The politics and impact of party leader visits in Australia

ABSTRACT

Local visits by party leaders are an integral part of election campaigning around the world. In the age of media-driven elections, do they matter? How are local contests selected for a visit? Most importantly, what effect do they have on the vote? Using a unique dataset covering local visits by the major party leaders during the 2013 and 2022 Australian federal election campaigns, we address these questions. We find that marginality is the main criterion for selecting a seat for a visit, followed by its geographical location. Leader visits in 2013 mattered more to the vote than in 2022. We explain the differences between the two elections and the four leaders by the relative standing of each of the leaders among voters. Overall, we conclude that the electoral impact of a leader visit is modest and context-dependent.

KEYWORDS:

• Leaders
• Elections
• Voting
• Australia
• Political behaviour

The campaign trail is traditionally an integral part of the theatre of an election. Political leaders get ‘out on the stump’ and want to be seen as normal people, sharing voters’ problems, conveying empathy and offering solutions. This personal side to election campaigning has become more important in recent decades with the decline in party loyalties and the associated increase in the personalization of politics (Garzia, Ferreira da Silva, and De Angelis 2021; McAllister 2007). Local visits by party leaders typically involve activities such as ‘meet and greet’ street walks, talks to sympathetic community groups, and tours of local enterprises and businesses (Middleton 2021). The impact of these activities extends far beyond the local area and the constituents who are directly affected. The media images that are generated fill the national television evening news and carry over to the newspapers (and increasingly and more expeditiously to the social media).

Do these extensive and orchestrated local visits by a party leader during an election campaign matter to the election outcome? What factors decide if a seat warrants a visit? In short, what are the politics (and specifically the impact) of leader visits? This paper uses a unique dataset covering the 2013 and 2022 Australian federal elections to examine the politics and impact of major party leader visits. We examine how seats are chosen for a visit, and estimate the electoral impact of a visit. Our results show that, not surprisingly, there is a non-random, strategic approach to the selection of seats visited by the major political leaders. At the same time, electorate visits by either of the two major party leaders appear to have little significant electoral impact, and what effect they do have is mediated by the popularity of the leader in question.

The paper proceeds as follows. The next section examines the current state of research on leaders and election campaigns, while the second section examines the Australian context and how political institutions shape campaign activities. The third section outlines the data, variables and methodology used in the analysis, while the fourth section presents the main findings. In the conclusion, we discuss the implications of our findings for understanding election campaign dynamics, the role of the major party leaders and their impact on the vote.

Leaders and election campaigns

Why do leaders invest such a lot of time and energy visiting local election contests during an election campaign? The thinking behind a leader’s visit is relatively simple: the stature of the political leader’s office or the popularity of the incumbent can have spill-over effects on the electoral support for the candidate being visited. Expressed most directly in the US literature, the logic of a visit is that the local candidate can often ride on ‘the coattails’ of an incumbent president or challenger (Cohen and Powell 2005). This value-adding effect occurs either by the leader helping an incumbent to retain their seat when they are at risk of defeat, or by aiding a challenger who might otherwise fall short of the required votes.

What determines whether or not a leader visits a seat? The strategic considerations for choosing which candidates to select for a leader’s visit are many and differ between parliamentary and presidential systems. Six main factors have been identified in the literature.

First, studies in parliamentary systems identify the marginality of the constituency as the most important consideration and whether the leader’s visit can make a difference to the local candidate’s election prospects. In presidential systems, the relevant consideration is the marginality of the electoral unit, which in the case of the US is the state (Cohen, Krassa, and Hamman 1991; Keele, Fogarty, and Stimson 2004; McClurg and Bryan 2009). A marginal constituency (or state) may also bring into play local factors, such as the competitiveness of the local contest and perhaps also the competitiveness of the state or region, and the risk of an incumbent being defeated by a challenger.¹ Other things being equal, a more marginal seat that carries a greater risk of defeat will be more likely to host a leader’s visit.

