Rose Institute Q&A: CA’s Top-Two Election System (part 3)

 II. Looking at Election Outcomes

What are the pathways for the top-two system to change how politics works?

There are at least five possible ways by which the top-two system can change politics.  These suggest what kinds of election outcomes should be examined.      

  1. The mix of candidates. The top-two system can encourage more moderate, pragmatic, or competent candidates to run for office.  Such candidates could be enticed into running based on the premise that the new electoral rules would increase their chances of winning: if they do win, politics would change because the officeholders would change.  Skeptics can point to several interesting lines of scholarly research that suggest this effect may not occur because: it may be difficult to encourage such candidates to run for office under any circumstances;[1] parties may coordinate informally to shape the field, rendering the formal rules useless;[2] and parties may issue endorsements that help secure the election of their preferred candidates in contested elections.[3]      
  1. The primary electorate.  By allowing for the full participation of unaffiliated voters and by potentially encouraging more participation overall, the top-two system could change the composition of the electorate (who the voters are and what they want).  A different type of electorate could either help different types of candidates win or provide incentives for the same types of candidates to behave differently.  This effect could address the ideological extremity of primary electorates, a central concern of reformers.[4]  Some initial findings were not particularly optimistic that the top-two system changed the electorate,[5] but this question requires careful and continued evaluation (see: “How should we count turnout?”).     
  1. Voter choice in the primaries. Even without changes to the composition of the electorate, the behavior of the electorate could change.  For example, Republicans in overwhelmingly Democratic districts could work with independents and centrist Democratic voters to support moderate Democrats in the primary (and then in general elections).  Some of the earliest work on the top-two focused on this very possibility, although it appears that voters, at least initially, tended to vote “sincerely” in the primary—for a candidate they identify as having views most similar to theirs.[6] 
  1. Same-party general elections. When two candidates of the same party place first and second in the primary, the general election now has the function of a partisan primary: choosing which candidate of that party will represent the district.  Yet: all of the voters can participate, in the higher-turnout general election, with voters of the other major party and all of the independents and third-party voters eligible to participate as well.  So even if nothing changes at the primary stage, the same-party general elections can be a pathway to change the incentives of politics.  The questions tend to be: do they happen often enough, and under the right circumstances, to meaningfully change politics?[7]
  1. Satisfying the public’s values. There is a rapidly-growing literature in public administration regarding the consequences of administrative burdens, the demands policies and institutions place on citizens, and the ways citizens learn about their position in society based on their experiences.[8]  The top-two, for independent voters, represents full inclusion in political life with neither formal barriers to participation nor additional requirements to request a party ballot.  Similarly, for voters of all types, the top-two may appeal to their sense of fairness, increase their sense of political efficacy, or otherwise satisfy their expectations about what a good process ought to be.[9]  In short: people might simply like it.  This is a kind of good unto itself, but also has other potentially positive downstream effects in a society otherwise deeply polarized and with a rising sense of populist dissatisfaction.

How will we know if the top-two system has worked?It is hard to assess whether the top-two has “worked,” in part because people have different definitions of what it means for the system to “work,” and in part because some political changes are difficult to evaluate.  This section focuses on the observable outcomes that can feed into those five pathways for influencing the nature of politics in California.

It is also important to note that we do not actually need to observe any of these events occurring for the potential of them occurring to influence politics.  That is, even if roughly the same number of people vote, and roughly the same kinds of candidates win elections, those same politicians could act differently once in office.  Why?  If the top-two procedure threatens their electoral prospects in a new way—or relieves pressure on them, with the threat of being “primaried” by partisan activists diminished—the elected officials may adapt to the new circumstances by changing their behavior.

How often, and under what circumstances, do same-party general elections occur?

Of the potential pathways for influencing politics, the risks and opportunities provided to candidates by same-party general elections come with the fewest demands on voters and can only occur with some version of nonpartisan rules.

