and losers implies that voters understand institutional changes. Voters have sufficient information to vote sensiblyon institutional change initiatives or at least receive effective cues that allow them to vote in response to how an institutional change might affect them (Lupia 1994; Bowler and Donovan 1998). If it is possible to show that decisions are consistent with patterns of who wins and loses under different rules, then we mayinfer that voters are reasoning in a quite sophisticated wayabout how institutions work. In terms of recent term-limit initiatives, for example, this means that manyvoters make decisions consistent with the (political) gains and losses that might result from the proposals. Voters supporting the partyout of power will thus support term limits as a wayof opening new opportunities (competitive, open seats) for access to the legislature. As evidence, Donovan and Snipp (1994) found with California's Prop 140, and Maglebyand Patterson (1996) found with a term-limit initiative in Utah, that voters aligned with the majorityparty in a legislature were more likelyto oppose term limits, while partisans of the minoritywere more likelyto be in favor.
It is also possible to view voters as thinking about institutions instrumentally but not necessarilyon narrowly self-interested grounds. In place of a concern for who wins and who loses under various institutional reforms, some voters mayhave ideological predispositions in favor of certain institutional arrangements that we maydistinguish—at least in principle—from self-interested concerns. The overlap between self-interest and ideologyis, of course, a strong one. Differentiating between the two effects can be enormously difficult. Nevertheless, if voters make decisions on taxes, vouchers, environmental regulations, Prohibition, bonds, and term limits based on assessments of how the measure affects their group, party, community, or themselves (Bowler and Donovan 1998), we may reasonably expect that theywould use similar criteria when assessing proposals such as Proposition 198.
If the link between understanding and vote choice is at issue when discussing specific targeted policies such as environmental or educational reforms, it is even more critical in considering votes on institutional reforms such as Prop 198. If voters have a hard time understanding individual issues, theymayhave an even harder time understanding the impact of changing institutions. The link between voter understanding of institutions and vote choice is thus of substantive importance in developing an understanding of vote choice anchored in a self-interested model.
An alternative to the self-interested hypothesis is that votes to change institutions mayreflect discontent and alienation with politics and society in general, rather than with the actual role of the institution targeted by
As a variant on the alienation argument, we note that demands for institutional change could be driven by short-term factors, most notablyanger about politics and politicians, economic uncertainty, or short-term reactions to scandal (Karp 1995; Bowler and Donovan 1998, 69–70). Although these opinions maybe systematic, theymaynot necessarily involve choice based on the consequences of the proposed change in political institutions. For example, voter dissatisfaction with parties and legislatures could be merelyan episodic, media-driven response to corruption or adverse coverage of "wasteful" government (Tolchin 1996) which spills over into votes in favor of changing institutional arrangements. Correlations between measures of these sentiments and support for reform maysimplyreflect an "across the board" distaste for the status quo, rather than instrumental voting with regard to particular institutions. Likewise, economic uncertaintymight cause some voters to lash out at institutions, regardless of the institution's relationship to economic conditions.
Proposition 198, then, presented voters with a chance to introduce specific institutional reforms. Thus, it provides an example of institutional reform which we mayuse to examine some of these more general arguments concerning voter responses to political reform efforts.
PROPOSITION 198: CHANGING THE RULES FOR PRIMARY ELECTIONS
Relative to some of the other items on the March 1996 ballot, Proposition 198 was a low-visibility issue with verylittle surrounding campaign. For the average voter, information about the blanket primaryproposal would have been buried under about $80 million in commercials for several tort reform initiatives,[4] and bymedia coverage of the Republican presidential primary. Of the $1 million spent on Prop 198, $950,000 was spent by the yes side, most of which went to covering the costs of acquiring signatures.
This does not mean that most voters lacked information that would allow for low-level instrumental voting. California's publiclyprovided ballot pamphlet includes a series of explicit arguments about who would win and who
In the case of Proposition 198, the ballot arguments targeted specific groups of voters who would win and lose from the proposed reform. Arguments in favor were signed bya former chair of the state's political watchdog agency(the Fair Political Practices Commission), a U.C. Berkeley professor (former IGS Director Eugene Lee), and two well-regarded state senators, Bay Area Republican Becky Morgan and San Diego Independent Lucy Killea. Proponents of the reform submitted arguments that contained specific information about consequences, stressing the inherent virtue of broadening the choices available to voters: "California's closed primary election system limits voters' choices to candidates within their own party, and excludes 1.5 million independent voters from voting in primaryelections at all. It favors the election of party hard-liners, contributes to legislative gridlock, and stacks the deck against more moderate problem-solvers" (California Secretaryof State 1996).
Proponents also sought to portraythe initiative as an attempt to take power from Democratic and Republican partyleaders. Their theme of "party hard-liners" was repeated at several points: for example, "Hard-liners in both major political parties oppose the Open Primarybecause it will weaken their power and the power of special-interest groups which support them. … The partychairmen don't want you to have a choice unless it matches theirs. If all voters are allowed to participate, theyfear theywill lose their power, as will the hard-liners and special-interest groups who support them" (California Secretaryof State 1996).
To some extent, these arguments constitute a call for greater freedom for independents and moderates to split their tickets and cast votes across partylines. Since specific proponents named in the ballot pamphlet were identifiably nonpartisan (at least symbolically) and independent, their arguments for open primaries could be credible with such voters. Lacking institutional constraints, independents, weak partisans, and minor-party voters maybe more likelythan partisans to divide their choices across party
Similarly, California's minor-party identifiers, who comprise about 5 percent of registered voters, mayalso evaluate the initiative in a positive light. The argument that closed primaries limit participation might appeal to registered minor-party voters who realize the constraint of voting onlyfor their own party's candidates in all races. In general elections California's minor-party voters mayopt for a minor-party candidate in some races, and major-party candidates in others.[5] Blanket primaries would allow them to engage in similar ticket splitting in primarycontests and so should lead them to favor the reform.[6] We test hypotheses about how the blanket primaryappealed to independents, weak partisans, and minor-party voters below.
Proponent arguments also suggested that closed primaries produced candidates that manyvoters were dissatisfied with and suggested that reform would produce more satisfactory candidates as in the following example: "Party registration in most California legislative districts heavilyfavors one partyor the other. In these so-called ‘safe' districts, the winner of the majorityparty's primaryelection is virtually guaranteed victoryin the general election, regardless of how extreme the candidate's views. In these districts, voters in the minority party have no real voice in the selection of their representative" (California Secretaryof State 1996).
One implication of this argument is that voters who did not like the candidates produced bya closed primarysystem would support the opportunityto vote for candidates from other parties in the future. This hypothesis can be evaluated bytesting whether partisans who disliked their own party's candidates in a closed nomination process favored change to an open primary. In the context of the March 1996 elections, we expect this group to include Republicans dissatisfied with their party's presidential candidates (Buchanan and Dole). Since the Democratic primarywas not contested, it is not possible to test the effects of Democratic voter dissatisfaction with the nomination choices within their party.
Intentionallyor not, opponents of Proposition 198 clearlysignaled their partisan credentials with the elites selected to sign their ballot arguments. These included the state chairmen of the Democratic and Republican parties, and prominent partisan figures, including conservative Republican commentator Bruce Hershensohn and former Democratic candidate for Governor John Van de Kamp. The name of former Democratic Assembly boss Jesse "Big Daddy" Unruh was even linked to one of the opponents (as director of the Unruh Institute at USC).
