Issuu

CMC-Rose Institute Poll: Political Attitudes in California and New York in a Time of Crisis for their Governors

September 21, 2021

Kenneth P. Miller Rose Professor of State and Local Government Director of the Rose Institute of State and Local Government Claremont McKenna College

J. Andrew Sinclair Assistant Professor of Government Claremont McKenna College Corresponding author: asinclair@cmc.edu Report prepared with the assistance of: Nohl M. Patterson, CMC ’22, Student Manager Adhitya Venkatraman, CMC ’22, Associate Student Manager And the CMC Rose Institute Student Staff

Introduction This document presents initial results from the CMC-Rose Institute Poll of September 2021, a survey measuring political attitudes of registered voters in California and New York at a time of crisis for their governors. The early days of the coronavirus pandemic focused attention on the governors of New York and California, Andrew Cuomo and Gavin Newsom. Already leading politicians of the Democratic Party with many years of experience holding public office, governors of two of the nation’s largest states, and potential future presidential contenders, the emergency put them at the center of the nation’s virus response. Governor Cuomo’s press conferences became such a phenomenon that he was awarded an Emmy and passed Newsom to become the most-followed governor on Twitter.1 Newsom, like Cuomo, “won widespread praise, not only for his response to the virus crisis, but for his articulation of an alternative vision to Donald Trump and the Republican’s approach to government.”2 By the end of March 2020, a Siena College poll found that 87% of New Yorkers approved of Governor Cuomo’s handling of the Coronavirus, with an overall approval rating of 71%.3 Governor Newsom’s approval ratings had increased in June to 65%, up over 10 points from the start of the pandemic.4 The prominence of these, and a handful of other Democratic governors, was possibly “reshaping the party’s pipeline of prospective presidential candidates for years to come.”5 In any event, it appeared that Newsom would easily win reelection for a second term, and Cuomo a fourth term, in the 2022 gubernatorial elections. Nevertheless, by early 2021, both were “facing the most intense scrutiny of their respective political careers in the midst of the coronavirus pandemic, as both grapple with missteps and allegations of wrongdoing that threaten their future.”6 By the end of August 2021, Cuomo had resigned ahead of an impeachment inquiry and Newsom was facing an unexpectedly tight recall election, with Newsom holding “only a slim margin” in the Berkeley IGS poll of late July.7 Newsom ultimately managed to decisively defeat the recall on September 14th, rescuing his political career at a time former Governor Cuomo found himself out of the Executive Mansion in Albany for the first time in a decade. Phil Willon, Jennifer Lu, and Sara Stoudt. “In the coronavirus crisis, Newsom uses social media to raise awareness of the pandemic – and his profile.” Los Angeles Times, May 10, 2020, https://tinyurl.com/uyvjswu7. 2 Todd S. Purdum, “Gavin Newsom’s Nation-State.” The Atlantic, April 21, 2020, https://tinyurl.com/52kzb7zm. 3 Siena College Research Institute. “87% of NYers Approve of Cuomo’s Handling of the Coronavirus.” March 30, 2020, https://tinyurl.com/suud6hcy. 4 Jill Cowan. “How Do Californians Feel About Governor Newsom?” New York Times, June 4, 2020, https://tinyurl.com/3br89374. 5 David Siders. “How the coronavirus is shaping the 2024 presidential race.” Politico, April 2, 2020, https://tinyurl.com/ywb5k6ec. 6 Reid Wilson. “On The Trail: Cuomo and Newsom – a story of two embattled governors.” The Hill, February 27, 2021, https://tinyurl.com/yyu6zka8. 7 Mark DiCamillo. “Gubernatorial recall election much closer among voters considered most likely to participate than among the overall electorate.” July 27, 2021, https://escholarship.org/uc/item/27x9k5qt. 1

Our survey involves interviews in both California and New York in the weeks leading up to the 2021 California recall election, a period directly following the resignation of New York’s Governor Cuomo. Unlike other surveys of the California recall election, our study focuses on comparing these two states and understanding which sorts of political institutions people in those states favor. New York and California have much in common. Both are large, Democratic-leaning states with a substantial divide between the urban and rural population. Nevertheless, the two states have very different political institutions. New York does not have a recall option, nor does it use California’s top-two primary system, among other differences. The first section of this report presents our findings for the California recall. These initial results test the reliability of our data: we are close to the correct California recall numbers and the correct (but inoperative, since the recall failed) replacement election numbers as well. In our data, as in the election, Governor Newsom wins the recall and the highest-profile Republican candidate, Larry Elder, leads the replacement field. The second section examines preferences over political institutions: having the recall option and using the nonpartisan top-two primary. Both turn out to be popular in both California and New York, although only California uses them. We examine the relationship between concepts of populism and institutional preferences as well. The third section looks at how voters assess the outcomes of politics. Voters are deeply concerned about the state of politics, seeing “the other side” as a significant problem. Voters are also concerned about issues like homelessness, Covid-19, and disasters linked to climate change. Yet, in the California recall election, it is mostly Republicans voting for the recall and Democrats voting to keep Newsom in office. As we explore in this section, voters tend to place blame on the opposite party, meaning that many dissatisfied Democrats still voted to keep Newsom anyway, since they viewed Republicans and Republican policies as source of their dissatisfaction. We conclude with some brief thoughts about the differences between the recall election and the top-two primary as well as the different fates of New York’s Governor Cuomo and California’s Governor Newsom. Methods This survey was implemented online by YouGov between August 30, 2021 and September 10, 2021. It includes a sample of 2000 registered California voters and 1675 registered New York voters. It includes post-stratified weights, computed by YouGov, that take into account gender, age, race, education, and the 2016/2020 presidential vote. The overall margin of error for the California registered voters sample is ±2.53%, with a slightly larger margin of error for the weights produced just to match likely voters. The margin of error in the New York registered voter sample is ±2.83. 2

The California Recall Election

Our survey accurately reflects what transpired in the recall election. We had 1822 “likely” voters among our 2000 California respondents, and have focused our analysis on that group, applying the survey weights YouGov provided for them.8 In our data, 59.8% of the likely voters preferred to retain Governor Gavin Newsom in office, voting “No” on the recall. Election returns from the Secretary of State’s office have “No” obtaining approximately 63.4%, with some vote remaining to be counted.9 These results are very similar to Newsom’s 61.9% in the 2018 general election against Republican John Cox and President Joseph Biden’s 63.5% in the 2020 presidential election against former President Donald Trump. The recall election results reflect the partisan divide in California.