Understanding how often, and under what circumstances, the same party elections occur can help us understand how the top-two procedure functions in practice.

How often.About one out of six general elections covered by the top-two rule features two candidates of the same party.  While the exact percentage varies from year to year, so far it has always been more than 10% and less than 20% of all of the elections covered by this procedure in each year.  Most of these same-party elections are Democrat-on-Democrat contests, although not all of them.

Where and when.Generally, same-party general elections are most likely to occur for open seats (elections without incumbents) and in highly partisan districts.[10]  The reason we see more Democrat-on-Democrat races is that California has more heavily Democratic districts than equivalent Republican ones; with the Republican Party continuing a long decline over many years, Democrats dominate many areas of the state.

In a recent broad study of American primary elections, Shigeo Hirano and James Snyder argue that the benefit of primary elections – relative to other kinds of nominating systems – is that they “are contributing to the US electoral system especially in uncompetitive areas, where they are most needed.”[11]  While the authors do not reach this conclusion themselves, it is a natural extension of their argument: the top-two system in California is doing the same thing, in bringing competition to uncompetitive areas, but sometimes moving that contest from the primary to the general election.


Do same-party elections promote competitiveness?


The top-two creates competitive general elections in safe seats.In particular, there have been competitive Democrat-on-Democrat general elections in districts that would otherwise be very safe seats and effectively won in the primary (see “How should we count turnout?” for an example).  It is very rare for candidates to be entirely unopposed in the general election, as well, although not all same-party general elections are equally competitive.  An important question is whether such competitive same-party elections happen often enough, or whether parties are able to ensure that their preferred candidates avoid them.[12]

Political scientists Gerald Gamm and Thad Kousser have demonstrated that states with cross-party competition enact different types of policies than states dominated by one party.  Competition encourages legislators to pass “broad bills” that serve community interests rather than “particularistic policy” to benefit more narrow interests.[13]  The problem is that it can be difficult to create cross-party competition in states with substantial majorities favoring one party.  Occasionally, candidates from the minority party win elections—for example, Republican Arnold Schwarzenegger in California—but these tend to be extraordinary exceptions.[14]   An important ongoing question is whether increased competition through same-party general elections is enough to encourage this kind of broad-based policymaking.

Does the top-two facilitate anti-competitive manipulation?  In both partisan primaries and the top-two system, candidates and parties have incentives to behave strategically.  This is part of why election rules matter: any procedure a society adopts for aggregating its preferences can influence which outcomes and officeholders are chosen.[15]

In particular, under the top-two rules, a candidate of the majority party has an interest in clearing the field to prevent a same-party general election because those elections are riskier and more competitive when they do occur.  If the candidate and party organization fail to keep out competitors, they may hope that enough voters will stick with a candidate from the weaker party in the primary to generate a more traditional Democrat-on-Republican general election race.

In most Democrat-on-Republican general elections, the winners obtain their party’s district presidential vote share plus or minus a few percentage points.  The party label does a lot of the work.

There is some anecdotal evidence about strategic behavior.  In the 2022 primary, one emerging story is the apparent preference of the Democratic Party’s endorsed candidate (or, at least, some of his supporters) for attorney general, incumbent Rob Bonta, to face a Republican instead of a Republican-turned-independent, Anne Marie Schubert.[16]  The effectiveness of such an effort can be evaluated after the primary; with two Republicans potentially splitting the Republican vote, encouraging Republicans to ditch one for the other runs the risk of accomplishing the reverse of the intended goal and actually helping Schubert.[17]

In some recent research, political scientist Jesse Crosson found that “the key to the system’s effectiveness lies in reformers’ ability to find ways to encourage more same-party competition.”[18]  Some candidates did seem able to avoid these kinds of same-party races.

