Biased Representation:

How Compulsory Voting and Campaign Finance Interact to

Influence Government Responsiveness

Sarah Steinberg
Political Science Honors Thesis
May 2016
Contents
Acknowledgements…………………………………………………………………….…..……..3

Chapter 1: Introduction………………………………………………………………...…………4

Chapter 2: Data and Methods……………………………………………………………………21

Chapter 3: Data Analysis…………………………………………………………………...……41

Chapter 4: Design of Case Studies………………….…….….….………………………………63

Chapter 5: Australia…………………………………………………………………….…..……75

Chapter 6: Brazil…………………………………………………………………………….….113

Chapter 7: Conclusions…………………………………………………………………………158

Appendix 1:……………………………………………………………………………………..168

Appendix 2:……………………………………………………………………………………..170

Appendix 3:……………………………………………………………………………………..172

Appendix 4:……………………………………………………………………………………..173

Bibliography:……………………………………………………………………….…….….….175

2
 Acknowledgements

I would like to express many thanks to all those who helped me throughout this process. First,
Professor Bertram Johnson provided invaluable support and guidance from the moment the
project began until the very end. He directed my research, edited countless drafts, and gave me a
great deal of advice about how to navigate the whole process. I could not have done this without
him. Second, I would like to thank Professor Jessica Teets for being an important second-pair of
eyes on the project throughout the year. Her insights helped me stay focused on what was most
important and clarify the broader points I was trying to make. In addition, Professor Michael
Kraus deserves a thank you for providing me with a place in his senior seminar class to research
and write about important background information for my case study on Brazil. Mohamed
Ayman and Emily Beneroff sent me related and relevant articles that I was able to incorporate
into my final piece. Jacinta Bailey directed me towards credible news and public opinion sources
in Australia, and Tatsatom Gonçalves helped me find information about Brazilian campaign
finance and legislative activity. Lastly, to all of the friends and family members with whom I
have had extended conversations about my research, thanks for listening!

3
 Chapter 1: Introduction

In the 2012 U.S. presidential and congressional elections, candidates, political parties,

PACs and independent groups spent a reported total of $6.3 billion (opensecrets.org). Many

predict that the money spent in the 2016 cycle will be far higher—the two nominees could each

nearly double what the nominees spent in 2012 (Hunt 2015). In order to stay competitive in an

election, candidates are highly dependent on campaign donations and independent expenditures

made by Super PACs on their behalf. Because of this reliance, special interests with deep

pockets allegedly have a type of access and influence that the average citizen lacks. Americans

perceive this bias: “[S]hadowy elites” appear to exercise disproportionate influence and the

American people sense that their representatives are not really working for them (Fukuyama

2014). The public overwhelmingly wants to reform the system—a Gallup poll in 2013 found that

79% of adults favored imposing limits on contributions and expenditures in races for the U.S.

House of Representatives and Senate (Saad 2013). Yet, every year, efforts to control spending

have been unsuccessful and elections are only getting more and more expensive.

At a Cleveland town hall meeting in March 2015, an attendee asked President Obama

how to solve the problem of big money influence in politics. Instead of making the common but

far-fetched pitch for overturning Citizens United v. FEC (2010) or passing a constitutional

amendment restricting contributions and Super PAC money, Obama floated a different idea:

In Australia and some other countries there is mandatory voting. It would be

transformative if everybody voted. That would counteract money more than
anything. If everybody voted, then it would completely change the political map in this
country because the people who tend not to vote are young, they’re lower income, they’re
skewed more heavily towards immigrant groups and minority groups, and they’re often
the folks who are scratching and climbing to get into the middle class and they’re
working hard. And there’s a reason why some folks try to keep them away from the
polls… So that may end up being a better strategy in the short term. (Yan 2015)
[Emphasis Added]

4
Given that we are unlikely to see any significant changes in campaign finance soon, President

Obama is looking for creative ways to incorporate non-voters into the political process as a

means to increase political equality and improve representation.

Research Question

My aim in this project is therefore to answer the following question: Does compulsory

voting affect the quality of representation as defined by responsiveness in democracies regardless

of their campaign finance regime? Are its results contingent upon the level of regulation and

oversight of election spending, or does it make no difference when controlling for the strictness

of the campaign finance system?

We begin with the question of whether the U.S. political system is unrepresentative

because of campaign donor influence, because of especially low voter turnout, or both. As

President Obama notes, our low and significantly skewed voter turnout may introduce an elite

bias into our government. In the 2012 presidential election, only 53.58% of eligible voters cast a

ballot. In the 2014 mid-term elections for Congress, the national turnout was a meager 32.98%

(Idea.int). 1 Minorities, the poor, and less educated citizens are historically severely

underrepresented at the polls, which has possibly led to their underrepresentation in government.

