Consistency or Selectivity? Patterns of Participation in British Elections

Consistency or Selectivity? Patterns of Participation in British Elections
Lawrence LeDuc (University of Toronto, CANADA)
Jon H. Pammett (Carleton University, CANADA)
Increasingly, elections in Britain involve multiple levels of government. Voters are called
upon to cast votes in national (Westminster) elections, European Parliament votes, local
elections, votes for regional assemblies (Scotland and Wales), the London mayoralty, and
(in 2011) a national referendum on electoral reform. Turnout varies widely among these various
levels. In the 2010 general election, turnout for the UK as a whole was 65%, – up slightly from
the level of the previous (2005) election (61%). Turnout however in other elections in Britain has
tended to be much lower. In the 2009 European Parliament election, it was just under 35%
nationally. Local elections generally draw about the same percentage (table 1). The most recent
(2012) election for mayor of London drew 38% of eligible voters to the polls. Clearly, voter
participation in elections at these levels is substantially lower than is typical of that in
parliamentary elections at the national level.
Turnout in multi level elections in other countries generally follows similar patterns. In
Germany, turnout in the 2009 European Parliament election (43.3%) was higher than in the UK,
but slightly lower than the average for all European countries in that year. It was however much
lower than turnout in the 2009 German federal election (70.8%). In German state elections,
turnout is typically somewhat lower than in federal elections (Stock-Gissendanner, 2009). In the
recent (2012) election in North Rhine Westphalia for example, turnout was 59.6%. Canadian
provincial elections have historically had turnout levels which are sometimes (but not always)
lower than those found in federal elections (Wesley, 2010; LeDuc and Pammett, 2010). In the
most recent (2012) provincial election in Ontario for example, turnout was 49.2%, while in the
federal election held the previous year turnout in Canada’s largest province was 61.5%. In some
provinces (e.g. British Columbia, Manitoba) turnout in provincial elections has often tended to
be about the same as that in federal elections, while in a few others (e.g. Newfoundland,
Saskatchewan, New Brunswick), provincial turnout has historically been higher than in federal
elections – a pattern which Studlar (2001) argued made Canada an exception in comparisons
with similar elections in other federal countries. In a study of the 2003 Ontario provincial
election, Cutler (2006) found that it exhibited most of the characteristics commonly associated
with first order elections. A similar argument has been advanced by Wolinetz and Carty (2006)
regarding Canadian provincial elections more generally.
As in Britain, turnout in local elections is generally lower than in elections at other levels,
but it also can be highly variable. In Germany, the mean turnout for mayoral and council
elections in North Rhine Westphalia was just over 59% – about the same as in the recent landtag
election – but it varied from a low of about 45% in some jurisdictions to instances registering as
high as 80% (Stock-Gissendanner, 2009). In Canada, turnout in Ontario municipal elections has
typically varied from the mid-30% range to just over 50% in some instances (Kushner et al,
1997). In the most recent (2010) mayoral election in Toronto, turnout was 53.3% – a substantial
increase from the 39.3% registered in the previous (2006) election in that city.
1
There is a tendency to think about turnout in elections at these different levels in a
unidimensional manner. More voters will turn out for a general election and pass up the
opportunity to vote in a local or regional contest. But, although the aggregate figures are
suggestive, it is entirely possible that there are voters who choose levels of the electoral system
at which to participate, and that some may opt to participate at one level rather than another. In
other words, if electorates at these different levels of government are not the same, it should not
be assumed that the electorate in a provincial or local election level is simply a subset of the
national electorate. To date, this has been a difficult hypothesis to test, because election studies
typically focus on the single level to which a particular study is directed. The BES will examine
voting behaviour primarily at the national level, European Election Studies at the European level,
and so on. Questions are rarely asked in any survey about voting in local elections. While some
studies may include recall questions about voting in a previous election, it is difficult to collect
data on participation at different levels in a single survey because elections occur at different
times and most large scale election studies are oriented mainly toward national politics.
In May 2011, we sponsored questions in the BES to probe the voting history of
respondents in the previous European, regional and local elections, as well as the 2011 AV
referendum and prior parliamentary elections.1 A parallel study using a similar question set was
conducted at the time of the 2011 Canadian federal election (Pammett and LeDuc, 2012). The
2011 elections in Britain provided a unique opportunity to examine voting at different levels
because they combined a vote in a national referendum with regional legislative elections in
Scotland and Wales and local elections in most jurisdictions throughout England. These data
enable mapping the types of British voters/nonvoters by those who voted in all of these elections,
voted in none, or voted in only some of those for which they were eligible. Such mapping is
facilitated by the size of the 2011 BES sample (N=22134). The survey also included a sequence
of questions (table 2) asking respondents to rate the importance of voting at each of these various
levels. The issue of selectivity of level of electoral participation is the particular focus of this
paper. As a complementary question, we also consider the decision made by first time voters
with respect to level of participation. Will such voters choose to participate at their first voting
opportunity, or will they select the level at which they wish to be engaged?
