Electoral Studies 18 (1999) 217–233

Compulsory voting, party stability and electoral

advantage in Australia1
M. Mackerrasa, I. McAllisterb,*
a
Department of Politics, University of NSW, Canberra, ACT 2600, Australia
b
Research School of Social Sciences, Australian National University, Canberra, ACT 0200, Australia

Abstract

Australia has the oldest and probably the most efficient system of compulsory voting among
the established democracies. The main reason for its introduction in the 1920s was to increase
turnout, a goal it has achieved without difficulty. Compulsory voting has also made a major
contribution to the long-term stability of the party system. However, the compulsory voting
system has had two disadvantages. As in other countries which have introduced compulsory
voting, historically there has been a very high level of invalid votes, although the levels are
now less than they were prior to 1984, as a result of various electoral reforms. Second, the
system disadvantages rightwing parties and advantages leftwing and minor parties. Survey
evidence from the 1996 federal election suggests compulsory voting reduced the Liberal–
National coalition’s first preference vote by some 5 per cent, compared to the coalition’s
expected vote under a voluntary system. If the system is ever removed, it is likely to be for
this reason, rather than because of any difficulties in administering it, or because of any philo-
sophical objection to compulsion among voters or politicians.  1999 Elsevier Science Ltd.
All rights reserved.

Keywords: Elections; Australia; Voting; Parties

* Corresponding author. Tel.: ⫹ 61 2 6249 2340; Fax: ⫹ 61 2 6247 8522; E-mail: Ian.McAllister@an-
u.edu.au
1
Previous versions of this paper were presented at the American Political Science Association meetings,
Washington DC, August 1997, and the European Consortium for Political Research Joint Sessions, Univer-
sity of Warwick, March 1998.

0261-3794/99/$ - see front matter  1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 6 1 - 3 7 9 4 ( 9 8 ) 0 0 0 4 7 - X
218 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233

No modern democracy has shown greater readiness to experiment with various

electoral methods than Australia. Voting is compulsory in State and Common-
wealth elections, preferential voting is widely used, and proportional represen-
tation applies in the election of the Tasmanian Assembly and the Commonwealth
Senate. The character of the party battle and the behaviour of the voters are affec-
ted by the compulsory franchise and preferential voting. Louise Overacker (1952).

The level of popular participation in national elections is usually regarded as a

sensitive barometer of the health of democratic institutions. Political systems that
attract high levels of participation are seen as more legitimate and more responsive
to the needs of the electorate compared to systems with lower levels of participation.
As a consequence, various electoral devices have been introduced to increase turnout,
the most notable of which is mandatory or compulsory voting. Compulsory voting
has been shown to increase turnout significantly, although it is not a necessary con-
dition for high levels of participation, as the case of Malta, Sweden and other coun-
tries demonstrates (Powell, 1986; Hirczy, 1994, 1995). Nevertheless, compulsory
voting is a cost-efficient institutional remedy to the problem of low turnout in the
advanced democracies.
Comparatively little attention has been devoted to the circumstances surrounding
the introduction of compulsory voting among the established democracies that have
adopted it—Australia, Belgium, Italy, Luxembourg and parts of Austria—and the
view of voters towards it.2 Perhaps more surprisingly, few studies have examined
the consequences of compulsory voting for the operation of the electoral system as
a whole, or for the party system (but see McAllister and Makkai, 1993; Power and
Timmons Roberts, 1995). In this paper, we examine compulsory voting in Australia,
including the reasons that were given for introducing it, the debates about its utility,
and the ways in which it is administered. We also analyse the consequences of
compulsory voting for invalid votes, party advantage, and party stability.

Compulsory enrolment and compulsory voting

As much as any of the established democracies, Australia has represented a vast,

democratic experiment. At the beginning of the 20th century, James Bryce com-
mented that Australia had “travelled farthest and fastest along the road which leads
to the unlimited rule of the multitude” (Bryce, 1921, 181). By the criteria of adult
suffrage, Australia was the first truly democratic state, achieving that status in 1903,3
with New Zealand second in 1907, and the remaining contemporary liberal democ-
racies following in the years after the First World War (Aitkin and Castles, 1989,

2
In addition, compulsory voting is used extensively in Latin America, most notably in Brazil (Power
and Timmons Roberts, 1995), but also in Argentina, Bolivia, Chile, Costa Rica, the Dominican Republic,
Ecuador, Guatemala, Mexico, Panama, Paraguay and Peru.
3
This date is subject to the caveat about the long delay in granting the vote to the Aboriginal population
on the same basis as whites.
 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233 219