This topic comes up often, especially since political stories that involve intrigue and strategy are much more interesting than ones that do not.  Still, the topic should be considered in light of three larger points of context:

  1. Most “manipulations” aim to produce the circumstances that would have occurred if the state used a partisan primary: a Democrat-on-Republican race in an uncompetitive district. If the problem is that the top-two system does not produce enough same-party general elections, reverting to a partisan system that produces none is not a solution.
  2. These same party elections do occur; when they occur, they can produce more competitive elections in places that otherwise would not have them; and elections in seats that ought to be competitive remain very competitive.
  3. Strategic behavior happens in other kinds of primaries and happened in California before Proposition 14. The top-two procedure makes some kinds of strategic behavior difficult—e.g., voters “raiding” in a party primary[19] and campaigns supporting third-party candidates in an effort to siphon off votes from major-party general election candidates.[20]

Any discussion of this topic needs to compare strategic behavior in the top-two to strategic behavior in other systems.  A famous example of attempted strategic behavior comes from California’s party primary era: in the 1966 gubernatorial election, incumbent Democrat Pat Brown apparently viewed one of his two likely Republican opponents—the moderate Republican mayor of San Francisco—as a more serious threat than the other one, an inexperienced actor associated with the more right-wing of the Republican Party.  Brown tried to influence the Republican primary, seeking to damage the Republican moderate by recirculating to journalists a prior conviction for violating a milk-pricing law when he had operated a dairy.[21]  And the actor?  Ronald Reagan.  It turned out he was not easily defeated—for governor or, later, for president.

It is certainly the case that political scientists should continue to evaluate the strategic response of candidates and parties to the new institutions.  Crosson put it well: this issue highlights “the importance of understanding how political parties and incumbents will react to institutional reforms, particularly when those reforms are not likely to benefit them.”[22]  A balanced assessment of the top-two will keep both points in mind: there is some kind of strategic response, but it is a response to the threat posed by competitive same-party general elections.  


Can a party be “shut out” of an election it ought to win?

Yes, although this happens only rarely.

Before the 2018 California primary, an editorial writer for the Washington Post warned:

[I]n several competitive Southern California districts in Republican hands, so many Democrats are running that party leaders fear the Democratic vote will end up badly splintered. That could mean no Democrat makes it to the November ballot in those districts, which would be an unexpected self-inflicted blow to the party’s hopes of taking control of the House.

The author warned that the top-two was a “cautionary tale about how good intentions alone are not enough.”[23]  This was a common concern in advance of the 2018 elections and has been raised as a possible objection to the top-two procedure.  Some additional information may help put this threat in perspective.

There is one famous exampleMost of the concern about the majority party being shut out of the general election is based on the outcome of a single contest, the Congressional District 31 election from 2012 (see: “What happened in CD31 in 2012?”).  The state has now held hundreds of elections under the top-two. For a party to be left off the general election ballot for a seat it ought to win, a lot of things have to go wrong for that party’s best candidate.  It can happen but is an extraordinary event.

Any party can win a competitive seat.The House races that caused so much anxiety in 2018 all ended up with Democrat-vs-Republican general elections, and the Democratic Party handily won control of the U.S. House.  Yet, what if one of those races had ended up as a Republican-on-Republican general election?  The thought experiment is worthwhile: these were marginal seats, competitive between parties.  That means, in states with traditional party primaries, it would not be surprising if either party won the seat.  While the path to the result would have been unusual in California, the result itself—by definition—would not have been.  It is not a shocking event for a Republican to win a competitive seat; either major party can win a competitive seat.  It is really only in the safe seats for one party that a victory for the other party would produce something truly unusual and—from the perspective of representation—possibly concerning (of which, as above, there is one notable example: CD31 in 2012).

Parties lose winnable seats with other primary types too.In terms of representation, a strange event did happen in 2017: Alabama elected a Democrat, Doug Jones, to the United States Senate.  Despite the endorsement of many national Republicans, including President Trump, incumbent (appointed) Senator Luther Strange lost to the more populist Roy Moore in the Republican primary.  The electorate rejected Moore—barely—in favor of Jones in a state in which Trump beat Clinton by about 25 percentage points.  Political scientist Andrew Hall has noted that nominating extremists in partisan primaries can sometimes produce moderation—because periodically the extremists lose seats that a party ought to otherwise win.[24]  These are rare events too, but can happen.