Figure 1.1, above, breaks down the recall vote by “party identification,” how voters choose to describe themselves. Independent voters reporting leaning towards one party or the other are included with voters of that party, as past political science research has tended to The unweighted results are actually quite similar to the weighted ones. Unweighted, “No” wins in this group with 62.2% of the vote. Applying the weights, “No” wins with 59.8% of the vote. It is “best practice” to apply survey weights to this kind of data, though, so we provided weighted data throughout. 9 This is from the California Secretary of State’s election returns as of 9/20/21. https://electionresults.sos.ca.gov/returns/governor-recall. 8

support the idea that “leaners” are very similar to partisans. Almost 90% of Democratic identifiers planned to vote “no” in a state dominated by that party (57% of the likely voters overall). Approximately 90% of Republican identifiers planned to vote “yes,” a mirror image of the Democratic totals, although with a much smaller group of voters (making up only 29% of the electorate). While true, non-leaning, independent voters did favor the recall, this group is not adequately large or uniform to make up for the Democratic Party’s advantage in California. Although Newsom’s victory makes the second question moot, our survey also accurately captured both Republican Larry Elder’s lead and the substantial abstention in the replacement election. Overall, Elder led a divided field, obtaining in our survey 48.3% of the vote among the specific candidates we listed, with Democrat Kevin Paffrath in a distant second. This mirrors the early returns; Elder had 47.4% of the actual vote – but with only 2,809,638 votes out of 10,601,811 cast on the first question, or 26.5% when abstention is permitted.10 In our survey data, allowing for abstention (the most popular response), Elder had 25.5% of the vote. Figure 1.2 presents our results for the second question, split by party; Elder obtained support from two-thirds of Republicans.

Current vote totals as of 9/20/21.

Democrats overwhelmingly intended to abstain or declined to choose from among our alternatives.11 We included a follow-up question asking respondents, no matter how they voted on the second question, to say which person they thought would make the best governor of California. For that question, we included Newsom on the list and forced a choice among the candidates listed. Newsom was the most popular choice by far, with Elder in a distant second place.12 Table 1.1: Do you agree with any of these statements (in favor of/opposing) recalling Newsom? Respondents could select as many as applied, so column percentages do not add to 100%.

Reasons for. Newsom is corrupt or hypocritical. Newsom is not very likeable. Newsom supports bad policies. Newsom abuses his authority. I like a different candidate better. A conservative governor is better. Newsom does not listen to people. People, not politicians, should be in charge. None of the above. Reasons against. The recall is just partisan politics. The recall process if flawed/unfair. I am worried about who might replace him. Newsom has done a good enough job. Newsom favors policies I like. Newsom has done his best. A liberal governor is better. Elected officials should get their full term. None of the Above

Recall Vote No: Retain Yes: Recall Percent of Percent of those voting those voting “No.” “Yes.” 6 71 9 41 4 73 4 72 3 35 2 47 4 63 8 41 75 3 Percent of Percent of those voting those voting “No.” “Yes.” 64 4 51 5 70 13 63 4 48 1 55 2 37 4 27 3 3 75

Total Percent of likely voters. 32 22 31 31 16 20 28 21 46 Percent of likely voters. 40 32 47 39 29 34 24 17 32

Our survey included: Doug Ose, Kevin Kiley, Kevin Faulconer, Ted Gaines, Caitlyn Jenner, Larry Elder, Kevin Paffrath, John Cox, and Brandon Ross. We included the full ballot descriptions, so respondents would know that Paffrath and Ross were Democrats. It seems likely that many respondents selecting “some other candidate” instead of making a specific choice ultimately did not select any candidates in the actual election itself. 12 Newsom’s percentage in this question – 49%, to Elder’s 24% -- reflects his advantage, but may underestimate his support, as not every respondent may have read the question carefully enough to realize Newsom was included. 11