Surveys of course tend to over report voting participation, and recall questions are
especially prone to such distortion. The BES is no exception to this tendency. However, the large
number of cases in the 2011 BES provides a greater opportunity to track the movement of nonvoters than is available in most election surveys. While voting may be over reported, non-voting
reports are better trusted. Further, a report of having voted in an election at one level combined
with a report of not having voted at a different level is more likely to reflect the respondent’s best
recall of those events. The marginals for voting reports in the various elections inspire some
confidence in the data, as these accurately reflect the directional differences in turnout at the
different levels, even as they may overestimate voting more generally. In other words, a
respondent is more likely to report not voting in the 2009 European Parliament election than in
1
We are grateful to Harold Clarke, David Sanders, Marianne Stewart and Paul Whitely for the opportunity to insert
these items into the May 2011 wave of the BES, and to Joe Twyman for his assistance in formatting the questions.
We also acknowledge the support of the Social Sciences and Humanities Research Council of Canada for the larger
project on which this research is based.
2
Table 1. Turnout in British elections: 2003-2011
%
All UK
England
Scotland
Wales
London
2011
AV referendum, legislative (Scotland, Wales), local councils (England)
42.2
42.6
50.7
41.7
35.4
2010
Westminster parliament, local councils (England)
65.1
65.5
63.8
64.7
64.5
2009
European parliament, local councils (England)
34.7
39.1
28.5
30.4
33.5
2008
Local councils (England, Wales), London (mayor, assembly)
—
34.8
—
43.1
45.3
2007
Legislative (Scotland, Wales), local councils (England, Scotland)
—
36.7
52.4
43.3
—
2006
Local councils (England)
—
36.5
—
—
—
2005
Westminster parliament, local councils (England)
61.4
61.3
60.8
62.6
57.8
2004
European parliament, local councils (England, Wales), London (mayor, assembly)
38.5
38.8
30.9
41.4
37.0
2003
Legislative (Scotland, Wales), local councils (England, Scotland)
—
35.6
49.4
38.2
—
Sources: Electoral Commission [www.electoralcommission.org.uk]; UK Political Info [www.ukpolitical.info]; European Parliament
Information Office in the UK [www.europarl.org.uk/]; Plymouth University LGC Elections Centre [www.plymouth.ac.uk/elections]
3
Table 2. Sample question sequences on vote history and importance of voting at different
levels – May 2011 BES2
Thinking back to some past elections, please indicate which of the following elections you voted
in [for each election, please tick the box to the right of it that applies to you]:
Definitely voted
Did not vote
Not eligible then
Not sure/
Don't know
The 2005 British
General Election
The 2007 Election
for the Scottish
Parliament
The 2009 Election
for the European
Parliament
The most recent
local election
People are called upon to vote in elections at different levels of government - European, national,
regional, local, and sometimes referendums. Thinking generally and not about any specific
election, please indicate how important it is to you personally to vote in elections at that level, on
a scale of 1 to 7, where 1 is not at all important and 7 is very important [for each type of election,
please tick that box to the right of it that best represents your view].
1 Not at all
important
2
3
4
5
6
7 Very
important
Don't know
British general
elections
Elections for the
Scottish
Parliament
Elections for the
European
Parliament
Local elections
Referendums
2
Pre-election wave. Questions about voting in the AV referendum (and coincident elections) and the 2010 general
election were asked separately as part of the main study. The question sequence shown here for Scotland was
adjusted appropriately for respondents in England or Wales.
4
the 2010 general election, even if both in the aggregate are over reported. Nevertheless, to
provide a more accurate picture of the electorate surveyed here, we apply a corrective weight
based on the 2010 reports.3 We thus begin with a sample weighted to reflect the actual turnout in
the 2010 general election.
Voting in three elections
The tendency in Britain in recent years to schedule elections at multiple levels on a single
date makes our task somewhat more complicated. The May 2011 referendum also involved
elections for local councils in most parts of England as well as elections for the Scottish
Parliament and Welsh Assembly (table 1). Likewise, voters in many parts of England in the 2010
general election were also able to vote at the same time for local councils. In 2009, English local
elections were postponed for a month so that they could be scheduled in tandem with the vote for
members of the European Parliament, which took place throughout the entire country. Clearly,
this practice has the ability to boost turnout to levels that are at least slightly higher than they
likely would be if the elections were held separately. The turnout of 42% in the 2011 AV
referendum, for example, was about six percentage points higher than the average turnout in
local elections in Britain, suggesting that some potential non-voters were attracted to the polls by
the prospect of voting in the referendum. The variation can also be seen in the higher turnout in
Scotland, where referendum voters had the additional opportunity to elect members of the
Scottish Parliament. In London, where only the referendum question was on the ballot, turnout
was significantly lower than in other parts of the country.