208). But perhaps more importantly, Australians have been more willing to experi-
ment with democracy than many of their contemporaries: “in no other liberal democ-
racy, it seems safe to say, have the permutations and combinations of electoral reform
been as great” (Goot, 1985, 179). These experiments have ranged from the franchise
and the methods used to weigh votes, to the use of compulsory voting to increase
turnout. Despite many changes during the course of the century, the system of com-
pulsory voting—once introduced—has remained relatively immune to change.
A system of compulsory enrolment for all eligible voters was introduced for Com-
monwealth elections in 1911. It had the intended effect of greatly increasing the size
of the electorate: in April 1910 enrolment was 2 258 482; by the time of the May
1913 federal election it had risen to 2 760 216. The decision to make enrolment
compulsory was taken on the recommendation of the Chief Electoral Officer, R.C.
Oldham, who wanted to rationalise the electoral roll.4 Prior to that change the compil-
ing of an up-to-date electoral roll was thought by Oldham to be too cumbersome
and expensive as it occurred infrequently and did not operate effectively. This
change, which established the principle of compulsion and opened the door to com-
pulsory voting, was therefore driven by the bureaucracy, not by politicians. It was
a means of improving the accuracy of the electoral rolls and was proposed as a
simple ‘machinery measure’ for the benefit of electoral officials. The states quickly
followed suit, beginning with Queensland in 1914.
Later, compulsory voting was argued to be the natural corollary of compulsory
enrolment. Compulsory voting is, however, a different matter: several English-speak-
ing countries have compulsory enrolment, but have stopped short of introducing
compulsory voting. At present, the Australian compulsory enrolment provisions are
as well-enforced as in other English speaking countries, although there appears to
be less compliance in Australia. This seems to be due to the high number of immi-
grant non-citizens in Australia and also to a relatively high level of non-compliance
among the young. Research conducted for the Australian Electoral Commission
(AEC) has consistently found that young people are less likely to enrol (McAllister,
1985; Elliott and Shananhan Research, 1989). Recent surveys conducted for the
AEC5 have found that just 78 per cent of the eligible population aged 18 to 24 are
enrolled, compared to 93 per cent for the eligible population as a whole.6
At the state level, compulsory voting was first introduced in Queensland by the
Denham Liberal government in time for the 1915 general election. The measure
closely followed other major amendments to the electoral laws and appears to have

4
As Oldham (1911, 162) put it: “a thoroughly efficient roll can only be continually preserved under
a system of compulsory enrolment, or by the employment of an army of electoral inspectors to continu-
ously watch the movements of people throughout the Commonwealth”.
5
Monthly surveys conducted for the AEC since November 1997 have shown an average of 78.2 per
cent of 18–24 year olds are not enrolled. Our thanks to Brien Hallett for providing this information.
6
These figures explain how the International Institute for Democracy and Electoral Assistance (1997,
14) can show Australia as having a turnout of only 82.7 per cent during the 1990s, ranking it 20th in
the world in terms of turnout. Estimating turnout based on enrolment rather than eligible population puts
Australia in second place, having a turnout at federal elections during the 1990s of 96.2 per cent, just
behind Malta with 96.7 per cent.
220 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233

been the last desperate attempt of an unpopular government to save itself from defeat.
In the event, the attempt was unsuccessful as Labor won a sweeping victory at the
election. Over the next three decades the other states and the Commonwealth fol-
lowed suit. The system was introduced to Commonwealth elections in 1924, follow-
ing the report of the 1915 Royal Commission which viewed compulsory voting as
a logical extension of compulsory enrolment; the delay in its introduction was caused
by divisions within the Labor government, who could agree only on its application
to referendums (Goot, 1985, 192). In the states, Victoria introduced compulsory vot-
ing in 1926, New South Wales and Tasmania in 1928, Western Australia in 1936
and South Australia in 1941. There was no agreement on the introduction of compul-
sion for local government elections and it has applied in only a few states, some of
which have veered between voluntary and compulsory systems in the postwar years.
In all cases, the move to introduce compulsory voting was bipartisan. The new
system suited the convenience of all parties and was opposed by none. Normally
the bill to introduce compulsion was presented by a private member. This enabled
the major parties to allow an important shift in the system to occur without the risk
that any party machine would be blamed. In the case of the Commonwealth, the
entire 1924 debate in both houses took less than 3 hours and 26 minutes in the
Senate, and 52 minutes in the House of Representatives. While most of the arguments
advanced in favour of the change were directed at increasing turnout, it was also
suggested that imposing a fine of up to £2 would help to reduce the cost of mounting
elections. One member also suggested that it would produce fairer election compe-
tition by reducing the National Party’s advantage in car ownership among its mem-
bers, which it exploited by driving voters to the polls in rural areas.
The effect on turnout of the introduction of compulsion was immediate and dra-
matic (Table 1). In the six states and the Commonwealth, turnout increased by an
average of 23.2 per cent, ranging from 12.6 per cent in Queensland (where turnout
was relatively high anyway under voluntary voting) to 37.8 per cent in South Aus-
tralia. The average level of turnout under the system of compulsory voting was 91.6
per cent. In effect, the increase in turnout was significant, which was the goal of the
politicians who introduced it. There was also a flow-on effect to turnout in state

Table 1
Federal and state turnout, before and after the introduction of compulsory voting

Election prior to Election after Increase

Year introduced Jurisdiction
compulsory voting compulsory voting (per cent)

1914 Queensland 75.5 88.1 12.6

1924 Commonwealth 58.0 91.3 33.4
1926 Victoria 59.2 91.8 32.6
1928 New South Wales 82.5 94.9 12.4
1928 Tasmania 81.9 95.0 13.1
1936 Western Australia 70.1 91.6 20.5
1941 South Australia 50.7 88.5 37.8

Source: McAllister et al. (1997).

