Four observations on electoral reform

Peter Loewen, PhD

Director, School of Public Policy and Governance
Associate Professor, Political Science
University of Toronto
In this short note, I make four observations. The overall objective of these observations is to induce some
pause among members of the committee and their colleagues. I hope that you would reflect on and give
equal weight to the known benefits (and drawbacks) of our current system as you do the known and
unknown benefits and drawbacks of other systems.
My observations are the following:
1.
2.
3.
4.

There is a potential upside to electoral reform, but it seems limited.

The downsides to electoral reform are unknown and potentially substantial.
Canadian democracy already functions well.
For most of the problems ailing our democracy, there are potential fixes at hand which do not
require fundamental institutional change.

Taken together, these observations suggest that the committee should not engage in wholesale reform of
our electoral system. Instead, I argue, it should consider and recommend smaller, targeted reforms which
might address the problems which currently beset our political system.
Observations
1. Potential upside to a change in our electoral system
The balance of cross-national evidence suggests that changing our electoral system would have modest
effects on some aspects of our political life namely voter turnout and perceptions of fairness. It is less
clear that it would eliminate incentives for strategic voting, a common strike against our current system. It
is equally unclear that electoral reform towards a PR system would increase the match between actual
government policy output and citizens preferences. I take these conclusions from the evidence given in
Andr Blais submission, which I found the most thorough, convincing, and evenhanded. By Blais telling,
shifting to a PR system will likely have the following effects, net the effects of several other relevant
factors:

Voter turnout will be on the order of 3 percentage points higher in a PR system.

Strategic voting the abandoning of ones preferred choice to vote for another party for tactical
reasons will be as frequent. However, its nature may change.
Citizens will express greater sense with the fairness of an election. However, their overall
democratic satisfaction will not be enhanced.
Governments will not better reflect the average policy preference of voters, though there will more
more representation of diverse viewpoints in Parliament.

Taken together, these estimates suggest that there is some upside to electoral reform. It is, however,
modest and not nearly as sweeping as advocates might claim.
What Blais does not note, Leslie Seidle and others have in their presentations: electoral reform would
likely increase gender balance in our Parliament. This is an unalloyed good.
Of course, advocates of PR electoral systems might argue that such studies somehow underestimate the
salutary effects of PR. I think it is a reasonable objection that cross-national, econometric estimates do
not tell the whole story. A common alternative approach is to point to several laudatory features of various
countries, picking some from one country and some from another. Such evidence should be regarded

sceptically, especially as it is often not selected in a systematic and unbiased manner. Even worse, it
occasionally ignores that beneficial features and outcomes in various countries are mutually incompatible.
A reasonable alternative approach is to look at another country which is most similar to our own and
which has experienced a change in electoral systems. By observing the pre-reform and post-reform
averages on several outcomes of interest, we can say something about how electoral reform might
change the politics of a country.
New Zealand provides such a case, and for obvious reasons. It shares a common colonial heritage with
Canada. It has a long history of uninterrupted democratic rule, with power alternating between a small
number of single parties who regularly commanded majority governments. In 1996, after a series of
referendums, New Zealand moved to an MMP system. Seven elections have been held under this
system.
To generate an estimate of the effects of electoral reform in New Zealand, Ive compared the seven
1
election averages before and after reform . Four outcomes of interest have been calculated:

The effective number of legislative and parliamentary parties (Lasko and Taagepera 1979).
The average number of parties in government.
3
Official voter turnout.
4
Citizen assessments of satisfaction with democracy before and after the reforms.

Summary statistics for each of these measures for each available year can be found in Table 1, below.
What changed after New Zealands electoral reform? There was a relatively impressive growth in the
number of effective parties. The average number of effective electoral and legislative parties (i.e. votes
and seats) before reform was 2.9 and 2.0, respectively. After reform, these numbers of grown to 3.6
effective electoral parties and 3.2 effective legislative parties. There was likewise a growth in the average
number of parties in government, all attributable to the two party minority coalitions negotiated in the first
four parliaments following reform. However, in the last two elections these have given way to single party
minority governments which survive on supply arrangements. In sum, there is evidence that reform in
New Zealand has broadened the number (and indeed the ideological range) of parties in their Parliament.
It was, in a more limited way, increased the formal influence of small parties on governments.
What did not change after reform in New Zealand? Turnout certainly did not increase. The mean turnout
in six elections before reform was 88%. In the seven elections after reform, it was 80%. Given that turnout
has been declining around the world, however, perhaps the more relevant metric is the trend before and
after reform. In other words, was reform able to arrest the decline? Theres little evidence of this. There is
a slight uptick in turnout in the first post-reform election, some 3 percentage points. By 2014 turnout had
declined another 10 percentage points. This should give pause to those who claim that PR has any largescale effects on voter turnout. Finally, citizens appear no more satisfied with democracy in New Zealand
after reform than they were before. We have only one data point prior reform. Then, the average