In spite of this consolidation of elections, there is still sufficient variation across time and
place to permit some analysis of selective voting participation. Turnout in the 2010 general
election (65%) was substantially higher than turnout at any of the other levels, and varied less
among different regions of the country (table 1). Turnout in Scottish Parliament elections tends
to be consistently around 50% of the electorate as evidenced by the last three elections for that
body, even though each of these was combined with one other electoral task. Turnout in Welsh
Assembly elections tends likewise to be consistent over time at about 40%. Turnout in European
Parliament elections is lower, even when these are combined with local elections as in 2009 and
2004. When they are not, as was the case in Scotland and Wales in 2009, turnout declines even
further. Thus, combining local or European elections with those at a “higher” level tends to pull
turnout up towards that of the higher level. When these elections occur separately, either in a
particular year or a particular region, turnout reverts toward the mean of that level. We thus have
enough variation across time, region and levels to provide an initial picture of electoral migration
between these different types of electoral contests.
Table 3 displays the pattern of responses with respect to voting in the first three elections
considered here – the 2011 AV referendum (and coincident elections), the 2010 general election,
and the 2009 European Parliament election. Slightly less than half of the post-election wave
3
Reported non-voters in 2010 are assigned a weight of 4.058 while voters are weighted at 0.712. Those not eligible
to vote in 2010, either by reason of age or citizenship, are assigned a weight of 1.0. Non eligibility is based on both
self reports and year of birth. For purposes of this analysis, those answering “don’t know; not sure” are counted as
non voters.
5
Table 3. Reported voting in the 2011 AV referendum, 2010 general election and 2009
European Parliament election
% voted in …
All Britain
England
Scotland
Wales
All three elections
44.9
43.9
53.3
46.0
National, EP
3.4
3.6
2.1
2.1
National, AV*
12.7
12.8
12.4
10.5
EP, AV*
3.7
3.6
4.8
3.3
National only
4.8
5.2
2.1
3.9
EP only
2.0
2.1
1.0
2.1
AV* only
6.3
6.2
6.8
7.5
None
22.2
22.6
17.6
24.6
18196
15544
1682
970
**N=
* Includes voting in other 2011 elections scheduled concurrently.
** Weighted N (see footnote 2). Respondents eligible to vote in all three elections only.
…………………………………………………………………………………………………
respondents (45%) report voting in all three of these elections. The number is higher in Scotland
(53%), likely because of the simultaneous scheduling of the Scottish Parliament election, which
has typically attracted about this level of participation. About one in five respondents (22%)
indicated that they did not vote in any of the three elections. As might be expected based on the
aggregate turnout figures, a substantial number of those who reported voting in the referendum
and the 2010 general election (12.7% nationally) indicated that they did not vote in the elections
for members of the European Parliament in 2009.
Lest we think however that voters simply decided to skip the EP election, there are small
subsets of the sample who voted in that contest but not in one of the other two (3.4% and 3.7%
respectively in table 3). And there is even a small component (2.0%) who report voting only in
the European contest. While such small numbers may simply be indicative of random fluctuation
or personal reasons for not voting in a particular election, they may also suggest a degree of
selectivity with regard to the level at which a voter feels that it is worthwhile to participate. This
is the more likely explanation for the slightly larger group (4.8% nationally) who indicated that
they voted only in the 2010 general election and did not participate in the other two contests.
More surprising however are the 6.3% of the sample who indicated that they voted in the 2011
AV referendum, but did not participate in either the 2010 general election or the 2009 European
6
Table 4. Importance of voting in elections at different levels, by voting participation
in three elections
[means scores on seven point scale: 7 = “very important”]
Voted in …
National
European
Regional*
Local
Referendums
All three elections
6.8
6.1
6.7
6.6
6.6
National, EP
6.5
5.4
6.1
6.1
6.0
National, AV**
6.6
4.0
6.2
6.0
6.1
EP, AV
6.3
5.5
6.4
6.0
6.2
National only
6.1
3.7
4.9
5.1
5.0
EP only
5.6
5.0
5.6
5.1
5.4
AV only
5.3
3.7
5.1
4.7
5.3
None
3.9
2.9
3.0
3.4
3.5
All respondents
6.1
4.9
5.9
5.7
5.8
England
6.1
4.9
—
5.7
5.8
Scotland
6.2
5.1
6.2
5.9
5.9
Wales
6.0
4.8
5.4
5.3
5.5
* Asked in Scotland and Wales only.