Opponents largelyportrayed the blanket primaryas a bad thing in and of itself, likening the reform to "letting UCLA's football team choose USC's head coach!" (California Secretaryof State 1996). Compared to proponents' claims, arguments against the initiative were less clear about the likely consequences of the proposal, beyond claiming (as did the proponents) that "special interests" were somehow involved. Consider the following claims:
Proposition 198 is an invitation to political mischief. This would be a dream come true for the politicians, political consultants and special interests who will use specialized targeting to manipulate the political system to the benefit of their hand-picked candidates.
Who supports Proposition 198?
Proposition 198 is a cynical attempt by a few ambitious politicians who cannot win the support of their own political party. So now they want to change the rules to serve their own self-interest. (California Secretaryof State 1996)
Both proponents and opponents claimed that special interests would somehow benefit: For proponents, special interests would benefit bythe defeat of Prop 198, while opponents argued that special interests would gain if it passed. One difference was that proponents identified the "special interests" as partyleaders and those who backed the political parties. All of this should have sent signals to strong partisans that the initiative would be bad for their party's interests. For others, the arguments and elites listed both for and against the measure mayhave signaled that the initiative was supported bypartisans whom theyopposed. In the analysis below, we examine evidence about how voters responded to these signals.
HYPOTHESES AND DATA
Our hypotheses may be grouped into two broad categories corresponding to the two arguments about voter motivation advanced above. One group of hypotheses tests the degree of self-interest—and hence understanding of institutional effects—that underpins vote choice. The other group tests whether opinion toward institutions is driven not so much byself-interest as bygeneral affect and disaffection.
If voters are acting out of self-interest, we expect those with strong partisan attachments to dislike the idea of reform which would allow voters with weak or even rival partyattachments to help decide who their nominees should be. We also expect political activists—those who make contributions to parties and attend political functions—to be resistant to a reform that opens their party's nomination process up to rank-and-file voters from any party.
Conversely, we expect supporters of minor parties and voters who identify themselves as independents to be supportive of a measure which gives them a wider choice in primaryelections. Furthermore, some major-party supporters mayremain dissatisfied with their party's nominees and so may favor a reform which widens the choices open to them. Indeed, the individuals who proposed the reform in the first place fall into this latter category.
We have mixed expectations, however, concerning the impact of the proposal upon minoritygroups. On the one hand we might expect minorities to support the proposal. Opening up the nomination process would give candidates more incentive to cultivate ethnic voting blocs—especially in attempts to encourage voting across partylines and along ethnic lines. On the other hand, we might expect minorities to regard political parties as appropriate vehicles for bringing minority candidates into the political system and, hence, to be wary of attempts to change that process.
These instrumental motivations are readily operationalized with standard surveyresponses. It is much more difficult, however, to find measures which tap into non-self-interested motivations. Nevertheless, some survey questions allow us to test how some general and less narrowly self-interested forms of voter discontent might also affect support for the Proposition 198 reform. General disaffection from the political parties maygenerate support for reform proposals. So, too, mayeconomic hardship. Voters who are generallyfed up—either with the political establishment or with their current lot in life—mayvote for reform proposals not necessarilyout of any deep understanding of the reform proposal itself, but simplyout of frustration with the status quo.
Age effects, too, mayrepresent something of a less-than-instrumental response to institutional change, since older voters maysimplybe more resistant to change of anykind.[7] Education is also included in our models, since we expect that voters with different cognitive abilities reason differentlyabout politics (Sniderman, Brody, and Tetlock 1991). Voters with more education maywell be more supportive of reforms that increase choices in elections, since theywill have an easier time processing information in a more demanding decision setting like a blanket primary.[8]
If the variables in this second group (age, economic hard times, and general disaffection from the parties) are the onlysignificant predictors of vote choice, this would suggest a more affect-driven, and consequentlyless rational, understanding of institutions byvoters. On the other hand, if our measures of self-interest are the onlysignificant predictors, then the evidence would tend to support the view that citizens reason instrumentally about reforming institutions.
Our tests are conducted with data from a Los Angeles Times poll (Los
Angeles Times 1996). The Times poll included attitudinal questions that allow us to develop refined tests of the basis of support for the blanket primary.
RESULTS
Table 3.3 presents results from a basic model of partisan-based instrumental voting on the open primaryproposal. Two additional models build on the basic model byaccounting for attitudes toward the parties in general (using a question that measures whether a respondent thinks neither party has good ideas), and for concerns about personal economic well-being (specifically, the respondents' concerns over being laid off). The full models also include a measure of political activism. Each model includes several dummyvariables reflecting partystatus and/or strength of attachments to party, as well as controls that account for potential attachments to status quo institutions (age and race). Republicans who were dissatisfied with their choices in the GOP presidential primaryare represented bya separate dummyterm (see the appendix for coding details).
As expected, strong partisans of both major parties opposed the initiative. Voters registered with minor parties were also significantly more likely to support the proposal, as were Republicans dissatisfied with the range of choices provided to them in their own presidential primary. The effects of strong partystatus, for both Democrats and Republicans, and for Republican voters who were unimpressed bytheir party's nominees, were consistent in each estimation. Overall, these results demonstrate that voters who would probablybe least likelyto make use of an open primary(bycrossing partylines) were less supportive of Prop 198. In each estimation, Anglo voters were not anymore supportive of the measure than minorityvoters, other factors held constant. Younger voters, whom we assumed to be less attached to established institutions, were found to be more supportive of the change to blanket primaries.
More interesting are independent voters, who probablyhad the most to gain under an open primaryin terms of increased access to election decisions. Contraryto the instrumental voting thesis, theywere not especially supportive of the measure relative to everyone else. Since this ran so counter to our expectations, we estimated a model that used an interactive term to distinguish between the "thoughtful independent" as opposed to the "disinterested independent." The third model in table 3.3 builds on the basic model byincluding a term representing the interaction between independence and interest in order to test if the most politically interested independents are more supportive of the open primaries. As with the other specifications, the effect of independent identification is insignificant.
Even when a wide range of variables are included in the second and
| Base Model | Plus General Attitudes | Plus "Thoughtful Independents" | |
|---|---|---|---|
| SOURCE: "California Pre-Primary Survey," Los Angeles Times Poll 372 (March 1996). NOTE: Sample is limited to registered voters. *Significant at p < .05 **Significant at p < .01 | |||
| Constant | 0.38 | 0.16 | 0.19 |
| (0.33) | (0.35) | (0.35) | |
| Anglo | 0.22 | 0.23 | 0.23 |
| (0.16) | (0.16) | (0.16) | |
| Minor-party registered | 0.35* | 0.33* | 0.34* |
| (0.19) | (0.19) | (0.20) | |
| Education | 0.02 | 0.04 | 0.04 |
| (0.03) | (0.03) | (0.03) | |
| Age | −0.009** | 0.009** | 0.009** |
| (0.003) | (0.003) | (0.004) | |
| Strong Democrat | −0.60** | −0.63** | −0.62** |
| (0.17) | (0.18) | (0.18) | |
| Strong Republican | −0.72** | −0.70** | −0.69** |
| (0.17) | (0.17) | (0.17) | |
| Republican dissatisfied with GOP nominees | 0.40** | 0.41** | 0.40** |
| (0.18) | (0.18) | (0.18) | |
| Independent | −0.12 | −0.12 | −0.24 |
| (0.18) | (0.19) | (0.23) | |
| Independent*Interest | 0.30 | ||
| (0.36) | |||
| Interest in politics | −0.07 | ||
| (0.14) | |||
| Activist | −0.22* | −0.22* | |
| (0.13) | (0.14) | ||
| Neither major party has best ideas | −0.09 | −0.09 | |
| (0.22) | (0.22) | ||
| Fear of layoff | 0.08* | 0.08* | |
| (0.04) | (0.04) | ||
| N | 1,039 | 1,039 | 1,039 |
| Percentage correct | 59 | 59 | 59 |
| −2LL Model | 1,391 | 1,385 | 1,384 |
| Improvement (X2) | 48** | 54** | 55** |
Table 3.4 reports the effects of some keyvariables on the probability that a respondent would vote yes on Prop 198. When all variables from the full logit model (column 3) in table 3.3 are held at their mean values, a voter would have a .51 baseline probabilityof voting yes on Prop 198. With these things held constant, respondents who most feared layoffs had a .59 probability of voting yes, while those fearing layoffs the least had only a .47 probabilityof voting yes.