This comprises the 1975, 1978, 1981, 1984, 1987, 1990, and 1993 elections pre-reform and the 1996,
1999, 2002, 2005, 2008, 2011, and 2014 elections post-reform.
2
The effective number of parties is a weighted measure used to calculate the distribution of votes or
seats across parties. A larger number suggests a more even distribution of votes or seats across a larger
number of parties.
3
I exclude the 1978 election from turnout calculations owing to widely reported problems with voter
enrolment.
4
For this measure, I am limited to the data provided in the New Zealand Election Studies from the year
1993 on. That study employs a standard question on how satisfied citizens are with the functioning of
democracy in their country. The variable has been recoded so that a 0 indicates the lowest level of
satisfaction and 1 indicates the highest.

satisfaction score was 0.63 on a 0-1 scale. In the election immediately after reform, this dropped to 0.49.
It has never approached the high observed in the last election under FPTP.

Table 1: Pre- and post-reform outcomes of interest in New Zealand

Effective
number
Effective
Democratic
of
number of
Election
Turnout
satisfaction,
electoral
legislative Parliamentary
year
rate
0-1
parties
parties
outcome
1975

83%

1978

2.6

1.9

Majority

2.9

2.0

Majority

Government
outcome
Single party
majority
Single party
majority
Single party
majority
Single party
majority
Single party
majority
Single party
majority
Single party
majority

Parties in
government
1
1

1981

91%

2.9

2.1

Majority

1984

94%

3.0

2.0

Majority

1987

89%

2.3

1.9

Majority

1990

85%

2.8

1.8

Majority

1993

85%

0.63

3.5

2.2

Majority

1996

88%

0.49

4.4

3.8

Minority

1999

84%

0.49

3.9

3.5

Minority

2002

77%

0.54

4.2

3.8

Minority

2005

81%

0.54

3.0

2.9

Minority

2008

79%

0.52

3.1

2.7

Minority

2011

74%

0.48

3.2

2.9

Minority

2014

78%

0.53

3.3

3.0

Minority

88%

0.63

2.9

2.0

1.0

80%

0.51

3.6

3.2

1.7

Pre-reform
mean
Post-reform
mean

Coalition
majority
Coalition
minority
Coalition
minority
Coalition
minority
Single party
minority
Single party
minority
Single party
minority

The move to MMP in New Zealand returned some of the benefits which reform advocates claim. It
increased the number of competitive parties in elections and in the parliament. It also increased
somewhat the number of parties in government. It did nothing, however, to increase voter participation or
to increase evaluations of democracy. I should also note that there is no obvious trend towards kindler,
gentler politics or policy in this country.

1
1
1
1
1

2
2
2
2
1
1

Taken together with the cross-national evidence, the most judicious reading of the data is electoral reform
is likely to have some positive effects, albeit quite limited.