What about Republicans in statewide races in California?Significantly, the 2016 and 2018 U.S. Senate elections in California featured two Democrats.  Is that a problem?  A variety of arguments exist for and against the value of same-party general elections for these kinds of statewide offices.  Those discussions ought to take place, though, understanding the context: President Trump won only about a third of the vote in 2016 and 2020 in California and the Democrat-on-Republican statewide races of the top-two era have not been competitive.  The state as a whole meets most definitions of a safe Democratic seat; in presidential elections, California has been as safely Democratic as Alabama has been safely Republican.

This feature of the top-two is a significant difference between top-two and “Final Four/Five.”  The new “Final Four” system in Alaska and the proposed “Final Five” reform essentially remove the risk of a majority shut-out by advancing more candidates to the general election.  The trade-off is that the primary becomes much less important.  There is a longer discussion of this issue at the end of this document (see: “How does top-two compare to final-four or final-five?”).   


Do independent or third-party candidates often win elections?

Independent and third-party candidates have not been very successful under the top-two system, although they were not very successful before it either, and in some ways the top-two rule provides opportunities for credible third parties that have not yet been realized.

Have independent or third-party candidates won elections with the top-two? 

Yes, but it is a rare event.  In 2020, former Republican Chad Mayes won the election for State Assembly District 42 as an independent (“No Party Preference”) candidate.  Mayes, the former Republican Assembly leader, had left the party along with fellow Assemblymember Brian Maienschein; Maienschien, though, became a Democrat and won his seat under that label.[25]  Mayes was involved in an effort to create a third party (“New Way California”), which has not yet achieved electoral success.

Have independent or third-party candidates often won elections in the decade before the top-two?

No.  The difference between California’s old and new primary systems is the stage of the process when most third-party candidates are defeated.  Under the old system, third-party candidates would appear on the general election ballot.  With the new top-two system, third-party candidates occasionally make it to first or second place and appear against a Republican or Democrat in the general election, but most third-party or independent candidates are defeated at the primary stage.

Despite the growth of the No Party Preference voters in California, Republican and Democratic candidates have dominated politics in the state during both the eras of partisan and nonpartisan primaries.  This pattern is largely the consequence of having single-member districts, which tend to support a two-party system.[26]  For that reason, some reformers would like to create multi-member districts in order to support more parties.[27]

Are there opportunities for third parties in California? 

Yes, although these opportunities have not been realized so far.  The top-two system allows voters dissatisfied with the likely winner to coalesce around an alternative, in part because there are only two finalists.  A district could have, for example, a Democratic finalist and a “not Democrat” alternative.  That is, broadly, how the (former Republican) independent Assemblymember Chad Mayes was elected in Assembly District 42 in 2020: his presence in the three-way primary race narrowly pushed out the Democratic candidate and created a Republican-vs.-independent general election.  Mayes won handily, 56%-44%.

Nevertheless, it is difficult to move beyond the American two-party system and the difficulty is not new.  For example, progressive Republicans in the age of Hiram Johnson struggled with whether to stay in the Republican Party or break away to form a new party.[28]  The alternative is using a party label while creating an identifiable subgroup brand (such as “progressive” or “Tea Party”); there are exceptionally low barriers to candidate affiliation with political parties[29] in the United States, so candidates of various types can be successful while using a single party label.


Has voter turnout increased?

Increased turnout was one of the main expectations associated with Proposition 14.[30]  Unfortunately, as with the election of more moderate or collaborative legislators, this effect is difficult to evaluate.  There are several issues to consider: overall participation in the primary and general elections, participation in specific contests, and the meaningfulness of the participation.  This answer provides a brief overview; see the additional questions at the end for a longer explanation (“How should we count turnout?”).

