Content uploaded by Luke Mansillo
Author content
All content in this area was uploaded by Luke Mansillo on Sep 19, 2023
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=frfs20
Regional & Federal Studies
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/frfs20
Australian state and territory elections: regional
incumbents matter
Rodney Smith, Luke Mansillo & A. J. Brown
To cite this article: Rodney Smith, Luke Mansillo & A. J. Brown (2023) Australian state and
territory elections: regional incumbents matter, Regional & Federal Studies, 33:4, 421-439, DOI:
10.1080/13597566.2023.2240234
To link to this article: https://doi.org/10.1080/13597566.2023.2240234
View supplementary material
Published online: 01 Aug 2023.
Submit your article to this journal
Article views: 54
View related articles
View Crossmark data
Citing articles: 1 View citing articles
Australian state and territory elections: regional
incumbents matter
Rodney Smith
a
, Luke Mansillo
a
and A. J. Brown
b
a
Discipline of Government and International Relations, The University of Sydney, Sydney,
Australia;
b
School of Government & International Relations, Griffith University, Brisbane,
Australia
ABSTRACT
Recent research is ambiguous about the status of Australian regional elections,
seeing them as conforming to the second-order election model but also as
affected by regional politics. We clarify this ambiguity, drawing on aggregate
and individual level data to explore the variable impact of national and
regional incumbents on regional elections. Although national incumbents
seem to affect Australian regional elections, under some circumstances
regional incumbent parties are able to electorally outperform their national
incumbent counterparts. We suggest that Australia’s uncoordinated national
and regional election cycles and federal distribution of policy responsibilities
both help to focus voter attention on the performance of regional
incumbents. The way that regional incumbents manage key policy issues,
including Covid-19 in recent years, appears to matter for their electoral
support, making Australian regional elections more than second-order events.
ARTICLE HISTORY Received 4 July 2022; Accepted 20 July 2023
KEYWORDS Incumbency; second-order election effects; Australia; federalism; election cycles
Introduction
After the Liberal Government in South Australia lost the March 2022 state elec-
tion, the result was interpreted in two contradictory ways. The Liberal Prime
Minister Scott Morrison denied that his national government had anything
to do with his state counterpart’sdefeat:‘This election was being fought on
state issues. The federal election will be fought on federal issues’(Morrison,
quoted in Butler 2022). For Morrison, Australian state and national elections
were separate events. The Labor Party, by contrast, was keen to portray its
victory in South Australia as a judgement on the Prime Minister and his national
government: ‘Scott Morrison was a drag on the Liberal vote here’,claimedthe
South Australian Labor Senator, Penny Wong (quoted in Butler 2022).
1
© 2023 Informa UK Limited, trading as Taylor & Francis Group
CONTACT Rodney Smith rodney.smith@sydney.edu.au
Supplemental data for this article can be accessed online at https://doi.org/10.1080/13597566.2023.
2240234.
REGIONAL AND FEDERAL STUDIES
2023, VOL. 33, NO. 4, 421–439
https://doi.org/10.1080/13597566.2023.2240234
These conflicting interpretations of state and territory elections are
common not only among Australian political leaders, who deploy them for
strategic reasons, but also among political scientists. In their recent analysis
of Australian regional elections in this journal, Dain Bolwell and Richard
Eccleston (2018) exemplify this contradiction. Their initial aggregate quanti-
tative analysis emphasizes a ‘second-order’explanation of Australian regional
elections, in which the fortunes of state and territory parties are largely deter-
mined by which party governs nationally. When the Liberal National Party
Coalition (LNP) is in office at national level, its regional counterparts lose
votes at subnational elections. When the Labor Party is in power nationally,
its state and territory branches will suffer at subnational elections (2018,
263–264).
2
In the second half of their article, by contrast, they undertake a
qualitative analysis of specific recent state and territory elections in which
the performance of regional incumbents and oppositions is seen as critical
to subnational election outcomes (2018, 264–270). The result of their
approach is ambiguity about whether and why regional government and
politics matter for regional elections in Australia.
Bolwell and Eccleston’s contribution is a welcome one, since academic
attention to Australian elections is overwhelmingly focused on the national
level (see McAllister 2011). A major edited collection is produced on each
national election (see, for example, Gauja, Sawer, and Simms 2020) and
numerous articles appear on aspects of nation-wide party campaigns and
electoral behaviour. This national focus does include some research on the
impact of different regions on national election outcomes (see, for
example, Economou et al. 2020; Martinez i Coma and Smith 2018);
however, research on voting in subnational elections in the states and terri-
tories remains under-developed. Individual subnational elections in particular
jurisdictions do sometimes attract analysis. More comprehensive research
that attempts to identify patterns across all Australian subnational elections
is rare. To our knowledge, the most recent systematic research on Australian
subnational elections prior to Bolwell and Eccleston’s was a collection edited
by Jeremy Moon and Campbell Sharman 15 years earlier (Moon and Sharman
2003).
This paper builds on Bolwell and Eccleston’s analysis. Like them, we focus
on the two main competing explanations for regional election results found
in the comparative literature: that regional elections are second-order events
in which votes are used ‘to express satisfaction or disappointment toward
national parties’or, alternatively, that regional government and politics do
matter to regional election outcomes (Schakel and Dandoy 2013a,6–10).
We improve and extend Bolwell and Eccleston’s analysis in three ways.
First, we provide a more systematic discussion of the three factors that the
comparative literature suggests should move Australian subnational elec-
tions closer to or further away from being second-order phenomena:
422 R. SMITH ET AL.
territorial cleavages, regional authority and electoral timing (Schakel and
Dandoy 2013a). First, Australia lacks territorial socio-cultural cleavages.
Second, it has asynchronous national and regional election cycles. Third,
while Australia has a relatively centralized federal system, the states and ter-
ritories retain significant and visible policy responsibilities. The first factor
suggests that Australia is not a good candidate for findings that regional
incumbency matters for regional vote choice, while the second and third
factors suggest that regional incumbency does matter. We do find regional
incumbency effects, which point to the need for a more complex understand-
ing of Australian federalism and regional voting.