We also asked the survey respondents to evaluate several commonly discussed reasons for favoring or opposing recalling Newsom (included in Table 1.1). Respondents were asked to select all of the statements with which they agreed. These choices reflected the partisan and ideological nature of the contest as well. Among voters favoring recalling Newsom, more than 70% agreed that Newsom was “corrupt or hypocritical,” supported “bad policies,” and “abuses his authority.” Only 35% of the recall supporters selected “I like a different candidate better” – an affirmative, rather than a negative, judgment. While such voters tended not to select any of the reasons for opposing the recall, it is notable that 13% did agree that they were worried about who might replace Newsom – but voted for the recall anyway. Voters who disapproved of Newsom tended to strongly disapprove of him. In the whole California registered voters sample, 24% strongly approved, 32% somewhat approved, 14% somewhat disapproved, and 30% strongly disapproved. Among Republican identifiers, 75% strongly disapproved, with only 14% somewhat disapproving. The dislike for Newsom among Republicans was certainly more zealous than his support within his own party: among Democrats, only 38% strongly approved while 46% somewhat approved. Nevertheless, they would vote overwhelmingly for him, and few crossed over into either disapproval category. Among voters opposing recalling Newsom, the most commonly selected reason was also negative: “I am worried about who might replace him,” at 70%. Newsom did have some positive support, though, with 63% also agreeing that he had “done a good enough job.” The least popular reason was an anti-recall principle (“elected officials should get their full term,” 27%), although 51% thought this particular recall process was flawed or unfair. Most of the ‘no’ voters did not agree with any of the reasons for recalling Newsom, although 9% did concede that Newsom “is not very likeable.” In pre-election polling, Elder was the clear leader among the replacement alternatives, and much of the media coverage described the recall as a choice between Elder and Newsom. On both the recall and replacement questions, preferences split among party lines. In a state with a considerable advantage for the Democratic Party, that meant the election was not competitive. The 2021 recall differed from the 2003 recall in several respects. First, Schwarzenegger had both a unique brand and a more centrist set of ideological positions than Elder. Second, the state was considerably more Republican in 2003. Third, Davis had lost more support among Democrats, who also had a serious replacement option on the ballot. In many ways, the 2021 recall election had more in common with the 2018 gubernatorial election or 2020 presidential election than the 2003 recall, despite the unusual structure of the ballot.

Populism and Political Institutions

The 2021 California recall spurred a considerable amount of discussion about the existence of the recall option in California. One perspective on the recall rule is that it is one of the few remaining ways in which the state’s Republican Party can remain relevant. While there are other perspectives on the recall, at least this narrative would suggest Democrats should wish to abolish it or substantially alter the procedure. The New York Times suggested the Newsom gubernatorial recall election would generate a campaign to “recall the recall.”13 Voters in New York do not have the power to recall elected officials. Former Governor Cuomo was forced out in August by an impeachment threat, but only following conduct so egregious as to lose the support of a substantial number of legislators of his own party.14 They were following the preferences of their voters: in our data, 63% of Democratic voters preferred to have Cuomo resign instead of stay. Yet, they were reliant on legislators to deliver that judgment. We asked voters in both California and New York about their preferences: do Californians favor the recall? Do New Yorkers wish they had it? California’s recall procedure is the result of progressive-era reforms in the state. Over a century ago, California’s progressives pushed not only for the recall but also for the initiative and referendum process as well as the direct primary.15 Some of the heirs of that tradition, acting for many of the same reasons favored by the progressive reforms of an earlier century, pushed for the adoption of the nonpartisan top-two primary system in California in 2010. Proposition 14 passed, transitioning nearly all offices in California to this new system beginning in 2012. While simple partisan identity explains a great deal of the vote choice in the 2021 California recall, the relationship between partisanship and the rules voters want for politics is less straightforward. One of the goals of our project is to consider other types of attitudes; in particular, we are interested in populist attitudes present in both parties, although perhaps in different ways. In this section, we report our findings on the relationship between these other attitudes and what sorts of political institutions voters want. The recall and the top-two primary were both intended to tighten the relationship between the views of citizens and the outcomes of politics. While we describe in this section how both rules have substantial support in California – and, in fact, in New York, which has neither rule – we also show how they draw support from different types of voters. A discussion about reforming the recall process will benefit from considering it as part of a state’s electoral system. See Thomas Fuller, Maggie Astor, and Conor Dougherty. “As California Votes, It Rethinks Its Tradition of Direct Democracy.” New York Times, Sept. 14, https://www.nytimes.com/2021/09/14/us/californiarecall-constitutional-amendment.html. 14 See Luis Ferré-Sadurni and J. David Goodman. “Cuomo Resigns Amid Scandals, Ending Decade-Long Run in Disgrace.” New York Times, Aug. 10, https://www.nytimes.com/2021/08/10/nyregion/andrew-cuomoresigns.html. 15 See, among other places, Kevin Starr’s Inventing the Dream: California Through the Progressive Era. New York: Oxford University Press, 1985. 13

Within California: 70% of voters want to retain the recall procedure. As illustrated below in Figure 2.1, 70% of the 2021 California recall voters wished to continue to allow recall elections, a substantial majority. As might be expected, nearly every voter (94%) choosing to recall Newsom favored the process. A comfortable majority (54%) of voters choosing to retain Newsom in this most recent recall, though, also thought the process should be available in the future.

While a recall process was broadly supported by the public, some voters may be both in favor of having recalls and reforming the existing recall system. Our question does not address various potential recall reforms, such as higher signature thresholds or alternative replacement procedures. As shown in the previous section, 51% of the “no” voters on the recall also said the recall process was “flawed/unfair.” Of the voters who voted “no” on the recall and who described the process as “flawed/unfair,” 45% still favored allowing the recall process. Among those voting “no” on the recall who did not describe the process that way, 63% favored allowing the recall process.

California and New York: Allowing recall elections is popular in both states. To compare preferences over allowing recall elections in the two states, we shift to using registered voters in both (since New York did not have an election in September of 2021). Despite not having a recall procedure, the percentage of registered voters favoring the recall in New York (66%) is similar to the percentage in California (71%).16 Both states have similar partisan breakdowns over allowing recall elections as well, as shown in Figure 2.2. With Democrats tending to hold statewide office, nearly all Republicanidentifying registered voters prefer allowing recalls, along with most independents. Nevertheless, in both states, allowing recalls is still more popular than prohibiting them among Democratic-identifying registered voters. (As before: independent and third-party “leaners” are included with their preferred major party.)

The California number is not precisely identical for registered voters (70.6%) and likely voters (70.1%).