The significant impact of personal economic uncertaintydoes raise the possibilitythat some voters, generallyfed up with "something," would vote for anyproposed reform that would change the institutional status quo, regardless of the real relationship between the reform and the voter's disenchantment. Even though voters who feared layoffs were more likely to support Prop 198, there is no clear relationship between rules about primaries and an individuals' job prospects.
Having said that, we move on to note indications that manyvoters were behaving in a manner consistent with thoughtfulness and an understanding of the consequences of changing institutions. This is most obvious in the opposition of strong partisans to the change. As shown in table 3.4, coefficients from table 3.3 predict that strongly identifying Democrats had only a .36 probabilityof voting yes on Prop 198, which is substantially lower than the .51 baseline probabilityof support. Republicans present a slightly more complex picture in that their regard for their party's nominee had a significant impact on their vote on the measure. Republicans satisfied with their nominee had onlya .38 probabilityof supporting the measure, while those who were dissatisfied had a .48 probability of voting yes. Clearly, party stalwarts reallydid not like this proposal to change the primarysystem. Although this pattern of findings is in line with the self-interest thesis, we should note that, overall, our results illustrate that a mix of motivations underlaythe desire to support Proposition 198.
We have proposed elsewhere (Bowler and Donovan 1998, 62–64) that different types of voters reason differently about ballot measures, with a
| NOTE: Probabilities of voting yes given highest score on relevant independent variable when all other variables set at their mode or mean value. a Probabilityof a yes vote with all variables in table 3.3, column 3, set at mode or mean values. | |
| All respondents | |
|---|---|
| Baseline probabilitya | .51 |
| Minor-party registration | .63 |
| Strong Democratic identification | .39 |
| Activist | .48 |
| Strong Democratic activist | .36 |
| Strong GOP dissatisfied with GOP candidates | .48 |
| Strong GOP satisfied with GOP candidates | .38 |
| Fearing layoffs | .59 |
| Not fearing layoffs | .47 |
Conversely, if less-educated voters have fewer of the cognitive resources which facilitate processing elite cues and ballot pamphlet information into conclusions about the open primary, support for Prop 198 among those voters could be structured less bypartisan-instrumental reasoning and more byfactors requiring less detailed information about the specific policy choice. Narrow economic self-interest (Bowler and Donovan 1998, 103), general economic concerns, and general concerns about the state's economy(Bowler and Donovan 1998, 74) have all been found to playa larger role among less-educated voters in decisions about initiatives.
Following this logic, in table 3.5 we report reestimations of the second model from table 3.3 (which includes partisan variables and measures of general dissatisfaction), with the sample divided into two sets of voters: the highlyeducated and those not highlyeducated. It is important to note that the direction of the effects of education displayed in table 3.3 might reflect that well-educated voters were simplymore supportive of Prop 198. This effect could hold at the same time that education also caused differences
| Not Highly Educated | Highly Educated | |
|---|---|---|
| SOURCE: "California Pre-Primary Survey," Los Angeles Times Poll 372 (March 1996). NOTE: Sample is limited to registered voters. *Significant at p < .05 **Significant at p < .01 | ||
| Constant | −0.20 | 1.52** |
| (0.32) | (0.42) | |
| Anglo | 0.40** | −0.03 |
| (0.20) | (0.27) | |
| Minor-party registered | 0.59** | −0.04 |
| (0.26) | (0.31) | |
| Age | −0.004 | −0.02** |
| (0.004) | (0.006) | |
| Strong Democrat | −0.38* | −1.04** |
| (0.23) | (0.29) | |
| Strong Republican | −0.67** | −0.69** |
| (0.24) | (0.26) | |
| Republican dissatisfied with GOP nominees | 0.54** | 0.23 |
| (0.24) | (0.28) | |
| Independent | −0.14 | −0.10 |
| 0.24) | (0.30) | |
| Neither major party has best ideas | −0.12 | 0.03 |
| (0.30) | (0.34) | |
| Fear of layoff | 0.12** | 0.03 |
| (0.05) | (0.07) | |
| Activist | −0.27 | −0.15 |
| (0.18) | (0.20) | |
| N | 606 | 435 |
| Percentage correct | 58 | 63 |
| −2LL Model | 805 | 569 |
| Improvement (X 2) | 34** | 31* |
Table 3.5 also demonstrates variation in support within each education group that is consistent with the idea that variation in cognitive resources (education) causes citizens to reason differentlyabout ballot initiatives. Well-educated voters reflect instrumental reasoning, while less-educated
| NOTE: Probabilities of voting yes given highest score on relevant independent variable when all other variables set at their mode or mean value. a Probabilityof a yes vote with all variables in table 3.5 set at mode or mean values. | |
| Voters witd low education levels | Probability |
| Baseline probabilitya | .50 |
| Strong Democrat | .42 |
| Strong Republican satisfied with GOP candidates | .35 |
| Strong GOP dissatisfied with GOP candidates | .48 |
| Fearing layoffs | .61 |
| Not fearing layoffs | .43 |
| Voters with high education levels, | Probability |
| Baseline probability a | .55 |
| Strong Democrat | .36 |
| Strong Republican | .45 |
| Strong GOP dissatisfied with GOP candidates | .50 |
| Fearing layoffs | Not significant |
Strong partisans in the more-educated group, for example, were more likelyto decide on the initiative on the basis of partisan-based reasoning than were strong partisans in the less-educated group. The effect of strong party identification is substantially larger for Democrats among the better educated. This is illustrated in the predicted probabilities reported in table 3.6. Among the well educated, the probabilityof a strong Democrat voting yes was only .36, far lower than the .55 baseline probability of support among other well-educated voters. The effect is significant among the less educated, yet much smaller. In this group, strong Democrats had a .42 probabilityof supporting Prop 198, compared to the .50 baseline for all less-educated voters. Conversely, our indicator of generalized dissatisfaction that has no specific grounding in election rules—a fear of layoffs—structured support onlyamong the less educated. Less-educated voters fearing layoffs had a .61 probability of supporting Prop 198, compared to a .43 probabilityfor other less-educated voters who did not fear layoffs.
DISCUSSION
In California and other direct democracystates, voters have opportunities to constantlymold and shape their institutions. The initiative device ensures that institutional rules governing political processes are anything but stickyand permanent. Since institutional reform initiatives allow change in the verymanner in which politics can be conducted, theyhave consequences well beyond their adoption. This chapter demonstrates that voting on the blanket primary initiative appears to have been structured bya significant component of instrumental reasoning about the consequences of changing the rules regarding voting in primaryelections. Those whom we expect to be least likelyto take advantage of a more open process, and thus to be motivated to oppose changing the rules, were in fact more opposed to Prop 198.
