A review of comparative research on proportional representation vs. winner-take-all voting systems
Read the full collection of evidence for proportional representation below, or for a more condensed version (which doesn’t include everything), click through the slide show summary.
SLIDESHOW: Click the “next” arrow ( >) to advance manually the next slide.
Why Proportional Representation?
This document summarizes results from research comparing the performance of the two main families of voting systems: winner-take-all and proportional representation (PR). We already know that PR is a way of ensuring that all votes count and delivering more representative election results. The research cited below goes further, by demonstrating the impact of PR on the policy choices made by governments. This research shows that PR outperforms winner-take-all systems on measures of democracy, quality of life, income equality, environmental performance, and fiscal policy.
Comparing Winner-Take-All to Proportional Systems
Substantial comparative research has been conducted on the impact of winner-take-all systems vs proportional systems. This document covers a wide range of indicators because the choice of electoral systems has wide-ranging implications for how citizens relate to their governments and how government policies are considered and implemented.
Here are some of the main reasons we might expect a proportional system to have an impact:
- PR gives equal value to every vote and for this reason is likely to lead to increased government accountability to citizens and greater voter satisfaction. Some of the impacts of this can be seen below in the section titled Measures of Democracy.
- A shortcoming of winner-take-all systems is that small shifts in electoral preferences are sufficient to generate large shifts in power. Legislators can lose their seats at any time and power can easily shift from one majority government to another. The resulting political instability has significant implications:
- it leads to a high level of uncertainty;
- it encourages the hyper-partisanship observed in countries using such voting systems;
- it encourages politicians to focus on immediate political advantage at the expense of long-term issues; and
- it contributes to policy-lurch when one government replaces another.
The introduction of a proportional system does away with this amplification of small changes in voter preferences, leading to a greater level of political stability, less partisanship, a longer-term perspective and increased policy coherence over time.
- PR empowers ordinary citizens, by making every vote count and allowing for a wider range of views to be represented in the legislature. This can reduce inequality and improve access to social services over time, and, according to Salomon Orellana (2014, summarized here), could have a salutary effect on how a country deals with diversity more generally. Orellana argues that increased opportunities for diversity and dissent allow PR countries to outperform winner-take-all countries in four areas:
- policy innovation
- mitigating the pandering of politicians in the pursuit of voters by promising quick-fix solutions
- increasing the political sophistication of the electorate
- limiting elite control over decision making.
The electoral system we currently use everywhere in Canada belongs to a family of electoral systems based on the election of one representative in each riding on a winner-take-all basis. This type of system is sometimes called a “majoritarian” system, although with several parties in the running, the winner is often elected with only a plurality of the vote (more votes than any other candidate). We thus prefer the expression “winner-take-all” to refer to electoral systems in which candidates are elected based on a plurality or a majority of the vote. This includes most notably our own first-past-the-post (FPTP) system, and systems like those used in France or Australia, which use run-off or instant runoff mechanisms to ensure that the winner enjoys majority support. Winner-take-all models tend to produce majority governments by a single party, whether that party achieves a majority of the vote or not.
By comparison, proportional systems are designed to ensure that the composition of the legislature reflects how citizens voted to the maximum extent possible. In between, are semi-proportional systems which fall someplace between the two, of the sort we have in Italy, Greece and Japan. The degree of proportionality that is achieved can thus vary quite considerably from one country to the next.
The distinguishing feature of most winner-take-all voting systems is that representatives are elected one at a time in single-member districts or ridings. This is what allows the major parties to win the bulk of the seats, getting more seats than their overall share of the vote. In contrast, proportional systems elect representatives in multi-member districts, making it possible to allocate seats proportionately to the share of votes in the district as a whole. The larger the district in terms of seats, the more proportional the result is likely to be.
Carey and Hix (2011) have observed how this plays out by demonstrating the impact of district size on the most widely used measure of proportionality, the Gallagher index (GI). They find that increasing the districts size from 1 to 6 or 7 suffices to ensure a quite high level of proportionality (2011: Figure 3).
In the research that follows, most researchers use a simple dummy variable to distinguish between winner-take-all, PR or semi-proportional systems. Some have tried using the Gallagher index instead, but the GI does not measure the proportionality of the system so much as the proportionality of particular election results. As such the GI can vary from election to election. A better alternative could be the average district size, as used to good effect by Baird et al. (2007) and Laura Wills-Otero (2009: 41).
Arend Lijphart, a world-renowned political scientist, spent his career studying various features of democratic life in different countries. In his landmark study (2012), he compared 36 democracies over 55 years. Lijphart uses his own classification system, speaking of “majoritarian” and “consensual” democracies. Winner-take-all vs. PR voting systems are the major differentiating factor of these two categories.
Lijphart found that consensual democracies outperformed winner-take-all ones on 16 out of 17 measures of sound government and decision making, including:
- the quality and independence of the public service,
- the quality of policy making,
- the rule of law, and
- the level and control of corruption and state capture by elite interests.Lijphart qualifies consensual democracies as “kinder, gentler democracies.” Among his findings about countries using proportional systems:
- Voter turnout was higher by 7.5 percentage points, when contextual factors are taken into account.
- Government policies were closer to the view of the median voter.
- Citizens were more satisfied with the performance of their countries’ democratic institutions, even when the party they voted for was not in power.
- There was a small increase in the number of parties in Parliament.
- The share of women elected as legislators was 8 percentage points higher.
- Scores were higher on measures of political participation and civil liberties
Research by other authors has yielded similar results. Lijphart’s finding that proportional systems lead to governments that better reflect the views of the median voter was confirmed by McDonald, Mendes and Budge (2004), who looked at 254 elections in 20 countries. Blais and Loewen (2007) found that citizens in countries with Pr felt their elections were fairer and that elected representatives were more responsive to the electorate.
Blais, Morin-chassé and Singh (2017) found that citizens are “sensitive to deficits in representation” and that satisfaction with democracy is lower when the party one voted for does not get seats in proportion to the popular vote. Recent experimental work by Plescia, Blais and Högström (2020) showed a drop in support for voting rules in Austria, England, Ireland and Sweden when results were disproportional. This held true for voters supporting larger parties as well as small parties. The authors observed that large-party voters also want a fair system, even if it disproportionality advantages their party.
