REJECTING COMPROMISE LEGISLATORS FEAR OF PRIMARY VOTERS Sarah

REJECTING COMPROMISE: LEGISLATORS’ FEAR OF PRIMARY VOTERS Sarah E. Anderson University of California, Santa Barbara Dan Butler University of California, San Diego Laurel Harbridge-Yong Northwestern University

Motivation: Gridlock • Lack of common ground (Polarization) • Fail to pursue common ground (Agenda control) • Reject common ground/compromise Proposal Preference SQ

Book Overview • Legislators reject half-loaf offers because they fear voter punishment • Fear punishment from subset of primary voters • Survey responses from legislators • Roll-call behavior • Survey of voters • One solution: More of the negotiation process in private

Why Might Legislators Reject Half-loaf Offers? • (from pre-analysis plan placed with EGAP) • Voter Punishment • H 1: Legislators who believe that their voters are very likely to punish legislators engaging in policy compromise will be less likely to vote yes on compromise. • Moral framing • H 2: Legislators who receive arguments for policy change framed in moral terms will be less likely to vote yes on compromise. • Denying the opposition party credit • H 3: Legislators will be less likely to support a compromise when the bill sponsors come from the opposing party or are bipartisan than when the sponsors are all from the legislator's party. • Future power • H 4: Legislators who expect that, as a whole, legislators serving in the future will be more likely to share their preferences will be less likely to vote yes.

2014 Survey of State Legislators • Survey state legislators on how they would vote on a gas tax proposal. • Measure necessary information (SQ, ideal point). • Asked • Gas tax preference • Vignette for survey experiment • Sample: • 257 (5% response rate) • Representative of majority status, gender, legislative professionalism, and term limits (But more Ds than Rs) • Results are robust to IPW.

Gas Tax Vignette Suppose that [two Democratic legislators / two Republican legislators / a Republican legislator and a Democratic legislator] in your state proposed a bill that would only affect the state gas tax, and would set the new state gas tax to [X] cents per gallon, a [Y] cent [increase/decrease]. Recall that the current gas tax in [STATE] is about [SQ GAS TAX] cents per gallon and you stated a preference for [THEIR STATED PREFERENCE] cents per gallon. In proposing the bill, one of the sponsoring legislators argued: [MORAL ARGUMENT TEXTS] Would you vote for this bill if it were introduced in your legislature? Yes No

23% of Legislators Reject Half-Loaf Offer

Politicians Fear Punishment from Subset of Primary Voters • Confirm Gas Tax Study: City Officials Experiment • AMOS: 1, 440 officials • Spring 2016 • Confirms finding: Fear of punishment predicts rejection • Politicians fear punishment from subset of primary voters • Survey responses from legislators (In-person surveys at NCSL) • Roll-call behavior (Size of Tea Party Predicts Rejection of Compromise) • Survey of voters • Summary: Politicians reject half-loaf offers because they fear punishment from a subset of primary voters

Testing Solutions: The Negotiating Process • No Labels and a focus on shared goals • Negotiation in public versus private

No Labels and a Focus on Shared Goals

NCSL 2016

Shared Goals Experiment and Compromise • Experimental Treatment • Use the pen and paper next to you to write down one reason why this issue is a top priority. • Legislator Smith shares this goal and provides own reason. • Which of two tools for addressing this issue would you prefer? Control moved immediately to preferred means

Negotiation Vignette Legislator Smith, a Republican, said he would prefer to secure Social Security and Medicare by decreasing benefits to recipients. Would you be willing to publicly support and vote for a proposal where 50 percent of the policy change was covered by decreased benefits to recipients and 50 percent was covered by increasing taxes to cover costs if Legislators Smith also did so?

Agreeing on Goals Has No Effect on Compromise

Focus on Shared Goals May Not Solve Problem of Fearing Voter Punishment • One-third of the public views compromise as selling out one’s principles (ANES 2016). • Among primary voters, only 29% think the ends justify the means (Anderson, Butler, Harbridge-Yong 2017 survey).

NCSL 2017: Negotiating in Public vs. Private

Public vs. Private Experiment and Compromise 1. Pick an area of the budget that should get a greater share than it does now, and one that should get a smaller share than it does now. 2. Random assignment to one issue. 3. Negotiation Vignette [Random assignment to public or private]

Negotiation Vignette: Public Versus Private As the details of the bill dealing with the spending are being finalized, the key Republican legislators on the other side of the issue (those for keeping last year's level of spending), ask to discuss the issue during a public meeting in order to work out a compromise. The press has been paying close attention to [chosen area] spending and the local paper is expected to attend and report on the meeting. The press is interested in reporting on the details of negotiations, including the back and forth of proposals and counterproposals.