It is, of course, difficult to generalize about popular motives for changing institutions from the single case of the blanket primary initiative. We cannot saywith certaintythat the constituency approving the blanket primaryis the same as that approving other institutional reforms such as term limits, campaign finance regulations, or new electoral systems. Nevertheless, our main findings are consistent with studies of mass support for such governance proposals. Term-limit initiatives that were largelyopposed by major partyelites and partisans of the majority party were supported byindependents, weak identifiers, and minority partisans (Donovan and Snipp 1994; Bowler and Donovan 1998). Referendums proposing change from plurality elections to proportional representation, while opposed by governing-party elites and their strong identifiers, have also been passed with support from weak partisans and minor-party voters (Banducci and Karp 1999). If a common thread exists in these patterns, it is that proposals striking at the power of established parties receive support from citizens who might be (or perceive themselves to be) disadvantaged byrules that give power to such parties.
We began our chapter with a discussion of normative issues surrounding institutional change. Although a standard theme in this literature emphasizes longevityas a source of legitimacy, results from this and other studies of popular support for institutional change might cause us to reconsider this assumption if change can be said to bring new legitimacyto the political process. Institutions are practices that stand the tests of time and endure, yet endurance can threaten legitimacy if practices come to be at odds with the experiences and preferences of a majorityof citizens. In an era when parties have diminishing contact with citizens and when (major) parties claim a declining share of registration and popular votes, it is not surprising that the legitimacyof rules that maintain "strong" parties are occasionally challenged. As an example, consider the likelihood of the contemporary
| Anglo | 1=Anglo; 0=not |
| Minor-party registered | 1=registered voter of minor party(including those who decline to state); 0=not |
| Education | 9-point scale, low to high |
| Age | in years |
| Strong Democrat | 1=Strong Democratic identifier; 0=not |
| Strong Republican | 1=Strong Republican identifier; 0=not |
| Republican dissatisfied with GOP nominees | Question: "Are you planning to vote for your can didate mostly because you like him and his policies, or mostlybecause he is the best of a bad lot, or mostlybecause you feel like sending a protest message?" Coding: 1=if a Republican and answer was "best of a bad lot" or "a protest message"; 0=other response or not a Republican |
| Independent | 1=self-identified independent; 0=not |
| Activist | 1=respondent gave moneyto a political campaign, volunteered, or attended political meeting; 0=did not do so |
| Neither major party has best ideas | Summary value=sum of "Neither" responses to the following question: "Regardless of which candidate for President you happen to prefer right now, which candidate, if any, do you think has the best ideas for handling the following issues: Bill Clinton or Bob Dole?" The seven issues were taxes, economic problems, affirmative action, balancing the budget, illegal immigrants, environmental issues, and foreign affairs. |
| Fear of layoff | Question: "Thinking about the next twelve months, how likely do you think it is that you or someone in your household will lose a job or be laid off—very likely, fairly likely, fairly unlikely, or very unlikely?" |
NOTES
1. Subject matter categorization is always fraught with difficulties and involves a certain amount of double counting. Since our concern here is to arrive at some general overview of the number of institutional reform proposals, we accept that the Secretaryof State categorizations will provide a rough but reasonable estimate for our current purposes. [BACK]
2. Recent campaign finance reform measures such as Prop 208 and Prop 212 were counted under "Campaign Reform." [BACK]
3. This logic is not to be confused with the assumption that alienation is reflected in "negative voting," or voting no on elite proposals. [BACK]
4. The Secretaryof State cannot determine the amount spent on the three March tort initiatives (200, 201, 202) since all spending was done bycommittees contesting multiple initiatives placed on the March and November ballots. Over $83 million was spent bymultiple proposition committees that reported supporting or opposing 200, 201, 202 and other November 1996 initiatives. The committees must report contributions but are not required to report how funds were allocated to specific initiative contests. [BACK]
5. Minor-party adherents are especiallyloyal to "their" candidates in the relativelyless important races like Secretaryof State. Theyare more likelyto support major partycandidates in the more prominent gubernatorial and senate races (Donovan, Bowler, and Terrio 1999). [BACK]
6. For example, a loyal Green could support most Green Party candidates while still having the abilityto support a far more credible leftist or environmental Democratic candidate for the Senate or for Governor in an open primary. [BACK]
7. Alternatively, younger voters may simply be more receptive to any change, regardless of content. [BACK]
8. It has been shown, for example, that the well educated are more supportive of proportional representation (Banducci and Karp 1999) and more likelyto engage in strategic voting in complex settings like Mixed Member Proportional election systems (Banducci et al. 1998). [BACK]
9. Alternative specifications that looked at the effect of having a general concern that the state's situation was poor (sociotropic evaluations) had no significant impact on support. [BACK]
REFERENCES
Abramson, Paul. 1979.
“Developing Party Identification: A Further Examination of Life-Cycle, Generational, and Period Effects.”
American Journal of Political Science76: 502–21.
Banducci, Susan; Todd Donovan; and Jeffrey Karp. 1999.
“Proportional Representation and Attitudes about Politics: Results from New Zealand.”
Electoral Studies18, no. 4: 533–55.
Banducci, Susan, and Jeffrey Karp. 1999.
“Perceptions of Fairness and Support for Proportional Representation.”
Political Behavior21, no.3. 217–38.
Banducci, Susan; Jeffrey Karp; Jack Vowels; and Todd Donovan. 1998. “Strategic
Bobo, Lawrence, and Frank Gilliam. 1990.
“Race, Sociopolitical Participation, and Black Empowerment.”
American Political Science Review84: 377–97.
Bowler, Shaun, and David Denemark. 1993.
“Split Ticket Voting in Australia: Dealignment and Inconsistent Votes Reconsidered.”
Australian Journal of Political Science28: 19–37.
Bowler, Shaun, and Todd Donovan. 1998. Demanding Choices: Opinion, Voting, and Direct Democracy. Ann Arbor: Universityof Michigan Press.
Bowler, Shaun; Todd Donovan; Max Neiman; and Johnny Peel. 1999.
“Elite Attitudes about Direct Democracy.”
Paper presented at the Western Political Science Association meeting, Seattle, WA.
Bridges, Amy. 1997. Morning Glories: Municipal Reform in the Southwest.Princeton, NJ: Princeton University Press.
"California Pre-Primary Survey."1996. Los Angeles TimesPoll 372 (March).
California Secretaryof State. 1996. Official Ballot Pamphlet: Primary Election. Sacramento.
California Secretaryof State. 1998. A History of the California Initiative Process. Sacramento.
Campbell, Angus; Philip Converse; Warren Miller; and Donald Stokes. 1960. The American Voter. Chicago: Universityof Chicago Press.
Citrin, Jack; Beth Reingold; and Evelyn Walters. 1990.
“The ‘Official English' Movement and the Symbolic Politics of Language in the United States.”
Western Political Quarterly43: 553–60.
Clingermayer, James, and B. Dan Wood. 1995.
“Disentangling Patterns of State Indebtedness.”
American Political Science Review89: 108–20.
Dalton, Russell. 1984.
“Cognitive Mobilization and Partisan Dealignment in Advanced Industrial Democracies.”
Journal of Politics46: 264–84.
Donovan, Todd, and Shaun Bowler. 1997.
“Direct Democracyand Minority Rights: Opinions on Anti-Gayand Lesbian Ballot Initiatives.”
In S. Witt and S. McCorkle, eds., Anti-Gay Rights: Assessing Voter Initiatives. Westport, CT: Praeger.
Donovan, Todd, and Shaun Bowler. 1998.
“Responsive or Responsible Government?”
In S. Bowler, T. Donovan, and C. Tolbert, eds., Citizens as Legislators: Direct Democracy in the American States. Columbus: Ohio State University Press.
Donovan, Todd; Shaun Bowler; and Tammy Terrio. 2000.
“Support for Third Parties in California.”
American Politics Quarterly28, no.1(October): 50–71.
Donovan, Todd, and Joseph Snipp. 1994.
“Support for Legislative Term Limits in California: Group Representation, Partisanship, and Campaign Information.”
