Do Open Primaries Improve Representation? An Experimental Test of California's 2012 Top‐Two Primary

• Date: 01 May 2016
• Author: Douglas J. Ahler,Jack Citrin,Gabriel S. Lenz
• Published date: 01 May 2016
• DOI: http://doi.org/10.1111/lsq.12113

DOUGLAS J. AHLER

JACK CITRIN

GABRIEL S. LENZ

University of California, Berkeley

Do Open Primaries

Improve Representation?

An Experimental Test of

California’s 2012 Top-Two Primary

To improve representation and alleviate polarization among US lawmakers,

many have promoted open primaries—allowing voters to choose candidates from any

party—but the evidence that this reform works is mixed. To determine whether open

primaries lead voters to choose ideologically proximate candidates, we conducted a

statewide experiment just before California’s 2012 primaries, the first conducted under a

new top-two format. We find that voters failed to distinguish moderate and extreme can-

didates. As a consequence, voters actually chose more ideologically distant candidates

on the new ballot, and the reform failed to improve the fortunes of moderate

congressional and state senate candidates.

According to a large body of research in American politics,

citizens’ political preferences are not polarized but their choices over

who represents them are (Bafumi and Herron 2010; Fiorina, Abrams,

and Pope 2005). Despite politicians and other party elites being more

ideologically polarized than at any time in the past century (Bonica

2014; McCarty, Poole, and Rosenthal 2008), ordinary citizens still tend

to claim to be ideological moderates and hold ideologically heterodox

bundles of positions (Baldassarri and Gelman 2008; Broockman 2016;

Fiorina and Abrams 2009). On its face, such a divergence between

voters’ preferences and the choices on offer is potentially worrisome for

democratic representation (Bafumi and Herron 2010; Fiorina and

Abrams 2009). But the ramif‌ications of elite polarization are potentially

more far-reaching and disturbing. In recent years, gridlock over routine

nominations in the Senate spurred Democrats to “go nuclear” and change

the f‌ilibuster rules. Partisan brinkmanship led Standard and Poor’s to

downgrade the nation’s credit rating for the f‌irst time in history. And

LEGISLATIVE STUDIES QUARTERLY, 41, 2, May 2016 237

DOI: 10.1111/lsq.12113

V

C2016 Washington University in St. Louis

ideological warfare produced the f‌irst federal government shutdown in

17 years. Since voters say they want effective governance above all else

(Fiorina, Abrams, and Pope 2005), the rise of legislative intransigence as

a consequence of rigid ideological divisions further signals a breakdown

of representation.

Reforming primary elections to make them more “open” is one

proposed solution for this worrisome state of affairs. Open primaries

take several forms but generally allow voters to cast ballots for candi-

dates of more than one party. The boldest recent reform is the top-two,

or runoff, format. Designed to reduce partisan control of the nominat-

ing process, it places candidates with any party label (or no party

label) on a single ballot, with the top two vote getters, regardless of

party, then competing in the general election. On the assumption that

voters prefer candidates more proximate to their own ideology and

that in the aggregate they tend to prefer centrist policies, reformers

theorize that these changes will benef‌it moderate candidates who,

once elected, will be more willing to compromise. Proponents also

argue that open primaries may increase participation by moderate vot-

ers (though see McCarty 2011, 365). This reasoning is consistent with

political scientists’ models of primary systems (Aranson and

Ordeshook 1972; Coleman 1971, 1972; McGann 2002) and was the

pitch California voters heard in 2010 from Abel Maldonado, the

author of the Top-Two Primaries Act that won 54% in a referendum

and took effect in the June 2012 primaries.

1

Many appear to have concluded that open primaries are likely to

improve ideological congruence between voters and candidates, thereby

yielding less ideologically extreme legislators (Burden 2004; Fiorina,

Abrams, and Pope 2005; Hacker and Pierson 2005; Kanthak and Morton

2001; Mann and Ornstein 2012). Indeed, McCarty observes, “It seems

almost a logical certainty that opening primary elections to more nonpar-

tisan and independent voters should have a moderating effect on politics

by increasing the chance that moderate candidates get nominated”

(2011, 363). But for all the apparent certainty, empirical evidence is

mixed.

On one side, a handful of studies f‌ind that open primaries do mod-

erate political outcomes. For example, Members of Congress’ (MCs)

roll-call votes from 1982 to 1990 appear to hew more closely to their dis-

tricts’ ideological leanings in states with semi-closed or open-primaries

than in states with closed party primaries (Gerber and Morton 1998).

And, examining California’s f‌irst attempt at primary reform in 1998,

Gerber (2002) concludes that moderates were more likely to advance to

238 Douglas J. Ahler, Jack Citrin, and Gabriel S. Lenz

the general election in state legislative races in 1998 than in 1996,

controlling for other characteristics of the contests.

On the other hand, several studies fail to f‌ind that open primaries

moderate politicians. Analyzing state legislators’ roll-call votes from

1996 to 2006, McGhee et al. (2014) conclude that open primaries are not

associated with reduced legislative polarization at the state level.

McGhee (2014) reaches similar conclusions about MCs’ roll-call voting.

According to several studies, ideologically extreme congressional candi-

dates fail to fare better in closed primaries compared to open primaries

(Brady, Han, and Pope 2007; Hall and Snyder 2013).

2

In another study

of the 1998 California blanket primaries, Bullock and Clinton (2011)

f‌ind that MCs elected in 1998 were on average no more moderate in their

roll-call voting than those elected in 1996 under the closed-primary sys-

tem, though MCs in competitive districts may have moderated slightly.

Studies on primary reform thus have produced mixed f‌indings, but

they have also relied solely on cross-sectional or pretest-posttest obser-

vational designs, which have well-known shortcomings for causal

inference (Campbell and Stanley 1963). Consequently, inferences from

these studies about how primary reforms affect polarization may be vul-

nerable to alternative explanations. To overcome these limitations, we

turned to an experiment. In a large survey of registered California voters

conducted just prior to the June 6 California primary, we randomly

assigned half the sample to vote with a ballot identical to the one they

would see in the actual top-two primary (treatment), and the other half to

the traditional ballot they would have seen had the referendum (Proposi-

tion 14) failed (control). With this design, we assess whether the reform

led voters to choose candidates closer to their claimed ideologies—that

is, whether the reform improved proximity voting—and whether it

helped moderate candidates for the US Congress and the California State

Senate.

The survey in which we conducted the experiment is one of the

most comprehensive studies of congressional primaries in a state, ena-

bling analysis of a range of factors potentially related to the success or

failure of primary reform. We therefore begin the article not with the

experimental results themselves but with an empirical analysis of

whether the assumptions reformers often make are plausible. In particu-

lar, we examine whether district electorates are indeed more moderate

than partisan primary electorates and whether voters have the knowledge

necessary to pick proximate (and therefore often moderate) candidates.

Previewing the f‌indings, voters appear to know so little about the

candidates’ positions that, even if they wanted to, they could not

intentionally cast a ballot for their district’s moderate candidates. They

239Do Open Primaries Improve Representation?