Turnout and participation
One would expect turnout to be higher in PR countries, where voters can more confidently cast a vote that will help elect a representative or party of their choice. This increases the incentive to vote for most people. Milner (2014) adds that PR makes it easier to vote, because parties pay more attention to their long-term policy stance, compared to parties’ focus on wedge issues under FPTP. One might counter that it’s harder for voters to hold individual parties to account when they are part of coalitions, as tends to happen under PR. However this balances out, the evidence supports the positive link one would expect between PR and electoral turnout.
The above graphic illustrates the higher turnout ratios observed in PR countries within the OECD. Looking at the most recent figures from the International Institute for Democracy and Electoral Assistance database (IDEA 2022), and setting aside non-comparable countries where voting is compulsory, one finds that the top turnout countries in recent parliamentary elections are all PR countries, followed by the US, with the UK, Canada and France trailing behind. Canada is in position 28 out of 33 OECD countries.
Of course, turnout rates can vary from country to country or from election to election for any number of reasons. Frank and Martínez i Coma (2021) provide the most up-to-date compendium of research results to date and the most comprehensive testing of political, institutional, and socio-economic factors associated with turnout. They find that 10 studies out of 19 on the subject found a statistically significant positive relationship between turnout and PR (Table 1, row 3). Only 1 found a negative relationship.
Frank and Martinez i Coma pursue this line of research further and conduct their own work on the subject, considering 70 different variables in 579 elections in 80 democracies from 1945 to 2014. Testing a wide range of model specifications, they find proportional representation to be a robust predictor of turnout accounting for a 5.2 percentage point boost in turnout (Table 5 and p. 20).
Among other robust predictors are compulsory voting and the concurrency of presidential elections at the same time as parliamentary elections — which might explain why the US has a somewhat higher turnout rate than Canada.
Looking at the impact of PR on voter turnout after New Zealand switched to PR in 1996, Vowles (2013) tackles the turnout issue from another perspective. He looks at turnout in a single country before and after PR was brought in. He points out that turnout in New Zealand was in a declining trend when PR was introduced and that the transition to PR was unable to arrest that trend. However, he comes to some interesting conclusions about the composition of turnout.
He posits that the switch from FPTP to PR might most significantly improve youth turnout because new voters are most likely to be responsive to change, in contrast to older generations. This is indeed what he finds (Vowles, 2013: 10-12). He also considers different types of ridings. Looking at the incentive to vote, he argues that the value of voting would be higher in competitive ridings and lower in safe ridings. On those grounds, we should expect turnout to increase the most in formerly safe ridings once PR is implemented. This is also what he finds. In highly competitive ridings, he finds the opposite: turnout in such ridings was actually higher under FPTP than with PR (2013: 15-18). However, the overall effect of PR on turnout appears to have been positive.
Concentration of Power
Winner-take-all voting models systematically favour the leading party, leading to the formation of single-party majorities based on as little as 37% to 39% of the popular vote. In comparison, PR tends to diffuse power to parties in proportion to the popular vote. Single-party majorities are rare in PR countries, where coalition or minority governments are the norm (see Horne et al, 2022, Figure 2).
The concentration of power that results in non-proportional voting systems makes it possible for one party to ram its agenda through unimpeded and creates opportunities for gaming the democratic process itself:
- In the US, the process of gerrymandering and the packing of the Supreme Court are good examples of this.
- In Canada, we have seen reversal of political financing reforms both federally and in Ontario and the reversal of key measures in the Ontario Municipal Elections Act that can best be understood based on the partisan interests of the governing party.
- In weakly proportional countries such as Poland, Hungary, and Turkey, the concentration of power in one party in recent years has allowed that party to manipulate the media or the justice system in its favour. Examples such as these illustrate that PR systems with a majoritarian bias are also subject to risks of subversion.
Looking at data covering 170 countries between 1975 and 2004, Lehoucq and Kolev (2015) have found a positive connection between the proportionality of electoral systems and the quality of elections.
They complemented this cross-country approach with a detailed study of electoral fraud in Costa Rica between 1901 and 1948. During this period, less populated districts used first-past-the-post while more populated ones used PR. Focusing on a database of 1300 accusations of fraud covering 175 elections, they found that first-past-the-post led to more fraud than proportional systems. The authors explain that “In FPTP systems, candidates face more incentives to prevent opposition voters from casting ballots, to tamper with voter rolls and even the ballot box, and to manipulate the tally of the vote. Winning, by even one (questionable) vote, is preferable to being shut out of parliamentary representation.”
Risks of extremism
A question that often arises in discussions about electoral reform is whether proportional representation might make it easier for extremist actors to gain disproportionate levels of influence.
If one looks at European countries with PR systems, one observes, as one would expect, that parties of both the far left and far right tend to win seats in proportion to their share of the vote. This is less likely to happen in winner-take-all countries, due to the built-in bias against small parties.
However, the implications of this need to be properly understood. Does one really want to deprive the vote to citizens whose views are considered extremist?
An important question in this regard is whether one would prefer to have extremists proportionately represented in the legislature or making themselves heard in the streets, as happened with the Freedom Convoy in the streets of Ottawa in the winter of 2022.
Another question is whether winning seats translates into disproportionate power for extremist views and whether the risks of extreme shifts in policy are any greater under PR. As a practical matter the extreme left and extreme right parties are very rarely accepted in government coalitions power at all. Experience in countries with proportional representation shows that if a party’s policies are too far out of line with mainstream opinion, the other parties will cooperate to exclude them from power. This has happened in Sweden, Germany and the Netherlands.
At the same time, the tendency under PR is that extremist parties end up moderating their views, as this is the only way they have any hope at all of ever being a part of government (Norris, 2023).
Meanwhile, having big-tent parties under winner-take-all may not prevent one of those parties from being captured by extremist forces, as has happened under the Republicans in the US, and achieving high levels of power indeed.
Taking environmental policy as an example covered later in this Review, recent research showed that right-wing extremists were more successful at reversing climate change policies in winner-take-all countries than PR countries (Lockwood, 2022).
In short, the evidence is clear that PR will help all small parties win seats, whether one judges those parties “extremist” or not. However, political instability seems to be highest in winner-take-all countries and lower in countries with proportional representation. The evidence suggests that the danger of extremists achieving a high level of political power is highest in countries that concentrate power in a small number of parties.