Journal of Politics56: 492–501.
Donovan, Todd; Tammy Terrio; and Shaun Bowler. 1997.
“Loyal Attachments and Protest Voting for Third Parties in State Elections.”
Paper presented at the Western Political Science Association meeting, Tucson, AZ.
Dubois, P.; Floyd Feeney; and E. Constantini. 1991. The California Ballot Pamphlet: A Survey of Voters. Report prepared for the Secretaryof State of California. Sacramento.
Gerber, E.; A. Lupia; M. McCubbins; and D. R. Kiewiet. 2000. Stealing the Initiative. Upper Saddle River, NJ: Prentice Hall.
Gibson, James, and Gregory Caldeira. 1998.
“Changes in the Legitimacyof the European Court of Justice: A Post-Maastricht Analysis.”
British Journal of Political Science28: 63–91.
Hensler, D., and C. Hensler. 1979. Evaluating Nuclear Power: Voter Choice on the California Nuclear Initiative. Santa Monica, CA: Rand Corp.
Hibbing, John, and Elizabeth Theiss-Morse. 1995. Congress as Public Enemy. Cambridge: Cambridge University Press.
Hibbing, John, and Elizabeth Theiss-Morse. 1998.
“Too Much of a Good Thing: More Representative Is Not Necessarily Better.”
PS: Political Science and Politics31: 28–31.
Jesse, Eckhard. 1988.
“Split-Voting in the Federal Republic of Germany: An Analysis of the Federal Elections from 1953 to1987.”
Electoral Studies7:109–24.
Karp, Jeffrey. 1995.
“Explaining Public Support for Legislative Term Limits.”
Public Opinion Quarterly59: 373–91.
Lupia, Arthur. 1994.
“Short Cuts versus Encyclopaedias: Information and Voting Behavior in California Insurance Reform Elections.”
American Political Science Review88: 63–76.
Lupia, Arthur, and Mathew McCubbins. 1998. The Democratic Dilemma: How Citizens Learn What They Need to Know. New York: Cambridge University Press.
Magleby, David. 1984. Direct Legislation: Voting on Ballot Propositions in the United States. Baltimore, MD: Johns Hopkins University Press.
Magleby, David, and Kelly Patterson. 1996.
“Political Knowledge and Term Limits: Can Angry Citizens Be Educated?”
Paper presented at Western Political Science Association meeting, San Francisco, CA.
March, James, and Johan Olson. 1984.
“The New Institutionalism: Organizational Factors in Political Life.”
American Political Science Review78: 734–49.
Moncrief, Gary, and Joel Thompson. 1991.
“The Term Limitation Movements: Assessing the Consequences for Female (and Other) State Legislators.”
Paper presented at Western Political Science Association meeting, Seattle, WA.
Mondak, Jeffrey. 1995.
“Elections as Filters: Term Limits and the Composition of the U.S. House.”
Political Research Quarterly48: 701–27.
Polsby, Nelson. 1983. Consequences of Party Reform. New York: Oxford University Press.
Riker, William. 1982. Liberalism against Populism. San Francisco CA: W. H. Freeman.
Schrag, Peter. 1998. Paradise Lost: California's Experience, America's Future. New York: The New Press.
Sears, David O., and Jack Citrin. 1982. Tax Revolt: Something for Nothing in California. Cambridge, MA: Harvard University Press.
Sniderman, Paul; Richard Brody; and Philip Tetlock. 1991. Reasoning and Choice: Explorations in Political Psychology. New York: Cambridge University Press.
Stanley, Harold, and Richard Niemi. 1991.
“Partisanship and Group Support, 1952–1988.”
American Politics Quarterly19: 189–210.
Tolbert, Caroline. 1998.
“Changing the Rules for State Legislatures: Direct Democracyand Governance Policies.”
In Shaun Bowler, Todd Donovan, and C. Tolbert, eds.,Citizens as Legislators: Direct Democracy in the American States. Columbus: Ohio State University Press.
Tolchin, Susan. 1996. The Angry American: How Voter Rage Is Changing the Nation. Boulder, CO: Westview Press.
Wolfinger, Raymond, and Fred Greenstein. 1968.
“The Repeal of Fair Housing in California: An Analysis of Referendum Voting.”
American Political Science Review2:753–69.
4. Context and Setting
The Mood of the California Electorate
Mark Baldassare
In this chapter, we analyze the mood of California voters during the 1998 primary. We use the Public Policy Institute of California (PPIC) Statewide Surveys conducted in April and May 1998, each with a total of two thousand adults, to gauge the political, social, and economic attitudes that provide the backdrop to voters' ballot choices (Baldassare 1998a, 1998b).
Voters nominated high-profile political insiders in both parties' gubernatorial races. The winners, Attorney General Dan Lungren and Lieutenant Governor Gray Davis, were conventional candidates, seemingly the kind who would have won in a closed primary. With the two-term Governor, Pete Wilson, termed out of office, voters chose the next-highest-ranking Republican and the highest-ranking Democrat in statewide elected office to run against each other in November. The losers in the governor's race, businessman Al Checchi and U.S. Representative Jane Harman, were the wealthy political outsiders who promised to shake up the Sacramento political establishment. In the U.S. Senate primary, the outcome was similarly predictable. Republican millionaire businessman Darell Issa lost to State Treasurer Matt Fong. The next-highest-ranking Republican in state office would face Democratic incumbent Barbara Boxer in the November election.
Why did voters make these choices? Why were Californians suddenly enamored with partisan favorites and the people who were holding high state offices? It was hard to imagine it was because they were feeling a great deal of trust and confidence in elected officials and state government. Instead, our survey evidence suggests that it was because the voters did not want to risk making changes. Some have argued that the rules of the blanket primary may have also helped incumbents (see chapter 7 by Salvanto and Wattenberg in this book), and they apparently did, as it was the current
CONFIDENCE RETURNS
California had gone from among the worst of economic times to some of the best in the brief time span of four years. State residents were in a very upbeat mood while they were pondering their ballot choices in the spring of 1998.
When asked what direction the state was headed, about six in ten said California was moving in the right direction, and only a third said California was going in the wrong direction. Residents in the three major regions accounting for most of the state's voters (that is, the Los Angeles metro area, the San Francisco Bay area and the Central Valley) all had a positive outlook. Latinos, whites, and other racial and ethnic groups were all equally positive. Four years earlier, opinions were almost the reverse of what we found in the spring of 1998. In a California voter survey that I conducted for KCAL-TV News in the spring of 1994, only 32 percent thought the state was headed in the right direction, while 56 percent believed it was going the wrong way (see table 4.1).
The good feelings about the state of the state were rooted in very positive perceptions of the California economy. Six out of ten said the state's economy was in excellent or good shape today, while a third said it was in fair shape, and only one in ten said it was poor. Positive ratings were highest in the San Francisco Bay area at 70 percent, but about half in the Los Angeles and Central Valley regions also thought the California economy was in excellent or good shape. About half of Latinos felt the state's economy was in excellent or good shape, compared with six in ten whites and other races. Most important, few in any region or demographic group thought the economy was in trouble.
Ratings of the highly coveted "quality of life" in the state had also turned positive. When asked to think about the quality of life in California, seven in ten residents said the quality of life in California was going either very well or somewhat well, while only three in ten thought that things were going somewhat badly or very badly. There were no differences in ratings of the quality of life across the major regions or racial and ethnic groups. This was a remarkable change from the public's position only four years earlier. At that time, 37 percent of KCAL-TV News survey respondents thought that things were going well, while 63 percent thought they were going badly.