Under the widely accepted delegate model of representation, representatives are supposed

to serve as proxies for their constituents’ preferences and values in a nation too large for direct

democracy. According to the model, representatives:

(1) [look] out for the good of a part (the interests of their electoral constituents), (2) [are]
defined by a third party (their constituents’ rather than their own judgment), and (3) [are]

1
These percentages are turnout of the Voting Age Population (VAP), which in the United States includes all citizens
above the age of 18. They are based on records kept during the election and the most recent census data.

5
 more responsive to sanctions (in particular, the hope of reelection). (Rehfelt 2009)2

However, representatives may have incentives to be delegates for only those who show up at the

polls or those who donate large sums of money crucial to reelection.

Compulsory voting is best defined as mandatory attendance at the polls or mandatory

submission of a ballot, which may or may not have on it a valid vote.3 Countries have

implemented compulsory voting for various reasons; one reason is to eliminate widespread

demographic bias that leaves large portions of the populations without advocates for their

interests in government. The concept of mandatory voting is not new—it has been around since

the rebirth of representative democracy in the modern era, appearing first in Georgia’s State

Constitution of 17774 and later and more prominently in Belgian laws in 1892 (IDEA.int).

Nevertheless, scholarly studies of the policy have focused on its efficacy in increasing turnout,

its morality, and its effects on reducing economic inequality.5 No comparative studies have been

done to examine its effectiveness in improving the quality of representation as defined by a

representative body’s responsiveness to its constituency.

Further, there are no studies examining the interaction between mandatory voting and

campaign finance regulations that test the relative effects of the two institutions on

responsiveness. This merits consideration, especially given that many political scientists who

have studied the influence of money in politics have found that the relationship between
2
This stands in contrast with the trustee model of representation, which says that representatives make decisions with
the greater, long-term national interest in mind even if it stands in conflict with ones own constituents’ short-term
interests.
3
Because of the secret ballot, it would be next to impossible to enforce an actual vote and most countries with
compulsory voting do not attempt to do so, sometimes allowing voters to mark boxes such as “none of the above,”
“against all,” or “abstain” (Birch 2009).
4
Virginia also had compulsory voting in the 18th century, but neither Virginia nor Georgia ever collected the penalty
fines. North Dakota and Massachusetts’s constitutions allow for a compulsory voting policy, but their legislatures
have not ever passed one (Lijphart 1997).
5
Its morality is highly contested; some believe that it violates basic principles of democracy and others believe that
it is simply weights the democratic values of participation and equality above individual freedom. Supporters of the
policy contend that democracies do mandate certain actions such as paying taxes, jury duty, military conscription
and some school attendance—occasional voting, they argue, is much less imposing (Lijphart 1997).

6
campaign finance regulation and influence is not as clear or strong as one might expect

(Ansolabehere 2003). Some find no effect at all (Smith 1995), and others have found that

campaign finance introduces biases into government and policymaking (Gilens 2012; Bartels

2008; Roscoe and Jenkins 2005). According to Gilens, elected representatives are responsive

only to the policy preferences of the wealthy (2012).

If campaign donations do not introduce bias into our government, maybe increasing

turnout regardless of the campaign finance regime can improve representation. If they do,

compulsory voting may or may not correct the money bias, depending on how much influence

campaign donations really have relative to turnout. Compulsory voting may merit further

consideration if the institution actually solves the representation deficit regardless of the level of

campaign finance regulation. On the other hand, if its effects are either negligible or dependent

upon other institutions, arguments in favor of compulsory voting are less persuasive.

Previous Studies on Representation, Compulsory Voting and Campaign Finance

Quality of representation is a tricky concept to measure given its somewhat subjective

nature and the complexity of how elected officials make their decisions. Political scientists have

used a variety of methods to try to do so. Some look at descriptive representation, a measure of

how demographically similar the legislature is to the nation (Pearson 2010). As John Adams

wrote, “A representative legislature…‘should be an exact portrait, in miniature, of the people at

large, as it should think, feel, reason and act like them’” (Pitkin 1967, 60). Others look at the

amount of casework a legislator does for his constituents to measure representation and

responsiveness (Abdel-Samad 2009). Some look at the roll call voting of representatives and

whether they match the preferences of certain groups, such as voters and non-voters (Griffin and

7
Newman 2005; Peltzman 1984; Miller and Stokes 1962). Others compare policy outcomes with

the preferences by income group (Gilens 2012) and still others look at how ideological

polarization in Congress reflects polarization in the population at large (Fiorina 2009).

