WEEK 9

Issues, Leaders, and Campaigns

Objectives

• Understand the case for and against issue voting in Canadian elections and identify periods where issue voting proved important

• Describe evidence in favour of leadership effects in campaigns and how this applies in Canadian elections

• Identify important components of political campaigns and how they affect (or do not affect) election outcomes

• Explain the enlightenment thesis and its applicability in the Canadian context

Issues

Do Issues Matter?

• One enduring debate: do individual issues matter in shaping vote choice and election outcomes?

• Issue voting requires people to:

• Form an opinion on an issue

• Identify a party's stance on an issue

• Pick the party closest to one's own viewpoint

The Case for Issue Voting

• People may choose parties that are on average closest to them (proximity voting)

• Some issues may be highly salient in a political campaign, priming them to evaluate politicians on the basis on these issues (remember priming effects?)

• Returning to the GST example: imagine you have some preference for or against tax cuts

  • You observe a blizzard of news coverage on the Conservative Party tax cut promise

  • Your opinions about tax cuts then get linked to your voting decision, for or against the Conservative party

  • If more people support the CPC position, they are net winners on that issue

The Case Against Issue Voting

1. Asks a lot of citizens:

1. Form their opinions on campaign issues

2. Identify party positions

3. Link their preferences to their vote

2. Projection effects: people assume their preferred party holds their issue positions

3. Follow the leader dynamics: people change their issue positions in line with their partisan identity

• Back to our GST example:

  • You have an ambiguous position on tax cuts, but you don’t like they Conservative Party and identify with the Liberal Party

  • You observe a blizzard of news coverage on the Conservative Party tax cut promise

  • You oppose the GST tax cut to remain in line with the Liberal Party position

Evidence

How do we typically observe issue voting?

1. We ask people questions about their views on important issues

2. We examine the relationship between views on this issue and vote choice

3. Control for Z variables (party ID, ideology, demographics)

2006 Canadian Federal Election

• Sponsorship scandal a high-profile issue due to the Gomery Inquiry

• CPC did a slow rollout of policy promises that dominated the election campaign

• Including gun registry abolition, GST cut, daycare tax rebate

• Conservatives began behind, but surged again in final weeks

• Corporate taxes, daycare, same-sex marriage, and the environment mattered in 2006

• Conservatives had no issue edge, and lost 4 points on the environment and daycare

• Liberals gained a couple of points in their support for higher corporate taxes (from NDP)

2008 Canadian Federal Election

• Policy agenda dominated by the Liberals and the Green Shift - a proposed carbon tax with offsetting rebates and tax cuts

• Conservatives reframed issue as tax grab

• Another major issue, at least in Quebec, was funding of the arts, with the CPC proposing cuts

• Green shift bolstered the CPC by 2 points at the expense of the Liberals

• Liberals gained that back at the expense of the NDP and CPC on environmental spending

• But Liberals lost 4.5 points on health care to benefit the CPC and NDP

• Overall, the issues cost the Liberals 4.5 points, mostly to the net benefit of the NDP

2011 Orange Wave

• 2011 election saw remarkable collapse of the Liberals and Bloc and rise of the NDP to official opposition status

• Big Issues: purchase of fighter jets, corporate tax cuts (both opposed by Liberals and NDP), omnibus crime package and repeal of the gun registry

• Inside of Quebec, sovereignty was a major issue pushed by the Bloc (to their detriment)

• Massive surge in NDP in 2011 election. Issue voting at work?

• Corporate tax attitudes, environmental and health spending => NDP voting, and effect increased over the campaign

• Explains 1/3 of the NDP surge in Quebec

• Expense of the Bloc - broader part of NDP success at winning over Bloc's left-wing

Caveats

• Which issues are going to matter in which election? How can we know? Sort of a fishing expedition

  • Why no CPC issue effects with robust policy platform in 2006?

  • Why sudden emergence of NDP in Quebec in 2011, when leader and policy mostly unchanged?

• Do issues really cause voting decisions or are we not fully accounting for confounding variables?

• Are the effects meaningful in size? It is possible they can matter in razor thin elections

• Rarely the one big issue that determines an election (but there are some exceptions)

Exception: Sponsorship Scandal

• Liberal Party was caught major corruption scandal in advance of the 2004 election

• A program was created to raise awareness of federal government spending in Quebec

• = slush fund where money was distributed to Liberal allies for little or no work

• Scandal broke before 2004 election and Gomery Inquiry investigated in 2005 with public hearing

• Gidengil et al. (2012) find that the scandal cost the Liberals 5.5 points in 2004 to the benefit of the CPC - a roughly 10 point swing

•  scandal was less of an influence in 2006: 3-point loss for Liberals and 3-point gain for Conservatives

• If not for the scandal, Liberals likely would have win a minority government

Leaders and Candidates

Party Leader Evaluations

• Party leaders are the most visible actors

• News coverage heavily determined by leader-centered campaign events

• Reinforced by leader debates

• To what degree to leader evaluations actually matter in voting decisions?

Case for Importance of Leadership

1. Makes voting decisions easy - perhaps a heuristic for less informed voters

2. News focus almost exclusively on party leaders - may prime citizens to weigh leadership strongly in vote decision

The Presidentialization Thesis

Could leadership matter more now than in the past? (The Presidentialization Thesis)

• Centralization of power in Prime Minister's office

• Rise in news media sensationalism and horse race coverage

• Declining partisanship in many countries and weaker parties

• Rise of populism

Case Against Party Leaders

1. Polarization in Canada means ideology, values and partisanship matter more than in the past

2. If you are a strong, committed partisan, this shapes your leader evaluations

3. Rarely are leader evaluations systemically different than party evaluations (aside from NDP)

Measuring Leader Evaluations

• Classic measurement strategy is 0-100 feeling thermometer ratings for parties and their leaders

• Analytic strategy: are these reported feelings associated with vote choice after controlling for Z variables (party ID, ideology, demographics, etc.)