As Jonathan Rodden (2019: 2) puts it with regard to the US, ,
The contemporary United States is often described as a society divided along a single, overarching left-right political dimension, with Democrats on one side and Republicans on the other. The parties are described as two hostile tribes… where group membership now goes well beyond shared party loyalty and political preferences, and has become a form of social identity … in which members of the outgroup are viewed with disdain and enmity…
Here in Canada, current levels of animosity between the Conservative Party of Canada and the Liberal Party of Canada, seems to be moving in a similar direction.
This raises questions about the extent to which this is a function of our FPTP electoral systems in these two countries.
Paradoxically, much of the cross-country research seems to point the other way, suggesting that polarization is higher in PR countries. The logic of this is easy enough to understand. If there are only two major parties as tends to happen under winner-take-all systems, you would expect both of those parties to gravitate towards the centre in an effort to form a majority government. To the extent that third parties manage to win seats, they will tend to carve out niches for themselves further left or right on the political spectrum. Since this is easier to do under PR, you end up with results that appear more polarized in PR countries.
Rodden (2019) questions this model, which he considers to inconsistent with what country experts and pundits mean when they talk about polarization. He questions the logic of a model that is based on a single left-right definition of a party’s position relative to another. In practice, voters share a much wider range of political perspectives that cannot be reduced to left-right perspectives, so there is no single “centre” position for any party and the average distance that voters feel from parties may be higher in a FPTP system than under PR. The way to test for this is to look more closely at the full range of responses from voters about how distant they feel from different parties. Doing this for the US, he finds a great deal of heterogeneity and considerable voter alienation from one or the other of the two political parties.
Polarization from this perspective is manifestly strongest in FPTP countries such as the US, the UK and Canada and weakest in PR countries.
Somer and McCoy (2018) arrived at a similar conclusion that the most extreme cases of polarization manifest themselves in majoritarian electoral systems.
Horne et al. (2022: 6) come to the same conclusion, as well, observing that “out-party” dislike declines dramatically as a function of proportionality. One reason for this is that voters tend to have more positive feelings about other parties that have previously been part of the same coalition or opposition group of the kind one finds in PR countries. These positive feelings tend to endure even after a coalition no longer exists.
The link between the electoral system and polarization is clearly visible in the following graphic, in which countries are positioned from left to right depending on the average number of seats per electoral district, logged. FPTP countries like Canada lie along the left axis, with a value of 0 (the log of 1 being zero). The more seats per district as you move to the right, the more proportional the electoral system. As the graphic shows, the degree of enmity for other parties (out-party dislike) declines significantly and unmistakingly from left to right.
The choice of electoral system also affects how politicians behave. Nemoto and Pinto (2019) studied the nature of political discourse in New Zealand’s Parliament before and after the adoption of PR in 1996. Analyzing over 800,000 parliamentary speeches by MPs from 1987 to 2016, they found a marked decrease in the anger and hostility in MP’s speeches overall after 1996, most significantly in the tone of ruling party MPs towards smaller parties who might one day be coalition partners with them.
Another aspect of polarization is how divided citizens are from each other. To the extent that FPTP concentrates power in one group of voters as opposed to another, we would expect frustration to grow in the out-group, feeding populist mobilization among the disaffected of the sort displayed in the gilet-jaune movement in France, under Donald Trump in the US and in Canada’s 2022 trucker’s convoy.
In extreme cases, the result can degenerate into civil war. A remarkable study in this regard was conducted by Marta Reynal-Querol (2002). She looked at 138 countries from 1960 to 1995, including roughly equal numbers of countries with majoritarian parliamentary systems, proportional parliamentary systems and presidential systems. Looking at 68 instances of civil war over this period, she found that all of these wars occurred in majoritarian or presidential systems. No country with a proportional parliamentary system had suffered a civil war.
Polarization that divides society into warring camps increases the level of uncertainty (Baker et al., 2020) and makes it difficult to have evidence-based conversations about serious issues, including life-and death issues such as pandemic management (Beramendi and Rodden, 2022).
As we shall see later, PR countries have higher levels of social cohesion in other respects as well.
Diversity of Representation
An obvious consequence of multi-member districts is that parties will want to put forward a diversified slate of candidates to reach a wider range of voters. This leads to greater diversity of representation beyond that of party representation.
One manifestation of this is in the ability of younger people to be elected as legislators. Using data from 128 countries, the Inter-Parliamentary Union (2016: 7) found that youth representation in legislatures was 15-20 times higher under proportional systems compared to majoritarian systems.
Stockemer and Sundström (2018), looking at 107 countries, concluded that “In absolute terms, PR systems have 5 percentage points more young deputies aged 35 and under and 7 percentage points more deputies aged 40 and under than plurality systems For mixed systems we find similar percentages than for PR.”
Lawyers, who tend to be relatively affluent white males, make up a large share of elected representatives in different countries (Joshi, 2019: section 1.1). Joshi argues that this may be particularly true in single member districts (SMDs) of the sort used in the US, the UK and Canada.
One way to test this is to look at countries using mixed member systems including both SMDs and party lists to elect multiple members. Joshi looks at three such countries: Germany, Japan and South Korea. As expected, he finds the share of lawyers to be considerably larger in SMDs compared to party list arrangements: 32.6% on average compared to 19.8% under the party list part of the system (calculated over 11 elections from Table 2). The way representatives are elected has clear implications for this dimension of diversity in our legislatures.
The tendency of PR countries to elect more women compared to winner-take-all countries is also well established in the comparative literature.
Looking at 168 countries from 1992 to 2010, Jennifer Rosen (Rosen, 2011) found that developed countries with PR elected more women by a margin of about 10 percentage points compared to majoritarian countries — a bit more in countries with closed list PR, a bit less in countries with open lists. Countries with mixed systems outperformed majoritarian countries by 6 points.
Contrary to expectations, she finds that quotas do not have a significant impact in developed countries. However, she gets different results for developing countries, where she finds a considerably smaller PR bonus and a much larger effect attributable to the use of quotas (Rosen, 2011: 314, model 2 coefficients).