Our second contribution is to deepen Bolwell and Eccleston’s conclusion
that Australia has a national party system by exploring minor parties in
more detail. They lump all minor party and Independent candidates together
into a single group. We distinguish between the two most significant left- and
right-wing minor party challengers since the mid-1990s to see whether these
new party developments have disrupted Australia’s nation-wide party com-
petition. We conclude that the Australian party system remains essentially
national.
Third, we use aggregate data to re-analyse the effects of national incum-
bency on regional elections, introducing regional incumbency as a poten-
tially important influence on regional voting. Rather than assuming that a
party’s incumbency at national level will inevitably lead to a loss of votes
for that party at regional level elections (Bolwell and Eccleston 2018, 263–
264), we argue that regional incumbency affects whether or not such a loss
occurs. We supplement this finding with survey data collected during the
Covid-19 crisis, which indicates that voter perceptions of regional govern-
ment policy performance affect regional voting.
The absence of territorial cleavages and “non-state wide”
parties
Although some scholars have attempted to make a case for territorial diver-
sity in Australia (see Aroney, Prasser, and Taylor 2012), the consensus view is
that none of the Australian states or territories has sufficiently distinctive
socio-economic, linguistic or cultural features for territorial parties to mobilize
effectively. This was true when the Australian party system formed from the
1890s to the 1920s and it remains the case (Aitkin 1982; Fenna 2019; Kemp
1978; Smith 2001). As Alan Fenna writes: ‘From the outset, Australian politics
has been dominated by nation-wide political parties, which from 1909 were
clearly divided on a Left–Right axis. There have been no regional parties to
inject a centrifugal dynamic into Australian federalism’(Fenna 2019, 45).
The Australian party system and electoral politics at national and regional
level have been dominated by a competition between the centre-left
REGIONAL & FEDERAL STUDIES 423
Australian Labor Party and the centre-right Liberal National Party Coalition.
Since the mid-1990s, however, minor parties have challenged these estab-
lished major parties and the vote share of the major parties has declined
(Moon and Sharman 2003; Marsh 2006). Bolwell and Eccleston trace these
developments to some extent, but they treat all minor parties, as well as Inde-
pendent candidates, as part of a single non-major party grouping (2018, 263).
Most of the minor parties and Independents have had very little impact;
however, two types of party within this grouping –the Greens, and a
family of right-wing parties, including Pauline Hanson’s One Nation Party
(ONP) –have achieved significant support. Given the success of these two
very different challenger parties –the Greens to Labor’s left (Crowe 2018;
Jackson 2016) and One Nation and others outflanking the LNP to the right
(Leach, Stokes, and Ward 2000)–it makes sense to separate them from the
amorphous ‘minor’grouping in order to explore whether their electoral
achievements have been nationwide or regionally specific.
There are reasons to expect both these types of challenger party to have
had variable regional strength. In the case of the Greens, from the 1980s a
number of different ‘green’parties formed independently to compete in
regional elections. They gradually merged but only became a truly nation-
wide party organization in 2003, when the Western Australian Greens for-
mally joined the ‘Australian Greens’. The Greens have remained a
decentralized party with a highly localized and federalized constitutional
structure (Jackson 2016, 15, 49–71; Miragliotta and Sharman 2012). As
Jackson (2016, 20) notes, ‘The party culture of the Australian Greens is …
marked by state and local differences’, including different capacities for elec-
toral success at the regional level.
The right-wing family of parties is regionalized in a different way. It gained
prominence following the shock success of ONP at the 1998 Queensland state
election (Leach, Stokes, and Ward 2000). Since then, the family has included
the United Australia Party (formerly the Palmer United Party) and Katter’s Aus-
tralian Party, and a range of less visible parties, including the short-lived
Fraser Anning’s National Conservative Party, the Yellow Vests Australia (for-
merly the Australian Liberty Alliance), Rise Up Australia, the Australia First
Party and Australians Against Further Immigration. Power in these parties
centres on their leaders, on whom the parties rely for public recognition
(Economou and Gazharian 2018;Kefford and McDonnell 2018). It is notable
that the leaders of the three highest profile parties –Pauline Hanson, Bob
Katter and Clive Palmer –are all based in Queensland.
To test whether the emergence of the Greens and the right-wing parties
has in fact decreased the congruence of election results between national
and regional levels in Australia, we follow Bolwell and Eccleston’s lead and
use one of Schakel’s(2011) measures of national-regional congruence of
the vote: a region’s pattern of voting support for different parties at a national
424 R. SMITH ET AL.
election compared with the same region’s voting pattern at the following
regional election. We use 1997 as a starting point, since this was the time
at which both the Greens and right-wing challenger parties began to gain
electoral traction. Our data set includes 55 state and territory elections
from January 1997 to the May 2022 national election.
Table 1 shows the average first preference votes for the Greens and right-
wing parties at regional lower house elections, comparing these with their
average first preference votes at the previous national lower house elections
in the same region.
3
It also compares the average incongruence scores for
each state and territory when all the non-major parties and Independents
are grouped together and when the Greens and right wing parties are separ-
ated out, to see whether separating out these parties lowers the congruence
scores.
The results in Table 1 confirm the variable strength of the Greens across
regions. The Greens strongest regional election results have occurred in Tas-
mania. The Tasmanian Greens regional vote is also distinctive in that it is on
average 5.27% higher than the Green vote across Tasmania in national elec-
tions. Different electoral systems at regional and national levels may play a
role here. The single transferrable vote (STV) variant of proportional represen-
tation used in Tasmanian state lower house elections reduces the threshold
for electoral success compared with the single member electorates used
for national lower house elections (Smith and O’Mahony 2006). For Tasma-
nians, this may make voting Greens more attractive at regional level, since
their vote is more likely to help elect a Greens MP than it is at the national
level.