Beyond questions about how to count turnout, it can be very hard to distinguish the effects of changes to the primary system from changes in other political institutions.  For example, starting in 2012, California moved all ballot propositions off the primary ballot to the November general election ballot, removing an important reason for voters to participate in primaries.[31]

Overall participationThe Secretary of State’s office releases turnout numbers that reflect the total number of registered voters who cast a ballot in the election.  This type of participation varies considerably from year to year, even without changes to the primary laws.  Political circumstances or the heightened interest attendant to presidential election years can explain much of the variation.

California used typical partisan primaries between 2002-2010.  In that period, the high and low turnout figures both came from the same year because, in 2008, California split its presidential primary from the primary election for the other offices.  In this period, participation ranged from a high of 57.7% (February 2008) to a low of 28.2% (June 2008).  In 2004, when both the presidential and non-presidential primaries took place simultaneously, the participation rate was 44.3%.  In the more typical midterm years in 2002, 2006, and 2010 participation was, respectively: 34.6%, 33.6%, and 33.3%.

After the adoption of the top-two, for 2012-2020, primary participation ranged from a low of 25.2% (June 2014) to a high of 47.7% (June 2016), a gap of just under 23 percentage points.  The 2020 combined presidential and regular primary also had a high turnout, 46.9%; that figure was a much higher number than the combined primary from 2012, at 31.1%.  Both the 2016 and 2020 numbers were slightly higher than the participation in the 2004 presidential primary.  Participation in the Trump-era midterm at 37.5% (June 2018) was much higher than the last Obama-era midterm’s 25.2% (June 2014), a gap of more than twelve percentage points.  The 2018 midterm in California had a higher participation rate than the 2002, 2006, and 2010 midterms; a national study by the Bipartisan Policy Center identified top-two rules as having the highest participation rates that year.[32]

After the modest turnout in 2012, the low participation in 2014 caused alarm.  Yet Los Angeles Times reporter Mark Barabak presented another possible explanation for that year’s low turnout: “relative contentment + a sense of predestined outcome = little incentive to vote.”[33]  The 2014 election was the off-year for a U.S. Senate election and the statewide ticket was headed by relatively popular incumbent Democrat Jerry Brown, facing no meaningful opposition.  It is not clear that the higher turnout in the more turbulent years of 2016, 2018, and 2020 is necessarily an indicator of healthier politics.

To sum up: overall participation varies a great deal year-to-year based on national context, the existence of simultaneous presidential primaries, and the state political mood.  It is not the case that turnout immediately spiked with the adoption of Proposition 14; it is the case that the 2014 election had the lowest turnout of the 2002-2020 elections; and it is the case that the 2016, 2018, and 2020 elections had relatively high turnout compared to the period using more typical partisan primaries in 2002-2010.

Participation inequality.  The participation numbers reported by the Secretary of State’s office, while accurate, can also be misleading.  The aggregate turnout reported by the Secretary of State reflects the number of ballots, not what voters did with those ballots.  Not all participation is equal.

Before the adoption of the nonpartisan top-two, unaffiliated voters had to ask to vote on a party ballot; if they did not request a party ballot, they could still vote for all the nonpartisan offices and the ballot propositions.  Many unaffiliated voters did not choose a party ballot.  In every year using the old system, then, a substantial number of voters did not participate in the primary for the offices to which the top-two system applies.  In 2010, for example, 8.5% of the total primary electorate did not vote on a party ballot including those offices.[34]  That fact shrinks participation estimates: removing non-party voters causes total participation to drop to 30.5% that year.  Since voters of the American Independent, Green, Libertarian, and Peace and Freedom Party were also excluded from the Democratic and Republican primaries – and, in the end, these parties did not elect any candidate to public office – the participation in the major party primaries amounted to 29.6%.  While not very different from the reported 33.3% participation rate that year, this gap still represents many people in a state as large as California.