** AV includes voting in other 2011 elections scheduled concurrently.
…………………………………………………………………………………………………
vote. While this may be partly due to the fact that the survey was being conducted at the time of
the 2011 vote, it may also be suggestive of the unique attraction of the opportunity to vote in a
national referendum on electoral reform. Given that the increase in turnout in the referendum
was about six percent over that of a typical local election in Britain, we may here be identifying a
small group of voters who specifically selected the referendum as a worthwhile voting
opportunity.
Further light can be shed on the issue of selectivity by the question asked in the 2011
BES regarding the perceived relative importance of voting in elections at different levels. As is
seen in table 4, there is some indication that voters are at least partly strategic with respect to
where they decide to become engaged. Those who vote in most or all elections tend to rate all
levels as “important”. Those who do not participate in elections at all rate voting as less
important regardless of level. However, the rankings begin to diverge for more selective groups
of voters. Those who participate in European Parliament elections rate them more important than
those who do not, even though the rankings for this level of elections tend to be lower overall.
Those who vote only in national elections believe these to be “important” and simultaneously
7
rate voting at other levels lower. Respondents in Scotland believe voting in Scottish Parliament
elections to be as important as in national elections, while Welsh respondents rate voting for
members of the Welsh Assembly as somewhat less important but still higher than European
elections. The European votes are ranked as relatively unimportant by most respondents in
comparison to all other levels. One might well surmise that the consistently low rate of
participation by British voters in European elections is in part due to the widespread feeling that
these elections are simply not very important. Combining them with elections at other levels may
well boost turnout somewhat, but it does little to change the perception of them as second (or
third) order contests.
Potential outcome effects
If there are voters who are selective with respect to level, it may be because they believe
that casting a vote at a particular level may produce a more desirable outcome at that level.
Liberal Democrat voters, for example, might be more inclined to participate at the European
level because of proportional representation at that level. Similarly, LD voters might have been
attracted to voting in the AV referendum because the adoption of AV would benefit their party,
or Conservative voters may have turned out in greater numbers in an attempt to defeat it. Table 5
indicates the vote choice in the 2010 general election and the AV referendum for each of the
distinctive participation patterns shown in table 3. The Conservatives fared slightly better among
those voting in all three elections and also among those voting in the 2010 election and the
referendum but not in the EP election. Overall, however, there is little evidence that selectivity
on the part of voters has had much effect on election outcomes. An analysis by party
identification (data not shown) finds that Liberal Democrat identifiers were slightly more likely
to report voting in all three elections or in the AV referendum, and they were of course more
likely than adherents of the other parties to vote YES in the referendum. However, the analysis
suggests that these variations in participation patterns had little overall effect on the outcome of
the referendum or the 2010 election. The percentage voting YES on AV varies only slightly
among the different groups of consistent or selective voters shown in table 5.
…………………………………………………………………………………………………
Table 5. Direction of vote in the 2010 general election and 2011 AV referendum by
participation patterns in three elections
Voters in… [ N ]
… vote in 2010 national election (%)…
Cons
Labour
Lib-Dem
Other
AV referendum
All three elections [ 8091 ]
38.0
26.9
22.8
12.2
39.6
National, EP [ 610 ]
35.6
29.5
23.3
11.6
—
National, AV [ 2274 ]
38.2
25.4
26.5
9.9
36.5
National only [ 858 ]
35.2
26.7
29.8
8.3
—
EP, AV [ 653 ]
—
—
—
—
37.2
AV only [ 1100 ]
—
—
—
—
39.5
% YES
8
Voting at four levels
While we are able to examine participation in three different elections using the data
shown in tables 3-5, the inclusion of the 2011 AV referendum introduces an unwelcome
complication. First, the referendum was a one-off event – not strictly comparable with elections
at other levels at other times. Taking place throughout the entire country, the referendum might
be thought of as a “national” contest, but it is clearly different than a general election and
exhibits more of a second order quality. The fact that it was held concurrently with local
elections in most parts of England makes it also in some respects an adjunct to the local contest.
And, the concurrence of this election with votes for legislative assemblies in Scotland and Wales
again changes the perception that voters may have of the relative importance of this election in
those regions. The additional recall data collected in the pre-election wave of the 2011 BES can
give us some further insights into the behaviour of voters in elections at the different levels.