A second consideration is the resources required for a visit, both by the leader and their staff as well as by the local party. The time the political leader has available, recognising the substantial personal burden of travel, will be a major consideration (Cohen, Krassa, and Hamman 1991; King and Morehouse 2004; Mellen and Searles 2013). There is also the opportunity cost of diverting time away from other activities, such as attending media events or national activities such as policy launches. For the local party, organising, conducting and delivering a leader visit can be an unwelcome burden on already stretched campaign resources. National campaign offices often decide on a visit at short notice and this typically involves allocating scarce local resources (including getting local members and supporters to attend the event and be seen) to ensure the visit is a political success (Middleton 2019).²

The leader’s personal popularity and longer term political prospects is a third factor that can have a major bearing on the choice of a visit. A popular leader will be in demand to visit in the hope that their luster will spread to the local candidate (Doherty 2007; Hoddie and Routh 2004; Lang, Rottinghaus, and Peters 2011; Mellen and Searles 2013). For example, an electorate visit by Bob Hawke was much sought after in the 1983 Australian election during the peak of his popularity (Butler 1983, 157). Equally, an unpopular leader can be seen as vote-subtracting. Many British Labour candidates viewed Jeremy Corbyn in the 2017 UK general election when he was often dissuaded from visiting a constituency due to his unpopularity among many voters (Middleton 2019). During Donald Trump’s presidency, his controversial personality and politics meant Republican candidates had decidedly mixed views about whether a visit by the president would help their cause (Abramowitz and Panagopoulos 2020).

A fourth consideration is how the leader views their own career prospects. A leader who sees themself occupying the position for a lengthy period may be motivated to visit as many local contests as possible (Cohen and Powell 2005; Sellers and Denton 2006). Such campaigning can be feedstock for the leader’s own political ambitions since they must win the most seats to form government. As such, the time and effort spent by a political leader on the campaign trail is essentially an investment in their own self-interest and long term political career. At the very least, assisting a local candidate at an election can help to deliver that person’s support if they are elected. In the US, presidential visits that assist a local candidate who is elected to Congress can help to further the president’s policy agenda by garnering support on the floor of the House of Representatives (Cohen, Krassa, and Hamman 1991; Herrnson and Morris 2007; Hoddie and Routh 2004; McClurg and Bryan 2009).

The dynamics of the local contest represents a fifth factor that may influence the decision to visit an electorate. Candidates with a strong local following, whose grass-roots campaigning is dominated by local issues, may not want the political leader to visit if it means drawing attention to possibly unpopular national issues. Similarly, a visit by one leader may prompt a visit from the opposing leader, again an unwanted distraction for a candidate campaigning on local issues. A visit by a leader can also have the counterproductive effect of mobilising supporters of the other party and diverting scarce resources to deal with the threat (Abramowitz and Panagopoulos 2020; Sellers and Denton 2006). And not least, the potential of a local contest to generate positive local media coverage is important, in contrast to often more critical national media coverage that can drown out the local message (Charnock, McCann, and Tenpas 2009; Cohen and Powell 2005; King and Morehouse 2004).

Finally, there is the importance of the media environment. The personalization of politics, driven largely by television, has resulted in leader-centric media strategies during election campaigns (Deacon et al. 2017; Mughan and Aaldering 2018). Since voters form their impressions of the leaders based on what they see in the mass media, a local visit, with its associated visual images, can convey a more effective political message than the release of a policy document (Lenz and Lawson 2011). The media environment and how it can be leveraged around a local visit by the leader is therefore another potentially important consideration in the choice of an electorate for a visit.

Does a visit by a leader attract additional votes? Most of the research addressing this question has come from the US and has found that a visit by a sitting president during a mid-term congressional election can add value to a candidate’s electoral support, whether incumbent or challenger (Cohen and Powell 2005). Such visits may impact on the national result if there are a sufficiently large number of close local contests (Holbrook 2002). Other scholars have questioned efforts to measure the electoral effect of a discrete leader visit. One problem is ‘the permanent campaign’, so that political leaders engage in constant election campaigning across the electoral cycle making it difficult to partial out the electoral impact of any specific visit or visits. Another problem is that leader visits may be motivated by the leader’s own electoral agenda, so visits to local candidates during the campaign might simply be part of the ordinary political process and not explicitly aimed at influencing the election (Doherty 2007; Eshbaugh-Soha and Nicholson-Crotty 2009).