Since the adoption of the top-two election system, voters face fewer formal barriers to primary participation but potentially have reasons to abstain in same-party general elections.  Voters without a candidate on the ballot from their party may decide to skip voting in the contest.  This phenomenon has been a major focus of analysis for scholars of the top-two election system.  Estimates for the “roll-off” of “orphans” in same-party general elections can be as large as 7%.[35]  Abstaining is also not a random event: “orphaned” voters are less likely to abstain if they identify a candidate as being ideologically closer to their own views.[36]  It is possible that some of this roll-off could be countered by mobilizing unlikely voters, although campaigns have not necessarily done so.[37]

To sum up: the old system made it more difficult for nonpartisan voters to participate in the primary, and many did not; in the new system, some voters choose to skip some races in the general election.  This is an example of the trade-offs inherent in choosing electoral systems.

Meaningful participation.  As mentioned earlier, Proposition 14 created a nonpartisan top-two election system, not just a “top-two primary.”  This distinction creates some confusion about the appropriate turnout comparisons.  For example, in the 2014 primary, participation was low—at 25.2%.  It is possible to compare that percentage to the other primary elections and say that participation was lower because 25.2% (in 2014) is a lower number than 33.3% (in 2010).  Nevertheless, the 2014 primary also had different political consequences than the 2010 primary: in some races, the 2014 primary did not decide which candidate would win the election, even in safe Democratic seats.  Since some safe Democratic seats had competitive same-party general elections that year, in those districts, the correct comparison might be between the 2014 general election turnout and the 2010 primary election turnout.

Furthermore, if we are thinking about political incentives, it is important to recognize that elected officials do not always know in advance when they will face a serious challenge in their districts, but instead assess potential types of threats.  In safe seats, elected officials have to recognize the risk that they will face the larger general election turnout in a same-party contest at some point in the future.  Even if the total electoral risk is small, it is still real.  In 2014, incumbent Democratic Assemblyman Raul Bocanegra easily won 62.5% of the primary vote in AD39.  In the primary, Democrat Patty Lopez placed a distant second with only 23.6%.  Yet, in the general election, Lopez defeated Bocanegra 50.5%-49.5%, in one of the more shocking defeats of the top-two era—and in a general election in which more than twice as many votes were cast as in the primary.[38]


[1] In a recent book, Andrew Hall argues that Congress would have polarized even if the most moderate available candidate won every House election because the candidates willing to run for office have grown more ideologically extreme.  Since the benefits from officeholding are uneven and potentially small, more moderate citizens are disinclined to run for public office.  See Andrew B. Hall, Who Wants to Run?: How the Devaluing of Political Office Drives Polarization, Chicago Studies in American Politics (Chicago: The University of Chicago Press, 2019),

[2] Hans J. G. Hassell, The Party’s Primary: Control of Congressional Nominations (Cambridge, United Kingdom: Cambridge University Press, 2018).

[3] Thad Kousser et al., “Kingmakers or Cheerleaders? Party Power and the Causal Effects of Endorsements,” Political Research Quarterly 68, no. 3 (2015): 443–56.

[4] See Ronald D. Hedlund and Meredith W. Watts, “The Wisconsin Open Primary 1968 to 1984,” American Politics Quarterly 14, no. 1–2 (January 1, 1986): 55–73,; Barry C. Burden, “The Polarizing Effects of Congressional Primaries,” in Congressional Primaries and the Politics of Representation, ed. Peter F. Galderisi, Marni Ezra, and Michael Lyons (Lanham, Maryland: Rowman & Littlefield, 2001), 95–115.