While most elections in recent years have involved some combination of levels, contests in
previous years have contained other permutations than those of 2011. The 2007 legislative
assembly elections in Wales did not involve a vote at other levels, while that in Scotland
included only local contests. The 2006 and 2008 elections, where these took place, were entirely
local. Further, by including a larger number of electoral contests in the survey question, we are
also able to better assess consistency of voting. Will those who report voting in all three of the
elections shown in table 3 also report having voted in earlier elections at the same or at other
level(s)? Or might greater selectivity emerge as different sets of choices are presented?
In table 6, we include all of the elections covered by the recall questions for those
respondents who were eligible to vote in all of those contests. The percentage that report
consistent voting changes relatively little, even as more elections are included in the analysis. It
is clear that there is a core of voters whose participation over time and across level is fairly
consistent. However, there is also a modest but persistent degree of selectivity, particularly with
regard to national elections and also extending to the 2011 AV referendum. There also appears to
be a small amount of selectivity in Scotland and Wales with regard to regional legislative
elections, although these numbers are too small to inspire confidence. Selectivity with respect to
European or local elections begins to disappear when more elections at multiple levels are
included in the mix. In other words, there appear to be few voters who participate only in
European Parliament elections or in local contests exclusively over any extended period of time.
A clue to the explanation of consistency in voting is found in the responses to items in the
survey dealing with attitudes toward the act of voting. Those who vote consistently in elections
at all levels are more likely to feel that they have a “duty” to vote, to feel satisfaction in the act of
voting, and to express feelings of “guilt” in not voting (table 7). They are also more likely to pay
attention to politics and to hold a party identification that is at least “fairly strong”. None of these
patterns is surprising, but they suggest that it is the attitude toward the act of voting generally
that is driving behaviour rather than attention to the level of the election or the choices presented.
9
Table 6. Reported voting in combinations of six previous elections*
% voted in …
All Britain
Scotland/Wales
All elections*
42.3
45.8
National only
4.8
3.4
National, AV
3.8
3.6
Local, AV
3.8
3.9
AV only
3.0
3.6
EP only
0.2
0.3
Local only
0.4
0.2
Regional only
—
1.8
Regional, AV
—
3.3
16.8
14.3
17431
2569
None
**N=
* Seven in Scotland and Wales. Elections include the 2011 AV referendum (and coincident elections),
the 2010 and 2005 UK general elections; 2009 and 2004 European Parliament elections, a previous local
election, and the 2007 legislative elections in Scotland and Wales. See question format in table 2.
Respondents eligible to vote in all elections only.
…………………………………………………………………………………………………
Nevertheless, we have established that there is a degree of selectivity in the electorate,
although smaller than the wide variations in turnout between elections at different levels might
suggest. This is most visible with respect to national politics. In both the analysis of the last three
elections (table 3), and in the longer term pattern (table 6), there exists a group of voters whose
orientation is specifically to national elections. This may account for the somewhat higher
turnout in the AV referendum, which was after all a national event even though it took place
concurrently with local elections in England and legislative assembly elections in Scotland and
Wales. There is also a somewhat smaller group of voters, particularly in Scotland, who orient
themselves specifically to regional elections. This may also partly account for the higher turnout
in Scotland in the AV referendum, which was there combined with regular elections for
members of the Scottish Parliament. No such groups however appear to exist for other levels,
such as local elections or those for members of the European Parliament. These elections appear
to be genuinely “secondary” in the minds of voters, and they tend to be thought of as less
important. When elections at these levels stand alone, and are not held in combination with those
at a “higher” level, turnout tends to fall off sharply.
10
Table 7. Attitudinal items, by frequency of voting
# elections voted: 2004-2011
0
1
2
3
4
5
6+
Duty to vote (% str. agree)
1.0
3.8
10.8
22.1
35.1
47.6
60.9
Feel guilty not vote (% str. agree)
0.1
0.9
4,2
9.6
18.4
28.7
42.0
Feel satisfaction if vote (% str. agree)
0.3
1.5
3.7
7.6
12.7
19.6
26.8
Weak or no party identification (%)
64.5
61.7
52.7
48.9
43.5
34.1
29.9
Attention to politics (mean / 10)
3.5
4.6
5.1
5.6
6.2
6.6
7.0
2686
1267
1017
1205
1953
1466
7533
N=
…………………………………………………………………………………………………
Young voters
A second objective of this paper is to examine the influence that the level of the election
may have on new voters as they reach the age of eligibility to vote. For those who came of voting
age for the first time in the year of a general election (e.g. 2005 or 2010), the decision to vote or
not would appear relatively straightforward. However, for those whose first voting opportunity
occurred in the year of a European Parliament election (2004, 2009) or (in Scotland and Wales)
those for regional assemblies (e.g. 2007), the decision may be quite different. Some newly
eligible voters may decide to take the first voting opportunity presented, regardless of level,
while others may wait for a more “meaningful” opportunity to cast their first vote.