 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233 221

elections, even where voting remained voluntary: in New South Wales, for example,
state turnout rose from 69.1 per cent in May 1925 to 82.5 per cent in October 1927.7
In general, ordinary citizens accepted the new system with little complaint and there
were comparatively few instances where voters were fined for not attending the polls
without sufficient reason; as a consequence the system has been easy to administer
ever since it was introduced.
By any standards, compulsory voting is popular among voters. Going back to the
earliest opinion poll which asked voters’ views on the topic, conducted in 1943,
never less than six out of every 10 voters have supported compulsory voting (Fig.
1). Support for the system increased gradually during the 1940s, 1950s and 1960s,
peaking at 76 per cent in a survey conducted in 1969. This gradual increase is prob-
ably a reflection of the large number of voters who have grown up under the system,
together with the absence of any political debate concerning its advantages or disad-
vantages. Support declined slightly in the 1970s and early 1980s, but in recent years
has strengthened. Surveys which permit the respondents to register the strength of
their opinion for or against the system indicate that voters who favour compulsory
voting have stronger views than those who oppose it (McAllister and Makkai, 1993,
28–29). As Aitkin (1982, 31) notes, compulsory voting “is in no sense seen as an
imposition on the electorate and resented by it”.

Fig. 1. Attitudes to compulsory voting, 1943–96. Question wordings differ slightly from survey to sur-
vey. Sources: Aitkin (1982, 31); 1979 Australian National Political Attitudes Survey; 1987–96 Australian
Election Study Surveys.

7
In October 1930, the first NSW state election held under compulsory voting, turnout rose further, to
94.9 per cent.
222 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233

Debates about compulsory voting

Colin Hughes, a long-standing scholar of Australian elections who was the Aus-
tralian Electoral Commissioner from 1984 to 1989, studied all the parliamentary
debates of the period when compulsory voting was introduced (Hughes, 1968). He
listed 15 arguments used in favour which can be summarised as follows:
1. Democratic government means majority rule and the expression of an opinion by
a majority of electors.
2. Voting is analogous to other duties society requires of citizens, such as giving
evidence in court proceedings, jury service, paying rates, compulsory education
or military service.
3. Voting is the most important civic duty, yet the burden is extremely light: once
or twice every 3 years.
4. The voter is not compelled to vote for anybody; he or she can always spoil the
ballot paper.
5. Compulsory voting is a necessary corollary of compulsory enrolment without
which the experience and penalties of compulsory enrolment would be pointless.
6. Compulsory voting would stop the growing demand of voters who had ‘got into
a loose way of voting’ for motor-car transport to the polls.
7. Social pressure applied by the trade union movement had already enforced com-
pulsory voting on Labor supporters.
8. Turnout figures were too low.
9. The quality of legislation coming from legislatures elected by a minority vote
would deteriorate.
10. Compulsion would enforce political education.
11. The franchise had been fought for, and therefore should be used.
12. As individual liberty consists in exemption from legal control, so political liberty
consists in participation in legal control.
13. Those who most readily criticise legislation are the least zealous in exercising
the franchise; they would be taught to be good democrats by becoming respon-
sible for public acts.
14. Compulsory voting would help to cleanse the rolls by checking on non-voters.
15. Compulsion would emphasise the responsibilities of electors.
The arguments used against compulsory voting during the 1920s and 1930s mostly
asserted that compulsion was wrong in principle and that it ‘is not the democratic
norm’. However, since the opponents of compulsion were defeated, their arguments
were not widely aired within the electorate. More interesting is the fact that some
Liberal politicians now advocate abandonment of the system and in recent years a
significant number of liberal policymaking bodies have endorsed a return to voluntary
voting, notably the party’s highest forum, the Federal Council, which passed motions
in 1988 and 1993 calling for a repeal of the current system (Jackman, 1997). While
many Labor politicians argue that this is pure promotion of the electoral interests of
the Liberal Party, it is worth noting that some ‘general interest’ arguments are cur-
rently used. The most important of these arguments is the belief that ‘a growing
 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233 223

number of Australians now doubt’ the integrity of the electoral system which
(allegedly) is rife with fraud.
More recently, a minority report (consisting mainly of Liberal and National
members) of a parliamentary committee charged with examining all aspects of the
conduct of the March 1993 federal elections argued that compulsion did encourage
electoral fraud. The minority report argued that because the law forces people to
vote, minimal barriers to enrolment and voting exist. “Thus, in practice compulsory
voting underpins a system which has very few checks in place to prevent and detect
fraudulent enrolment and voting” (Joint Standing Committee on Electoral Matters,
1994, 157). The report further went on to suggest that the way in which the system
is applied was undermining popular trust in it, as well as “being used as part of a
push for a de facto first-past-the-post voting system.”
Because voting is compulsory, politicians and electoral officials have gone to con-
siderable lengths to make the system voter friendly. Indeed, Australia probably is
the most voter-friendly country in the world. If Australians felt more strongly about
election results than they do then there probably would be more abuse of the system.
Allegations of such abuse are frequent, but factual evidence is rare. The desire to
ensure that the system is fair to voters (as distinct from political parties) has prompted
judges (sitting in a Court of Disputed Returns) frequently to order fresh elections
on the ground that the number of disenfranchised voters exceeds the majority of the
winning candidate. Typically, a re-election for a constituency occurs about once
every second year.
The most recent example of this phenomenon was the state seat of Mundingburra
following the Queensland state election on 15 July 1995, when Labour was returned
with a one seat majority.8 However, in Mundingburra the Labor candidate had won
by just 16 votes over his Liberal opponent. The Liberal loser petitioned the Court
of Disputed Returns which found that 22 postal votes had not been cast by soldiers
serving as peace-keepers in Rwanda because ballot papers had not been received by
them in time. A fresh election was ordered, the Liberal candidate won easily and
the government of the state changed as a consequence.
The major issues of Australian compulsory voting involve the following questions.
How seriously is the system enforced? Further, what level of fine does the elector
pay who neglects the democratic duty to vote? The main answer is that the system
is seriously enforced, contrary to the assertions of Jackman (1987). The federal Com-
monwealth Electoral Act outlines the administrative arrangements for compulsory
voting,9 and section 245 deals with sanctions against non-voters: “(1) It shall be the
duty of every elector to vote at each election. (2) The Electoral Commissioner must,
after polling day at each election, prepare for each Division a list of the names and