4
If that is the case, then it is strange that the pattern does not also
occur in the Australian Capital Territory, the only other Australian region to
use an STV system for lower house elections. The average regional Green
vote in the Australian Capital Territory is 1.47% lower than the Australian
Capital Territory Green vote at national elections. Party traditions and trajec-
tories might be more important than electoral systems in explaining Tasma-
nia’s distinctive regional Greens vote. The campaigners who formed the
Tasmanian Greens initially focused on building electoral support around
local issues and targeted seats in the regional parliament, achieving early suc-
cesses. The ACT Greens, by contrast, gave more equal emphasis from the
outset on regional and national electoral contests (Jackson 2016,52–54,
67–68).
Table 1 also shows that the right-wing challenger parties’success in
regional elections has been considerably stronger in Queensland (9.2%)
than elsewhere. This is to be expected, given the origins of the three main
right-wing parties in that state and the continued dominance of Queensland
leaders within the parties. Queensland is also the only state or territory where
the right-wing parties register a higher vote in regional elections than in
national elections; however, the difference is small (2.05%).
REGIONAL & FEDERAL STUDIES 425
Table 1. Green and Right-Wing Challenger Party Vote Shares with Two Measures of Party National-Regional Incongruence (1997–2022).
State or
territory
State or
territory
elections
Average state
or territory
Green vote (%)
Average difference
from previous national
Green vote in state or
territory(%)
Average state
or territory
right-wing vote
(%)
Average difference from
previous national right-
wing vote in state or
territory(%)
Incongruence 1: ALP v
LNP v all non-major
candidates (%)
Incongruence 2: ALP v LNP v
Greens v right-wing v all
other non-major candidates
(%)
New South
Wales
1999–2019 8.54 1.08 2.22 −1.41 6.91 8.43
Victoria 1999–2018 9.05 0.38 0.16 −1.68 6.50 8.05
Queensland 1998–2020 5.92 0.90 9.24 2.05 11.61 15.73
Western
Australia
2001–2021 8.50 −1.25 2.96 −0.49 10.64 11.88
South
Australia
1997–2022 6.15 −0.23 0.62 −1.25 6.05 7.61
Tasmania 1998–2021 14.72 5.27 0.62 −2.14 11.13 13.62
Australian
Capital
Territory
1998–2020 10.69 −1.47 0.00 −1.11 9.10 12.10
Northern
Territory
1997–2020 2.54 −5.17 0.22 −2.46 6.94 12.78
Sources: Calculated from official returns of the Australian Electoral Commission, state and territory electoral commissions.
426 R. SMITH ET AL.
Despite these regional differences, separating the Greens and right-wing
parties vote shares from the rest of the non-major candidates produces
only small increases in the overall incongruence of election results at national
and regional levels for any state or territory (see Table 1). The largest increase
in incongruence is found in the score for the Northern Territory, where
average incongruence increases 5.84%, from 6.94 to 12.78. The Australian
party system has become more complex in recent decades, as Labor and
the LNP have both lost electoral support; however, the system has retained
a relatively high level of congruity between national and regional electoral
contests.
National and regional electoral timing
The absence of territorial diversity and of non-state-wide parties does not
exhaust the possible sources of regional electoral diversity. Australia’s pat-
terns of national and regional election timing provide one basis for expectat-
ing regional elections to be more than second order events. The comparative
literature argues that regional elections will be subordinate to national ones
under two conditions. The first is vertical simultaneity, when national and
regional elections are held simultaneously, in which case a single national-
oriented logic of voting will prevail, with voters casting votes for the same
party nationally and regionally. The second is horizontal simultaneity,
where all regional elections are held simultaneously but on a date that is
different to that of the national poll. Here the literature expects stronger
second order effects, as some voters punish a national incumbent party by
voting against its regional counterpart (Fabre 2010;Jeffery and Hough
2006, 249–251; Schakel and Dandoy 2013b, 284–287).
Australia’s national and regional electoral laws, some of which are consti-
tutionally entrenched, make vertical and horizontal simultaneity almost
impossible. Under the Australian Constitution, national elections in Australia
must occur every three years, although the Prime Minister may request and
be granted an earlier election from the Governor General. The national
polling day could potentially clash with one of the regional polling days;
however, national elections would take precedence in any potential clash
of dates (Fabre 2010; Orr 2019,85–87). The normal three-year electoral
cycle at national level means that no region is likely to hold its election in
the same year as a national poll more than once a decade. Tasmania
excepted, the states and territories now have fixed-term elections every
four years. The fixed years and dates vary from jurisdiction to jurisdiction.
Only Queensland and the Australian Capital Territory hold elections at
roughly the same time as each other. There are provisions for earlier state
or territory elections if necessary –for example, if no government can main-
tain the confidence of the parliament –but these have never been invoked.
REGIONAL & FEDERAL STUDIES 427
These independent electoral cycles further lessen the likelihood that Aus-
tralian regional polls will be second order events held in the shadow of
nationwide electoral contests (Fabre 2010; Schakel and Dandoy 2013a, 13).
Instead, as the final sections of Bolwell and Ecclestone’s article suggest
(2018, 264–270), Australian regional elections tend to be genuine contests
in their own right. They focus mostly on regional policies and politics, and
garner limited media coverage elsewhere in the country.
Regional autonomy
Australia’s federal constitution provides the strongest basis for expecting
state and territory elections to have their own regional dynamics, since it
gives the state governments exclusive or shared responsibilities for key
public policy areas. The powers of territory governments are more formally
constrained than those of the states; however, in practice they exercise
similar levels of policy autonomy. The history of Australian federalism is
often represented as one of increased centralization. Fenna (2019, 48), for
example, writes:
Australian federalism was decentralized in design, as evidenced by the high
number of policy fields left wholly, or almost wholly, in the exclusive jurisdiction
of the States. However, it was transformed in the century following Federation
by a persistent and pervasive expansion in the role of the Commonwealth. The
pattern has been irregular, and varied in tempo across policy fields, but centra-
lization has been substantial in many policy fields.