Australia provides the perfect petri dish for comparing different electoral systems, because the two Australian houses of parliament use different systems: an STV-based PR system for the Senate and a winner-take-all system using ranked ballots (AV) for the House or Representatives. Kaminsky and White (2007) looked at elections in both chambers over a 61 year period and found that more than two and a half times more women were elected to the Senate than the House of Representatives. As of 2019, the share of women was 51% in the Senate as against 31% in the House of Representatives (International Parliamentary Union, 2021).
The Australian states of Tasmania and the Australian Capital territory — the only ones to also use PR-STV — have also elected relatively large proportions of women to their legislatures in recent years (BBC, 2018 and Burch, 2018, Appendix A).
Women in winner-take-all countries like Canada, the US and the UK currently make up about 30% women in Parliament in comparison to PR countries such as New Zealand, Sweden, Spain and Switzerland which have recently elected from 42% to 48% women (Inter-Parliamentary Union, 2021).
One of the biggest debates about PR is whether it leads to political and policy instability. This section looks at two facets of this question:
- the frequency of elections, and
- the issue of policy lurch
As we shall see, the evidence strongly favours PR over winner-take-all systems.
Frequency of elections
Regarding the frequency of elections, the comparative research shows little difference between PR and FPTP countries. Looking at elections from 1945 to 1998, Pilon (2007: 150) points out that countries using FPTP averaged 16.7 elections, while countries using proportional systems averaged only 16.0 elections. The difference between these two groups of countries is negligible.
Pilon points to other data that shows a somewhat shorter government life-span in PR countries (1.8 years as opposed to 2.5 years in FPTP countries), but discounts this result because it is heavily influenced by Italian experience mainly involving what would elsewhere only be considered as cabinet shuffles (2007: 147). He concludes that this type of instability is “not a problem for PR systems in western countries” (2007: 151).
However, there are other ways to assess the link between electoral systems and stability. Countries with winner-take-all systems that tend to oscillate from left to right, are characterized by policy shifts largely unrelated to underlying voter preferences, and cannot be said to satisfy the test of stability terribly well. Indeed, it is not unusual for one government to simply undo what the previous government has done, in a process of “policy lurch” in winner-take-all systems such as Australia, the US, the UK and Canada.
In New Zealand, dramatic policy shifts enacted by the National and Labour governments in the 1970s and 1980s were one of the main reasons for voter disaffection with the first-past-the-post system prior to the introduction of MMP in 1996. Although the government has continued to be dominated by the National and Labour parties at different moments since then, the need of the government to secure the support of other parties to enact legislation and the stronger voice of smaller parties in the House are felt to have considerably tempered the policy lurch phenomenon (Shaw 2016).
In Canada, we need look no further than the efforts by the Harper government to undo much of the Liberal government legacy that they inherited and the subsequent undoing of conservative policies by the Liberal government following the 2015 election.
After the Liberal Government came to power in 2015, they published a list of budget cuts enacted by the previous government, many of which were aimed at programs established by previous Liberal governments. The Liberal Government elected in 2015 then spent much of its first year in office undoing these cuts (Tasker, 2016; see Various media reports, 2015/2016, for more examples).
Policy lurch is a normal feature of our electoral system because small shifts in the political winds lead to significant shifts in power, as the country goes from a majority government at one pole of the spectrum to another.
Provincially, the cases of Alberta and Ontario stand out. Taking power back from the NDP in 2019, Jason Kenney spent his first 100 days in government undoing many of the policies brought in by Rachel Notley. In anticipation of the 2023 election, Rachel Notley is currently promising to spend her first 100 days undoing the policies of Jason Kenney.
Ontario is likewise replete with examples of policy lurch since since Doug Ford’s Progressive Conservatives replaced the Liberals in 2018. The most dramatic of these was the abandonment of the province’s cap and trade program on climate change and the broken contracts that accompanied this policy shift.
However, policy lurch is but one of many ways that the electoral system can impact on public policy. In a recent article, Filippetti and Vezzani (2022) ask just that question: what institutional conditions are required to ensure that public policy speaks to the general public interest as opposed to special interests and addresses a time frame longer than the typical election cycle?
There are several reasons to believe that winner-take-all systems would deviate from these two ideals. The most important is likely to be the extreme sensitivity of the system to small shifts in voter preferences, encouraging politicians and parties to cater to targeted minorities of the voting public, by focusing on “wedge issues,” swing voters and swing ridings.
Once a government is formed by a single party, a winner-take-all system makes it easier for lobbyists to win special consideration, due to the concentration of government power that results. This stands in contrast to the sort of “consensus-based” governance systems identified by Lijphart (2012) in which every vote counts and politicians have to work together in making policy.
We explore three areas of public policy in what follows:
- Funding for public research,
- Fiscal and economic policy, and
- Environmental stewardship.
Filippetti and Vezzani (2022) look closely at public funding for public research, the considerable benefits of which are widely recognized in policy circles — something that has become more obvious than ever in the health sphere as the world struggles with the COVID pandemic.
Funding for public research is a perfect example of a policy sphere that speaks to the general public interest while requiring a long-term policy perspective.
Using a comparative approach and a panel data set of 1,228 observations, the authors obtained robust and highly significant results under the following conditions (Table 3):
- systems of proportional representation;
- parliamentary, as opposed to presidential systems;
- bicameral legislatures; and
- the presence of encompassing civil society organisations.
The logic in each case is that these features encourage greater emphasis on the public good, providing a better institutional environment in support of science-based decision-making and innovation.
When considering the economic performance of countries using different systems, Carey and Hix (2009 and 2011) found that countries with moderately proportional systems were more fiscally responsible and more likely to enjoy fiscal surpluses.
Orellana (2014) found that proportional systems tend to have higher surpluses or lower deficits than less proportional systems, and lower levels of national debt. Orellana’s regression analysis predicts a surplus of 0.05 percent of GDP for fully proportional countries, against a deficit of 2.9 percent of GDP in winner-take-all countries. The projected national debt is 65.7 percent higher in winner-take-all countries than in those with fully proportional systems, meaning the cost of servicing the debt will be higher.
Researchers agree that single-member districts tend to pull in the direction of narrowly targeted programs benefiting small geographic constituencies. Conversely, multi-member districts and proportional representation induce politicians to provide benefits for broad groups of voters.