[5] Seth J. Hill, “Institution of Nomination and the Policy Ideology of Primary Electorates,” Quarterly Journal of Political Science 10, no. 4 (December 17, 2015): 461–87,

[6] See Douglas J. Ahler, Jack Citrin, and Gabriel S. Lenz, “Do Open Primaries Improve Representation? An Experimental Test of California’s 2012 Top-Two Primary: Do Open Primaries Improve Representation?,” Legislative Studies Quarterly 41, no. 2 (May 2016): 237–68,; Jonathan Nagler, “Voter Behavior in California’s Top Two Primary,” California Journal of Politics and Policy 7, no. 1 (2015),; R. Michael Alvarez and J. Andrew Sinclair, Nonpartisan Primary Election Reform: Mitigating Mischief (New York, NY: Cambridge University Press, 2015).

[7] This is one of the main conclusions here: Alvarez and Sinclair, Nonpartisan Primary Election Reform; see also: Jesse Crosson, “Extreme Districts, Moderate Winners: Same-Party Challenges, and Deterrence in Top-Two Primaries,” Political Science Research and Methods 9, no. 3 (July 2021): 532–48,

[8] See, among others Donald Moynihan and Pamela Herd, “Red Tape and Democracy: How Rules Affect Citizenship Rights,” The American Review of Public Administration 40, no. 6 (November 1, 2010): 654–70,; Pamela Herd and Donald P. Moynihan, Administrative Burden: Policymaking by Other Means (New York: Russell Sage Foundation, 2018).

[9] For some of these matters, see Betsy Sinclair and J. Andrew Sinclair, “Primaries and Populism: Voter Efficacy, Champions, and Election Rules,” Journal of Political Institutions and Political Economy 2, no. 3 (November 12, 2021): 365–88,

[10] J. Andrew Sinclair et al., “Crashing the Party: Advocacy Coalitions and the Nonpartisan Primary,” Journal of Public Policy 38, no. 3 (September 2018): 329–60,  This paper uses data only through 2016, but the trends hold through 2020.

[11] Shigeo Hirano and Jr Snyder James M., Primary Elections in the United States, Political Economy of Institutions and Decisions (Cambridge: Cambridge University Press, 2019), 307,

[12] For a discussion, see Crosson, “Extreme Districts, Moderate Winners.”

[13] Gerald Gamm and Thad Kousser, “Broad Bills or Particularistic Policy? Historical Patterns in American State Legislatures,” The American Political Science Review 104, no. 1 (2010): 151–70; Gerald Gamm and Thad Kousser, “Life, Literacy, and the Pursuit of Prosperity: Party Competition and Policy Outcomes in 50 States,” American Political Science Review 115, no. 4 (November 2021): 1442–63,

[14] Not everyone gets to be the Terminator before running for public office.  Notably, when Schwarzenegger was reelected in 2006, his personal brand was much better liked than the Republican Party as a whole; although the public sent Schwarzenegger back to Sacramento, it also sent substantial Democratic majorities in the legislature and Democrats for most statewide offices.

[15] William H. Riker, Liberalism against Populism: A Confrontation Between the Theory of Democracy and the Theory of Social Choice (Prospect Heights, Ill: Waveland Press, 1982).

[16] Soumya Karlamangla, “A Lesson in California’s Top-Two Primary System,” The New York Times, May 24, 2022, sec. U.S.,

[17] See Jack Pitney quoted in Alec Regimbal, “California AG’s Supporters Could Put His Election Bid in Danger,” San Francisco Chronicle, May 24, 2022,  The issue is that the California Republican Party has endorsed Nathan  Hochman.  The pro-Bonta group appears to be trying to help a different Republican, Eric Early.  If they succeed in shifting votes from Hochman to Early, that may split the Republican vote, and help Schubert make second place in the primary.

[18] Crosson, “Extreme Districts, Moderate Winners,” 547.

[19] Raiding – voting insincerely for a weak candidate in the other party’s primary – is generally understood to be quite rare in party primaries.  For a discussion in the context of California’s experiment with the blanket primary, see  Cain and Gerber, Voting at the Political Fault Line.  It is worth noting, though, that the top-two system means voters always put something at risk with any kind of insincere primary voting behavior because a vote for anyone other than their preferred candidate risks the primary defeat of their preferred candidate.  That is not always true in partisan primaries.