While electoral and political contexts have varied across studies of political leader visits, US modelling suggests a visit can add anywhere between 1 and 2 percentage points to the vote of the local candidate (Cohen, Krassa, and Hamman 1991). Probability models have found presidential visits to congressional candidates can lift the likelihood of the latter winning by more than 40 percent in competitive elections (Herrnson and Morris 2007). In Britain, research has found that constituency visits by the Conservative and Liberal-Democrat leaders during the 2010 general election added 1.2 and 1.5 percentage points, respectively, to their vote shares in the seats visited, both effects being statistically significant (Middleton 2015). Counterfactuals indicate well-targeted strategic visits can make the difference between winning and losing overall, Truman’s defeat of Dewey in the 1948 US presidential election being a case in point (Holbrook 2002).

As noted earlier, visits by an unpopular leader are often discouraged by local candidates as representing a drag on the party vote. US research provides some evidence to support this contention. For example, visits by relatively unpopular George W. Bush to Republican candidates in the 2002 and 2006 midterm Senate elections reduced the latter’s vote (McClurg and Bryan 2009). Similarly, visits by the unpopular vice president Dick Cheney (a Republican) during the 2000 presidential election increased Democrat support by as much as 0.5 percentage points in some states (Hill, Rodriquez, and Wooden 2010). More recently, campaign visits by Donald Trump during the 2018 midterm elections at best had no impact on, and may well have adversely affected, support for Republican senatorial candidates (Abramowitz and Panagopoulos 2020).

The politics of leader visits during an election campaign are complex. The calculations differ significantly between parliamentary and presidential systems since the institutional incentives are different. In both types of systems, national and local factors come into play in the decision whether to visit a seat, as does the popularity and career aspirations of the leader, the types of issues being played out during in the election campaign, and the potential for media attention (both positive and negative). This complexity makes the use of visits during a campaign a delicate strategic decision. There is also consistent evidence that visits from a popular leader can increase their party’s vote, while visits by an unpopular one can have the opposite effect. In the next section we examine some of the factors which determine how these decisions are made in the Australian context.

Australian election campaigning

Election campaigning in Australia differs in two important respects to campaigns conducted elsewhere. The first is the system of compulsory voting which means all electors can be influenced by a campaign, rather than just those who turnout to vote which is the case in voluntary voting systems (Birch 2016). Since the two functions of political parties during an election are mobilization and conversion (Dalton, Farrell, and McAllister 2011; Hillygus 2010), and in Australia mobilization occurs through the electoral system, the parties concentrate their resources on conversion. As a result of compulsory voting, then, election campaigns in Australia exhibit much less local level activity compared to campaigns in voluntary voting systems (Karp, Banducci, and Bowler 2008; Ward 2003). This necessarily affects the number and extent of leader visits compared to other countries, but means that they are more likely to be targeted towards seats that have the greatest potential and represent the best use of the party’s resources.

Compulsory voting also affects local level campaigning in two other ways. First, around one in five voters who attend the polls would not turnout under a voluntary voting system (McAllister 2011, 24–25). This sizeable group of voters tend to be less interested in and less informed about politics. Consequently, they are more open to short term influences on the vote, such as the personalities of the party leaders rather than the detailed policies of the parties. In principle, then, these voters should be more open to influence by leader visits and the types of personalized activities that visits deliver. A second consequence of compulsory voting is the requirement that the electoral system is as voter friendly as possible to avoid the need to coerce (and possibly fine) voters to turnout (Dassonneville et al. 2023). One aspect of this is early voting and Australia has been a pioneer in its use, such that around half³ of all voters now cast their ballots prior to election day, one of the highest rates in the world (McAllister and Muller 2018). Early voting arguably complicates party strategies in organising visits, since many voters will already have cast their ballots when a visit occurs.

The second way in which Australian election campaigns differ from those conducted elsewhere, and which is partly a consequence of compulsory voting, is media coverage. The Australian media is highly centralized and partisan, and because general elections are held every three years under compulsory voting, there is intense political coverage (McNair et al. 2017). While the political use of social media has increased dramatically (Bruns and Moon 2018; Cameron and McAllister 2022, 8), the established media remain the main source of political information for the majority of voters (Carson and Zion 2020). The conduct of election campaigns are ‘inordinately “leader-centric”’ (Strangio and Walter 2020, 107) and in general elections the two main party leaders are the perennial faces of the campaigns, with constant daily media attention on their travel to and activities in different seats and states (Errington 2015, 72; Mills 2020).