8
The result in seats (in a Legislative Assembly of 89 members) was Labor 45, National 29, Liberal
14, with one Independent.
9
Each of the six states and both the Australian Capital Territory and the Northern Territory have
provisions the same in principle as the federal one. The terms ‘Commonwealth electoral division’ and
‘State electoral district’ are the Australian equivalents of the British constituency, the American con-
gressional district and the Canadian riding.
224 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233

addresses of the electors who appear to have failed to vote at the election. (3) Subject
to subsection (4), within the period of three months after the polling day at each
election, each Divisional Returning Officer must send by post to each elector whose
name appears on the list prepared under subsection (2), at the address mentioned in
that list, a penalty notice.” The section continues through 18 subsections. No person
reading section 245 in full could doubt that compulsory voting really is enforced
in Australia.
If a voter is recorded as not voting the AEC will send a ‘please explain’ letter.
The voter is given an option to pay a $A20 penalty in which case the matter rests.
If what the AEC regards as a valid reason for non-voting is provided, then this will
result in no further action. For example, valid reasons would be being overseas, tried
to vote but failed, or membership of a religious order which prohibits voting.10 If
the reason is disputed it will be taken to court where the fine increases to $A50, in
addition to legal costs, if the person is convicted. In practice, the proportion of voters
required to pay penalties or taken to court never exceeds 1 per cent of the electorate
and is normally much less. For example, following the 1993 federal election, the
AEC investigated 490 230 cases of persons who appeared not to have voted, there
being a total of 11 384 638 enrolled voters at that election. Of the 490 230, fines of
$A20 each were paid by 23 230 electors who had voted (or 4.7 per cent of all non-
voters). The remainder gave valid reasons for not voting, save for 4412 who went
to court (or 0.9 per cent of all non-voters). The AEC has no information on what
happened thereafter to them.11

The consequences of compulsory voting

Invalid voting

The most significant consequence of compulsory voting is a large proportion of

invalid or spoiled votes at elections. Australia has one of the highest levels of invalid
votes among the established liberal democracies. The level is comparable to what
was found in the Netherlands in the two decades after the Second World War, which
had a compulsory voting system until 1971 (Irwin, 1974), but substantially less than
Brazil, where it has ranged as high as 40 per cent in recent elections (Power and
Timmons Roberts, 1995). These votes (called ‘informal’ in Australia) are set out for

10
One curious example is the Governor-General, the head of state. Since the head of state cannot vote
for him or herself, but voting is compulsory, the Governor-General receives a ‘please explain’ letter like
other non-voters. By convention, the Governor-General informs the AEC that it is inappropriate for him
or her to vote, and the matter rests.
11
The figures following the 1996 federal election are similar. A total of 519 684 voters did not vote,
of whom 29 129 paid a fine, and the remainder gave a valid reason for not voting, save for 8782 who
were summoned to appear in court.
 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233 225

the two houses of the federal parliament in Table 2, grouped by the type of electoral
system in operation at the time. The Senate results are the more striking. From 1901
to 1917 Senate informal voting was low, averaging just 5.6 per cent. However, that
was during the period of voluntary voting with a simple ballot. The introduction of
preferential, soon to be followed by compulsory, voting in the early 1920s raised
informality to almost one in every 10 votes cast—until something was done about it.
In the early 1980s the federal politicians worked on a plan to simplify the preferen-
tial vote, involving a major change to the Senate ballot paper format. From the
December 1984 election onwards the voter could cast an ‘above-the-line’ Senate
vote with a single mark of approval. (This is the number ‘1’ but a tick or cross is
accepted.) That would be deemed to be a full listing of preferences, as registered
by the party within 24 hours of the close of nominations. Thus, the system would
remain ‘ordinal’ in terms of the counting process while, inevitably, over 90 per cent
of voters could be lazy and treat the Senate vote as though it were ‘categorical’.
Meanwhile, the House of Representatives vote would continue to employ a fully
preferential system. The effect of this change was substantially to reduce the number
of informal votes at Senate elections. The Senate informal vote now averages 3.6
per cent.
The assertion that the design of the electoral system means that ‘over 90 per cent
of voters could be lazy’ needs detailed justification. Voters can still mark their own

Table 2
Informal voting by house and electoral system

Electoral system Informal votes (per cent)

Mean Std Dev

Senate
1901–17 First-past-the-post/voluntary voting 5.6 2.6
Preferential block majority/voluntary voting/optional
1919–22 9.0 0.6
preferences
Preferential block majority/compulsory
1925–31 8.8 1.6
voting/optional preferences
Preferential block majority/compulsory
1934–46 9.8 1.2
voting/compulsory preferences
Proportional representation/compulsory
1949–83 8.9 1.9
voting/compulsory preferences
Proportional representation/compulsory voting/ticket
1984–96 3.6 0.8
preferences
Total 7.7 2.7

House of Representatives
1901–17 First-past-the-post/voluntary voting 2.5 0.7
1919–22 Preferential/voluntary voting 4.0 0.7
1925–96 Preferential/voluntary voting 2.8 1.1
Total 2.8 1.0

Source: As for Table 1.