Despite this history of centralization since 1901, Australia’s regional govern-
ments retain considerable powers in policy areas that are highly visible to
voters.
One indication of these ongoing regional powers is Australia’s compara-
tively high score on the Hooghe, Marks, and Shakel (2010) Regional Authority
Index, which indicates little change in the autonomy of Australia’s states and
territories across four periods from 1950 to 2006. The Regional Authority
Index has been criticized for an over-reliance on institutional factors, too
little attention to policy (Dardinelli et al. 2019a: 5) and inadequate measures
of fiscal autonomy (Fenna 2019, 38). According to Dardinelli et al’s alternative
policy-oriented seven point ‘de/centralisation index’(2019a;2019b)–in
which a score of 1 means completely centralized, 4 evenly shared and 7 com-
pletely decentralized policy responsibilities –Australia has seen the most
rapid centralization out of the six federations they studied (Dardanelli et al.
2019b, 199–201).
5
Nonetheless, in 2010, Australia scored 4 or higher for
both legislative powers and policy delivery in the key policy fields of school
education, policing, the environment, civil law, criminal law, natural resources
and elections. In addition, it scored 4 or above for policy delivery in health
and agriculture (Fenna 2019,33–34). This suggests that the states and
428 R. SMITH ET AL.
territories retain significant policy responsibilities, even where these are now
shared with the national government. As Fenna notes: ‘Whether it be in edu-
cation, housing, health care, environmental protection, infrastructure or a
range of other areas of governance that were originally state matters, the
two levels of government are now inextricably intertwined’(2021, 200).
From a voter’s perspective, the performance of an Australian regional gov-
ernment may remain highly visible even if that government’s policy respon-
sibilities are now shared with the national government. Debates over school
education, health and policing remain staples of Australian state and territory
election contests, while balancing natural resource usage and environmental
protection has become a prominent election issue in recent decades. The
states and territories may be increasingly responsible for administering
rather than legislating in these areas; however, implementation and adminis-
tering of policy often counts far more for voters than the source of legislation.
As Dardinelli et al (2019: 9) concede,
Although [administrative] autonomy is less consequential than legislative
autonomy –which is one of the defining features of federal systems –it is
still significant because it can result in de facto policy-making that adapts
central policy objectives to local preferences or otherwise alters, frustrates, or
enhances the central government’s objectives.
It is also worth noting that policy areas that have been highly contentious in
recent Australian regional elections, such as urban planning and infrastruc-
ture development, are missing from Dardinelli et al’s de/centralization index.
In addition to these ongoing policy challenges, the Covid 19 crisis lifted the
profile of Australia’s regional governments. The impact of Australia’s federal
structure on the management of Covid-19 has been discussed extensively
elsewhere, often in a comparative context (see, for example, Boin, McConnell,
and ‘t Hart 2021; Bromfield and McConnell 2021; Murphy and Arban 2021;
Prasser 2020; Schnabel and Hegel 2021). A National Cabinet that included
the state and territory leaders was created to lead and coordinate the pan-
demic response. The Commonwealth Biosecurity Act 2015 governed the
national response but the states and territories also exercised considerable
highly visible powers, particularly through their public health acts. Regional
governments controlled access to schools, workplaces, shopping and
leisure activities, movement within and between regions, lockdowns and
curfews, social distancing rules, contact tracing, policing, hospitals and the
local distribution of vaccines and other medical supplies. Quarantine, a con-
current power under the Constitution that the Commonwealth could have
managed, was handed to the states. The state and territory governments
often took different approaches to these issues. Initial cooperation turned
to disagreement and conflict between regions and between the regions
and national government. This conflict included a constitutional court case
REGIONAL & FEDERAL STUDIES 429
against the Western Australian government that the Commonwealth briefly
supported. The usual media focus on the Prime Minister was matched if
not exceeded by a focus on the regional premiers, health ministers and
chief medical officers, who hosted daily media briefings watched by large
audiences (Bromfield and McConnell 2021; Murphy and Arban 2021;
Prasser 2020; Twomey 2021).
The underlying pattern of responsibilities within the Australian federation
and the more recent impact of Covid 19 in drawing attention to the policy
performance of state and territory governments suggest that Australian
regional elections will reflect much more than the second order conse-
quences of national incumbency. Regional incumbents will also matter.
Patterns of national and regional incumbency
Despite the ‘very high level’of party congruence between Australian national
and regional elections found by Bolwell and Eccleston (2018, 262) and
repeated in Table 1, the parties that have held government at national and
regional levels in Australia have often been incongruent. The pattern in
Figure 1, which covers 1997–2022, is typical of the long periods of incum-
bency at national and regional levels that have characterized Australian poli-
tics since federation in 1901 (Moon and Sharman 2003). The LNP dominated
national government over the 26 years under review, governing from March
1996 until November 2007 and then from September 2013 to May 2022. At
regional level, Labor enjoyed longer incumbencies than the LNP in all
states and territories except Western Australia. In almost half (44.6%) of the
regional elections held in this period, the incumbent party of regional govern-
ment differed from the party governing nationally. Periods of strong congru-
ence of incumbency over this period have been brief. For the first nine
months of Labor national government in 2007–2008, the party also governed
in all eight states and territories. The peak of LNP congruence was both
weaker and shorter, just two months between 31 September and 3 December
2014, when six of the eight regional governments were LNP. Periods of
Figure 1. Congruence of incumbents at national and regional levels, 1997–2022. Key,
Red: Labor, Blue: Liberal National Party Coalition and equivalents. Source: University
of Western Australia (n.d.) and official returns of electoral commissions.
430 R. SMITH ET AL.
maximum incongruence have been both stronger and longer. After Labor
won the South Australian election in March 2002, for example, the national
LNP government faced eight Labor state and territory governments for the
next five years and nine months.