Do election campaigns in Australia influence the vote? The answer to this question is yes, but in line with the international research the effects are modest – around 1 or 2 percentage points at most – and affect different groups of voters in different ways (Forrest and Marks 1999). Research has also shown digital campaigning is increasingly important in shaping the vote (Gibson and McAllister 2015), so how visits are publicized in the social media can influence voters. Leader visits in Australia may, in principle, have more effect due to the high levels of political personalization. This means the visual media often give more campaign coverage to leaders than to parties and their policies (Denemark, Ward, and Bean 2007). In principle, then, campaigns matter to the vote and within the spectrum of campaign activities aimed at voter conversion, leader visits form a significant part.

The international and Australian research on leader visits leads to four expectations about their patterns and effects. First, in terms of which seats the leaders select for a visit, leader visits are most likely to be aimed at marginal seats, which may either be seats held by the leader’s party (which they may risk losing) or those held by the opposing party (which they want to win). Second, seats that have the greatest potential for media coverage will be more likely to receive a visit.⁴ Third, it might be expected that those seats held by an incumbent from the governing party would be most likely to receive visits. For the challenger, they must win seats from the government if they want to form a majority, and for the incumbent, ‘sandbagging’ their seats is important to preserve their majority. Fourth, since the purpose of a visit is to increase the vote for the leader’s party, we would expect a visit to increase the vote for that party. These expectations lead to the following hypotheses:

H1 Visits by a party leader will be more likely to occur in a marginal seats compared to safe seats.

H2 Seats where voters consume more media will be more likely to receive visits compared to seats where voters consume less media.

H3 Seats held by the governing party will be more likely to receive visits by both party leaders compared to seats held by the opposition party.

H4 Visits by a party leader will increase the vote for the party in those seats compared to seats that do not receive a visit.

Data, measurement and method

Data. Two datasets are used, covering the 2013 and 2022 federal elections. The 2013 dataset covers visits by prime minister Kevin Rudd (leader of the Labor Party) and opposition leader Tony Abbott (leader of the Liberal Party) between Sunday 4 August 2013 and election day on Saturday 7 September 2013, a period of 35 days. The 2022 data was compiled from media reports of the campaign trails of prime minister Scott Morrison (leader of the Liberal Party⁵), and opposition leader Anthony Albanese (leader of the Labor Party) during the 2022 election campaign. The data series covers 41 days over the period Sunday 10 April 2022, the day the election date was formally announced, to election day on Saturday 21 May 2022.

Measurement. The coding and means of the variables used in the 2013 and 2022 analyses are shown in Table 1. The two main dependent variables are leader visits and vote swing. Leader visits is the number of leader visits to electorates; in 2013, 43 percent (or 65) of the 150 electorates had at least one visit from a party leader during the election campaign, compared to 57 percent (or 85) of the 151 electorates in 2022. Vote swing is the percentage change in the two party preferred vote in the election compared to the previous election, expressed as the swing to the Coalition and away from Labor. In 2013 the Coalition had a swing of 3.69 percentage points in its favour, with the Coalition government led by Tony Abbott replacing the Labor government first elected in 2007. In 2022, the Coalition experienced a swing against it of 3.56 percentage points, resulting in the replacement of the Coalition government led by Scott Morrison with a Labor government led by Anthony Albanese.

Table 1. Variables, definitions, means.

		Means
Variables	Coding	2013	2022
Vote swing	Percent two party preferred vote swing to Coalition	3.69	−3.56
Total party leader visits	Number	0.81	1.86
Liberal leader visits	Number	0.41	1.00
Labor leader visits	Number	0.41	0.87
Marginality	Percent two party preferred	2.47	2.53
Media consumption	Zero to 10 scale
Urbanization
Outer metropolitan	1 = yes, 0 = no	0.32	0.28
Provincial	1 = yes, 0 = no	0.14	0.17
Rural	1 = yes, 0 = no	0.27	0.25
Gender	Percent females	50.61	49.74
Educational qualifications	Certificate III or higher	46.42	46.46
Language other than English	Percent	20.08	19.95
Home owner	Percent	69.08	68.06
(N)		(150)	(151)

See text for definitions of variables.

The marginality of the electorate is a continuous variable which measures the percentage change in the two party preferred vote required for a seat to change hands based on the result of the previous election. Media consumption combines level of attention to election news in the newspapers, television, radio, and internet⁶ in each electorate and is scored from zero (no attention) to 10 (maximum attention). Preliminary analyses modelled the separate effects of the four different media sources but since none of them were statistically significant, a composite measure of overall media consumption in each electorate is used.