2. The downsides to electoral reform are unknown and potentially substantial

The current discussion on electoral reform has largely avoided discussions of the potential downsides of
reform. In my reading, there are four general downsides which we should consider. I present them below,
and consider their potential application to Canada. Before presenting them, I should note that I think they
are perhaps unlikely to obtain. Put differently, in expectation the effects of negative outcomes are likely
small. Given that the benefits are as well (see 1 above), we should still give these downsides due
consideration.
The first potential downside is the continuation of small, regional parties. There is a reasonable normative
debate over whether it is desirable for small, regionally focussed parties to continually contest elections.
Indeed, many critics of FPTP point out that in its Canadian application it gives outsized rewards to
regionally-focussed parties. I take this to be a criticism of such parties. We should then note that the
incentives for such parties to continue to contest elections without a geographic expansion of their appeal
is greater under a system of PR than under our current system.
The second potential downside is giving outsized influence to small political parties. Systems of
proportional representation increase the number of political parties by lowering the barriers of entry for
less popular parties. When this is accompanied by government that requires the support of small parties
whether through majority coalitions, minority coalitions, or single-party minorities who rely on supply
arrangements, as in the New Zealand case small parties can be given disproportionate influence over
policies.
A hypothetical example is illustrative. Imagine a parliament which returns a major party with 40% of the
vote, and two smaller parties each holding 5% of the vote. A majority coalition of these parties is thus able
to control 100% of government power with 50% of the vote. In some important senses, this is more
normatively desirable than what is common in our current system, where frequently a majority
government elected on 40% of the vote holds 100% of government power. But if we look more deeply into
such arrangements, we are likely to find the small parties each holding somewhere on the order of 10% of
5
cabinet seats. The question is then raised: what kinds of disproportionality of power do we want? Do we
want it given to parties which have demonstrated broader support in the population? Or do we want it
redistributed to parties which have more marginal support?
The third potential downside is increased incentives for political actors to exploit social divisions. On this
matter, Gary Coxs Making Votes Count is the definitive work. The proliferation of political parties is a joint
function of two factors: the social divisions or cleavages present in a country and a permissive electoral
system (i.e. one that does not punish small parties). In this sense, the multiplication of parties is evidence
of such exploitation. Proponents of PR may argue that Canada is immune to such exploitations. We are
6
not, they will say, Israel or Italy , so we should not expect such divisions. The comparative evidence is
well worth considering on this point. In Table 2, I have compiled a list of the 15 Western countries with the
7
th
largest foreign born populations. Here, Canada ranks 4 highest. Inspection shows that the list is
comprised of countries we would largely consider free, democratic, tolerant, liberal, and developed. And
yet the evidence suggests that among these countries, those with PR systems are more likely to have

See Gamsons Law (Carroll and Cox, 2007), which suggests a distribution of cabinet portfolios within a
government according to each partys contribution of seats to the government.
6
Though please see the next section in which I consider the performance of Canadian democracy. Given
the
7
I also used a cut-off of a population of 1 million or more, so as to exclude financial have states such as
Luxembourg, Monaco, and Lichtenstein.

parties who exploit the social division of immigration, and who are more successful in doing it. The
8
average vote share of the largest anti-immigrant party in countries with PR lower-house legislative
elections is 8.7%. In majoritarian countries, it is 3.5%. More starkly, the average seat share of antiimmigrant parties in PR countries with large foreign born populations is 10%. In majoritarian countries it is
0.10%.
The fourth potential downside is government instability. Again, there is a normative debate over whether it
is desirable for governments to frequently change hands. Surely, alternation is necessary and healthy for
democracy. On the other hand, constant and protracted negotiations run the risk of blurring
accountability, inviting trade-offs which result in government policies that resemble nothing any party
promised in an election, and ineffective governance.
Table 2: Anti-immigrant party performance by foreign born population and electoral system

Country

Foreign born
population

PR (lower
house)

Anti-immigrant vote
share, most recent
election

Seat share

Year of
election

Party

Switzerland

28

Yes

22

35

2015

Swiss People's
Party

New Zealand

28

Yes

Australia

28

No

2016

One Nation

Canada

20

No

2015

Austria

17

Yes

21

22

2013

Ireland

16

Yes

2016

Sweden

16

Yes

13

14

2014

Swedish
Democrats

Belgium

16

Yes

2014

Flemish interest

Norway

14

Yes

16

17

2013

Progress Party

Spain

13

Yes

2016

United States

13

No

2012

Germany

13

Yes

2013

National
Democratic Party

United Kingdom

12

No

13

2015

UKIP

Freedom Party

France

12

No

2012

Front National
(second round
vote)

Netherlands

12

Yes

10

10

2012

Wilders Party for

Freedom

The comparative evidence on the duration of governments in different electoral systems is clear (see
Blais et al 2007). In PR systems, governments on average are more than one year shorter in duration
than in the average plurality system. The effect of electoral system, once we net out other exogenous
factors, is about 200 days. All of this electoral system effect, however, is a result of the type of

I define as anti-immigrant a party which makes reduced legal immigration a major part of its policy
platform or policy in office. I have noted the parties fulfilling this criteria in each relevant country.