[20] Third-party supporters may object to this characterization; advocates for instant runoff voting (IRV) also can observe that IRV mitigates this problem.

[21] Ethan Rarick, “When Dems Meddled in 1966 GOP Race, the Result Was Reagan,” San Francisco Chronicle, April 14, 2002,; Rick Perlstein, Nixonland: The Rise of a President and the Fracturing of America (New York: Scribner, 2008), 72; Miriam Pawel, The Browns of California: The Family Dynasty That Transformed a State and Shaped a Nation (New York: Bloomsbury Publishing, 2018), 134.

[22] Crosson, “Extreme Districts, Moderate Winners,” 547.

[23] Dan Balz, “California’s Open Primaries Are a Cautionary Tale about Political Reform,” Washington Post, June 3, 2018,

[24] Andrew B. Hall, “What Happens When Extremists Win Primaries?,” The American Political Science Review 109, no. 1 (2015): 18–42.

[25] Laurel Rosenhall, “Another California Republican Defection: Former Party Leader Bails on the GOP,” CalMatters, December 5, 2019, sec. Politics,

[26] Gary W. Cox, Making Votes Count: Strategic Coordination in the World’s Electoral Systems, Political Economy of Institutions and Decisions (Cambridge, U.K. ; Cambridge University Press, 1997),

[27] For a discussion on this point, see Lee Drutman, Breaking the Two-Party Doom Loop: The Case for Multiparty Democracy in America (New York, NY: Oxford University Press, 2020),

[28] George E. Mowry, The California Progressives (Chicago: Quadrangle Books, 1963).

[29] Giulia Sandri, Antonella Seddone, and Fulvio Venturino, Party Primaries in Comparative Perspective (London ; Routledge, 2016),; Reuven Y. Hazan and Gideon Rahaṭ, Democracy within Parties: Candidate Selection Methods and Their Political Consequences, Comparative Politics (Oxford ; Oxford University Press, 2010); J. Andrew Sinclair, “Party Nominations and Electoral Persuasion,” in The Oxford Handbook of Electoral Persuasion, ed. Elizabeth Suhay, Bernard Grofman, and Alexander H. Trechsel (Oxford University Press, 2020), 862–84,

[30] Eric McGhee, “Open Primaries” (San Francisco, CA: Public Policy Institute of California, February 2010),

[31] An outstanding observation in Thad Kousser, “The Top-Two, Take Two: Did Changing the Rules Change the Game in Statewide Contests?,” California Journal of Politics and Policy 7, no. 1 (February 5, 2015),

[32] John C. Fortier et al., “2018 Primary Election Turnout and Reforms,” 2018,

[33] Mark Barabak, “Poor California Primary Turnout Explained: Contented People Stay Home,” Los Angeles Times, June 3, 2014, sec. California,

[34] Counting unaffiliated voters and voters affiliated with minor parties denied official state recognition (they are reported in the same column in the Secretary of State’s Statement of Vote.)

[35] Shawn Patterson, “Estimating the Unintended Participation Penalty under Top-Two Primaries with a Discontinuity Design,” Electoral Studies 68 (December 2020): 102231,; Daniel D. Bonneau and John Zaleski, “The Effect of California’s Top-Two Primary System on Voter Turnout in US House Elections,” Economics of Governance 22, no. 1 (March 2021): 1–21,

[36] Colin A. Fisk, “No Republican, No Vote: Undervoting and Consequences of the Top-Two Primary System,” State Politics & Policy Quarterly 20, no. 3 (September 2020): 292–312,

[37] Seth J. Hill and Thad Kousser, “Turning Out Unlikely Voters? A Field Experiment in the Top-Two Primary,” Political Behavior 38, no. 2 (June 2016): 413–32,

[38] Bocanegra would come back to win the seat in 2016, only to be forced to resign in 2017 amid sexual harassment allegations.