The biggest economic threat on the state's horizon in 1998 was the Asian
| All | Crossover Voters | |
|---|---|---|
| SOURCE: Baldassare 1998a and 1998b. NOTE: Crossover voters are the Republicans who say they will vote for a Democrat in the gubernatorial primary. | ||
| "Is California going in the right or wrong direction?" | ||
| Right direction | 55% | 58% |
| Wrong direction | 36 | 35 |
| Don't know | 9 | 7 |
| "The economy in California today is …" | ||
| Excellent | 11% | 13% |
| Good | 46 | 48 |
| Fair | 33 | 33 |
| Poor | 10 | 6 |
| "The quality of life in California today is going …" | ||
| Very well | 13% | 15% |
| Somewhat well | 57 | 62 |
| Somewhat badly | 21 | 20 |
| Very badly | 9 | 3 |
Californians were generally upbeat about their own finances, and they were much more optimistic about improving their economic fortunes than they had been four years earlier. Almost 40 percent said they were better off financially now than they were in 1997, while half said their financial conditions were the same, and only 12 percent said their financial state had become worse. More than 40 percent expected to be better off financially a year from now, half expected their conditions to be the same, and only one in sixteen expected to be worse off in 1999. Four years earlier, only 20 percent said they were better off than last year, and only one-third expected to be better off in the next year. In the 1998 survey, 87 percent described their current standard of living as "comfortable" or "more than comfortable." A similar number described themselves as having "just
The same attitudinal trends were evident among crossover voters, that is, Republicans who said they would vote for a candidate in the competitive Democratic gubernatorial primary.[1] Most crossover voters said that California was going in the right direction (58 percent), that the California economy was in excellent or good shape (61 percent), and that the quality of life was going very well or somewhat well (77 percent). Forty percent said their own finances were improving and would get better in the next year, while few had serious concerns about job loss.
The "angry voter" in the 1994 election had been replaced by the "status quo voter" in the 1998 election. This mood swing was a major factor in determining voter choices, as it took the appeal out of calls for economic change and political reform. Candidates with a track record became more attractive to voters, while those who were espousing a shift in government were viewed as a potential threat to the economic good times.
ISSUES CHANGE
Californians were focused on different policy issues now that the economy had improved. Education had become one of the biggest concerns of the public. While the fear of crime continued to be a worry, immigration was no longer a pressing issue.
In an open-ended question, we asked residents to name the most serious problem facing California today. Crime and education topped the list of public policy concerns, with these two issues named by nearly half of all respondents (see table 4.2). No other policy issue was named by more than 10 percent of respondents, including immigration and the economy. By contrast, the economy and crime (29 percent each) were named as the biggest state problems in the 1994 KCAL-TV News survey. Just before the November election in 1994, as a result of the attention generated by Proposition 187, immigration had joined these two issues as a big concern. Few had thought of education as a top issue in the turbulent social and economic times of the 1994 elections.
Follow-up questions in the 1998 PPIC Statewide Survey indicate that many Californians were concerned about the quality of public schools. Forty-six percent said that the quality of public schools was a "big problem" in the state today. One in three described the issue as somewhat of a problem. Only one in seven said it was not a problem at all. Public school parents were only slightly less likely to rate education as a big problem. Residents of Los Angeles and the San Francisco Bay area were most likely to rank
| All | Crossover Voters | |
|---|---|---|
| SOURCE: Baldassare 1998a and 1998b. | ||
| "Most serious state problem is …" (open-ended) | ||
| Crime | 28% | 26% |
| Education | 20 | 25 |
| Immigration | 7 | 10 |
| Economy | 5 | 2 |
| "How much of a problem is … in California today?" | ||
| Rated crime a "big problem" | 66% | 71% |
| Rated public schools a "big problem" | 46 | 44 |
Californians' estimates of spending per pupil and student achievement were consistent with their deep concerns about the quality of public schools. Almost half were aware that their state spent less per pupil than other states on public schools. A quarter thought that California spending was on par with the national average, while only one in seven thought the state spent more on public schools than other states. A little more than half were aware that the state's student test scores were below the national average. A third thought that student performance was on par with the national average, while only one in ten believed that the state's test scores were above average.
Surprisingly, two in three Californians ranked crime as a big problem in the state, despite published reports that crime was decreasing. This is because nearly half of Californians believed that the crime rate had increased in recent years, another one in four thought it had stayed about the same, and only one in four thought it had decreased, as it actually had. Still, two in three Californians said they felt safe walking alone in their neighborhood at night, and only one in six described their area as "very unsafe." Perceptions of safety from local crime had actually improved somewhat from 1994.
While few Californians ranked immigration as the most pressing problem, people in all regions believed that the immigrant population of the state was still growing. Seventy-three percent said that the overall immigrant population had risen over the past few years, with nearly half maintaining that it had grown "a lot." Californians were deeply divided, however, about the impacts of immigration. Nearly as many said they perceived immigrants as a benefit to California because of their contributions to the economy as said they were a burden to California because of their use of public services. These perceptions differed sharply along racial and ethnic lines, with most
The opinions of crossover voters on state policy issues were similar to those of the general public. Most mentioned crime and education as the top issues, while few named immigration and the economy. There was considerable awareness that the state was below average in student test scores and per pupil spending. Most thought that crime was increasing in the state, but most also felt that their own neighborhoods were safe. Eighty percent recognized that the immigrant population was increasing, though they were evenly divided in describing this trend as positive or negative.
With the economy in high gear, candidates from both parties focused their messages on improving the schools and controlling crime. The state budget was in a surplus, and thus voters were not moved as much by issues such as tax and spending cuts, limiting funds for immigrants, or efforts to create more jobs and rekindle the economy. Again, this mood seemed to favor candidates in the center of the political spectrum who represented the political status quo.
POLITICAL APATHY
Most Californians did not find the June 1998 primary to be very inspiring. In fact, elections today represent non-events in the lives of many state residents.[2] Keeping this public apathy in perspective helps us to understand the importance of the political and economic climate in determining the choices voters make in state elections.
In both of the PPIC Statewide Surveys before the primary, fewer than one in five of respondents expressed a great deal of interest in politics, about half said they had a fair amount of interest, and one in three expressed little or no interest in politics (see table 4.3). Similarly, 75 percent of Americans interviewed by the Pew Research Center in 1996 reported a great deal (25 percent) or fair amount (50 percent) of interest in politics. Consider the fact that the low political interest scores found in both of the 1998 PPIC Statewide Surveys were gathered with the backdrop of the Lewinsky-Clinton sex scandal in the spring, an event one would have expected to heighten interest in political news.
Political interest was not much higher among registered voters. About two in ten voters said they had a great deal of interest in politics, half had a fair amount of interest, and three in ten had little or no interest. Democrats and Republicans were alike in their low level of political interest, and both were more likely than independent voters to have a great deal or fair amount of interest in politics. Still, fewer than one in four voters in any of the political groups were highly interested in politics.
Only about a third of Californians said they followed what was going on
| All | Crossover Voters | |
|---|---|---|
| SOURCE: Baldassare 1998a and 1998b. | ||
| "How much interest would you say you have in politics?" | ||
| A great deal | 16% | 15% |
| Fair amount | 47 | 50 |
| Little or none | 37 | 35 |
| "Would you say you follow what's going on in government … ?" | ||
| Most of the time | 35% | 39% |
| Some of the time | 38 | 38 |
| Only now and then | 19 | 19 |
| Hardly ever; never | 8 | 4 |
Registered voters were twice as likely as those who were not registered to say they regularly follow government and public affairs. About four in ten registered voters said they follow government issues most of the time. There were no large differences across party lines. In all, fewer than half of the state's voters said they were closely attuned to the world of government and public affairs.