Problems:

• What exactly are you trying to measure? Feelings towards an individual? Policy? Party leader's personality? (validity problem)

• Can people make reliable distinctions in feelings by single degrees (39 degrees v. 40; reliability)?

• Encourages people to choose mid-point (50 degrees; validity)

Leader Trait Evaluations

• Trait evaluations may be a better bet

• Two dimensions:

  • Character traits: empathy, honesty, caring about others

  • Competence traits: intelligence, strength of leadership, competence

• People do not appear to make distinctions between leaders on these dimensions

• Evaluations of the same leader can vary by election

• Example: advantage for Layton on honesty, disadvantage for Martin that grew between 04 and 06

What Causes Leader Evaluations

• What causes leader trait evaluations, besides partisan identity?

• Right-leaning leaders tend to be advantaged on the competence dimension, and left-leaning leaders the character dimension

• Gender effects: female leaders do better on character dimension

• Much more work to be done

Does Leadership Matter

• Over 90% of voters cast a ballot in  line with their leader preferences

• But people tend to like the leaders of the party they support - partisan bias

• Very hard to causally estimate effect of leadership on voting decisions

• Bitter (2018) finds that both competence and character evaluations correlate with vote choice

• No increase over time, unstable over time and for party leader

• Effect of leadership evaluations on vote share and election outcomes depends on how badly advantaged certain leaders are relative to others

• NDP only gain about a point due to character evaluation edge

• Harper cost the CPC 7.5 points in 2006, but netted them 5.5 points in 2008

• Dion cost the Liberals 4 points in 2004

Campaign Effects

Components of a Campaign

A lot of moving parts in an election campaign, most of which we have small, if important effects on public opinion and vote choice:

1. Campaign messaging

2. Paid media

3. Earned media

4. Debates

Campaign Messages

Candidates have to decide which issues to campaign on or avoid, and their respective positions

• Valence issues are those where candidates share the same position (e.g. economic performance, crime, corruption)

• Directional issues (or position issues) are those on which the parties differ, usually along ideological lines

• On valence issues, incumbents will focus on them if conditions are good and avoid them if they are poor

• Candidates will focus on directional issues that are supported by most Canadians, and avoid issues where they are in the minority - sword v. shield issues

• Candidates especially like to talk about wedge issues that their party supporters agree on, but that divide the opposing party

Paid Media

• Most money is spend on excessive campaign advertisements - or paid media

• Ads are seen as beneficial because candidates have complete control of their messages

• Mixture of positive and negative - the latter seen to be more effective

• Recent research in the US shows important - but small - effects

• Dem ad advantage of 100 = 100 2pp increase in Presidential, 4-6pp in Senate, 6-9pp in Governor, and 8-9pp in House elections

• BUT: 95% of cases fall between a Dem ad advantage of 16 and a GOP advantage of 8 (presidential level)

• Implications: effects in practice are even smaller because parties are relatively evenly matched in ad buys

• Most research finds it is ad volume, rather than the message itself that matters

Earned Media

• Paid media is expensive, so candidates also rely on acquiring earned media from their coverage In the news

• Campaign speeches and events are geared towards earning coverage in local market

• Engage in processes of news management to disseminate their message (i.e. press releases, spin, sound bites)

• Tone of earned media tracks vote intention (but causal direction difficult to disentangle)

• Perhaps more important for small parties

• Earned media can also prime certain issues or leadership considerations (e.g. 1988 and free trade, 2011 and Jack Layton)

The Debates

• The biggest set of campaign events are the English and French language debates

• The details - how many? Who is invited? Format? - vary considerably election to election

• Journalists obsess about the mythical "knock-out blow" that wins an election

• For example: Brian Mulroney's "you had an option, sir" line during the 1984 election

• Debates can move vote intention in the campaign

• More important is the commentary that follows a debate - we see effects for debate watchers and non-watchers as well

• Effects usually dissipate before election day (exception 1988, but didn’t change outcome)

Do Campaigns Matter?

• It is often claimed by journalists and political commentators that "elections matter" .. But how? How much? And for whom?

• But: many people have party IDs and strong values and make a decision before the campaign even begins

• Leaders and key issues often set before campaign begins

• On the other hand, often lots of volatility in vote intentions

• Around 50% of Canadians report deciding either during the campaign or on Election Day

• A sizable share of Canadians do not report a party ID

Importance of Late Deciders

• Maybe we observe campaign effects primarily among those voters who hadn't decided before the campaign?

• Pre-campaign deciders much more stable in their vote choice (pre-post 1997 election, 84%) compared to those interviewed before or during their time of decision (26%, 46%)

• Political knowledge and interest only minor factors shaping decision delay - mostly lack of party ID

• PC's under Jean Charest roundly praised for debate performance - saw bump in the polls

• Only occurred among those who had not decided before the campaign

• Note: bump dissipated before Election Day for everyone

• Also found drop in intention for Liberals to benefit of Reform among late deciders

• Maybe cost Liberals as much as 7 points

• In a close election, these differences can matter

Unanswered Question

Some unanswered questions:

• Do these effects emerge principally through priming or persuasion?

• Is there a learning component to campaigns?

• What exactly causes these campaign effects? Can this be understood systemically?

• Is it all about a lack partisan attachment?