There is considerable variance in the voting opportunities presented to new voters over
the 2004-11 period. Those who attained voting age in the period leading up to the 2004 European
Parliament election had the opportunity to vote in that election, but our data indicate that fewer
than one in three did so (table 8). More of these voters waited until the general election of the
following year to cast their first vote, with slightly more than half of this cohort participating in
that contest. A similar pattern is found among the new voters coming of age in the period prior
to the 2009 European Parliament election. Interestingly however, voting participation in the 2010
general election also tends to be somewhat lower for those who came of voting age during the
period when only the opportunity to participate in a electoral contest at a different level was
available. In other words, a longer waiting period for the opportunity to participate in a general
election appears to also reduce voting. This first voting opportunity may be of considerable
importance if it represents the first act in developing a “habit of voting” which, once established,
might carry forward into subsequent elections (Franklin, 2004). A consistent pattern of voting,
established early, might also carry forward to elections at the non-national levels. Voting at the
next opportunity at a different level would then have a reinforcing effect on electoral
participation. Alternatively, a long wait between voting opportunities that are considered
meaningful might impair future participation at all levels.
11
Table 8. Reported voting of newly eligible voters: 2004-2011
Year of first eligibility ……
% voting in …
First electoral opportunity
2004
2005
2006
2007
2008
2009
2010
E
G
L
L R*
L
E
G
2004 EP election (E)
28.8
2005 general election (G)
52.6
55.2
2007 legislative assemblies (R)*
45.0
47.4
50.0
25.0
2009 EP election (E)
32.1
38.9
26.3
29.6
30.5
27.9
2010 general election (G)
45.1
55.2
40.4
34.9
44.9
42.4
57.1
—————————————
Any local election (L)
43.2
41.4
34.3
38.8
48.4
46.4
50.0
2011 AV referendum
42.5
44.6
36.2
52.2
53.3
37.0
68.2
239
198
188
180
98
92
28
N=
* Scotland and Wales only.
…………………………………………………………………………………………………
Some of the cohorts of young voters entering the electorate during the 2006-2009 period
did appear to seize the opportunity to participate in the various local elections which occurred
during those years, although the pattern at that level is mixed. For those entering the electorate in
2008 or 2009, when there were also elections for the London mayoralty and European
Parliament respectively, rates of participation in the local elections4 are reported at somewhat
higher levels (table 8). For the cohort entering in 2006 however, when the only elections
available were for English local councils, reported participation is lower. Nevertheless, for each
of these cohorts, reported participation in local elections is consistently higher than that reported
for the European level, even for those cohorts entering the electorate for the first time in a
European election year.
The pattern appears quite different for those cohorts who came of voting age for the first
time in the period leading up to a general election (2005 or 2010 in table 8). Here, reported
participation in the general election is much higher – above 50% for both the 2005 and 2010
elections. Those whose first voting opportunity occurs in a general election year are therefore
more likely to take the opportunity to participate, perhaps because of the greater importance
assigned to elections at the national level more generally or because of the enhanced
mobilization activities which tend to take place in general election years. In recent years, many
of these types of campaigns have been particularly directed at young voters.
4
The question asked referred to “the most recent local election”, and did not specify either time or place.
See table 2.
12
The higher voting rates of new voters coming of age in general election years seems also
to affect their participation in subsequent elections, including those at other levels. For the 2005
cohort, the reported voting rate for the 2005 general election (55%) is replicated at the 2010
election five years later. But the reported voting participation of this cohort in the European
election of the previous year is also higher than that found among the other groups of new voters.
This suggests that the “habit of voting”, once established, can and will carry over to participation
at other levels. But the habit must be initiated somewhere, and the evidence implies that it is
more likely to be established in a general election setting than at one or more of the various other
levels.
Within some constraints imposed by the nature of the 2011 BES data, the “habit of
voting” hypothesis can be tested more explicitly by comparing the subsequent behavior of voters
and non-voters within several of the electoral periods covered by the recall data. As is seen in
table 9, the behaviour of new voters, once established in either a national election or one at
another level, is likely to carry through to subsequent elections at both the same and other levels.