governments which PR systems produce. The most stable form of government is a single party majority.
Majority coalitions and single-party minorities each last about one year less than that standard. Minority
coalitions as has been the case in most of the post-reform period in New Zealand last nearly two
years less on average than single-party majorities. Whether such frequent deaths of governments are
desirable is a normative question. The empirics, however, are clear: proportional systems are associated,
on average, with less stable governments precisely because of the types of governments they produce.
It is worth noting, in addition, how these changes in government come about. In formal terms, there are
three ways a government can change. First, there can be an inter-election change of parties constituting
the government. In PR systems, this accounts for about a third of government changes. In plurality
systems, it is just 10%. Second, there can be an inter-election change in the head of government. This
accounts for changes in just 11% of PR cases and 17% of plurality changes. Third, there can be an
election. This accounts for just over half (55%) of changes in proportional systems, while it accounts for
three quarters (73%) in plurality systems. The implication is clear: PR systems are more likely to generate
government changes that are not initiated by an election, but by breakdown and replacement in a
parliament. For those who argue that PR will return as equally stable government as a plurality system,
this is a problem.
3. Canadian democracy functions well
My own reading of testimony to and questioning by the special committee has suggested to me that the
functioning of Canadian democracy has not been sufficiently appreciated. Certainly, there is much with
which we can take some issue. Our country has experienced one party dominance rivalled only by
Sweden and Japan. We have, as in most other countries in the world, experienced significant decline in
our rates of voter participation, though this saw a large correction in the last election. Perhaps most
importantly, we do frequently experience parties winning outsized majorities on much less than a majority
of the ballots cast. None of these are particularly good things. They are certainly well-rehearsed as
critiques.
What is noted much less frequently are at least four measures on which our democracy has performed
well. First, our democracy has experienced more than 40 federal elections and dozens of peaceful
transitions of power, both between leaders from different parties and within federal parties. While this is a
basic standard of democracy (Przeworski 2000), it is one that sets Canada apart from most other
democracies. Indeed, Canadas run of uninterrupted democratic rule is among the longest in the world,
surpassed by less than a handful of other countries.
Second, by the standards of their times, our elections have been freely and fairly conducted and our
franchise has been liberally composed. Save the Canadian Pacific scandal and relatively pedestrian
turnout buying in early elections, Canadas democracy has been a model of well-run elections.
Third, our democracy performs well in the political representation of minorities and indigenous peoples,
9
especially compared to our Anglo-American counterparts. More historically, our political parties have a
long track record of representing the broad diversity of our country whether linguistic, confessional, or
ethnic without the emergence of explicitly ethnic or confessional parties. I wish to note especially that
this has happened against the backdrop of founding groups and later waves of immigrants who at various
times viewed each other as unfit for common purpose and interaction. Put starkly, our country has long
held the potential to be a tinderbox of identity. For the most part, weve avoided all but the smallest of
fires.
On this, much is made of the point that we are not Italy or Israel. But this cannot mean that we are not a
country that is characterized by competing regional economies, often deep religious and ethnic
differences, and different ways of life. I assume that those who make this argument must mean that
despite having the makings of a deeply divided and dysfunctional polity, we are not one. Our electoral
system just might have something to do with that.

I refer readers to the testimony of Leslie Seidle on this point.

Fourth, our country has a long record of protecting the rights of minority groups. In more recent years, this
has largely been the work of the Charter. But before its advent, it is still the case that protections were
extended, often because of an electoral logic. At other times, they were extended because of the norms
of broad coalition building that are the norm within our political parties.
4. For most of the problems ailing our democracy, there are potential fixes at hand which do
not require fundamental institutional change.
I wish finally to encourage the committee to take at once a broad and modest approach to reforming our
democratic institutions. There are, to be sure, shortcomings in our current democratic system. Turnout is
lower than we might like. We have not yet approached an even balance of female to male Members of
Parliament. Party leaders seem perhaps too strong vis--vis Members of Parliament. Local party
members do not enjoy real control over the selection of candidates. Parliamentary committees are often
weak and have neither the time nor the capacity to properly study and deliberate over policy.
This list is hardly exhaustive; yet there are potential solutions at hand for all of these problems which do
not require a fundamental change to a central institution. Instead, the Committee and the Members
parties can explore a number of changes to parliamentary procedure, administrative law, and party rules
that could address some or all of these problems.
It seems more judicious to engage in a systematic and iterative process of improving our democratic
institutions as is now occurring in the appointment of Supreme Court justices, for example then it does
to engage in wholesale reform.

References
Blais, A., Loewen, P. and Ricard, M., 2007. The Government Life Cycle. Democratic Reform in New
Brunswick.
Carroll, R. and Cox, G.W., 2007. The Logic of Gamson's Law: Pre-election Coalitions and Portfolio
Allocations. American Journal of Political Science, 51(2), pp.300-313.
Cox, G.W., 1997. Making votes count: Strategic coordination in the world's electoral systems (Vol. 7).
Cambridge: Cambridge University Press.
Laakso, M. and Taagepera, R., 1979. The" Effective" Number of Parties:" A Measure with Application to
West Europe". Comparative political studies, 12(1), p.3.
Przeworski, A., 2000. Democracy and development: political institutions and well-being in the world,
1950-1990 (Vol. 3). Cambridge University Press.