To control for social background, five variables measure the social characteristics of electorates using the 2011 census (for the 2013 results) and the 2021 census (for the 2022 results). The degree of urbanization is measured by a four-category nominal variable, with inner metropolitan forming the excluded category. Gender is measured by the percent of females in the electorate, educational qualifications by the percent in the electorate who possess a certificate III qualification or higher, non-English speaking by the percent in the electorate who speak a language other than English,⁷ and home ownership by the percentage of home owners in the electorate.⁸

Method. Since the dependent variable is an interval variable we use ordinary least squares regression methods.

Results

Both major party leaders made extensive electorate visits around the country during the 2013 and 2022 election campaigns. Table 2 shows that during the 2013 election campaign, the Liberal leader, Tony Abbott, made 48 visits out of a total of 150 electorates, while Kevin Rudd, the Labor leader, visited 43 electorates. In 2022 Scott Morrison, the Liberal leader, made 66 electorate visits compared to 59 visits for Labor’s Anthony Albanese. However, Albanese lost a week of campaigning when he tested positive for COVID-19 in the early stages of the campaign and had to isolate for seven days as a result.⁹ Overall, each leader visited around four in every 10 of the federal electorates. Comparing the 2013 and 2022 patterns suggests the pace of election campaigning, at least in terms of the total number of leader visits, has increased in recent years. Evidence from the 2019 election bears this out, and in that election Scott Morrison made 75 electorate visits and Bill Shorten 68 visits (Smith 2020, 226–227).

Table 2. Leader visits, 2013 and 2022.

	2013		2022
Electorates	Rudd	Abbott	Morrison	Albanese
Not visited	107	102	85	92
Visited, of which:
One	30	38	28	24
Two	10	8	14	17
Three	2	1	11	6
Four	0	1	7	8
Five	1	0	2	1
Six	0	0	4	3
(At least one visit)	(43)	(48)	(66)	(59)

Number of visits by each party leader for each of the House of Representatives electorates (150 in 2013, 151 in 2022).

The second panel of Table 2 shows how many visits each of the electorates received. In both 2013 and 2022 and across both parties, a plurality of seats received one visit each. This was especially the case in Abbott’s 2013 campaign, when 38 of the 48 seats he visited were visited just once. By contrast, in 2022 the number of seats receiving two or three visits from both leaders was almost equal to the number of seats receiving one visit only. At the other end of the scale, in 2022 four seats received six visits from Scott Morrison (these were the marginal Coalition seats of Bass and Chisholm, the Labor marginal seat of Parramatta, and the safe Labor seat of Lyons). Morrison also visited his own safe Coalition seat of Cook four times during the campaign. Anthony Albanese visited three seats no less than six times (the marginal Coalition seats of Bass and Brisbane, and his own safe seat of Grayndler).¹⁰

Leaders and senior frontbenchers are more likely to concentrate campaign visits on target seats than on unwinnable seats. During the 2022 campaign, Albanese visited predominantly government-held marginal seats, but his week of COVID-enforced isolation allowed Labor to showcase some of its other frontbench talent. Morrison’s campaign itinerary was more problematic. While he visited various marginal Coalition seats in outer-suburban Queensland, as well as seats in western Sydney and Tasmania, a range of political and organisational problems presented significant constraints. These included: the late selection of candidates; factional conflict in the West Australian Liberal Party; anti-China rhetoric which limited the scope for campaigning in marginal seats with large proportions of Chinese-born voters; and a perception among voters in the ‘Teal’-contested seats that the Liberals lacked an effective climate change policy (Kefford and Mills 2023, 111). Moreover, unlike Albanese, Morrison was unable to use his main cabinet colleagues – Barnaby Joyce and Peter Dutton – due to their unpopularity.

To what extent was there asymmetry in the seats visited? If one leader visited an electorate what was the probability the other leader would follow suit? The data suggests there was a high degree of overlap in the seats visited. Of the 85 seats Morrison did not visit in 2022 and the 92 seats Albanese did not visit, no less than 66 of them did not receive a visit from either leader. In 2013 the pattern was similar, and of the 107 seats Rudd did not visit and the 102 Abbott did not visit, 85 were not visited by either leader. As already noted, there was also a greater intensity of visits in 2022 compared to 2013. In 2022, one lucky seat, the marginal Liberal seat of Bass, received no less than six visits from each of the two leaders. In 2013, just two seats – once again Bass and Petrie – received two visits only from each leader.