This pattern of high levels of party congruence at national and regional
elections and much lower levels of incumbency congruence deserves
further exploration. Does it reflect a tidal pattern, to use Bolwell and Eccles-
ton’s(2018) metaphor, in which a party’s success nationally produces an ‘ebb
tide’for the same party at regional level, leading it to eventual but inevitable
loss at regional elections? Or, as Figure 1 suggests, are some regional parties
able to swim against the tide, losing votes when the national incumbent
should give them an advantage (e.g. Labor in Tasmania in 2014 and South
Australia in 2018) and gaining votes when the national incumbent means
they should lose them (e.g. the LNP in New South Wales in 2007 and Labor
in the ACT in 2012)? To what extent does the performance of regional incum-
bents matter?
Exploring regional incumbency effects
To answer the questions posed above, we need to explore the possible
effects of regional incumbency in a systematic way. As noted earlier,
Bolwell and Eccleston (2018, 264–270) do not account for regional incum-
bency in their general explanation of regional elections and so risk unob-
served heterogeneity affecting their results. In addition, they employ a
naïve measure for the effect of national incumbency by simply recording
which party was in power at the national level. This induces specification
error. They do not compare the direction or magnitude of the swings experi-
enced by parties at national level with those experienced by their regional
counterparts (Bolwell and Eccleston 2018, 260–263). This approach gives an
incomplete picture of regional level versus national level party fortunes.
Bolwell and Eccleston would consider a party that suffered a 2.0% negative
swing in a regional election as evidence of second order election effects if
that party’s national counterpart was the incumbent, even if that counterpart
had endured a worse negative swing of 5.0% at the previous national
election.
Our approach to the data addresses these issues by comparing the primary
votes achieved by the two major parties in regional elections with the primary
votes those parties recorded in the same regions at the previous national
election. In Schakel and Dandoy’s(2013a, 20) terms, this is an NR-RR test of
election congruence. We analyse votes at 54 regional elections between
1997 and 2022 under three different incumbency conditions. The first con-
dition (Coalition-Coalition) includes elections where the LNP was the incum-
bent at both national and regional level (20 elections). The second condition
REGIONAL & FEDERAL STUDIES 431
(Coalition-Labor) includes elections where the LNP was the national incum-
bent but Labor governed regionally (23 elections). In the third condition
(Labor-Labor), Labor was incumbent at national and regional level at the
regional election (11 elections). The fourth possible condition (Labor as the
national incumbent and the Coalition in power regionally) occurred only
once and therefore cannot be analysed.
6
If the second-order election model applies to Australian regional elections,
we should observe similar slopes across the different conditions, because
national incumbency is what counts for regional elections. The slopes
should indicate lower support for the national incumbent party at regional
elections relative to the votes recorded by that party in the same regions
at the previous national election. Similarly, to the extent that the party in
opposition at the national level benefits from these ‘second-order’effects
at regional elections, the slopes for that party’s votes in regional elections
should remain above the votes that the party recorded in the same region
at the previous national election.
The results in Figure 2 do not support these predictions. When the
Coalition is in power nationally, regional Coalition incumbents are not
Figure 2. Predicted Regional Vote Shares under Different Incumbency Conditions,
1997–2022. Note: Estimates produced from a hierarchical model with random slopes
and intercepts. The orange lines are 45-degree lines; all data points would lie on
these lines if national vote shares perfectly predicted regional vote shares. The black
lines are regression lines summarising the estimated national-regional vote share
relationship. The blue and red lines are the simulations of this relationship generated
by the model.
432 R. SMITH ET AL.
consistently at an electoral disadvantage. In eight of 20 regional elections held
under this incumbency condition, Coalition regional incumbents achieved a
higher vote than the Coalition had at the previous national poll. The
Coalition-Coalition results in Figure 2 show that where Coalition incumbents
do well at national elections, gaining over 40% of the vote, regional Coalition
incumbents cannot keep pace with them. By the same token, as the Coalition
incumbent national vote falls below 40%, regional Coalition incumbents do
not suffer correspondingly but instead do increasingly better than their
national counterparts. Similarly, votes above 37% for the opposition ALP at
the national level are not matched by Labor oppositions at the regional level.
The Coalition-Labor and Labor-Labor incumbency conditions better
conform to ‘second-order’expectations. The estimated regional vote slopes
for Coalition regional oppositions when the Coalition governs nationally
and for Labor regional incumbents when Labor is in power nationally are
very similar. Both slopes show an increasing regional election disadvantage
to parties in these situations as national-level votes climb above 35%.
When the national Labor vote in a region reaches 50%, for example, the esti-
mated disadvantage for its regional counterpart is 11 percentage points,
while for the Coalition the difference is 13 percentage points. The advantages
to Labor regional incumbents when the Coalition is in power nationally and
to Coalition regional oppositions when Labor governs nationally are not as
pronounced. Nonetheless, the estimated regional Labor vote under the
Coalition-Labor incumbency condition is higher than Labor’s national-level
vote until the latter reaches 45%, while the estimated regional Coalition
vote under the Labor-Labor incumbency condition is higher than the
Coalition’s vote at the previous national election until the national-level
Coalition vote reaches 47%.
Discussion
The results in Figure 2 suggest that different combinations of national and
regional incumbency have varied effects on regional election results. While
the Coalition-Labor and Labor-Labor conditions roughly match the ‘second
order’model of regional elections, the Coalition-Coalition condition does
not. The distinctiveness of the Coalition-Coalition outcome may arise
because the experience of national Coalition government is so common –
it occurred in 19 of the 26 years under review –that regional Coalition gov-
ernments have more time to develop experience in successfully managing
relations with their national counterparts. The relatively short period of
Labor national government from 2007 to 2013 was also marked by
ongoing leadership instability and policy setbacks (Aulich 2014; Aulich and
Evans 2010), which probably made it more difficult for regional Labor govern-
ments to resist electoral spill-over from the national level.