The clearest test is provided by the cohort of voters who came of age at the time of the 2004
European Parliament election. That election of course coincided with local elections in England
and Wales and with the London mayoral election. Nevertheless, voting participation for the
country as a whole was only 38.5% (table 1). But for those new voters who did participate in that
election, their voting rate in subsequent elections at every level remained high, including the
2009 European election and the AV referendum (table 9). For those who passed up that first
opportunity to vote however, subsequent participation remained low, including a rate of only
35.1% in the general election of the following year. For 2004 EP voters, all reported voting in at
least one election of the following years, compared to 42.1% of 2004 non-voters who indicated
…………………………………………………………………………………………………
Table 9. Reported voting of newly eligible voters, 2004-2011 by vote/non vote in first
election of eligibility
Voters/non voters in election held during first year of eligibility…
2004
2005
2006-08
% voting in subsequent …
V
NV
V
NV
V
NV
2005 general election
95.0
35.1
2007 legislative assemblies*
92.8
30.0
85.7
10.0
2009 EP election
76.7
12.2
59.4
10.3
62.7
9.1
2010 general election
80.0
33.3
61.9
17.7
87.0
2011 AV referendum
78.2
Any local election
86.7
30.4
69.7
11.6
23.8
56.7
19.2
% not voting in another election
—
42.1
9.2
N=
60
147
97
2009
V
NV
10.9
83.3
21.0
81.2
21.6
82.3
19.2
85.5
20.4
85.6
17.5
51.0
—
64.7
—
37.2
79
125
220
25
62
13
that they did not vote in any subsequent election. This pattern is repeated for every subsequent
cohort of newly eligible voters. Voting in the first election available tends to forecast continued
electoral participation, while non-voting make future participation less likely. Further, with
respect to this pattern of continued participation, level is almost irrelevant, even though new
voters are more likely to exercise the first opportunity to vote if it occurs in a general election
year.
Conclusion
Turnout in the AV referendum was higher than would have been the case had the 2011
election been merely a vote for local councils. It was higher in Scotland because of the
coinciding Scottish Parliament election, but it may also have been boosted elsewhere because
some voters “selected” the AV vote as more worthwhile than other non-national elections such
as elections for members of the European Parliament or other local contests. There is also some
evidence of selectivity of participation with respect to votes for regional assemblies in Scotland
and Wales, and significantly more for national elections, which consistently generate higher
turnout than elections in Britain at any other level. Unlike evidence from Canada, where we have
conducted similar analyses using recall data (Pammett and LeDuc, 2012), selectivity of electoral
participation in Britain runs mainly in one direction.
The largest components of the British electorate tend to be either consistent voters or
consistent non-voters. Over time, those who vote in national elections will also participate in
other elections to varying degrees. There are however voters who will participate only in national
elections. To the extent that the AV referendum was a national event, it appears to have also
attracted some of these. European or local elections do not seemingly have the capability to
attract selective voters, but those for regional assemblies may have a modest degree of ability to
do so, as also do some provincial elections in Canada.
Although European or local elections do not appear to have the ability to attract and
maintain a distinctive cohort of selective voters, turnout in these will fluctuate based on the entry
and exit of voters who do participate in elections at other times and at other levels. In previous
Canadian work, we have referred to participants of this kind as “transient voters”, because they
move into and out of the electorate fairly easily, sometimes for reasons that have relatively little
to do with political motivation (Clarke et al, 1980; 1996; LeDuc and Pammett, 2010). For the
most part, these voters are little different than those who vote more consistently or more
frequently, although their entry into or exit from the active electorate may have some effect on
electoral outcomes, depending on the specific short term factors that motivate their participation.
But they are not a distinct or “selective” electoral cohort.
Newly eligible voters entering the electorate in a year other than a general election year
face the dilemma of exercising their first vote in an election perceived to be of less importance or
waiting until the next general election. Fewer will take the opportunity to vote in such an election
than will do so in a national election year, but those that choose to vote in the “lesser” contest
establish a habit of voting that has the potential to carry through to other levels and maintain
itself over time. This is also true for new voters in the general election years. However, failure to
14
establish the habit of voting early on tends to reduce future participation at all levels. Thus, the
typical four or five year time span between national elections may have the effect of further
diminishing the already low voting turnout of the young.
The increasing tendency in Britain to schedule different levels of elections concurrently
causes turnout to vary at a particular level more than it otherwise might, as turnout is affected
mainly by the higher level. However, this is mainly an administrative measure to save money
rather than a genuine stimulus to voting participation. British citizens have a rather clear
understanding of the ordering of importance of elections at different levels, and they almost
universally believe national elections to be “important”. Elections at other levels, and particularly
European elections, are consistently ranked lower in importance, and turnout tends to follow
these rankings. There are other discernible effects that come into play, such as the model of
electoral system employed at each level or the occasional inclusion of a referendum question on
the ballot. But there is mainly one British electorate, that is selective with respect to national
politics but transient in its participation at other levels.