Matsusaka looks at the “degree of congruence between policy choices and public opinion”

on the state level (2010, 133).6 My approach is similar to Matsusaka’s in that I look at macro-

level outcomes to determine whether they reflect the interests of constituents on the whole as

measured by public opinion surveys or the interests of special, concentrated groups that donate

significant amounts to campaigns. In this study, I determine whether the policy of compulsory

voting increases congruency between public opinion and policy outcomes, and if its effects on

congruency (what I call government responsiveness) are dependent on the amount or type of

campaign finance regulation.

Few would contest the evidence that compulsory voting works if the desired outcome is

nearly universal turnout. In countries where it is enforced, even weakly, turnout rates are

consistently high (Brennan and Hill 2014; Chong and Olivera 2008; The Electoral Commission

2006). There is no real consensus on how the policy affects other variables beyond turnout,

though there is a large body of literature on why and how turnout matters.

The theory that higher turnout would affect representation (either in a positive or negative

way) can be understood only if we look to the delegate model of representation. Delegate

representation can be measured empirically and descriptively. If a representative is responsive to

the views, needs and policy preferences of his or her constituents, he or she can be defined as

6
He also looks for correlations with different campaign finance and election laws such as the availability of direct
democracy, contribution limits, public funding and independent commission redistricting (2010). He finds: “states
chose the policy preferred by a majority of citizens (equivalent to the median voter outcome) 59 percent of the
time—only 9 percent more than would have happened with random policymaking. Majoritarian/median outcomes
were 18-19 percent more likely when direct democracy was available, and 11-13 percent more likely when judges
were required to stand for reelection. The likelihood of a majoritarian/median outcome was not correlated with a
variety of election laws, including campaign contribution limits, public funding of campaigns, and commission-
based redistricting” (Matsusaka 2010, 133).

8
highly representative. The literature suggests that delegate representatives whose seek to

represent their constituents’ interests in the United States are more responsive to those who show

up at the polls, though some would contend that politicians are responsive to an even smaller

subset of the population: the white, wealthy and well-educated (Gilens 2012). That additional

bias aside, the mantra seems to be “if you don’t vote, you don’t count” (Burnham 1987, 99). V.O.

Key wrote similarly, “the blunt truth is that politicians and officials are under no compulsion to

pay much heed to classes and groups of citizens that do not vote” (Key 1949, 527). A study in

2005 found that “voter preferences predict the aggregate roll-call behavior of Senators while

non-voter preferences do not” (Griffin and Newman 2005, 1206; McElwee 2015b). While the

consensus is not absolute, many theoretical and empirical studies come to similar conclusions,

finding that especially where the interests of voters and non-voters diverge, representatives favor

the policies of the voters who can help their reelection prospects (Fiorina 1974; Hill and Leighly

1992; Canes-Wrone, Brady, and Cogan 2002). There do not appear to be major differences

between voters and non-voters on foreign policy, ideological issues or issues concerning

partisanship; however,

Nonvoters are slightly more in favor of an increased government role in the domestic
arena. They are more likely to oppose curtailing government spending for health and
education services, and they are more likely to support government guarantees that
everyone has a job and a good standard of living. (Bennett and Resnick 1990, 789-793)

A study of 12 Western European countries between 1830 and 1938 found that “the gradual

lifting of socio-economic restrictions on the voting franchise contributed to growth in

government spending” (Aidt et al. 2006, 249; McElwee 2015b). Correspondingly, another study

found that turnout is positively correlated with the extent of government redistribution and also

predicts a reciprocal effect; when more people of different socioeconomic statuses vote,

redistribution in favor of lower income citizens becomes more likely, and when there is more

9
redistributive policy, lower-income citizens have a higher stake in government and show up

more frequently at the polls (Mahler 2008).

Arend Lijphart, former president of the American Political Science Association, wrote

famously on turnout in his piece “Unequal Participation: Democracy’s Unresolved Dilemma”

(1997). First, he contends, unequal turnout is biased against poor citizens. Second, turnout is

declining everywhere and is especially low in the United States. He proposes that compulsory

voting could offset this class bias by increasing political equality. This proposition has spurred

greater, targeted research in the field of representation and compulsory voting.