Seat visits in both 2013 and 2022 were therefore markedly non-random. The most obvious explanation for these patterns, expressed in the first hypothesis, is the marginality of the electorate at the previous election. The parties chose to target electorates held by their competitor considered winnable, or electorates they already held that they considered to be at risk of loss. To what extent did marginality at the previous election contribute to which seat was targeted to receive a leader visit? These estimates are shown in Table 3, based on an ordinary least squares regression analysis predicting the number of leader visits to each the electorate from the marginality of the seat and a range of social background variables.

Table 3. Explaining leader visits, 2013 and 2022.

	2013	2022
	Est	Est
Marginality	-.058**	-.144**
	(.014)	(.034)
Media consumption	.718	.216
	(.560)	(1.105)
Urbanization (inner metro)
Outer metro	-.513**	−1.597**
	(.260)	(.621)
Provincial	.139	-.732
	(.387)	(.883)
Rural	-.702*	−1.333
	(.413)	(.945)
Gender (female)	-.030	-.043
	(.138)	(.038)
Language other than English	-.007	-.006
	(.008)	(.018)
Education qualifications	-.002	.002
	(.017)	(.039)
Home owner	-.036**	.025
	(.015)	(.024)
Constant	4.138	4.067
	(6.129)	(3.902)
Adj R-sq	.213	.135
(N)	(150)	(151)

** p<.05, * p<.10. Figures are partial regression coefficients (Est) and standard errors (in parentheses) predicting the number of leader visits in an electorate.

The modelling reported in Table 3 shows the marginality of the seat is easily the most important determinant of whether or not the seat is visited by a political leader. This finding holds for both 2013 and 2022. As the margin of electoral safety increases and the risk of losing (and the possibility of winning) decreases, the probability of a visit by a political leader declines markedly. Looked at another way, the less safe (that is, more marginal) a seat the more likely it was to get a visit from a political leader in both elections. These findings therefore confirm our first hypothesis.

The second hypothesis predicts that seats will be chosen for a visit based on their potential to generate media coverage and therefore influence voters outside of the electorate. For this hypothesis to be confirmed, seats with higher levels of media consumption should attract more visits than seats with lower levels of media consumption. The results in Table 3 reject this hypothesis, and while the effects in both 2013 and 2022 are in the predicted direction, neither is statistically significant. We therefore reject H2.

The third hypothesis predicts that the leaders of both parties will be more likely to visit seats held by the government in order to either win office (for the challenger) or retain office (for the incumbent). This means both leaders should be more likely to visit Labor-held seats in 2013 and Coalition-held seats in 2022. This hypothesis is tested in Table 4. The results show that this prediction is supported in 2013 but not in 2022. In 2013 Kevin Rudd visited 36 percent of Labor-held seats, compared to 21 percent of Coalition-held seats, the difference being statistically significant at the 5 percent level. There is a similar, though slightly less marked, pattern for Tony Abbott’s visits which is statistically significant at the 10 percent level. In 2022 there is no clear pattern: Morrison was slightly less likely to visit Labor-held seats, contradicting the hypothesis, while Albanese was more likely to visit Coalition seats, supporting the hypothesis. However, neither estimate is statistically significant. We therefore reject H3.

Table 4. Incumbency and leader visits, 2013 and 2022.

	Percent of seats visited
	Coalition	Labor	(Significance)
2013
Rudd	21	36	(p = .045)
Abbott	25	39	(p = .069)
2022
Morrison	43	47	(p = .623)
Albanese	42	36	(p = .463)

Percent of leader visits to each electorate by party holding the electorate (150 in 2013, 151 in 2022).

It remains to test whether a leader visit improves the vote for the leader’s party. Table 5 answers this question by regressing whether or not an electorate received a visit from either a Liberal or Labor leader in 2013 and 2022 on the vote swing. In 2013 a visit by either leader significantly influenced the party vote. In the case of Tony Abbott, a visit resulted in a swing to the Coalition of 1 percent of the two party vote, net of other things in the model. A visit from Kevin Rudd actually increased the vote for the Coalition by 0.788 percent. In other words, each visit by a party leader was worth around 1 percentage point of the two party vote, and the difference between the two party leaders are minimal. However, paradoxically a visit by Kevin Rudd actually aided the Liberal vote. We discuss possible explanations for why this was the case below.