REGIONAL & FEDERAL STUDIES 433
Nonetheless, under all three conditions, at least some regional incumbents
managed to achieve more favourable electoral outcomes than their national
counterparts. The causes of these varied regional level outcomes deserve
more systematic attention. A close reading of the qualititative section of
Bolwell and Eccleston’s article that discusses recent regional elections
suggest a number of factors are important to the outcomes in one or more
cases. These factors include the regional premier’s popularity, regional leader-
ship instability, the general performance of regional parties, the longevity of
the regional government, regional policy initiatives, and government man-
agement in areas such as transport, power supply and mining (2018, 264–
270). We agree with Bolwell and Eccleston that regional politics and policy
are electorally important. Moreover, the specific factors that they identify
are the regional counterparts of factors found to be important in survey
research on voting behaviour at Australian national elections (see, e.g.
Bean 2018; Ratcliff, Sheppard, and Pietsch 2020). Our criticism of Bolwell
and Eccleston is that these factors need to be explored systematically at
the regional level, rather than being deployed in an ad hoc way.
We do not have the data to comprehensively test these potential factors
here. As noted earlier in this article, systematic survey research on Australian
regional elections is virtually non-existent. We can, however, draw on evi-
dence from the 2021 Australian Constitutional Values Survey (ACVS) to test
the impact of three factors –national incumbency, regional government
policy performance regarding Covid 19, and the longevity of regional govern-
ments –on regional voting intentions. The ACVS was fielded in February
2021, around one year into the Covid 19 response by Australian national
and regional governments. The survey was delivered by telephone to a stra-
tified national adult sample, which included 1395 eligible voters.
Voters generally judged the varied national and regional pandemic
responses to be successful. The 2021 ACVS respondents were asked: ‘And
overall, how well would you say each level of government has handled the
COVID-19 pandemic? The Federal government’then ‘And overall, how well
would you say each level of government has handled the COVID-19 pan-
demic? The [territory/state] government’. Two-thirds of voters thought
both levels of government had handled Covid-19 well. Around a quarter of
voters distinguished between national and regional governments’handling
of Covid-19 (see online Appendix, Table 1).
Respondents were asked how they would vote if an election for their state
or territory lower house ‘was held today’, with a follow-up question asking
which way undecided voters were ‘leaning’. The results of both questions
were combined into a single binary dependent variable (intended vote for
the regional incumbent 1; intended vote for any other candidate 0). We test
the relationship between regional policy performance and intended regional
incumbent vote using logistic regression. The initial model included four
434 R. SMITH ET AL.
independent variables drawn from the earlier discussion in this paper. The first
is whether the regional incumbent is LNP (coded 1) or ALP (coded 0). The
expectation derived from second order regional election theory is that this vari-
able will correlate negatively with intended vote for the regional incumbent,
since LNP regional incumbents will be punished by the presence of the LNP
national government. The second variable is how well respondents believed
their state or territory government had handled the Covid-19 pandemic. We
expect a positive relationship between voters’assessments of regional govern-
ment’s handling of Covid-19 and intentions to vote for the incumbent. Item
responses were transformed into a series of dummy variables, with the refer-
ence category ‘not at all well’. The third variable is the longevity of the incum-
bent regional government, measured in months from when it was first elected
to February 2021, the month the ACVS was fielded. This variable was included
to test Bolwell and Eccleston’s(2018) suggestion that regional governments
become less popular over time. The final independent variable was judge-
ments about the national government’s handling of Covid-19, since voters’jud-
gements about the national LNP government’sperformancemayhavespilled
over into their regional voting intentions.
Our key finding is that the strongest influence on intention to vote for a
regional incumbent is the voter’s assessment of how well that incumbent
dealt with Covid-19. Citizens who thought their regional government had
handled Covid-19 very well were over 13 times more likely to vote for the
incumbent party than citizens who thought Covid had not at all been
handled well. On this result, policy performance appears to matter for
regional incumbents. By contrast, policy performance on Covid-19 by the
national LNP government was unrelated to intended regional vote. There is
evidence for a weak second-order regional election effect: intention to vote
for a regional incumbent reduced when that government was LNP. A regional
government’s longevity had no effect on the likelihood of voting for them
(see online Appendix, Table 2).
Model 2 included controls for socio-economic variables, none of which
had a significant relationship with intended vote for regional incumbent.
We also ran a third model that included an interaction item for ‘Regional
incumbent LNP’and ‘Regional Covid-19 handled very well’. The item was
not significantly correlated with intended regional vote and is not shown.
Voter assessments of how well their regional governments had handled
Covid-19 remained very strong predictors of intended regional vote in
Model 2 (see online Appendix, Table 2).
Conclusion
Writing in the European context, Shakel and Dandoy point out that the
‘degree to which regional elections may be considered second-order varies
REGIONAL & FEDERAL STUDIES 435
substantively’across countries (2013a, 8; see also Schakel and Dandoy 2013b,
297–298; Jeffery and Hough 2006). This paper presents one of very few sys-
tematic comparative analyses of Australian state and territory elections to
be conducted in the last 20 years. The aggregate and individual level data
presented in this paper reveal Australian regional elections to be more
than second-order events. While the Australian party competition at national
and regional levels is highly congruent, Australia’s uncoordinated cycles of
national and regional elections and federal distribution of policy responsibil-
ities help to focus attention on each regional incumbent when their turn
comes to face the people. Regional incumbents are sometimes able to
resist the drag of national incumbency on their votes. The Covid 19 crisis
suggests that one means by which they can do this is through policy compe-
tence. Future survey research could extend this study to include a broader set
of policies, as well as the other factors suggested by Bolwell and Eccleston as
important for regional incumbent electoral success. To adopt Hough and
Jeffery’s words (2006b:9–10), more research needs to be done on exactly
what is ‘at stake’in Australian regional elections.