California and New York: Nonpartisan primaries are popular, but with different voters. California uses a nonpartisan “top-two” election system. With this rule, all voters choose between all candidates in the primary and the two candidates with the most votes advance to the general election, even if they are from the same party. New York continues to operate a partisan primary system, although it recently experimented with ranked choice voting within party primaries for New York City offices.17 While the impact of the top-two system is debated within political science, advocates offer a rationale very similar to that for the recall procedure: improving the control over, and performance of, elected officials.18 In both California and New York, the top-two system outpolls a partisan primary system, although the voters supporting each are quite different from the voters currently embracing the recall procedure. While Republicans like the recall, they tend to be more suspicious of the top-two, which is relatively more popular with Democrats, particularly in California.

For a summary of primary election types, see National Conference of State Legislatures, “State Primary Election Types,” Jan. 5, 2021, https://www.ncsl.org/research/elections-and-campaigns/primarytypes.aspx. 18 For example, Christian R. Grose, “Reducing Legislative Polarization: Top-Two and Open Primaries Are Associated with More Moderate Legislators.” Journal of Political Institutions and Political Economy, 1(2) 2020: 267-287. 17

Populism and Partisanship: distinguishable concepts, but with partisan shades. The definition and concept of “populism” remains under debate and development within political science. In recent literature, particularly studies focused on events in the United States over the last several years, there has been an emphasis on measuring different types of populist ideas, both in elite expression (what candidates say, or their rhetorical style) and in the attitudes citizens hold.19 Following advances in this literature, we included four questions that capture related, but still different, senses of the populist ideas citizens may have. The first question focuses on the idea that some powerful elite is denying ordinary citizens their “fair share” of the resources of society. The second captures anti-expertise attitudes. The third addresses corruption by politicians. The fourth focuses on the way the policies favored by “regular Americans” – the “true people” in the populism literature – are potentially frustrated. In advance of conducting the survey, we expected to see some populist attitudes, expressed to some extent, in both major party groups. Nevertheless, we also expected to see some partisan differences, if only in degree. We expected Democrats to express more financial concerns and Republicans to have more anti-expert sentiment and higher concerns that “regular Americans” aren’t getting their policy preferences. Just about everyone seemed likely to express some concern about the corrupting influence of money in politics. That is largely what we find, as illustrated in Table 2-1. Many respondents agree with the populist statements. In both states, the concern about money in politics is pervasive. In California, 72% of respondents either agreed or strongly agreed that “most politicians are corrupted by the money in politics.” In New York, 75% of respondents agreed or strongly agreed with that statement. The notion that “wealthy or politically powerful people” prevent “everyday Americans” from getting ahead obtains agreement from over half the respondents in both states, although that sentiment is more concentrated among Democrats. The view that if only “regular Americans” obtained their preferred policies that “the country would be a lot better off” got nearly half of the respondents as well; although a greater share of Republicans relative to Democrats agreed or strongly agreed, substantial proportions of Democrats indicated agreement. The question pitting the “wisdom of ordinary people” against the “opinions of experts and intellectuals” has the lowest agreement levels in both states. In California, only 28% overall agreed or strongly agreed with this statement; in New York, it was only 26%. Furthermore, 19

See, among others: Agnes Akkerman, Cas Mudde, and Andrej Zaslove, “How Populist Are the People? Measuring Populist Attitudes in Voters.” Comparative Political Studies, 47(9) 2014: 1324-1353; Kirk Hawkins and Levente Littvay, Contemporary US Populism in Comparative Perspective, New York: Cambridge University Press, 2019; Cas Mudde and Cristóbal Rovira Kaltwasser, “Populism and (liberal) democracy: a framework for analysis.” In Cas Mudde and Cristóbal Rovira Kaltwasser, eds., Populism in Europe and the Americas. New York: Cambridge University Press, 2012, 1-26; Eric J. Oliver and Wendy M. Rahn. “Rise of the Trumpenvolk: Populism in the 2016 Election.” Annals of the American Academy of Political and Social Science, 667(1) 2016: 189-206; Pippa Norris, “Measuring populism worldwide.” Party Politics, 26(6) 2020: 697-717.

the patterns of disagreement on this question sharply diverge: 28% of Democrats in California strongly disagreed with this view, as compared to only 4% of Republicans. While this question connections in an obvious way to contemporary debates over policies such as vaccine and mask mandates and other public health responses to Covid-19, it also gets at the heart of the assumptions underlying recall elections. As scholars of various forms of direct democracy observe, it is placing more power directly in the hands of the electorate.20 We expected expression of populist sentiment to go along with preference for allowing recall elections. Such a relationship is not likely to be observable among Republicans, who are overwhelmingly in favor of having the option of recall elections; in these heavily Democratic states, there is little variation in Republican opinion to explain. We theorized that populist attitudes, though, might explain why some Democrats would still want to allow for recall elections – as populist remedy for misrule. In the second table included below, Table 2-2, we show the percentages of Democrats – for those who favor prohibiting recalls, and those who favor allowing them – selecting each level of agreement on the populism questions. There are only very slight differences for the first question on financially getting a “fair share.” On the experts question, Democrats in California who want to prohibit the recall are a bit more likely to “strongly disagree” that they would rather put trust in wisdom of ordinary people relative to those who would allow it, who tend to merely “disagree” instead. In California, Democrats who want to allow the recall are a bit more likely to agree about the influence of money corrupting politics. There are very slight differences for the “regular Americans” question, as well. In New York, the differences between the columns are generally very small. In short: there are some limited indications that the Democrats favoring allowing recalls exhibit slightly greater populist attitudes on the opinions of experts and influence of money in politics. The differences between parties on these particular populism questions, and on support for the recall, is much larger than the differences among Democratic identifiers. After presenting Table 2.1 and 2.2, we turn to a different approach to measuring populism: support for populist politicians.