15
References
Bakvis, Herman (ed.) 1991. Voter Turnout in Canada. Toronto: Dundurn
Blais, André. 2006. “What Affects Voter Turnout?” Annual Reviews of Political Science
9: 111-25.
Clarke, Harold D., David Sanders, Marianne Stewart and Paul Whiteley. 2004. Political
Choice in Britain. Oxford University Press.
Clarke, Harold D, Jane Jenson, Lawrence LeDuc and Jon H. Pammett. 1980. Political Choice in
Canada. Toronto: McGraw-Hill Ryerson.
Clarke, Harold D, Jane Jenson, Lawrence LeDuc and Jon H. Pammett. 1996. Absent Mandate:
Canadian Electoral Politics in an Era of Restructuring. Toronto: Gage.
Cutler, Fred. 2008. “One Voter, Two First-Order Elections?” Electoral Studies 27: 492-504.
Cutler, Fred and J. Scott Matthews. 2005. “The Challenge of Municipal Voting: Vancouver,”
Canadian Journal of Political Science 38: 359-382.
Eijk, Cees van der, Mark Franklin and Michael Marsh. 1996. AWhat Voters Teach Us About
Europe-Wide Elections: What Europe-Wide Elections Teach Us About Voters,@ Electoral
Studies 15: 149-66.
Franklin, Mark N. 2004. Voter Turnout and the Dynamics of Electoral Competition in
Established Democracies Since 1945. NY: Cambridge University Press.
Heath, Anthony, Ian McLean, Bridget Taylor and John Curtice. 1999. “Between First and
Second Order: a Comparison of Voting Behavior in European and Local Elections in Britain,”
European Journal of Political Research 35: 389-414.
Hough, Dan and Charlie Jeffrey (eds.) 2006. Devolution and Electoral Politics. Manchester
University Press.
Jeffrey, Charlie and Dan Hough. 2009. “Understanding Post-Devolution Elections in Scotland
and Wales in Comparative Perspective,” Party Politics 15: 219-240.
Johnston, Richard. 1980. “Federal and Provincial Voting: Contemporary Patterns and
Historical Evolution,” in David J. Elkins and Richard Simeon, Small Worlds: Provinces and
Parties in Canadian Political Life. Toronto: Prentice-Hall.
Kushner, Joseph, David Siegel and Hannah Stanwick. 1997. “Ontario Municipal Elections:
Voting Trends and Determinants of Electoral Success in a Canadian Province,” Canadian
Journal of Political Science 30: 539–553.
16
LeDuc, Lawrence. 2003. The Politics of Direct Democracy: Referendums in Global
Perspective. Toronto: Broadview Press.
LeDuc, Lawrence, and Jon H. Pammett. 2010. “Voter Turnout,” in Heather MacIvor (ed.),
Election. Toronto: Emond Montgomery.
Marsh, Michael. 1998. “Testing the Second Order Election Model After Four European
Elections,” British Journal of Political Science 28: 591-607.
Milner, Henry. “Electoral Systems, Integrated Institutions and Turnout in Local and National
Elections,” Canadian Journal of Political Science 30: 89-106.
Nakhaie, M. Reza. 2006. “Electoral Participation in Municipal, Provincial and Federal
Elections in Canada,” Canadian Journal of Political Science 39: 363-390.
Pammett, Jon H. and Lawrence LeDuc. 2012. “Consistency or Selectivity? Patterns of
Participation in Canadian Elections,” paper to be presented at the annual meeting of the
American Political Science Association, New Orleans.
Saint-Germain, Maurice and Gilles Grenier. 1994. “Le parti québécois, le ‘NON’ B
Charlottetown et le bloc québécois: est-ce le mLme électorat?” Revue québécoise de science
politique 26: 161-78.
Stock-Gissendanner, Scott. 2009. “Voter Turnout in German Cities,” paper presented at the
ECPR General Conference, Potsdam.
Studlar, Donley T. 2001. “Canadian Exceptionalism: Explaining Differences Over Time in
Provincial and Federal Voter Turnout,” Canadian Journal of Political Science 34: 299-319.
Wesley, Jared. 2010. “Slack in the System: Turnout in Canadian Provincial Elections, 19652009,” paper presented at the annual meeting of the Canadian Political Science Association,
Montreal.
Wolinetz, Steven and R. Kenneth Carty. 2006. “Disconnected Competition in Canada,” in
Hough and Jeffrey (eds.), Devolution and Electoral Politics
Prepared for presentation at the EPOP annual conference.
Oxford, Sept. 7-9, 2012
17