Some recent papers have studied the influence of the electoral system on the composition of government spending. Perssons and Tebellini (2008) theorise that:
Multi-member districts and proportional representation diffuse electoral competition, giving the parties strong incentives to seek electoral support from broad coalitions in the population through general public goods or universalistic redistributive programs (e.g., public pensions or other welfare programs). In contrast, single-member districts and plurality rule typically make each party a sure winner in some of the districts, concentrating electoral competition in the other pivotal districts. Both parties thus have a strong incentive to target voters in these swing districts.
Most researchers agree that countries with proportional representation spend more on average on social expenditures as a percentage of GDP (social programs, education, pensions etc). For example, Lijphart (2012) found that countries with PR spent an average of 4.75% more on social expenditures than majoritarian democracies.
Orellana (2014) found the countries with proportional representation spend less on military expenditures.
Level of Government Spending
Research summarized by Perotti, Rostagno and Milesi-Ferretti (2001) found that countries with proportional systems had higher government spending overall (not just on social expenditures). This seemed to be related to the number of parties in government coalitions, and the challenges of satisfying the demands of several parties.
However, more recent research has challenged this finding. Martin and Vanberg (2013) studied 15 European democracies over 40 years. They concluded that there are ways of managing budgetary pressures by instituting appropriate budgetary rules. An example is to determine the overall budget in advance, so that increased allocations for a particular Ministry would have to come from another. In terms of winner-take-all vs. proportional systems, they conclude that there is no fiscal trade-off involved.
Research by Funk and Gathman (2010) came to a similar conclusion by looking at the effects of adopting proportional representation in Switzerland from 1890 to 2005, as more and more cantons came to adopt PR over time. They found that while PR increased social expenditures, overall spending did not increase:
Proportional systems shift spending toward broad goods (e.g. education and welfare benefits) but decrease spending on targetable goods (e.g. roads and agricultural subsidies). Our evidence does not suggest that proportional representation increases the overall size of government.
Turning to the issue of economic performance more generally, the correlation seems to depend on the sample being used. Lijphart and Orellana found no relationship between electoral systems and economic growth. However, when Knutsen (2011) looked at a much longer historical period involving 3,710 country-years of data covering 107 countries from 1820 to 2002, he found that proportional and semi-proportional systems produced an “astonishingly robust” and “quite substantial” increase in economic growth – a one percentage point increase – compared to winner-take-all systems. Knutsen suggests that this may be because PR tends to promote broad-interest policies rather than special interest policies; and because PR systems produce more stable and thus more credible economic policies. He concludes that PR and semi-PR systems generate more prosperity than winner-take-all systems.
Knutsen’s empirical results are corroborated by more recent work by Alfano and Baraldi (2015) who studied 91 countries from 1979–2010. The particularity of Alfano and Baraldi’s work is that it considers the degree of proportionality of electoral results using the Gallagher Index (GI) as an indicator. Highly proportional countries like Sweden and Germany have a GI between 1 and 3. Canada typically has a GI of about 12. According to Alfano and Baraldi’s results, a country with a GI of 2.3 could expect an annual growth rate one percentage point higher than a country with a GI of 14. A one percentage point difference is also what Knutsen found.
Alfano and Baraldi find a non-linear relationship between proportionality and rates of growth. The relationship appears as a curve with the growth rate reaching its maximum value under a GI of around 6.5 matching the level of proportionality that we would expect to achieve in any of the made-in-Canada systems being proposed for this country. The predicted growth rate drops sharply for a Gallagher index above 14.
What emerges from this body of research is that proportional systems categorically and strongly outperform non-proportional systems in terms of economic growth. It seems that even a modest level of proportionality is helpful. In Knutsen’s work, semi-PR and PR countries seem to achieve much the same result in terms of growth rates.
Opponents of proportional representation sometimes suggest that making votes count is “bad for business” because businesses will be overtaxed and investors will flee. However, the average corporate tax rate in PR countries was actually lower in 2017 according to figures from the OECD: 22.9% for PR countries vs. 27.2% for winner-take-all countries.
Looking at trade protection, which economists usually decry as being bad for the economy, in a study of 147 countries over 23 years, Evans (2009) found that winner-take-all systems have higher tariffs than proportional systems.
Taken together, these research results suggest that transitioning to a proportional voting system would be good for Canada’s economy. Nine out of ten of the wealthiest OECD countries are countries with proportional voting systems.
Numerous researchers have found a connection between PR and good environmental stewardship. One of the earliest was the work by Baird et al. (2007), which provides a valuable base in a number of ways.
First is their analysis of why FPTP systems might make inroads on environmental performance more difficult. They give two reasons: the need to reach a broad constituency for whom the environment may not be a common concern in order to win seats on a winner-take-all basis; and the tendency to marginalize smaller parties that might favour the environment (2007:48).
They argue that PR would avoid these biases and predict a positive relationship between the Yale Environmental Protection Index (EPI) and PR when controlling for per capita income, level of literacy, the presence of democratic institutions other than PR, transparency and the presence of post-materialism environmental values in the country.
This set of control variables makes for a compelling test of their hypothesis but makes it harder to disentangle the impact of PR itself from other variables. They find their results to be stronger when using a more informative measure of PR than the usual 0-1 dummy variable. Instead, they use the average size of electoral districts: 1 for single-member districts and higher numbers differing by country for countries using multi-member districts. This is a step up from the methodology used by most other researchers. They find that the use of this measure successfully brings out the impact of PR on environmental performance, which comes out as positive and statistically significant.
Their results indicate that an increase in the average district size from 1 to 15 (i.e. from first-past-the-post to a moderately proportional system), would increase a country’s EPI by 4.8 percentage points. To appreciate how significant this is, consider that Canada currently has an EPI of 50.0, coming in 49th out of 180 countries. With a moderately proportional electoral system, Canada’s expected EPI would be 54.8, putting it in 33rd place.
Multiple authors have come to similar results:
- Frederiksson and Millimet (2004) found that countries with proportional systems set stricter environmental policies.
- Cohen (2010) found that countries with proportional systems were faster to ratify the Kyoto protocol, and that their share of world total carbon emissions had declined.
- Lijphart (2012) found that countries with proportional systems scored six points higher on the EPI.