Negotiation Vignette: Public Versus Private As the details of the bill dealing with the spending are being finalized, the key Republican legislators on the other side of the issue (those for keeping last year's level of spending), ask to discuss the issue during a public meeting in order to work out a compromise. The press has been paying close attention to [chosen area] spending and the local paper is expected to attend and report on the meeting. The press is interested in reporting on the details of negotiations, including the back and forth of proposals and counterproposals. Democratic legislators

Negotiation Vignette: Public Versus Private As the details of the bill dealing with the spending are being finalized, the key Republican legislators on the other side of the issue (those for keeping last year's level of spending), ask to discuss the issue during a public meeting in order to work out a compromise. The press has been paying close attention to [chosen area] spending and the local paper is expected to attend and report on the meeting. The press is interested in reporting on the details of negotiations, including the back and forth of proposals and counterproposals. in a closed door meeting

Negotiation Vignette: Public Versus Private As the details of the bill dealing with the spending are being finalized, the key Republican legislators on the other side of the issue (those for keeping last year's level of spending), ask to discuss the issue during a public meeting in order to work out a compromise. The press has been paying close attention to [chosen area] spending and the local paper is expected to attend and report on the meeting. The press is interested in reporting on the details of negotiations, including the back and forth of proposals and counterproposals. in a closed door meeting but no press will be allowed at the meeting. However, the local paper is interested in reporting on the final outcome of the negotiation.

Negotiating in Private Increases Ability to Compromise but Decreases Willingness to Meet

Part II Recap • Emphasizing shared end goals does not increase willingness to compromise on means. • Conducting negotiations in private rather than public may increase willingness to compromise.

Legislators’ Perceptions of Retribution by Subsets of Primary Voters Makes Compromise More Difficult. • Has the potential to exacerbate legislative gridlock. • Solutions? (Focus of Future Field Experiments) • Private/public negotiations • Communicate the outcome • Social interactions

Additional Slides

Comparison of Study 1 Survey Sample to State Legislator Population Attribute % in Sample Female Republican Term Limits In Majority 28 43 27 63 Mean in Sample Squire Index 0. 17 % in Population 24 52^ 26 64 Mean in Population 0. 20 Note: ^ A chi-squared test indicates that the distribution in the sample is significantly different (at p<0. 05) than the distribution in the population. For all other attributes, we reject the null that the distributions are different (or that the mean in the sample is different from the mean in the population). The Squire Index measures the degree of professionalism across legislatures.

Predicting legislators’ roll call votes using their stated preferences Dependent Variable = Vote to Increase Gas Tax Preference for Proposal over Status Quo |SQ – Preference| - |Proposal. Preference| Republican Constant Observations R-squared (1) (2) 0. 047* 0. 033* (0. 011) (0. 010) 0. 450* (0. 064) 58 0. 249 -0. 476* (0. 104) 0. 725* (0. 081) 58 0. 455 Note: Data comes the vote on Wyoming House Bill 69 (in 2013) and New Hampshire Senate Bill 367 (in 2014). The dependent variable is a dummy variable that takes a value of 1 when the legislator voted to increase the gas tax and 0 when they voted against the gas tax increase. Both models are estimated using OLS regressions. Standard errors in parentheses. * p<0. 05 (one-sided)

City Officials Study

Which constituencies might worry legislators? • General election voters • Probably not • Harbridge and Malhotra 2011; Delli Carpini and Keeter 1996 • Primary voters • Probably • Abramowitz 2010; Hill 2017; Karol 2015 • Donors • Possibly • La Raja and Shaffner 2015; Bonica 2013

2017 National Conference of State Legislators

NCSL 2017 Participants by state Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware District of Columbia Florida Georgia Hawaii Idaho Illinois Indiana Iowa Kansas 3 (2) 7 (2) 6 (3) 14 (10) 9 (3) 3 (1) 1 (1) 2 (2) 3 (0) Kentucky Louisiana Maine Maryland 15 (8) 14 (10) 4 (4) 6 (6) North Dakota Ohio Oklahoma Oregon 2 (0) 8 (3) 4 (2) Massachusetts Michigan Minnesota Mississippi Missouri 28 (12) 6 (2) 5 (4) 13 (13) 12 (6) Pennsylvania Rhode Island South Carolina South Dakota Tennessee 6 (4) 1 (1) 6 (3) 9 (4) 11 (5) 4 (2) 5 (2) 25 (15) 4 (3) 2 (1) 6 (4) 0 0 Montana Nebraska Nevada Texas Utah Vermont 6 (0) 13 (8) 10 (8) Virginia Washington West Virginia Wisconsin Wyoming 10 (8) 4 (3) 1 (1) 9 (5) 3 (3) New Hampshire New Jersey New Mexico New York North Carolina 3 (3) 8 (5) 3 (3) 11 (10) 1 (0) 5 (2) 1 (0) 9 (3)

How likely is each group to punish a legislator for voting in favor of a compromise bill? Voters in your party’s primary election Campaign donors to your party General election voters who don’t donate money or vote in a primary Not at all likely 9. 5% Only slightly likely 32. 3% 15. 7% 33. 3% Somewhat Very likely 42. 9% 15. 2% 43. 8% 29% 11% 40% 20% 6. 2%

How likely is each group to punish a legislator for voting in favor of a compromise bill? Percentage of Legislators 50 40 30 Not at all likely Only slightly likely Somewhat likely 20 Very likely 10 0 Primary Donors General

In general, if you were to make compromises on policy, how much retribution would you face from voters?