226 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233

preferences for candidates ‘below the line’, although not many do so. In March 1993
‘below the line’ votes were 599 527 or 5.6 per cent of the total formal vote. In March
1996 they were 615 472 or 5.7 per cent. The number of such votes only ever becomes
significant in less populous states and where adverse publicity is given to party tick-
ets. This can produce a reaction of sympathy for a candidate seen to be disadvantaged
by the major parties, as occurred in Tasmania in 1996 when the Green candidate,
Dr Bob Brown, was elected in a circumstance in which ‘below the line’ voting rose
from 21.1 per cent in 1993 to 27.4 per cent in 1996.
The 1984 change to the Senate ballot paper has caused a number of observers to
argue that the Australian Senate is now almost a party list system of proportional
representation. This is incorrect: it remains (constitutionally and technically) a system
of direct election by the single transferable vote. However, the change did create
confusion at the first two elections under the new Senate ballot paper format, those
held in December 1984 and July 1987. In both elections, for the first time in Aus-
tralian electoral history, Senate informality was less than that for the House of Rep-
resentatives. In the three most recent federal elections the system has stabilised, with
an average informality level of 3.2 per cent for the Senate and 3.1 per cent for the
House of Representatives, a level that is still high by world standards, but low in
comparison with recent Australian elections.
The high levels of informal votes are also linked to four underlying factors which
do not occur elsewhere. The first is the complex nature of Australia’s electoral sys-
tem, at least judged against international standards. Table 2 shows, for example, that
at the federal level the Senate has used six different systems during the century,
while the House of Representatives has used three separate systems. By contrast, in
Britain, apart from changes in the franchise and other more minor changes relating
to the format of the ballot paper, the electoral system has remained the same over
the period (Blackburn, 1995). Related to this complexity is the frequency of elections
at the state and federal levels of government. Federal elections are held every 3
years, but usually more frequently, as are all state elections except New South Wales
where they must be held every 4 years, a change introduced in 1981. In practice, this
means that voters sometimes have to cast a ballot as often as once every 18 months.
A third factor increasing invalid votes is the increasing tendency to hold House
of Representatives elections in conjunction with Senate elections and constitutional
referendums, all of which have different rules for completing the ballot. For refer-
endums the elector writes ‘yes’ or ‘no’ in the space provided, while the Senate ballot
paper requires either one number above the line, or one or more numbers below the
line. Finally, there are the differences in the rules for casting a formal vote between
state and federal elections. Most Australians live in states with ‘optional preferences’
while the Commonwealth has ‘compulsory preferences’. When compulsory voting
is combined with these factors and a high proportion of non-English speaking immi-
grants, it is not surprising that the net result is one of the highest levels of spoiled
ballots found in any Western liberal democracy (McAllister and Makkai, 1993).
 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233 227

Party advantage

Levels of turnout have long been linked to different voting outcomes. High levels
of turnout are thought to favour leftwing parties, since their supporters generally
have lower socioeconomic status, which in turn leads to a reduced probability of
voting. In a study of national voting in 19 democracies, Crewe (1981, 253) found
“a connection between a country’s overall turnout and the electoral strength of those
of its parties established to represent the working class.” By contrast, low turnout
benefits rightwing parties, whose supporters are generally of higher socioeconomic
status and more likely to turn out to vote. Empirical support for these propositions
has been found in a variety of studies, though the net impact on the vote is usually
only in the order of several percentage points (McAllister and Mughan, 1986). In
Australia, compulsory voting has been shown to favour the Labor Party, since dispro-
portionately more of its supporters would be expected to abstain under a voluntary
voting system (McAllister, 1986).
Evidence to support these propositions can be found in the 1996 Australian Elec-
tion Study Survey, which asked respondents if they would have still voted if voting
had not been compulsory.12 More than two out of every three voters said that they
would definitely have voted, with a further 19 per cent saying that they probably
would have voted (Table 3, first column). In total, then, the survey suggests that a
shift to a voluntary system of voting could probably count on a turnout level of
around 87 per cent. This is a level which is currently found only among the top

Table 3
Voting by self-reported turnout under a voluntary systema

(Column per cent) 1996 vote

Labor Lib–Nat Democrat/Green Total (n)

Definitely would have

(68) 36 53 11 100 (1156)
voted
Probably would have
(19) 42 41 17 100 (315)
voted
Might, might not (5) 44 33 23 100 (92)
Probably not (5) 52 25 23 100 (87)
Definitely not (3) 45 28 27 100 (47)

a
The question was: “Would you have voted in the election if voting had not been compulsory?” 1996
vote is first preference vote for the House of Representatives.
Source: 1996 Australian Election Study Survey.