Thomas Cronin, Direct Democracy: The Politics of Initiative, Referendum, and Recall. Harvard University Press, 1989, 127. “The recall also raises questions about the competence, interest, and rationality of the average voter. It presupposes a voter with discriminating judgment and with sufficient attentiveness and information to be able to vote wisely on the question of whether or not to discharge a public official…”

Table 2.1: Populism. “Please indicate if you agree or disagree with the statement.” Displaying percentages within each category of self‐reported party identification, including “leaners” with their respective party, in California and New York. California New York Dem. Ind. Rep. Total Dem. Ind. Rep. Total N N % % % % % % % % Everyday Americans can’t get ahead financially because wealthy or politically powerful people won’t let them have their fair share. Stg. Disagree 3 9 24 10 176 3 13 26 11 153 Disagree 9 16 25 15 279 8 12 25 14 225 Neither 19 24 21 20 401 16 34 21 20 333 Agree 37 24 19 30 618 35 19 17 28 478 Stg. Agree 32 27 11 25 526 37 21 11 27 486 Total 100 100 100 100 2,000 100 100 100 100 1,675 I’d rather put my trust in the wisdom of ordinary people than the opinions of experts and intellectuals. Stg. Disagree 28 9 4 19 407 26 9 5 17 328 Disagree 30 17 14 23 483 32 18 15 25 440 Neither 24 37 37 29 564 27 44 37 32 511 Agree 13 20 28 18 346 10 19 25 16 249 Stg. Agree 5 17 18 10 200 5 11 19 10 147 Total 100 100 100 100 2,000 100 100 100 100 1,675 Most politicians are corrupted by the money in politics. Stg. Disagree 2 1 0 1 23 1 1 1 1 17 Disagree 9 3 4 7 130 7 3 4 6 99 Neither 26 12 14 20 401 21 19 11 18 304 Agree 40 37 39 40 794 44 38 37 41 697 Stg. Agree 23 47 42 32 652 26 38 47 34 558 Total 100 100 100 100 2,000 100 100 100 100 1,675 If regular Americans got the laws and policies they wanted the country would be a lot better off. Stg. Disagree 6 1 2 4 86 3 3 3 3 51 Disagree 13 14 6 11 234 13 10 10 12 199 Neither 40 34 31 37 756 39 39 32 37 631 Agree 28 29 34 30 570 30 27 27 29 479 Stg. Agree 13 22 28 18 354 16 20 27 20 315 Total 100 100 100 100 2,000 100 100 100 100 1,675

Table 2.2: Populism questions and recall preferences among Democratic identifiers. Displaying percentages within each type of recall preference in California and New York. Includes “leaning” identifiers with their respective party. California New York Prohibit Allow Total Prohibit Allow Total Recall Recall Dem. N Recall Recall Dem. N % % % % % % Everyday Americans can’t get ahead financially because wealthy or politically powerful people won’t let them have their fair share. . Stg. Disagree 3 3 3 30 3 3 3 26 Disagree 9 9 9 106 8 8 8 88 Neither 19 19 19 218 17 16 16 170 Agree 35 39 37 443 33 37 35 362 Stg. Agree 34 30 32 373 38 35 37 384 Total 100 100 100 1,170 100 100 100 1,030 I’d rather put my trust in the wisdom of ordinary people than the opinions of experts and intellectuals. Stg. Disagree 33 25 28 352 27 24 26 288 Disagree 25 34 30 360 32 32 32 332 Neither 24 23 24 275 27 28 27 266 Agree 14 13 13 131 9 11 10 99 Stg. Agree 4 5 5 52 5 4 5 45 Total 100 100 100 1,170 100 100 100 1,030 Most politicians are corrupted by the money in politics. Stg. Disagree 2 1 2 16 1 1 1 11 Disagree 11 8 9 100 8 6 7 75 Neither 31 22 26 296 22 21 21 223 Agree 36 44 40 485 42 46 44 456 Stg. Agree 20 25 23 273 26 27 26 265 Total 100 100 100 1,170 100 100 100 1,030 If regular Americans got the laws and policies they wanted the country would be a lot better off. Stg. Disagree 9 4 6 72 3 2 3 30 Disagree 14 12 13 158 13 13 13 136 Neither 38 42 40 475 38 40 39 398 Agree 26 29 28 315 28 30 30 303 Stg. Agree 13 12 13 150 18 14 16 163 Total 100 100 100 1,170 100 100 100 1,030

These four questions are not the only way to try to measure populism; another approach looks at leader rhetorical style and identity. It is possible to describe the progressive faction of the Democratic Party and the segment of the Republican Party favoring former President Trump as “populists.”21 One question on the survey presented a list of political leaders (drawn from a previous survey, which also used a similar strategy, in Oct. 2020). Respondents were asked to pick the person on the list who “comes closest to representing your views.” These politicians are then sorted into four groups.    

Progressive: Bernie Sanders, Alexandria Ocasio-Cortez, Elizabeth Warren, and Kamala Harris. Center-Left: Joe Biden and Hillary Clinton. Old GOP: Mitt Romney, Lisa Murkowski. Trump GOP: Donald Trump, Marsha Blackburn, Tim Scott, Tom Cotton.