Notes
1. In the Australian context, federal elections refer to elections at the national
level, while state and territory elections refer to regional elections. This may
be confusing for readers outside Australia, particularly in Europe, who com-
monly understand ‘state’or ‘state-wide’elections to refer to national contests
and sub-state elections to refer to regional contests. In this paper, we use
‘regional’and ‘state and territory’elections interchangeably.
2. ‘Liberal National Party Coalition’and ‘LNP’are used in this paper to refer to the
main non-Labor party or parties in all Australian jurisdictions but the terms are
(necessary) simplifications. The Liberal and National parties are separate parties;
however, they have operated in a formal Coalition at national level and in the
state of New South Wales since the 1920s. In the state of Queensland, the
Liberal and National parties were in a long-standing Coalition but amalgamated
in 2008, forming the Liberal National Party. In some states and territories,
notably Western Australia, the Liberal and National parties operate relatively
independently. In others, there is no National Party and the main non-Labor
party is the Liberal Party.
3. Australian national, state and territory elections all use versions of preferential
voting, in which voters are either required or permitted to express preferences
for more than one candidate by numbering ballot papers 1, 2, 3 etc. The Aus-
tralian national parliament is bicameral, as are five of the six state parliaments.
Queensland and the two territories have unicameral parliaments. Where there
are two houses, the lower house is the chamber that determines the govern-
ment. With the exception on Tasmania and the Australian Capital Territory,
lower house or unicameral elections involve contests for single member elec-
toral districts. Upper house elections are significant but, in line with most
research on regional elections (see Hough and Jeffery 2006a; Schakel 2011;
Dandoy and Schakel 2013), they are not discussed in this paper.
436 R. SMITH ET AL.
4. On average, three Green MPs have been elected to the 25 seat Tasmanian
House of Assembly at state elections since 1998, while no Green MP has
been elected to the Australian House of Representatives from the five Tasma-
nian electorates.
5. The other five federations were the United States, Canada, Switzerland,
Germany and India.
6. This was the 2013 Western Australian state election. Both the Labor and
Coalition votes were higher at regional than national level (ALP 33.13 percent
versus 31.18 percent; Coalition 53.15 percent versus 50.60 percent).
Disclosure statement
No potential conflict of interest was reported by the author(s).
References
Aitkin, Don. 1982.Stability and Change in Australian Politics. Canberra: Australian
National University Press.
Aroney, Nicholas, Scott Prasser, and Alison. Taylor. 2012.“Federal Diversity in Australia:
A Counter-Narrative.”In The Future of Australian Federalism: Comparative and
Interdisciplinary Perspectives, edited by Gabrielle Appleby, Nicholas Aroney, and
Thomas John, 272–300. Cambridge: Cambridge University Press.
Aulich, Chris, ed. 2014.The Gillard Governments. Melbourne: Melbourne University
Press.
Aulich, Chris, and Mark Evans, eds. 2010.The Rudd Government. Canberra: ANU Press.
Bean, Clive. 2018.“Changing Leaders, ‘Mediscare’and Business as Usual.”In Double
Disillusion: The 2019 Australian Federal Election, edited by Anika Gauja, Peter
Chen, Jennifer Curtin, and Juliet Pietsch, 235–253. Canberra: ANU Press.
Boin, Arjen, Allan McConnell, and Paul. ‘t Hart. 2021.Governing the Pandemic: The
Politics if Navigating a Mega-Crisis. Cham: Palgrave Macmillan.
Bolwell, Dain, and Richard. Eccleston. 2018.“Ebb Tide in Blue: Recent Sub-National
Elections in the Australian Federation.”Regional & Federal Studies 28 (3): 253–274.
https://doi.org/10.1080/13597566.2018.1465050.
Bromfield, Nicholas, and Allan. McConnell. 2021.“Two Routes to Precarious Success:
Australia, New Zealand, Covid-19 and the Politics of Crisis Governance.”
International Review of Administrative Sciences 87 (3): 518–535. https://doi.org/10.
1177/0020852320972465.
Butler, Josh. 2022, March 20.“South Australia Election: Labor Win Should Have
Morrison Government ‘Trembling’, ALP Says”.The Guardian.
Crowe, Shaun. 2018.Whitlam’sChildren: Labor and the Greens in Australia. Melbourne:
Melbourne University Publishing.
Dandoy, Régis, and Arjan H. Schakel, eds. 2013.Regional and National Elections in
Western Europe: Territoriality of the Vote in Thirteen Countries. London: Palgrave
Macmillan.
Dardanelli, Paolo, John Kincaid, Alan Fenna, André Kaiser, André Lecours, Ajay Kumar
Singh, Sean Mueller, and Stephan Vogel. 2019b.“Dynamic De/Centralization in
Federations: Comparative Conclusions.”Publius: The Journal of Federalism 49 (1):
194–219. https://doi.org/10.1093/publius/pjy037.
REGIONAL & FEDERAL STUDIES 437
Dardanelli, Paolo, John Kincaid, Alan Fenna, Lecours Kaiser, André Singh, and Ajay
Kumar. 2019a.“Conceptualizing, Measuring, and Theorizing Dynamic De/
Centralization in Federations.”Publius: The Journal of Federalism 49 (1): 1–29.
https://doi.org/10.1093/publius/pjy036.
Economou, Nick, and Zareh. Ghazarian. 2018.“Localism, Diversity and Volatility: The
2016 Australian Federal Election and the ‘Rise’of Populism.”Australasian
Parliamentary Review 33 (1): 134–155.
Economou, Nick, Zareh Ghazarian, Narelle Miragliotta, Will Sanders, Rodney Smith,
John Warhurst, and Paul Williams. 2020.“The Election in the States, Territories
and Regions.”In Morrison’s Miracle: The 2019 Australian Federal Election, edited by
Anika Gauja, Marian Sawer, and Marian Simms, 223–252. Canberra: ANU Press.
Fabre, Elodie. 2010.“Multilevel Election Timing—A Comparative Overview.”Regional &
Federal Studies 20 (2): 175–199. https://doi.org/10.1080/13597561003729897.