In response to Lijphart’s bold theory, scholars have looked at how compulsory voting

affects redistributive policies. If representatives are responsive to their voting constituents, one

should find that nearly universal turnout under compulsory voting pushes policy outcomes in the

direction of the preferences of the previous non-voters (poor and minority groups). In practical

terms, as discussed above, this would mean more economic redistribution. Some scholars have

found evidence of exactly this—Chong and Olivera’s study of 91 countries over the years 1960-

2000 found that compulsory voting improves income distribution. The researchers apply the

median voter theorem: when the median voter’s income is below the mean, aggregated

preferences will lead to more redistributive policy, and vice versa (2008). Another statistical

analysis reveals the same pattern: “It does appear that in Western Europe and Latin America, at

least, mandatory attendance at the polls promotes social equality” (Birch 2009, 131).

O’Toole and Strobl’s 1995 study found that compulsory voting was correlated with an

increase in government spending on health, housing and transfer policies and a decrease in

government spending on defense and economic services. In Australia, the implementation of

mandatory voting in 1924 changed the party composition of parliament in favor of the Labor

10
Party and led to higher pension spending at the national level (Fowler 2013). A recent working

paper that looked at the abolition of compulsory voting in Venezuela in 1993 found that the

repeal of the policy correlated with rising economic inequality in the years afterwards (Carey and

Hoiuchi 2015). One would presume that the government allowed the rising economic inequality

and policies that reinforced the trend because the new set of representatives in the national

legislature became less responsive to the population that had stopped voting—most notably,

those in the lower income-quartiles.

In contrast, other research suggests that compulsory voting would neither change

outcomes drastically nor be desirable. Brennan argues that voluntary voters actually vote for

what they believe is in the “national interest” because they know they cannot gain individually

from voting for one candidate or party over another. New voters, he contents, are more likely to

vote for ‘bad’ candidates. Groups that currently do not vote abstain for a reason—they are

“ignorant, misinformed, irrational and biased about political issues” (Brennan 2014, 44) and are

not scholars in political science, sociology or economics. They cannot discern good policy

platforms from bad ones, regardless of how noble their preferences are (Brennan 2014). As

Winston Churchill once purportedly said: “The best argument against democracy is a five-minute

conversation with the average voter.” 7

The theory that voluntary non-voters are less informed than voluntary voters has been

supported by various studies (Birch 2009), though other authors suggest that this might change

under a compulsory voting system. If citizens know they have to vote, some claim that they will

7
There are contradictory theories on how compulsory voting affects quality of government. If non-voters are indeed
terribly misinformed and under-qualified to judge policy platforms and their own interests, the quality of the
candidates and political discourse may suffer under a compulsory voting regime. The subject merits further
consideration. However, it does not affect this study of how nearly universal turnout influences the quality of
representation as defined by constituent-representative congruency and representative responsiveness to the average
voter.

11
pay more attention to political issues and therefore achieve higher political knowledge levels

(Lijphart 1997; Birch 2009; Brockington 2005). In addition, instead of focusing on likely

voluntary voters, civil society issue groups and political parties may spread their information to a

wider population, knowing that they all must show up to cast a ballot. Those who were

previously excluded will then have opportunities to become better informed.

Opponents of compulsory voting worry that voluntary non-voters would vote randomly if

compelled to cast a ballot, though this fear appears to be overstated. Because of the nature of

Australia’s preferential voting system, researchers have been able to study the frequency of

random voting, also known as the ‘donkey vote,’ which is when a voter marks 1 through 5 in

order down the ballot. Different studies have determined that somewhere between 2-3% of

ballots in Australia were either intentionally ruined or randomly assigned (Galston 2011), a

figure that may in fact be lower than in many countries with voluntary voting (Birch 2009).

Only one study has looked more directly at the connection between compulsory voting

and representation, through it uses a measure of perceptions of quality of representation. In this

statistical analysis, Birch found no statistically significance difference between countries with

sanctioned compulsory voting and countries with voluntary voting. This finding led Birch to

believe that changes in “electoral integrity” are dependent not just on the existence of

compulsory voting but also on other factors such as the social and political context in which the

institution exists (Birch 2009, 115). However, perceptions do not always align with reality,

especially when those perceptions were gathered in opinion surveys subject to various other

biases. For example, survey respondents may have interpreted ‘representation’ differently, or

may not actually know much about how their representatives perform in office. It is neither

universally accepted nor empirically proven that compulsory voting produces a higher

12
congruency between constituent views and needs and their representatives’ actions; therefore, the

subject warrants further consideration. My statistical analyses and case studies serve to examine

whether real representation is high as determined by actual policy outcomes. I look closely at the

political and social contexts of each country to see if other factors do interfere with the influence

of compulsory voting as Birch proposes.