This list was not intended to be exhaustive, but many people who follow politics ought to be able to fairly easily select an alternative. Among California likely voters, the vast majority of Republican identifiers simply select former President Trump (70%). Senators Romney and Murkowski are grouped together as defying the rest of the party on significant votes (Romney on impeachment, Murkowski on the confirmation of Justice Brett Kavanaugh); they command a very small share of the Republican support (combined, 6%). The rest of the Republican support scatters among the other candidates. For Democratic identifiers, the split is much more diverse. About a third select Senator Sanders or Representative OcasioCortez; only a quarter pick President Biden, and even among Democratic-identifiers more people picked former President Trump than selected Senator Romney (although these are very small fractions of the totals). The progressive faction – the faction identified most with left-leaning populist rhetoric – is more likely to support allowing recall elections (see Figure 2.4 on the following page). The way this variable is re-coded, however, hides an even starker result: 69% of those selecting Ocasio-Cortez favor allowing recalls and 62% of those favoring Sanders support it. Warren supporters are more like Biden and Clinton supporters, with 51%, 51%, and 41% supporting allowing recall elections respectively in that group. If Ocasio-Cortez and Sanders are the ‘most populist’ of the Democrats, in terms of elite rhetorical style, then there does seem to be a relationship in this sense. The support in the progressive faction, in both states (as shown in Figure 2.5), is important for establishing the popularity of the recall procedure: this faction makes up a substantial portion of each state’s voters and are more favorable towards the recall than the center-left group in both states. 21

See, for example, Eric Oliver and Wendy Rahn’ 2016 paper, “Rose of the Trumpenvolk,” cited earlier. This approach also mirrors forthcoming work, coauthored by Betsy Sinclair and J. Andrew Sinclair, in the Journal of Political Institutions and Political Economy, but focused on the top-two primary.

Problems and Political Institutions

The survey included multiple measures of problems and performance. One of the most interesting is an open-ended question early on in the survey that encouraged respondents to write down three important problems in their state and country. These responses can be quite revealing, not only about what the voter identifies as a problem, but also about how the voter views politics as a whole. For many: the politics are the problem. Many voters chose to write something about politics – not the outcome of policies, but about the political situation or environment as a whole. Democrats who listed some kind of political problem – a complaint about Republicans, polarization, misinformation, political violence, and so on – were about 10 percentage points more likely to favor prohibiting the recall election than their fellow Democratic identifiers who did not describe anything in this category.

For example, respondent “1565657027” listed as the three most important problems: “Republican racists,” “Covid-19,” and “Climate change and climate change deniers.” This Democrat wished to prohibit recall elections. Another respondent, “1565654409,” wrote that the three problems were “The Republican party,” “White Supremacists,” and “Fascists trying to steal our elections.” This Democrat also wanted to prohibit recall elections. While virtually all the Republican respondents favor allowing recall elections, and favored specifically recalling Newsom in this one, more Republicans listed ‘political’ problems in the free response section: 34%, to 23% for the Democrats. The Republican comments about political problems are largely the mirror image of the comments from Democratic-identifying respondents. For example, respondent “1563417835” wrote that the three big problems were “Immigration,” “Loss of constitutional freedoms,” and “Draconian dictatorship.” As with the Democrats, mentions of political issues – something about the nature of politics, or a specific complaint about Democrats being the problem – often mix with particular policy concerns as well. Respondent “1564601463” wrote that the three problems were the “Progressive socialist agenda,” “Federal and state governments are too large and powerful,” and “We had a great economy that is being destroyed by Biden and the Democrats.” Respondents did much more, though, than to just complain about the other party. Some specific policy issues were mentioned by many; one example of such an issue is homelessness. Example: Homelessness, a problem without an obvious political consequence. In recent years, the general issue of homelessness has provoked a great deal of conversation in California. This is a salient issue for many Californians: in our coding of the free-response question about listing up to three most important problems in the country and in the state, nearly 30% of Californians mentioned something related to homelessness. California respondents were three times more likely to mention this issue than New York respondents as well, with under 10% mentioning homelessness in our New York sample (see Figure 3.2, top of the following page). In California, homelessness also catches the attention of respondents across the political spectrum. Nearly 33% of Democratic identifiers mention it, as do nearly 25% of Republicans. While they may not agree on what to do about it, or who to blame for specific problems, this is a high-profile issue in the state. It is not quite at the level of health topics (including Covid19 and the related public health responses), which 37% of Californian respondents mention, but it is considerably more frequently mentioned than other subjects. For example, only 6% of California respondents mentioned something about foreign policy on a survey conducted during the final withdrawal from Afghanistan. 18

For an issue on the minds of so many voters, homelessness might be expected to produce an impact on the recall election. Using the theories of “retrospective voting” – people vote against incumbents when something is wrong – one might also expect voters concerned about this issue to vote against Newsom. Nevertheless, as Figure 3.3 (prior page) illustrates, it does not appear that concern on this issue likely had much of an impact in California’s 2021 recall. For Democrats and Republicans, the first reason that opinion on homelessness did not change the dynamics of the recall contest, outside of the impact it may have on party identification, is that very little beyond partisanship is needed to explain the recall vote. No matter what they thought about any subject, nearly all Democrats voted to retain Newsom. For Republicans, it is the reverse: no matter what they thought about any subject, the vast majority of Republicans voted to recall Newsom. Even beyond the partisanship story, there is not much within partisan categories to distinguish individuals interested in this issue from those who are not. For example, it is not as if Democrats worried about homelessness voted to recall Newsom while the other Democrats voted to retain him. If concern about homelessness do not do much to explain the recall vote, beyond partisanship, does it shape preferences over having recall elections at all? Are citizens frustrated about this issue more likely to want to have the recall option?