Fenna, Alan. 2019.“The Centralization of Australian Federalism 1901–2010:
Measurement and Interpretation.”Publius: The Journal of Federalism 49 (1): 30–56.
https://doi.org/10.1093/publius/pjy042.
Fenna, Alan. 2021.“Commonwealth-State Relations.”In Australian Politics and Policy:
Senior Edition, edited by Peter Chen, Nicholas Barry, John Butcher, David Clune,
Ian Cook, Adele Garnier, Yvonne Haigh, Sara C. Motta, and Marija Taflaga, 185–
203. Sydney: Sydney University Press.
Gauja, Anika, Marian Sawer, and Marian Simms, eds. 2020.Morrison’s Miracle: The 2019
Australian Federal Election. Canberra: ANU Press.
Hooghe, Liesbet, Gary Marks, and Arjan. Shakel. 2010.The Rise of Regional Authority: A
Comparative Study of 42 Democracies. London: Routledge.
Hough, Dan, and Charlie Jeffery, eds. 2006a.Devolution and Electoral Politics.
Manchester: Manchester University Press.
Hough, Dan, and Charlie. Jeffery. 2006b.“An Introduction to Multi-Level Electoral
Competition.”In Devolution and Electoral Politics, edited by Dan Hough and
Charlie Jeffery, 2–13. Manchester: Manchester University Press.
Jackson, Stewart. 2016.The Australian Greens: From Activism to Australia’s Third Party.
Melbourne: Melbourne University Publishing.
Jeffery, Charlie, and Dan Hough. 2006.“Devolution and Electoral Politics: Where Does
the UK Fit In?”In Devolution and Electoral Politics, edited by Dan Hough and Charlie
Jeffery, 248–256. Manchester: Manchester University Press.
Kefford, Glenn, and Duncan McDonnell. 2018.“Inside the Personal Party: Leader-
Owners, Light Organizations and Limited Lifespans.”The British Journal of Politics
and International Relations 20 (2): 379–394. https://doi.org/10.1177/
1369148117750819.
Kemp, David. 1978.Society and Electoral Behaviour in Australia. St Lucia: University of
Queensland Press.
Leach, Michael, Geoffrey Stokes, and Ian Ward, eds. 2000.The Rise and Fall of One
Nation. St Lucia: University of Queensland Press.
Marsh, Ian, ed. 2006.Political Parties in Transition? Sydney: Federation Press.
Martinez i Coma, Ferran, and Rodney. Smith. 2018.“The States and Territories.”In
Double Disillusion: The 2016 Australian Federal Election, edited by Anika Gauja,
Peter Chen, Jennifer Curtin, and Juliet Pietsch, 211–234. Canberra: ANU Press.
McAllister, Ian. The Australian Voter: 50 Years of Change. Sydney: UNSW Press.
Miragliotta, Narelle, and Campbell. Sharman. 2012.“Federalism and New Party
Insurgency in Australia.”Regional & Federal Studies 22 (5): 577–594. https://doi.
org/10.1080/13597566.2012.733943.
438 R. SMITH ET AL.
Moon, Jeremy, and Campbell Sharman, eds. 2003.Australian Government and Politics:
The Commonwealth, the States and the Territories. Cambridge: Cambridge University
Press.
Murphy, Julian R., and Erika. Arban. 2021.“Assessing the Performance of Australian
Federalism in Responding to the Pandemic.”Publius: The Journal of Federalism 51
(4): 627–649. https://doi.org/10.1093/publius/pjab026.
Orr, Graeme. 2019.The Law of Politics: Elections, Parties and Money in Australia. Sydney:
The Federation Press.
Prasser, Scott. 2020.“A Funny Thing Happened on the Way to the National Cabinet—
Out Goes Good Policy, One, Two, Three.”Australasian Parliamentary Review 35 (1):
141–156.
Ratcliff, Shaun, Jill Sheppard, and Juliet. Pietsch. 2020.“Voter Behaviour.”In Morrison’s
Miracle: The 2019 Australian Federal Election, edited by Anika Gauja, Marian Sawer,
and Marian Simms, 253–274. Canberra: ANU Press.
Schakel, Arjan. 2011.“Congruence Between Regional and National Elections.”
Comparative Political Studies 46 (5): 631–662. https://doi.org/10.1177/
0010414011424112.
Schakel, Arjan H., and Régis. Dandoy. 2013a.“Introduction: Territoriality of the Vote: A
Framework for Analysis.”In Regional and National Elections in Western Europe:
Territoriality of the Vote in Thirteen Countries, edited by Régis Dandoy and Arjan
H. Schakel, 1–26. London: Palgrave Macmillan.
Schakel, Arjan H., and Régis. Dandoy. 2013b.“Conclusion: Regional Elections in
Comparative Perspective.”In Regional and National Elections in Western Europe:
Territoriality of the Vote in Thirteen Countries, edited by Régis Dandoy and Arjan
H. Schakel, 275–299. London: Palgrave Macmillan.
Schnabel, Johanna, and Yvonne Hegel. 2021.“Explaining Intergovernmental
Coordination During the COVID-19 Pandemic: Responses in Australia, Canada,
Germany and Switzerland.”Publius: The Journal of Federalism 51 (4): 537–569.
https://doi.org/10.1093/publius/pjab011.
Smith, Rodney. 2001.Australian Political Culture. Sydney: Pearson Education.
Smith, Rodney, and John. O’Mahony. 2006.“The Cartel Parties Model and Electoral
Barriers.”In Political Parties in Transition? edited by Ian Marsh, 94–115. Sydney:
The Federation Press.
Twomey, Anne. 2021.“Who Is Responsible for Quarantine Under the Constitution?”
Australian Law Journal 95 (4): 253–256.
University of Western Australia. n.d.Australian Politics and Elections Archive 1856–2018.
Accessed at https://elections.uwa.edu.au/.
REGIONAL & FEDERAL STUDIES 439