What is impressive is the consistency in the responses to this question over the election cycle. The level of interest did not change very much as the primary grew closer, even among those defined as most likely to participate based on past voting. In April 1998, 9 percent of likely voters were following news stories about the upcoming elections "very closely," 43 percent said "fairly" closely, and 48 percent said "not too" or "not at all" closely. In May 1998, 13 percent were "very" closely following the election news, 48 percent said "fairly" closely, and 39 percent said "not too" or "not at all" closely.
There was also a high level of political apathy among the crossover voters. A third reported little or no interest in politics, and almost a quarter said they infrequently followed government and public affairs. About half said they followed news about the election "not too closely" or "not at all closely" in both April 1998 and May 1998.
The public's lack of interest in politics and election news has taken its toll on knowledge of even the most basic facts about state politics. In a December 1998 PPIC Statewide Survey, I asked an open-ended question, "California voters elected a new Governor on November 3. Could you give me the name of the new Governor of the State of California?" Fifty-three percent named Governor-elect Gray Davis, 5 percent gave other names and 42 percent said they were not sure about the name of the new Governor.
TELEVISION RULES
How does a mostly disinterested public learn about upcoming elections? Largely from television, which Californians ranked as their top source of information in response to the question "Where do you get most of your information about what's going on in politics today?" (see table 4.4). When we asked this survey question in both April and May 1998, 41 percent named television as the source of most of their political information, and 34 percent named newspapers. The remaining one in four residents said they got most of their political news from the radio (10 percent), talking to people (7 percent), magazines and the Internet (3 percent each), and other sources (2 percent). Even among registered voters, television had a slim lead over newspapers as the major source of political information. There were no differences in the sources of political information across party lines. Of those not registered to vote, more than half named television as their major source of political information, while only one in four named newspapers.
These results were similar for the crossover voters. Thirty-nine percent relied mostly on television, and 36 percent depended mostly on newspapers for political news. More than eight in ten said they recalled the television commercials for the gubernatorial candidates. When asked whose ads they had seen the most, most of those who had seen ads mentioned the Checchi ads.
There is little doubt that Californians are hooked on television news. State residents were much more likely to say they watched television news every day than to say they read a daily newspaper (59 percent to 45 percent). Even the registered voters were more likely to watch local television news every day than to read a daily newspaper, by a wide margin. Californians are less likely than adults nationwide to be reading a newspaper every day (45 percent to 51 percent), and they are also slightly less likely to watch local television news on a daily basis (62 percent to 59 percent), as reported elsewhere (Baldassare 2000). Overall, it appears that relatively few Californians make an active effort to gather election news and political information.
In addition to television news, Californians gathered much of their information
| All | Crossover Voters | |
|---|---|---|
| SOURCE: Baldassare 1998a and 1998b. | ||
| "Where do you get most of your information about what's going on in politics today?" | ||
| Television | 41% | 39% |
| Newspapers | 34 | 36 |
| Radio | 10 | 12 |
| Talking to people | 7 | 6 |
| Magazines | 3 | 3 |
| Internet, online services | 3 | 3 |
| Other | 2 | 1 |
In the end, the candidates who spent more money for their campaigns lost the major state races. The recall of political advertising is a testament to the power of television in reaching a broad audience with a political message. But the fact is that the record-setting spending on television commercials proved insufficient to propel the self-financed outsider candidates to victory.
DEEP DISTRUST
How did Californians regard their government and elected officials when they went to the polls in the spring of 1998? The survey evidence here is very consistent. Despite their rosier outlooks on the state economy and their personal finances, many continued to feel disillusioned with government
| All | Crossover Voters | |
|---|---|---|
| SOURCE: Baldassare 1998a. | ||
| "How much of the time can you trust the government in Washington to do what is right?" | ||
| Always or most of the time | 26% | 26% |
| Only sometimes | 62 | 61 |
| Never | 12 | 13 |
| "When something is run by the government, it is usually wasteful and inefficient." | ||
| Agree | 62% | 64% |
| Disagree | 38 | 36 |
| "Most elected officials care what people like me think." | ||
| Agree | 51% | 52% |
| Disagree | 49 | 48 |
| "Most elected officials are trustworthy." | ||
| Agree | 56% | 61% |
| Disagree | 44 | 39 |
In response to a question repeated from the National Election Studies (1996), only a quarter of Californians said the federal government could be trusted to do what is right either all of the time or most of the time (see table 4.5). Six in ten trusted the federal government "only sometimes," while 12 percent said they never trusted the federal government. The Pew Research Center in 1998 reported that 34 percent of Americans had a high level of trust in the federal government. So, Californians were expressing less confidence than the nation as a whole. Twenty-six percent of the state's voters trusted the federal government "always" or "most of the time." By comparison, a similar 26 percent of the crossover voters were also very trusting. Few in any voter group said they always or mostly trust the federal government.
Californians were not very trusting of government when it came to its efficiency and fiscal performance. Two in three residents believed that "when something is run by the government, it is usually wasteful and inefficient." A similar number of crossover voters held this view. In the Pew Research Survey in 1997, a similar 64 percent of Americans agreed that there was a lot of waste in government.
Nor do state residents see their elected leaders as particularly responsive to their needs. Only half of Californians agreed that "most elected officials
The lack of trust in elected officials was very closely replicated in another PPIC Statewide Survey question. Fifty-six percent of Californians agreed that "most elected officials are trustworthy," while 44 percent did not. The findings for crossover voters were, again, fairly similar. And, in a Pew Research Center Survey in 1997, 51 percent of Americans thought that most elected officials were trustworthy.
Californians' distrust in government is not limited to the federal level; they also showed little confidence in their state government. In a PPIC Statewide Survey conducted in the fall of 1998, only one in three Californians said they could trust the state government in Sacramento to do what was right either always or most of the time. About half of Californians thought that the state government wasted a lot of the money that was paid in taxes. Two in three residents saw the state government as pretty much run by a few big interests looking out for themselves. Californians are, in general, only a little less cynical about their state government than their national government.
Few Californians report a lot of confidence in their elected leaders. In the April 1998 PPIC Statewide Survey, only 11 percent said they had a great deal of trust and confidence in Governor Pete Wilson when it came to solving state problems. As for the California Legislature, only 4 percent had a great deal of trust and confidence in legislators to solve state problems. Their mayors and city councils (16 percent) were more trusted in solving city problems, though the number of Californians having a great deal of confidence in their Board of Supervisors (8 percent) to solve county problems was about the same as for the California Legislature. More Californians expressed a great deal of confidence in President Clinton (30 percent) than in the U.S. Congress (9 percent) in terms of handling national problems. For crossover voters, the results were very similar and the findings indicate a lack of confidence in all levels of government.
Californians gave mixed reviews to political leaders and legislative bodies. Most gave high ratings to President Clinton's overall job performance, despite ongoing investigations into his actions. Six in ten said he was doing an excellent or good job as President, though those ratings varied widely by party affiliation. Californians were much less generous towards Congress, the California Legislature and Governor Wilson, with one in three giving excellent or good ratings to each of them. The patterns in job ratings were, once again, very similar among the crossover voters. However, job performance ratings were somewhat higher than they were during the 1994 election.
MAKING CHOICES
Voters were in a cynical mood about politicians and government in the spring of 1998. Yet they liked the way things were going in the state. As a result, their preferences for candidates in the June primary were in the direction of making "safe" status quo choices, as is evident in their preferences for candidate qualifications and their ballot choices.