Table 5. Leader visits and the vote swing.

	2013		2022
	Est (SE)	Est (SE)	Est (SE)	Est (SE)
Visits by Liberal leader	1.000**	–	-.043	–
	(.451)		(.226)
Visits by Labor leader	–	.788**	–	-.309
		(.389)		(.219)
Marginality	.007	.002	.074	.063
	(.046)	(.046)	(.056)	(.052)
Media	−2.215	−2.479	2.968*	2.903*
	(1.749)	(1.765)	(1.677)	(1.663)
Urbanization (inner metro)
Outer metro	.788	.631	1.176	1.056
	(.823)	(.816)	(.970)	(.941)
Provincial	.592	.212	4.025**	3.910**
	(1.209)	(1.215)	(1.339)	(1.331)
Rural	1.452	.862	3.162**	3.021**
	(1.320)	(1.292)	(1.443)	(1.428)
Gender	.668	.702	.091	.095
	(.431)	(.433)	(.059)	(.058)
Language other than English	.071	.004	.089**	.087**
	(.056)	(.024)	(.027)	(.027)
Educational qualifications	.037	.035	-.213**	-.208**
	(.052)	(.052)	(.059)	(.058)
Home owner	-.006	-.003	-.037	-.030
	(.048)	(.048)	(.037)	(.037)
Constant	−27.383	−28.046	−7.416	−7.704
	(19.119)	(19.190)	(5.996)	(5.893)
Adj R-sq	.073	.067	.284	.294
(N)	(150)	(150)	(151)	(151)

** p<.05, * p<.10. Figures are partial regression coefficients (Est) and standard errors (in parentheses) predicting the percent swing to the Coalition in an electorate.

The comparable estimates for the 2022 election show a different pattern. A visit by Scott Morrison had no statistically significant effect on the Coalition vote, net of other things. Similarly, there is also no statistically significant effect for a visit by Anthony Albanese. By any standards, visits were much less important to voters in 2022 than in 2013.

The difference between the two elections may be the result of two changes in the election context. First, as we saw in Table 2, the scope and intensity of visits increased between 2013 and 2022. As visits have become less of a novelty for voters, they may have become normalized, so voters were less likely to be influenced by the presence of a party leader in their electorate. This may well explain the significant reduction in the impact of a visit by both leaders between the two elections.

The second change is the relative popularity of the leaders. In the 2013 election both Abbott and Rudd were relatively unpopular; in Rudd’s case, his ousting by Julia Gillard in 2010 left him bitter and he continually undermined Gillard during her period as prime minister. This resulted in a collapse in Rudd’s ratings among voters (Manning and Phiddian 2015).¹¹ For example, in 2007, the Australian Election Study found that 72 percent considered Rudd ‘honest’, but by 2013 only 45 percent held the same view (Cameron and McAllister 2022, 93–94). In the 2022 election Scott Morrison was also a highly unpopular figure due to a series of missteps during the 2019–2020 bushfires and the 2022–2022 pandemic (Taflaga 2023). Indeed, the Australian Election Study found Morrison was the most unpopular major party leader in the 35 year history of the survey (McAllister 2023). It is perhaps not surprising, then, voters did not respond favourably to a visit by him.

Conclusion

Campaign visits by the major party leaders to individual electorates are an important part of the theatre of Australian elections. For leaders and their strategists, it is a chance to show the leader is engaged and popular with ordinary people. A visit, properly managed, can generate positive advertorials on the nightly television news, in the newspapers and increasingly in the social media. The visual images of a visit, in a properly managed media environment, can help to ‘frame’ a policy announcement and reinforce its message. However, such visits also involve considerable strategic planning and soak up scarce resources. Using a unique dataset covering two national elections, we have addressed two questions about leader visits that have not hitherto been addressed in the Australian context: what determines the choice of an electorate for a visit?; and, what is the electoral impact of a visit?

Confirming the international literature, we find the marginality of the seat is the main consideration in deciding on a visit. Seats that require relatively small vote swings to change hands are more attractive to strategists than seats that require large swings. A secondary consideration is location, and rural seats and those in outer metropolitan areas are generally less likely to be selected for a visit compared to inner city seats. Such considerations obviously reflect the logistics of visits and the need to showcase the leader to as large a group of voters as possible, and to garner positive media coverage. There is also some evidence visits are directed to government-held seats. However, we find no evidence that local media consumption is a consideration in the choice of electorate.