12
Survey estimates of expected turnout will, as Jackman (1997) notes, underestimate the decline in
turnout due to voluntary voting. This is caused by the tendency of opinion surveys to under-sample
disinterested voters who would be less likely to vote under a voluntary system. It is difficult to estimate
the likely error caused by survey non-response in predicting turnout, but it is likely to be small. In any
event, our interest here is in the broad patterns rather than in precise estimates.
228 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233

quintile of established democracies (International Institute for Democracy and Elec-

toral Assistance, 1997). A further 5 per cent said that they were unsure whether they
would have voted, while a total of 8 per cent said that they either ‘probably’ or
‘definitely’ would not have voted under a voluntary system.
The second part of Table 3 shows the strong partisan bias that exists among these
various groups of voters. The results confirm that the less likely the voter is to say
that he or she would have voted, then the less likely he or she would be to support
the Liberal–National coalition, and the more likely to support either Labor or the
minor parties. For example, among those who definitely said that they would have
voted, 53 per cent reported voting Liberal–National in the federal election; by con-
trast, Liberal–National support drops to just 25 per cent among those who said that
they probably would not have voted if voting had been voluntary. Indeed, among
the small group who said that they would definitely not have voted, minor party
support almost exceeds support for the Liberals and Nationals.
These are substantial electoral effects, but what net impact would they have had
on the outcome of the election, given the relative sizes of the various turnout groups?
Table 4 suggests that the effect would have been significant, but not sufficient to
change the outcome of the election which, in any event, saw the defeat of a Labor
government which had ruled for 13 years. The first line of Table 4 shows voting
among those who said that they definitely would have voted, and is therefore the
same estimate as the first line in Table 3. The remaining lines of Table 4 re-estimate
the vote adding in the other four turnout groups. The estimates suggest that Labor
gained about 2.5 per cent of the first preference vote through compulsory voting,
while the minor parties gained just under 3 per cent. However, the Liberal–National
vote was particularly sensitive to turnout. If only those who were committed to
turning out had voted, the Liberal–National coalition would have attracted over 52
per cent of the first preference vote; compulsory voting meant that they actually
attracted just 47 per cent, a net loss of over 5 per cent.

Table 4
Party vote at different turnout levelsa

1996 vote

Labor Lib–Nat Democrat/Green Total (n)

Definitely would have

36.1 52.6 11.3 100.0 (1156)
voted
Adding:
Probably would have
37.2 50.2 12.6 100.0 (1471)
voted
Might, might not 37.7 49.2 13.1 100.0 (1563)
Probably not 38.4 47.9 13.7 100.0 (1650)
Definitely not 38.6 47.3 14.1 100.0 (1697)

a
Estimates are first preference vote for the House of Representatives.
Source: As for Table 3.
 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233 229

These are substantial effects, and while not large enough to change the outcome
of the 1996 election, would have the potential to determine the result of a closely
fought contest. The results also confirm other findings, which have suggested that it
is the vote of rightwing parties that is most sensitive to changes in turnout, while
the impact on the vote for leftwing parties is usually slight (McAllister and Mughan,
1986). The reason appears to be that higher levels of turnout benefit both leftwing
parties and minor or protest parties. Leftwing parties benefit electorally because high
turnout mobilises disproportionately more of their supporters, who might otherwise
abstain. Minor or protest parties benefit because high turnout also mobilises dispro-
portionately more swinging and uncommitted voters, who have usually defected from
the major parties.

Party stability

A third consequence of compulsory voting is a high level of party stability. That

is reflected in the large number of voters who identify with one or other of the major
parties, and in the preponderance of representatives of the major parties in the federal
and state parliaments. Australia has one of the highest levels of party identification
in the world. While partisan dealignment in Britain, the United States and some other
countries has seen many voters rejecting the major parties, and either abstaining or
changing their party from one election to the next (Wattenberg, 1994; Dalton, 1996),
Australia has largely been immune from these changes. Compulsory voting ensures
that voters cast a ballot and the act of voting means that they are forced to think,
however superficially, about the major parties. The frequency of elections at the state
and federal levels further reinforces the high visibility that parties enjoy within the
electorate. Although, as noted previously, compulsory voting also favours support
for minor and protest parties, the advantage they gain is small compared to the overall
impact on major party support.
A stronger likelihood of voting if the system were changed to voluntary voting
is associated with the direction and strength of partisanship. Fig. 2 shows that among
those who said they definitely would vote, just 12 per cent reported that they did
not have a partisanship in 1996. By comparison, just over half of those who said
that they definitely would not vote were non-partisans. Weak partisans are also more
numerous among those with less commitment to voting under a voluntary system,
the main difference being between those who were unequivocal about their intention
to vote, and the other respondents. The system of compulsory voting is therefore
one of the underpinnings of Australia’s traditionally high level of partisanship.
A further indication of the system’s impact on partisanship is the likelihood that
a voter will use a ‘how to vote’ card when he or she visits the polls. Under Australia’s
system of preferential voting, voters may direct their second and subsequent prefer-
ences to the candidates of other parties. The political parties make use of the system
by providing their supporters with ‘how to vote’ cards outside polling stations, which
advise voters to direct their later preferences to specific candidates and parties. These
are usually parties with whom reciprocal preference arrangements have been made,
or with whom they are aligned on important issues. Around half of all voters report
230 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233