Figure 3.4 on the previous page extends our analysis of preferences over recall institutions. There remains little to say about the Republicans beyond partisanship: nearly all Republicans favor allowing recalls. The Democrats are much more divided, with a majority preferring allowing recalls but a sizeable minority wanting to prohibit them. Within each category – prohibit or allow, for Democratic identifiers – the proportion concerned about homelessness is about the same (roughly one-third). The free response section provides some context for these findings. For example, respondent “1565738087,” a Democrat, identified the three most important problems as “homelessness,” “racism,” and “corporate profits” – and voted “no” on the recall, while still wanting to retain the ability to recall elected officials. This is a different kind of concern than what respondent “562353201,” a Republican voting “yes” in the recall, wrote: “Illegal [immigration],” “Homelessness and Crime,” and “Moral decay.” Despite the existence of attention on this issue, the public debate appears to be part of the larger ideological disagreements of the era, and thus just part of regular party politics.

Does interest in the homelessness issue correspond with preferences on other institutions, sharing similar goals of policy responsiveness and accountability? We can examine the relationship between thinking this is an important issue and preferences over primary election type (Figure 3.5, above). Since homelessness is an issue that may be perceived as unresolved, it may produce more interest in the top-two primary, a rule advocates suggested 21

would help stimulate problem-solving in Sacramento. The within-party differences are, though, quite slight. Much of the difference is at the party level. Example: Covid‐19, an issue on which voters split on substance and performance. Since the onset of the pandemic, few issues have been as important in American public life, and as salient in partisan politics, as the response to Covid-19. Initial praise for their response, often directly contrasted with the federal response provided by the Trump Administration, put both California’s Governor Newsom and New York’s Governor Cuomo in the national spotlight. As of this September, with new variants ranging and fights over mask and vaccine requirements ongoing in many states, it remained a top concern for voters in both California and New York.

It is sometimes difficult to tell where the issues related to Covid-19 end and other healthrelated issues begin: the pandemic impacts, and is influenced by, nearly every part of the U.S. health system. Counting all health-related comments – most of which were about Covid-19, or plausibly related – 45% of our New York sample thought this was among the three most important issues, as did 37% of our California sample (Figure 3.6). This measure may undercount interest in the topic, though, particularly among Republicans. For example, respondent “1564655889” wrote that the three most important problems were 22

“forest management,” “energy management,” and California’s “moron communist governor.” Since this is not an explicit mention of health, but is an explicit mention of a kind of political problem, this was coded as politics rather than health. Nevertheless, the respondent may very well have meant that Gavin Newsom was a “moron communist” in part because of mask, vaccine, or shutdown policies. Nevertheless, another question in the survey may shed some additional light on this topic. One question presented a list of eight topics and asked, “for each issue or area of activity listed below, how satisfied are you with how things have gone recently?” One of the items on the list was “the Covid-19 pandemic in your area.” We phrased the question that way to try to encourage Californians and New Yorkers to think about their states and regions rather than states like Florida and Texas, which were also in the news a substantial amount at that time. Figure 3.7 on the next page breaks down the responses to the Covid-19 satisfaction question by state and party identification group. Who was particularly dissatisfied? Note that dissatisfaction can come from a variety of sources: someone could be dissatisfied because local officials imposed mask mandates while someone else, living in the same place, could be dissatisfied because local officials were not zealously enforcing them. Nevertheless, despite the potential ambiguity, some trends appear in the data. In both states, a higher proportion of Republicans express dissatisfaction than Democrats. In New York, adding together both “dissatisfied” and “very dissatisfied,” 41% of Republicans gave a negative assessment. In California, that number is 55%. Compare those numbers to those from the Democrats: 31% in New York and 29% in California. Despite the differences between the parties, the levels of dissatisfaction remain high. According to recent statistics from the New York Times, about 68,000 Californians have died from Covid-19. New York, with a smaller population, has seen over 54,000 deaths. Beyond the deaths, the economic and social disruption have been significant adverse events, and there is a general expectation that citizens will vote out incumbents whose terms include plenty of bad news. Having roughly 30 percent of Democrats express dissatisfaction in heavily-Democratic states – now, with Democrats in power in Washington and in Albany and Sacramento – could be a real cause for alarm. Nevertheless, as we know, Newsom comfortably survived the recall election. Why?

The graph of dissatisfaction with Covid-19 and vote in the CA recall looks very similar to the same graph about homelessness: with nearly all Democrats voting “no” and nearly all Republicans voting “yes,” there just is not much room for any issue – even Covid-19 – to matter beyond the level of party identification.

In an alternative scenario, one in which Newsom lost the election, we might expect to see the 23.8% of likely-voter Democrats who are voting “No: Retain” Newsom (in Figure 3.8) to appear instead in the “Yes, Recall” line of the Democratic section of the graph. As with homelessness, or other policy areas, the dissatisfied Democrats are the key to California elections: they are the weak spot in the majority coalition. Why do dissatisfied Democrats not break for “Yes” on the recall in California? The answer likely lies in the partisan distribution of blame. We asked a follow-up question providing several different types of elected officials and asked respondents to approve or disapprove of their performance. If we focus on explicit disapproval – “disapprove” or “strongly disapprove” – the main trend is easy to spot in the data, broken down by partisanship in Table 3.1 on the following page.