In the April 1998 PPIC Statewide Survey, California voters surprised us when 44 percent said they preferred statewide candidates with a track record in elected office. Forty percent said they wanted candidates who were political newcomers. A month later, in the May 1998 PPIC Statewide Survey, there was an eight-point margin in favor of seasoned politicians over political outsiders (46 percent to 38 percent) when voters were asked about the candidate qualifications they preferred (see table 4.6). Democrats were more likely to prefer experienced politicians than were Republicans and independents. In contrast to most other questions in the PPIC Statewide Surveys, the responses of crossover voters differed substantially, with only 38 percent saying experience in elected office was the most important qualification for candidates, and 48 percent saying experience in running a business was most important. Still, across all groups of respondents, the degree of emphasis placed on experience in office is in stark contrast to their deep distrust of elected officials.
In another rebuke of outsider politicians, a third of the voters in the April 1998 PPIC Statewide Survey said they would be less inclined to vote for candidates who spent millions of their own dollars for political campaigning, while only 11 percent said they would be more inclined to vote for such candidates. About half said this would make no difference in their ballot choices. In the May survey, 53 percent said they favored the candidates who raised money from their supporters to pay for their political campaigns, while only 35 percent said they preferred candidates who can spend their own money. There were no differences across parties. These findings were consistent with the success of the candidates in the primary. Those who had gone the conventional route of collecting contributions to finance their campaigns won, while the self-funded outsiders lost.
Voters thus went into the June primary in a risk-adverse mood. They wanted choices that would maintain the status quo of good economic times and avoid slipping back into the deep recession of the not-so-distant past. With this in mind, there would be no one like Jesse "The Body" Ventura emerging as the candidate for Governor of California or U.S. Senator in the state's first-ever blanket primary.
| All | Crossover Voters | |
|---|---|---|
| SOURCE: Baldassare 1998b. | ||
| "People have different ideas about the qualifications they want when they vote for candidates for state wide office, such as Governor or U.S. Senator. Which of these is most important to you?" | ||
| Experience in elected office | 46% | 38% |
| Experience running a business | 38 | 48 |
| Other | 10 | 10 |
| Don't know | 6 | 4 |
The voters' wish list for candidates' qualities gave Lungren and Davis for Governor and Boxer and Fong for the U.S. Senate big advantages in June. Voters were unwilling to take a chance with politicians untested in statewide offices, such as Al Checchi, Jane Harman, and Darrell Issa, who each greatly outspent their rivals. Most Republicans stayed with Lungren and Fong, while most Democrats voted for Davis and Boxer. The voters had spoken. They didn't want a radical change in their state's elected officials to get in the way of the good times that were underway.[4]
CONCLUSION
California voters participated in their first-ever blanket primary in June 1998. In this election, voters were faced with a wide array of ballot choices. The blanket primary is an opportunity for political scientists to analyze voting trends for signs of strategic voting and other specific efforts to change the outcome of elections where party candidates are chosen. This was an unusual election for many reasons besides the change in primary rules. There were record amounts of money spent because of the entry of three millionaire candidates for Governor and U.S. Senator. The President faced a sex scandal and impending impeachment that had captured much of the attention that the media devotes to politics. The economy was booming in California after one of the darkest periods since the Great Depression.
Because of several unique features of the 1998 primary, we should not draw too many firm conclusions from this election. More observations would be needed to assess the fullest possible impact of crossover voting by party members and participation among independents. Still, it is worth speculating about the results of the 1998 election. The rules of the blanket primary are seen as reinforcing incumbency advantages and, thus, the political status quo (see chapter 7 by Salvanto and Wattenberg in this book).
Voters showed a strong preference for current state elected officeholders rather than political outsiders. In terms of candidate choices, then, the experience in 1998 may be typical of what we would expect from elections held under a blanket primary system.
This chapter focused on the impacts of political context and economic setting on California's blanket primary. I considered the two major state races, that is, the races for the Governor's office and the U.S. Senate. The outcome was highly conventional and one that would have been expected in a closed primary. The victorious Democrats were the highest-ranking Democratic state elected officeholder for Governor and an incumbent for U.S. Senator. The victorious Republicans were the two highest-ranking Republican state elected officeholders, since the Governor had been termed out of office. It appears that the good economy, in combination with the blanket rules that might have worked to reinforce incumbency advantages, produced an outcome that was dominated by experienced candidates who were already holding statewide offices. In contrast, the wealthy outsider candidates were defeated in the primary, and then the Democratic candidates won in the general election.
In sum, the voters' mood is an important key to understanding the ballot choices made in the June 1998 primary. Most Californians thought that things were going well in the state. There was a dramatic rebound from four years earlier in public attitudes towards the economy, quality of life, and personal finances. As a result, the state issues people cared most about changed from the economy and immigration to crime and the quality of education. Most Californians remained politically apathetic and cynical about government. They preferred the career politicians who were closely aligned with the major parties because these were the candidates providing the least risk that government would interfere with prosperity. In another context, the blanket primary rules may have had a different impact on the election. We would have to observe other elections under a blanket primary system in bad economic times to make that judgment. The 1998 primary was not one in which voters were looking for political tools that would provide them with significant electoral change.
NOTES
I wish to thank Jonathan Cohen and Ana Maria Arumi for research assistance. Bruce Cain and Elisabeth Gerber provided helpful comments on an earlier draft.
1. For the purposes of this analysis, crossover voters are defined as Republicans who said they would vote for Democratic candidates in the gubernatorial primary. In all, there were 333 crossover voters in our two preelection surveys, including 175 Republicans in April 1998 and 158 Republicans in May 1998. We did not include Democrats who voted for Republicans, since there were too few for separate analysis,
2. As another indication of low public interest in state elections, about three in ten adults eligible to vote went to the polls in June 1998, and about four in ten adults eligible to vote went to the polls in November 1998 (California Secretary of State 1998a, 1998b). [BACK]
3. The theme of voter distrust is also evident in policy preferences. Californians tend to be liberal on social issues and conservative on fiscal issues (i.e., "New Fiscal Populists"); in other words, they prefer that the government play a limited role in their lives (see also Baldassare 2000; Clark and Inglehart 1998; Schrag 1998). [BACK]
4. In the statewide surveys before the November 1998 election, more than half of the voters said that the candidate's stands on the issues mattered the most to them, while only two in ten said character was most important when they go to vote for Governor. After the election, Lungren's pollster said that the focus on character over issues in the gubernatorial campaign was a "strategic fatal error" (Skelton 1999). [BACK]
REFERENCES
Baldassare, Mark. 1998a. PPIC Statewide Survey: April 1998. San Francisco: Public Policy Institute of California.
Baldassare, Mark. 1998b. PPIC Statewide Survey: May 1998. San Francisco: Public Policy Institute of California.
Baldassare, Mark. 2000. California in the New Millennium: The Changing Social and Political Landscape. Berkeley: University of California Press.
California Secretary of State. 1998a. Statement of the Vote: June 1998. Sacramento.
California Secretary of State. 1998b. Statement of the Vote: November 1998. Sacramento.
Clark, Terry, and Vincent Hoffman-Martinot, eds. 1998. The New Political Culture. Boulder, CO: Westview Press.
California Secretary of State. Clark, Terry, and Ronald Inglehart. 1998.
“The New Political Culture.”
Pp. 9–72in Clark and Hoffman-Martinot 1998.
Gissinger, Steve. 1998.
“Top Gubernatorial Vote-Getters Spend the Least, Records Show.”
Orange County Register, August 4.
National Election Studies. 1996. National Election Studies. Ann Arbor: University of Michigan.
Schrag, Peter. 1998. Paradise Lost: California's Experience, America's Future. New York: New Press.
Skelton, George. 1999.
“Election Autopsy Shows Lungren Was DOA.”
Los Angeles Times, January 25