- Using data from the International Energy Agency, Orellana (2014) found that between 1990 and 2007, when carbon emissions were rising everywhere, the statistically predicted increase was significantly lower in countries with fully proportional systems, at 9.5%, compared to 45.5% in countries using winner-take-all systems.
- Orellana found that citizens in countries with proportional representation were more supportive of environmental action, and more willing to pay the costs associated with environmental protection. He found the use of renewable energy to be approximately 117 percent higher in countries with fully proportional electoral systems.
In sum, countries with proportional systems do more to protect the environment.
An issue of some concern in recent years has been the positioning of right-wing populist parties as opponents of strong climate change policy. Here too a remarkable difference can be observed in how countries with proportional systems have responded to such challenges compared to those with winner-take-all systems. Looking at the impact of the far right on climate policy in 31 OECD countries from 2007 to 2018, Lockwood and Lockwood (2022) found that in countries with proportional representation, far-right parties had no significant effect on climate policy.
Read Fair Vote Canada’s report on PR and Climate Policy.
As noted earlier, PR tends to empower ordinary citizens and one might expect that to be reflected in indicators of income inequality and of social policy outcomes. This expectation is also borne out in comparative research.
Lijphart (2012: 282) found that countries with proportional systems had considerably lower levels of income inequality.
Likewise, Birchfield and Crepaz (1998:192) found that “consensual political institutions [which use PR] tend to reduce income inequalities whereas majoritarian institutions have the opposite effect.” The results of the regression work they present were highly significant, with PR accounting for 51% of the variance in income inequality across countries. Birchfield and Crepaz explain this result in terms of the higher degree of political power of people in PR systems. In their words (1998:191):
The more widespread the access to political institutions, and the more representative the political system, the more citizens will take part in the political process to change it in their favour which will manifest itself, among other things, in lower income inequality. Such consensual political institutions make the government more responsive to the demands of a wider range of citizens.
Vincenzo Verardi, in a study of 28 democracies (2005), also found that when proportionality increases, inequality decreases. Iversen and Soskice (2006) found that PR is associated with greater efforts to promote income redistribution.
Bernauer, Giger and Rosset (2015) looking at 24 Parliamentary democracies, found that the political preferences of low income citizens are better represented in a more proportional system, and the political preferences of the rich are better represented in a winner-take-all system.
Investigating the broader impact of PR on society, Carey and Hix (2009 and 2011) looked at 610 elections over 60 years in 81 countries and found that PR countries garnered higher scores on the United Nations Index of Human Development, which incorporates health, education and standard of living indicators. Carey and Hix consider that the Index of Human Development provides “a reasonable overall indicator of government performance in the delivery of public goods and human welfare.” Lijphart found that countries with PR spent an average of 4.75% more on social expenditures than winner-take-all democracies.
Similarly, Altman, Flavin and Radcliff (2017) found robust evidence from 21 OECD countries that citizens in countries with proportional representation reported higher levels of life satisfaction, concluding that the real-world consequences for human well-being of PR and other features of democratic rule are substantial compared to other common predictors.
Patterson (2017) studied the experience of 179 countries with different types of democratic or autocratic regimes between 1975 and 2012 and found that countries with proportional systems most reliably outperformed other countries with regard to population health. Compared to “closed autocracies” at the other end of the spectrum, countries with PR had up to 12 more years of life expectancy on average, 75% less infant mortality, and lower mortality rates for most other age groups. In contrast, countries with winner-take-all systems were no better than closed autocracies in terms of health improvements over time. The authors conclude that democracies are not equally effective at holding leaders accountable and that broad representation of diverse interests appears to be what matters most to ensure positive health outcomes.
Christina Gathmann (2019) exploits a unique time-series database from Switzerland, where, between 1890 and 1950, eighteen of the country’s twenty-five cantons switched to proportional representation in a staggered way. She was thus able to exploit a data set of multiple before and after scenarios against which to assess the impact of changing over from a plurality to a proportional system.
She documents a “sizable shift in political participation and representation when cantons switched from a majoritarian to a proportional system… (that) gave the previously underrepresented working class a greater weight in the political process.” This led to changes in public spending, most significantly in the education sector, with ultimate benefits in terms of population health. She estimates that the electoral reform contributed between 11 and 17 percent to the observed decline in mortality during this period.
This is a sizable contribution considering the role of other factors, including major advances in sanitation and medical practices during this period. As she points out, “such sizable mortality reductions demonstrate that electoral rules and hence, the form of democratic representation have powerful consequences for population well-being.”
Michael J Rigby et al. (October, 2021) have studied how well 42 OECD or European Union countries managed the COVID-19 pandemic in the critical period between May 2020 and Nov. 2020.
Trying to explain the differences in COVID impacts across countries, they found that neither the health system, nor measures of social and economic inequality helped to explain these differences. What they did find was a clear linkage between successful management of the pandemic and political features including the presence of a coalition government, having a female prime minister and a proportional voting system.
These features were particularly evident among the countries who showed the lowest increase in COVID death rates per million people from May to November 2020. Among the top six, five had coalition governments, four had female PMs, and all six were PR countries. Among the top 14 performers, 12 had coalition or minority governments, including Canada. Similar results were obtained using the “resilience index” as an indicator (Table 6).
The authors argue that “coalition governments, by being more fragile and needing to keep political partners continuously convinced and thus committed, also have to be persuasive… Conversely, countries with first past the post elections can result in governments representing only a minority of the population, and single party majority governments can ride out criticism with comparative impunity over several years until the next election” (p. 21). They cite the US and the UK as examples where populist-style majoritarian governments have led to comparatively high mortality rates. They conclude that the way a government was led, and whether leaders based their decisions on scientific evidence and shared the reasoning with citizens, had notable effects (p. 2).
In their view, “majority governments, seeming strong but being more impervious to feedback, appear less effective in pandemic management” compared to coalition governments elected under proportional systems (p. 23).
Electoral systems can affect how citizens and government interact and how citizens relate to each other. Orellana provides a number of reasons why the diversity of views in PR systems can have an impact.
Prejudice, Tolerance and Changing Attitudes
Using data from the World Values Survey conducted between 1981 and 2010, Orellana found that citizens in countries with proportional systems tend to show less prejudice towards minority and marginalized groups. Countries with winner-take-all systems scored approximately 44 percent higher on the prejudice scale than countries with fully proportional electoral systems.