Table 3.1: Disapproval for handling Covid‐19. Displaying percentages within each party type “strongly disapproving” or “disapproving” of each elected official in both California and New York. Each question is separate, so respondents could approve or disapprove of all or none. Dem. %

Ind. %

Rep. %

Total %

11 8 83

53 49 47

83 76 13

38 33 58

16 7 84

51 43 39

76 77 14

39 33 57

California Disapprove Newsom Disapprove Biden Disapprove Trump New York Disapprove Cuomo Disapprove Biden Disapprove Trump

When it comes to evaluations of specific individuals, it’s clear that Democrats blame Republicans and Republicans blame Democrats. In California, 11% of Democrats disapproved of how Newsom had handled the Covid-19 pandemic. In New York, although Governor Cuomo had resigned by the time the survey went into the field, only 16% of Democrats disapproved of his performance on the pandemic. Democrats in both states gave President Biden excellent reviews, with only 8% in California and 7% in New York disapproving of his performance. In contrast, they overwhelmingly disapprove of President Trump: 83% in California, 84% in New York. To some extent, former President Trump stands in for the whole of the Republican Party; on other questions (not presented here) we find similar numbers for Republicans in Congress and the state legislature among Democratic identifiers. Republicans, on the other hand, reverse the numbers: 76% in California and 77% in New York disapprove of President Biden’s handling of the coronavirus. Governor Newsom is as unpopular with California Republicans as former President Trump is with California’s Democrats: 83% of Republican identifiers disapproved of Newsom’s performance. Cuomo in New York fared slightly better, with 76% disapproval. The mirror images within the party groups, and the somewhat even splits among the “true” independents, hides the reality of Democratic dominance. There are a lot more Democratic identifiers, particularly when independent “leaners” are included in those categories. In both states, among registered voters as a whole, disapproval of the Democratic governors was approximately twenty percentage points lower than disapproval of President Trump. 26

Table 3.2: Policy Preferences for Covid‐19 Response. Displaying percentages within each party type for policies respondent would “support having the government implement, if it felt it needed to.” Respondents could agree with as many as they would like, so the columns do not add to 100%. . California Encouraging individuals to get vaccines. Requiring vaccination to enter certain buildings or events. Requiring everyone to wear masks indoors. Shifting schools to remote teaching until children can get vaccinated. Closing restaurants, bars, and similar indoor businesses. Providing financial relief for those negatively impacted. None of the Above New York Encouraging individuals to get vaccines. Requiring vaccination to enter certain buildings or events. Requiring everyone to wear masks indoors. Shifting schools to remote teaching until children can get vaccinated. Closing restaurants, bars, and similar indoor businesses. Providing financial relief for those negatively impacted. None of the Above

Dem. %

Ind. %

Rep. %

Total %

The partisan differences in approval reflect substantial differences in preferred responses to Covid-19. We framed our question as “if novel coronavirus (Covid-19) cases start rising in your community again, which policies would you support having the government implement, if it felt it needed to?” The responses are in Table 3.2, above. In California, the share of Democrats supporting a vaccine mandate was 53 percentage points above the support among Republicans (73%, 20%). In New York that gap is 52 percentage points (77%, 25%). The proportions on the mask requirements are similarly skewed.

Even on the least popular of these responses – closing things down – 46% of California Democrats and 42% of New York Democrats would support them. For Republicans, the percentages are 12% and 11%. So, while there are many Democrats who express dissatisfaction and frustration, the policies pursued by the governors of these states, along with the governors themselves, remained quite popular within their critical base of support. Cuomo fell from power for other reasons, and Newsom beat back a recall largely focused on his Covid-19 policies (and the potential policies of his most likely replacement). Examples: Wildfires in California and Extreme Weather in New York In recent years, Californians have experienced large wildfires and New Yorkers a variety of extreme weather events; in the public discourse, these are often described as impacts of climate change. They are also simply adverse events, putting at risk lives, property, and a sense of stability. We asked Californians about “how things have gone recently” with “the risk of wildfires near you.” New Yorkers had exactly the same question but with the “risks from extreme weather near you.” Figure 3.9 (below) displays the results for both states. As with Covid-19 and homelessness, there are many unsatisfied voters across all party groups. Republicans and independents in California express the most dissatisfaction, and Californians express more concern than New Yorkers overall. Although over half of California’s Republicans expressed concern about this issue, these are not the voters likely to cast the pivotal vote to depose a governor of the Democratic Party. The threat for Governor Newsom would come from dissatisfied Democrats; although there are fewer of them than there are Republicans, it remains the case that 42% of Democrats expressed dissatisfaction. As with the other performance issues: if the Democrats are voting “no” and the Republicans are voting “yes” on the California recall, there is just not much variation of these sorts of issues to explain (Figure 3.10). While we did not ask the wildfire follow-up question about blame to all survey participants – only to those who said they were dissatisfied or very dissatisfied – the patterns are very similar to the Covid-19 question. Only about 40% of Democrats were dissatisfied on the wildfire question; of those, only about 20% disapproved of Newsom’s performance. In contrast, of the dissatisfied Democrats, over 70% disapproved of former President Trump’s handling of the issue.

Conclusion

It is not news to say that we live in a time of deep political division in the United States and in individual states. Our study adds to the mass of that evidence: to know someone’s partisan identification is to know a great deal about their political behavior, including with a high probability their vote on the 2021 California recall election. The deep partisan divide can make it difficult to hold elected officials accountable for their performance in office. Democrats do express some dissatisfaction with the way things are going, but voting to recall Newsom would have involved a real policy cost to implement accountability, and most Democrats seemed to have found Newsom acceptable enough. That seems to be particularly true given the likely replacement, Republican Larry Elder. Governors Newsom and Cuomo have experienced different fates: Newsom remains in office while Cuomo is out. To some extent, these outcomes seem to be a matter of alternatives: from a partisan perspective, Democrats in New York could safely encourage their legislators to throw Cuomo out of office because his replacement would be a fellow Democrat. By contrast, Newsom managed to prevent another serious Democrat from being on the ballot in California—and thus, for voters, recalling Newsom meant placing a Republican in the state’s top office. This difference makes the comparison between the recall and the top-two all the more important. The top-two primary system makes it possible for elected Democrats to face other Democrats, either while running to hold onto a seat or running for a higher office. Some of the features of the recall election that make it difficult to wield as an instrument of accountability are simply not present in the nonpartisan primary. Dissatisfied voters in California may find the regular electoral process holds out more hope for meaningful accountability opportunities than does the recall process.