The second question is the effect of leader visits on the vote. We find that a visit from a leader has a modest effect on the vote. In 2013, a visit from Tony Abbott increased the Liberal vote by around 1 percentage point, but paradoxically, a visit by Kevin Rudd also increased the Liberal vote. This apparently contradictory result may be explained by Rudd’s relative lack of popularity and the image of Labor disunity, so that voters reacted negatively to a local visit by him. This is in line with the international research which suggests that a visit by an unpopular leader can negatively impact the vote for that leader’s party.

In 2022, there are no significant electoral consequences of a visit by either Scott Morrison or Anthony Albanese. The absence of any effect for Morrison undoubtedly reflects his lack of popularity among voters and the existence of well-publicized events where some voters made overt displays of dislike for him. The lack of any impact on the vote from an Albanese visit from may well be a consequence of his attenuated campaigning due to his one week COVID isolation. Another explanation may be his relatively recent elevation to the Labor leadership, so voters had not formed a firm opinion of him.

In summary, leader visits can matter to electoral outcomes, but the effect is likely to be small or even non-existent. What seems to shape the effect on the vote is the popularity and personality of the leader in question, with a visit by an unpopular leader either having no effect on the vote or even a negative one. Showcasing a relatively popular leader can increase the party vote by around 1 percentage point, which is broadly in line with international research. This is a small electoral effect, but could result in a handful seats changing hands if properly managed and targeted. In a narrowly fought election, this could mean the difference between winning and losing.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Brent Davis

Dr Bent Davis is a Visiting Fellow in Political Science at The Australian National University, Canberra.

Ian McAllister

Ian McAllister is Distinguished Professor of Political Science at The Australian National University, Canberra. Our thanks to two anonymous referees from this journal for their constructive and helpful suggestions on an earlier draft of the paper.

Notes

1 This has generated a large literature in the US: see Charnock, McCann, and Tenpas (2009); Cohen, Krassa, and Hamman (1991); Doherty (2007); Eshbaugh-Soha and Nicholson-Crotty (2009); Herrnson and Morris (2007); Keele, Fogarty, and Stimson (2004); Lang, Rottinghaus, and Peters (2011); Mellen and Searles (2013); and Sellers and Denton (2006). For Britain, see Middleton (2015). For an alternate view, see McClurg and Bryan (2009).

2 Equally, a visit can generate financial donors for the local party as well as help to motivate grass-roots supporters (Cohen, Krassa, and Hamman 1991; Herr 2002; Herrnson and Morris 2007; Sellers and Denton 2006).

3 In the 2022 election, 51.7 percent of all votes were cast early, an increase of 10.5 percent on the 2019 election (Australian Electoral Commission 2022).

4 We measure the importance of the media through voters’ consumption of election news, using the AES surveys. Media markets are also an important potential factor determining the choice of an electorate for a visit. However, we were unable to investigate this due to the difficulty in matching overtime media market data to federal electorates.

5 Hereinafter referred to as the Coalition. We do not cover visits by the National Party leader since these are difficult to track and often do not generate much publicity.

6 The questions were: ‘How much attention did you pay to reports about the election campaign in the newspapers? … on television? … on the radio? … on the internet?’ The response categories are: a good deal, some, not much, none at all.

7 Preliminary analyses also included the proportion who were born in a non-English speaking country, but language proved to be a better predictor and is included here.

8 Preliminary analyses used a wider range of control variables, including age and birthplace. These had no substantive effect on the results and to reduce the number of variables in the model (bearing in mind the small number of cases and the standard errors) they were excluded from the final models.

9 Albanese was unable to undertake visits from 21 to 28 April, which was in the early stages of the election campaign.

10 Smith (2020: Table 11.1) notes that in 2019 the marginal NSW seat of Reid received five visits from each leader, and the marginal Tasmanian seat of Braddon also received considerable attention from the two leaders (Economou et al. 2020, 240).

11 When Labor was elected in 2007 the Australian Election Study found that Rudd was the most popular leader since Bob Hawke in 1987 (Cameron and McAllister 2022, 88).