In general, if you were to make compromises on policy, how much retribution would you face from voters? Independent Variables Punishment by General Election Voters Punishment by Primary Election Voters Punishment by Donors Intercept N R-squared Dep. Var. = Fear of Retribution Scale 0. 072 (0. 064) 0. 144* (0. 071) 0. 069 (0. 072) 0. 246* (0. 038) 205 0. 07

Legislators think primary voters want them to kill the compromise. • Vignette about a co-partisan member of Congress: • Compromise bill included half of what legislator and their party wanted. • Majority of legislator’s party opposing bill. • Legislator’s vote will be pivotal.

External validity: The Tea Party and the Boehner years Tea Party supporters “would not hear of compromise” and they sought to “loudly tell Republican officeholders to do what they want or face challenges from the right in the next election” (Skocpol and Williamson 2012). Eric Cantor’s loss conveyed the message that “negotiations or compromises can get you beat” (Rep. Lee Terry (R-NE)).

Tea Party Support Among Republican Constituents • 2011 -2015 (112 th -114 th Congresses) • GOP Majority and Tea Party relatively prominent • Republican members of the House • Cooperative Congressional Election Studies 2010, 2012, 2014 What is your view of the Tea Party movement; would you say it is very positive, somewhat positive, neutral, somewhat negative, or very negative, or don't you know enough about the Tea Party movement to say?

Identifying Compromise Votes • CQ Key Votes that required some Democratic votes to pass

Cong CQ Key Votes 112 Session 1 Session Vote # Bill # 46 HR 1 112 112 112 113 113 1 1 1 2 2 2 1 1 1 376 491 690 72 195 659 23 30 55 S 990 HR 1249 S 365 HR 3630 HR 4628 HR 152 HR 325 S 47 113 113 1 1 1 89 125 550 HR 933 HR 1765 HR 2775 113 113 113 1 2 2 640 21 31 61 258 HJRES 59 HR 3547 HR 2642 S 540 HR 4660 113 2 322 HR 4870 113 2 327 HR 4870 113 2 452 HCONRE 105 Issue at Stake in Vote Republican Vote (Yea-Nay) 110 -130 Fiscal 2011 Continuing Appropriations/F-35 Alternative Engine (Rooney Amendment) PATRIOT Act Extensions Patent Overhaul Debt Limit Payroll Tax Relief Extension Student Loan Interest Rates Tax Rate Extensions Disaster Supplemental Short-Term Debt Limit Increase Violence Against Women Act Reauthorization Continuing Appropriations Air Traffic Controller Furloughs Fiscal 2014 Continuing Appropriations and Debt Limit Suspension Fiscal 2014 Budget Agreement Fiscal 2014 Omnibus Appropriations Farm Programs Debt Limit Extension Fiscal 2015 Commerce-Justice-Science Appropriations/Medical Marijuana (Rohrabacher Amendment) Defense Appropriations/A-10 Aircraft (Miller Amendment) Defense Appropriations/Surveillance (Massie Amendment) 196 -31 168 -67 174 -66 146 -91 202 -30 85 -151 49 -179 199 -33 87 -138 203 -27 202 -12 87 -144 169 -62 166 -64 162 -63 28 -199 49 -172 155 -73 135 -94 180 -37 U. S. Military Forces in Iraq 113 114 114 2 2 1 1 1 1 507 563 109 224 374 569 579 665 673 705 HJRES 124 HR 83 HR 240 HR 2048 HR 2146 HR 597 HR 1314 S 1177 HR 22 HR 2029 159 -71 Fiscal 2015 Continuing Resolution/Training and Arming Syrian Rebel (Mc. Keon Amendment) Fiscal 2015 Omnibus Appropriations Immigration Deportations Patriot Act Revision Trade Promotion Authority Export-Import Bank Budget Deal Elementary and Secondary Education Surface Transportation Fiscal 2016 Omnibus Appropriations 162 -67 75 -167 196 -47 190 -50 62 -177 79 -167 178 -64 178 -65 150 -95

Tea Party support among constituents predicts rejection of compromise Independent Variables DV = Vote Yes % with Positive View of Tea Party -0. 72* (0. 25) Mean Ideology of Strong Republicans -0. 03 (0. 05) Mean Voter Ideology -0. 18* (0. 05) 113 th Congress -0. 07* (0. 02) 114 th Congress -0. 11* (0. 02) R-Squared 0. 033 N 7, 214 Note: Clustered standard errors (clustered by legislator), shown in parentheses. * Significant at p<0. 05.

Recap of Primary Voter Punishment • Legislators fear punishment for compromising from primary voters. • Voters who oppose specific compromises do punish • Tea Party supporters predict rejection of compromise

Who are these voters? Nationally Representative Survey – 2029 respondents

Not primary voters, not people who oppose compromise in general. Only people who oppose the specific compromise.

No Labels Experiment at NCSL 2016