Fig. 2. Likelihood of voting and partisanship. See Table 3 for question wording. Use of the ‘how to
vote’ cards is for House of Representatives. Source: As for Table 3.

making use of these ‘how to vote’ cards so their impact is substantial.13 Fig. 2 shows
that the most committed voters are those who are most likely to make use of the
cards and to follow their party’s wishes. Among those who are least likely to vote
only around four in every 10 say that they use the cards. Once again, the evidence
suggests that compulsory voting is strongly linked to party factors.
Other evidence to support the proposition is the party composition of the federal
House of Representatives, which illustrates party stability by virtue of the relative
absence of minor party representatives. Of the 2096 contests for the federal lower
house that have taken place between 1949 and 1996, only eight—or 0.4 per cent—
have been won by independents. However, four of the eight independents had pre-
viously been elected under a party label, and for various reasons resigned or been
expelled from their party; only four contests, then, have been won by independents
without a prior party association in the electorate (McAllister, 1997). While the elec-
toral system represents the major barrier for independents, compulsory voting fosters
widespread loyalties towards the major parties across the electorate, thus making it
much more difficult for independent and minor party candidates to attract the elec-
toral support they require for successful election.

13
In the 1996 Australian Election Study, 56 per cent of voters reported using a ‘how to vote’ card to
guide their vote in the House of Representatives.
 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233 231

Conclusion

The original purpose behind the introduction of compulsory voting in Australia

was to increase turnout, and it has been particularly effective in meeting that goal.
A comparative study of compulsory voting found that the Australian system had
been substantially more effective in increasing electoral participation compared to
Austria, where compulsory voting has operated for presidential elections since 1949
(Hirczy, 1994). Why has the system of compulsory voting been so effective and the
electorate so compliant in supporting it? The answer lies in Australia’s distinctive
utilitarian political culture, which in turn has its origins in the Benthamite outlooks
of the early white settlers (Collins, 1985). Hartz (1964) has argued that the cultural
development of Anglo-American colonial societies is determined by the values and
beliefs that were dominant during the period in which they ‘split’ from the host
society, Britain. In each case, the new society bears the imprint of the values and
beliefs of Britain at the time that the colonial ‘fragment’ was established as an inde-
pendent entity.
In the United States the split from Britain occurred in the late 18th century and
the fragment that took root was characterised by the libertarian ideals of the 18th
century Enlightenment. By contrast, Australia’s split occurred in the 19th century
so that the colonial ‘fragment’ was imbued with the utilitarian ideas of Jeremy Ben-
tham and his followers. As a result, Australian political culture has been avowedly
utilitarian in its orientation. Based on the assumption that the goals of the society
were more likely to be met by widespread participation in the electoral process, the
logical corollary was to enforce voting legally. Within the utilitarian culture of Aus-
tralia, compulsory voting represented a swift, pragmatic and effective solution to
what was seen as a major political problem.14
Although it was less clearly articulated at the time, a further purpose for introduc-
ing compulsory voting was to ensure the survival of the major political parties by
removing the need for them to mobilise the vote. This had the consequence of
enabling them to devote their resources to persuasion and conversion, rather than to
motivating a reluctant electorate to go to the polls, but compulsory voting also helped
to foster the stability of the party system. This has occurred through sustaining high
levels of partisanship, itself a result of frequent voting by the vast majority of the
electorate who maintain a ready party identification. Party stability has also occurred
by ensuring that the major parties dominate the legislatures. Frequent elections and
the generational transmission of partisanship, driven by the engine of compulsory
voting, have resulted in a remarkably high level of party stability.
The major disadvantages of compulsory voting, at least in the Australian system,

14
Although various attempts were made to introduce compulsory voting in the United States, none
have succeeded. Three states (North Dakota in 1897, Massachusetts in 1918 and Oregon in 1919) passed
constitutional amendments making obligatory voting possible, subject to enabling acts of the state legis-
latures. Popular referendums approved the North Dakota and Massachusetts amendments in 1899 and
1918, respectively, but Oregon rejected it in 1920 (Abraham, 1955). Our thanks to a reviewer for this
journal for bringing these facts to our attention.
232 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233

are the high levels of invalid votes, particularly in pre-1984 Senate elections, and
the inherent advantage gained by the Labor Party and the minor parties. The high
level of informal votes is now much reduced, although in the case of the Senate this
has been accomplished by introducing ticket preferences, and this has inevitably
begun to change the character of the Senate (Farrell and McAllister, 1995). Aus-
tralia’s level of spoiled votes now compares favourably with many other countries.
For example, the level of spoiled votes in the 1996 House of Representatives election,
at 3.2 per cent, was lower than in the most recent lower house elections in three of
the four European countries with compulsory voting—Italy (7.8 per cent), Belgium
(7.5 per cent) and Luxembourg (6.5 per cent)—but higher than Greece (2.0 per cent)
(International Institute for Democracy and Electoral Assistance, 1997).
The partisan effects that result from the system of compulsory voting, particularly
the inbuilt disadvantage that it accords the Liberal–National coalition, is more diffi-
cult to deal with. The potential electoral benefits of the system were undoubtedly a
motivation in Labor’s support for the introduction of compulsory voting in the 1920s.
Equally, while the conservatives realised that they might be disadvantaged, they
judged that by removing the need for voter mobilisation, more resources would be
available to devote to voter conversion. While the contemporary debate about com-
pulsory voting is couched in terms of moral principles such as choice, responsibility
and obligation, partisan motives are never far from the surface. If the system of
compulsory voting is ever removed in Australia, it is likely to be for partisan reasons,
rather than for any philosophical objections to the system, either by voters or poli-
ticians.