He found that citizens in countries with more proportional electoral systems tend to have higher levels of tolerance for homosexuality, abortion, divorce, euthanasia and prostitution; and a higher level of disagreement with the notion that men make better leaders.
Furthermore, their attitudes towards those issues tended to evolve more quickly than elsewhere. Over a roughly 25-year period, the share of the population demonstrating tolerance of homosexuality increased by 41 percentage points in countries using proportional systems but only 20 percentage points in single member district systems.
Law Enforcement and National Defence
Perhaps because PR mitigates pandering for votes based on quick fixes, both Lijphart and Orellana found that countries with less proportional systems tend to have more public support for punitive solutions to crime, and produce more punitive policy outcomes including higher incarceration rates and greater use of capital punishment. Orellana found that support for incarceration is approximately 28 percentage points higher in countries with winner-take-all systems. Confirming similar results by Lijphart, he found that the statistically-predicted incarceration rate for countries with fully proportional systems was 136 per 100,000 people compared to 246 per 100,000 in winner-take-all countries.
Relying on an indicator of privacy and surveillance produced by Privacy International for over 30 countries, Orellana found that countries with proportional systems scored 58% higher on the privacy index.
Looking at the average military expenditure as a percentage of GDP between 1988 and 2012 and data from the Stockholm International Peace Research Institute, Orellana found that the predicted level of military spending for countries with winner-take-all systems was more than twice as high as for countries with fully proportional systems (2.6% vs. 1.1% of GDP).
Leblang and Chan (2003) found that a country’s electoral system is the most important predictor of a country’s involvement in war, according to three different measures: (1) when a country was the first to enter a war; (2) when it joined a multinational coalition in an ongoing war; and (3) how long it stayed in a war after becoming a party to it.
Lijphart found that PR is strongly correlated with a lower degree of violent events, more political stability and a lower risk of internal conflict. Qvortrup and Lijphart (2013), looking at the same 36 countries studied in Lijphart’s earlier work, found “statistically significant correlations between the index of consensus democracy and a higher incidence of fatal domestic terror incidents in the period 1985–2010.” They also found that the risk of fatal terrorist attacks was almost six times higher in winner-take-all democracies.
In conclusion, the existing body of comparative research internationally leaves little room for doubt that PR is the better choice.
PR outperforms winner-take-all systems in almost every respect:
- higher quality of democratic life,
- prudent fiscal management,
- higher economic growth,
- better environmental management,
- reduced income inequality,
- higher levels of human development,
- greater tolerance of diversity,
- a less punitive approach to law enforcement,
- greater respect for privacy, and
- lower levels of conflict and militarism.
ACE Electoral Knowledge Network (updated regularly online). Encyclopaedia on Electoral Systems.
Alfano, M. & Baraldi, A. (2015). Proportional Degree of Electoral Systems Growth: A Panel Test. Rev. Law Econ. 2015; 11(1): 51–78
Altman, A, Flavin, Flavin, P. and Radcliff, B. “Democratic Institutions and Subjective Well Being.” Political Studies 2017, Vol. 65(3) 685–704.
Baird, Austin, Rachel Bodily, and Angela Merriam (2007). “Beyond Democracy: The Effects of the Electoral System on Environmental Performance.” Sigma: Journal of Political and International Studies: Vol. 25 , Article 6: 47-58.
Baker, Scott R., A. Baksy, N. Bloom, S. J. Davis and J. Rodden. “Polarized Elections Raise Economic Uncertainty.” Squarespace.com, Dec. 14, 2020.
Beramendi, Pablo and Jonathan Rodden. “Polarization and Accountability in Covid Times.” Frontiers in Political Science, Jan. 19, 2022.
Bernauer, Giger and Rosset (2015). “Mind the gap: Do proportional electoral systems foster a more equal representation of men and women, poor and rich?”International Political Science Review. 36-1: 78-98.
Birchfield, Vicki and Crepaz, Markus (1998). “The Impact of Constitutional Structures and Collective and Competitive Veto Points on Income Inequality in Industrialized Democracies.” European Journal of Political Research 34: 175-200.
Blais, A, Loewen, PJ (2007). “Electoral systems and evaluations of democracy.” In: Cross, W (ed.) Democratic Reform in New Brunswick. Toronto: Canadian Scholars Press, pp. 39–57.
Blais, Morin-chassé and Singh (2017). “Election outcomes, legislative representation and satisfaction with democracy.” Party Politics. Volume: 23 issue: 2, page(s): 85-95.
Burch, Joy (2018). “You can’t be what you can’t see – Women in the Legislative Assembly for the Australian Capital Territory.” 49th Presiding Officers and Clerks Conference, Wellington, New Zealand, 8-13 July 2018.
Carey, John M. and Hix, Simon (2009). “The Electoral Sweet Spot: Low-magnitude Proportional Electoral Systems. PSPE Working Paper 01-2009. Department of Government, London School of Economics and Political Science, London, UK.
Carey, John M. and Hix, Simon (2011). “The Electoral Sweet Spot: Low-magnitude Proportional Electoral Systems.” American Journal of Political Science 55-2: 383-397.
Cohen, Darcie (2010). Do Political Preconditions Affect Environmental Outcomes? Exploring the Linkages Between Proportional Representation, Green parties and the Kyoto Protocol. Simon Fraser University.
Evans, C. L. (2009). “A protectionist bias in majoritarian politics: An empirical investigation.” Economics & Politics, 21-22: 278–307.
Filippetti, Andrea and Vezzani, Antonio (2022). “The Political Economy of Public Research, Or Why Some Governments Commit to Research More Than Others.” Technological Forecasting & Social Change, 176 (2022).
Frank, Richard W. and Ferran Martínez i Coma (2021). “Correlates of Voter Turnout.” Political Behavior, May 11, 2021.
Fredriksson, P. G. and Millimet, D. L. (2004). “Electoral rules and environmental policy.” Economics Letters, 84-2: 237–44.
Funk, Patricia and Gathman, Christina (2010). “How do electoral systems affect fiscal policy? Evidence from state and local governments 1890-2005.” CESifo Working Paper.
Gathmann, Christina (2019). “Proportional Representation, Political Responsiveness and Child Mortality.” IZA Institute of Labour Economics. IZA Discussion Paper No. 12729. Oct. 2019.