Acknowledgements

Our thanks to conference participants for their comments and to two anonymous
referees from this journal for their constructive comments. The 1996 Australian Elec-
tion Study Survey was conducted by David Gow, Roger Jones and Ian McAllister
and funded by the Australian Research Council. Neither the original collectors of
the data nor the sponsoring agency are responsible for the analyses and interpretation
herein. We wish to thank Brian Hallett of the Australian Electoral Commission for
answering our questions so promptly.

References

Abraham, H.J., 1955. Compulsory Voting. Public Affairs Press, Washington, DC.
Aitkin, D., 1982. Stability and Change in Australian Politics, 2nd ed. Australian National University
Press, Canberra.
Aitkin, D., Castles, F., 1989. Democracy untrammelled. In Hancock, K. (Ed.), Australian Society. Cam-
bridge University Press, Melbourne.
Blackburn, R., 1995. The Electoral System in Britain. Macmillan, London.
Bryce, J., 1921. Modern Democracies, 2 vols. Macmillan, London.
 M. Mackerras, I. McAllister / Electoral Studies 18 (1999) 217–233 233

Collins, H., 1985. Political ideology in Australia: the distinctiveness of a Benthamite society. Daedalus
114, 147–169.
Crewe, I., 1981. Electoral participation. In: Butler, D., Penniman H.R., Ranney, A. (Eds.), Democracy at
the Polls. American Enterprise Institute, Washington, DC.
Dalton, R.J., 1996. Citizen Politics, 2nd ed. Chatham House, Chatham, NJ.
Elliott and Shananhan Research, 1989. Identifying barriers and motivations for enrolment for young
people. Elliott and Shananhan Research Report, Sydney.
Farrell, D., McAllister, I., 1995. Legislative recruitment to upper houses: the Australian Senate and the
House of Representatives compared. Journal of Legislative Studies 1, 243–263.
Goot, M., 1985. Electoral systems. In: Aitkin, D. (Ed.), Surveys of Australian Political Science. Allen
and Unwin, Sydney.
Hartz, L., 1964. The Founding of New Societies. Harcourt, Brace and World, New York.
Hirczy, W., 1994. The impact of mandatory voting laws on turnout: a quasi-experimental approach. Elec-
toral Studies 13, 64–76.
Hirczy, W., 1995. Explaining near-universal turnout: the case of Malta. European Journal of Political
Research 27, 255–272.
Hughes, C.A., 1968. Compulsory voting. In: Hughes, C.A. (Ed.), Readings in Australian Government.
University of Queensland Press, St Lucia.
International Institute for Democracy and Electoral Assistance, 1997. Voter turnout from 1945 to 1997:
a global report. International IDEA, Stockholm.
Irwin, G., 1974. Compulsory voting legislation in the Netherlands: impact on voter turnout. Comparative
Political Studies 7, 292–315.
Jackman, R.W., 1987. Political institutions and voter turnout in the industrial democracies. American
Political Science Review 81, 405–423.
Jackman, S., 1997. Non-compulsory voting in Australia? What the surveys can (and can’t) tell us. Paper
presented to the Annual Meetings of the American Political Science Association, Washington DC,
August 1997.
Joint Standing Committee on Electoral Matters, 1994. The 1993 Federal Election. AGPS, Canberra.
McAllister, I., 1985. Enrolment/population ratios. Australian Electoral Office Research Report No. 9, Can-
berra.
McAllister, I., 1986. Compulsory voting, turnout and party advantage in Australia. Politics 21, 89–93.
McAllister, I., 1997. Australia. In: Norris, P. (Ed.), Pathways to Power: Legislative Recruitment in
Advanced Democracies. Cambridge University Press, Cambridge.
McAllister, I., Mughan, A., 1986. Differential turnout and party advantage in British general elections,
1964–83. Electoral Studies 5, 143–152.
McAllister, I., Makkai, T., 1993. Institutions, society or protest? Explaining invalid votes in Australian
elections. Electoral Studies 12, 23–40.
McAllister, I., Mackerras, M., Boldiston, C., 1997. Australian Political Facts, 2nd ed. Macmillan, Melbour-
ne.
Oldham, R.C., 1911. Compulsory enrolment. Paper prepared by the Chief Electoral Officer. Common-
wealth of Australia Parliamentary Papers 2 (27), 161–162.
Overacker, L., 1952, The Australian Party System. Yale University Press, New Haven, CT.
Powell, G.B., 1986. Voter turnout in comparative perspective. American Political Science Review 80,
17–43.
Power, T.J., Timmons Roberts, J., 1995. Compulsory voting, invalid ballots, and abstention in Brazil.
Political Research Quarterly 48, 795–826.
Wattenberg, M.P., 1994. The Decline of American Parties, 1952–1992. Harvard University Press, Cam-
bridge, MA.