Gordon, S. B., & Segura, G. M. (1997). “Cross-national variation in the political sophistication of individuals: capability or choice?” Journal of Politics, 59-1: 126–47.
Horne W, Adams J, Gidron N. The Way We Were: How Histories of Co-Governance Alleviate Partisan Hostility. Comparative Political Studies. May 2022.
International Institute for Democracy and Electoral Assistance (IDEA, 2022). Voter Turnout Database.
International Institute for Democracy and Electoral Assistance (IDEA, 1999). Youth Voter Participation.
Inter-Parliamentary Union (2016). Youth Participation in National Parliaments.
Inter-Parliamentary Union (2021). Monthly Ranking of Women in National Parliaments (online).
Iversen, T., & Soskice, D. (2006). “Electoral Systems and the Politics of Coalitions: Why Some Democracies Redistribute More Than Others. American Political Science Review 100-2: 165–81.
Kaminsky, J., & White, T.J. (2007). Electoral systems and women’s representation in Australia. Commonwealth and Comparative Politics 45: 185-201.
Knutsen, Carl (2011). “Which Democracies Prosper? Electoral Rules, Forms of Government, and Economic Growth.” Electoral Studies 30: 83-90.
Leblang, D., & Chan, S. (2003). “Explaining Wars Fought By Established Democracies: Do Institutional Constraints Matter?” Political Research Quarterly 56-24: 385–400.
Lehoucq and Kolev (2015). “Varying the Un-Variable: Social Structure, Electoral Formulae, and Election Quality.” Political Research Quarterly Volume 68, Issue 2.
Lijphart, Arend (2012). Patterns of Democracy. Government Forms and Performance in 36 Countries. New Haven, CT: Yale Press.
Lockwood, Ben and Matthew Lockwood (2022). “How do right-wing populist parties influence climate and renewable energy policies? Evidence from OECD countries.” Global Environmental Politics, April 6 2022.
Martin, Lanny W. and Georg Vanberg (2013). “Multi-party Government, Fiscal Institutions and Public Spending.” The Journal of Politics, Vol. 75, No. 4 (Aug. 28, 2013): 953-967.
McDonald, M., Mendes, S. and Budge, I. (2004). “What are Elections for? Conferring the Median Mandate.” British Journal of Political Science 34: 1-26, Cambridge University Press.
Nemoto, Kuniaki and Pedro Franco de Campos Pinto (2019). “Civility and Hostility in Parliamentary Politics.” November 7, 2019.
Norris, Pippa (2004). Electoral Engineering: Voting Rules and Political Behavior. Cambridge University Press.
Norris, Pippa (2011). Making Democratic-Governance Work: The Consequences for Prosperity. HKS Faculty Research Working Paper Series RWP11-035, John F. Kennedy School of Government, Harvard University.
Orellana, Saloman (2014). Electoral Systems and Governance: How Diversity Can Improve Policy Making. New York: Routledge Press. Click here for a summary produced by FVC.
Patterson, A. (2017). Not all built the same? A comparative study of electoral systems and population health. Health and Place 47: 90-96.
Pilon, Dennis (2007). The Politics of Voting: Reforming Canada’s Electoral System. Toronto: Emond Montgomery.
Plescia, Blais and Högström (2019). “Do people want a ‘fairer’ electoral system? An experimental study in four countries.” European Journal of Research.
Qvortrup, M., & Lijphart, A. (2013). Domestic terrorism and democratic regime types. Civil Wars, 15-4: 471–485.
Reynal-Querol, Marta (2002). “Political Systems, Stability and Civil Wars.” Defence and Peace Economics, 2002, Vol. 13(6): 465–483.
Rigby, Michael J, Kinga Zdunek , Fabrizio Pecoraro, Marco Cellini, and Daniela Luzi (2021). “When Covid-19 first struck: Analysis of the influence of structural characteristics of countries – technocracy is strengthened by open democracy.” PLOS ONE, Oct. 4, 2021.
Rodden, Jonathan (2019). “Keeping Your Enemies Close: Electoral Rules and Partisan Polarization” in Frances Rosenbluth and Margaret Weir, eds.: Who Gets What: The New Politics of Insecurity, Cambridge University Press, 2021.
Qvortrup, M., & Lijphart, A. (2013). Domestic terrorism and democratic regime types. Civil Wars, 15-4: 471–485.
Rule, W. (1994). “Women’s under-representation and electoral systems.” Political Science and Politic, 27: 689-692.
Rigby, Michael J, Kinga Zdunek , Fabrizio Pecoraro, Marco Cellini, and Daniela Luzi (2021). “When Covid-19 first struck: Analysis of the influence of structural characteristics of countries – technocracy is strengthened by open democracy.” PLOS ONE, Oct. 4, 2021.
Rule, W., & Zimmerman, J.F. (1994). Electoral systems in comparative perspective: Their impact on women and minorities. Westport, Connecticut: Greenwood Press.
Shaw, James (2016). Presentation at the Green Party of Canada Convention, Ottawa, August 5-7, 2016 and associated discussion.
Tasker, John Paul (2016). “5 ways the budget turns back the clock on the Harper era.” CBC News, March 24, 2016.
Various media reports (2015/2016). “Burning down the Harper legacy serving Trudeau well,” Thestar.com; “Carolyn Bennett reinstates funds frozen under First Nations Financial Transparency Act,” CBC News; “Environment Minister seeks to boost Parks Canada after nearly $30 million in Tory budget cuts,” CBC News; “Fisheries minister in Vancouver to help reopen Kitsilano Coast Guard,” National Observer; “Liberals to reopen funding taps for women’s groups,” Thestar.com; “Liberals restore refugee health benefits cut by previous government,” The Globe and Mail.
Verardi, Vincenzo (January 2005). ”Electoral Systems and Income Inequality.” Economics Letters, 86-1: 7-12.
Vowles, Jack (2013). “Electoral System Change, Generations, Competitiveness and Turnout in New Zealand, 1963–2005.” British Journal of Political Science, May 31, 2013.
Wills-Otero, Laura (2009). “Electoral Systems in Latin America: Explaining the Adoption of Proportional Representation Systems During the Twentieth Century.” Latin American Politics and Society, Volume 5, Issue 3, Fall 2009: 33 – 58.