Regardless of its effects as a stand-alone policy, compulsory voting may not be the sole

factor that influences congruency between constituents’ preferences and their representatives’

actions. The campaign finance system too, among many other variables, may affect quality of

representation. Political campaigns cannot succeed without substantial amounts of money and

there may be electoral costs to ignoring the interests of voters and contributors alike. Though the

link between campaign contributions and legislative action is far from empirically proven, there

is good reason to believe that campaign finance affects representation.8 A meta-analysis of over

30 studies of the U.S. Congress found that “one-third of roll call votes exhibit the impact of

campaign contributions” (Roscoe and Jenkins 2005, 52). These results confirm what many

already suspected. It would seem that at least some campaign contributions are made to influence

votes and not simply to express approval of a candidate’s platform. The influence of money is

not limited only to votes on the floor of the legislature; elected representatives can work for

interest groups when drafting policy language, when testifying in committees and by pressuring

colleagues. While subtle, these types of influence also disrupt the link between the average

8
Part of the support for this claim is the biological principle of “reciprocal altruism,” which Francis Fukuyama
insists is “genetically encoded in our brains and emotions. A human being in any culture who receives a gift from
another member of the community will feel a moral obligation to reciprocate” (Fukuyama 2014, 478). Robert
Putnam describes a system of reciprocity that is essential to effective and good governance. Reciprocity, as Putnam
quotes from philosopher Michael Taylor, is “usually characterized by a combination of what one might call short-
term altruism and long-term self-interest: I help you out now in the (possibly vague, uncertain and uncalculating)
expectation that you will help me out in the future” (Putnam 2000, 134). According to these theories of human
behavior in society, interest groups may donate money to political campaigns and not expect an immediate return
favor; however, they expect that the donation will not go necessarily go unrewarded over time.

13
constituent and his or her representative. The fact that some people and organizations donate

heavily to both parties reinforces this perception. These donations cannot be purely ideological—

they are instead intended to secure some type of access or influence in policy regardless of which

side wins.

Campaign finance regulation can be present in a wide variety of ways and can differ both

in kind and in degree. Some countries opt for contribution and expenditure limits with disclosure

and transparency, some for public financing of campaigns, some for free media access for

candidates or parties and others for different combinations of policies. All types of regulation can

be strict, weak, or loosely enforced. Given that both compulsory voting and the level of

campaign finance regulation may have some effect on the same dependent variable (quality of

representation) the two may interact.

Some scholars believe that compulsory voting and campaign finance may be connected in

a different way: compulsory voting may directly reduce money in politics.9 This theory is

interesting, but it speaks of a secondary effect and alternate reason for the implementation of

compulsory voting. Such an effect will not the main focus of this study; however, it demonstrates

that scholars have previously speculated that the two policies may be connected somehow. The

nature of that connection, whether it be direct causation or correlation through a third variable,

has been unproven to date. I investigate whether the interaction of compulsory voting and the

existing campaign finance regime has some effect on quality of representation through elected

official’s responsiveness to the overall public interest. No other scholars have considered this

specific interaction. My findings serve to address this existing gap in the literature.

9
Lijphart predicts that compulsory voting could reduce the role of money in politics. With mostly everyone
guaranteed to vote, there would be no need to spend so much money inducing non-voters to show up at the polls
(Lijphart 1997). William A. Galston, a senior fellow at the Brookings Institute, agrees that compulsory voting could
lead to less election-related spending and reduce the incentive to discourage turnout through negative ads,
“improv[ing] not only electoral politics but also the legislative process” (Galston 2011).

14
Theories

Both compulsory voting and campaign finance can influence representation, and

democracies around the world have various combinations of these two institutions. The two may

interact in a predictable way to either improve or weaken quality of representation. In this section,

I outline key theoretical expectations of how these two variables may interact.

Countries may only experience a strong increase in quality of representation when

compulsory voting is combined with strict campaign finance laws. If nearly everyone votes and

big money special interests contribute and spend without limits, oversight or transparency, the

previous non-voters may be unduly influenced to vote for candidates who do not truly represent

their interests. Typically less informed about the issues and policy alternatives, non-voters may

be susceptible to negative attack ads and candidates who make unrealistic promises. Politicians

may know that they can avoid electoral accountability by throwing money at their campaigns and

misleading constituents in what Jane Mansbridge calls the “manipulation” of voters (2003).

Politicians may vote for policies that benefit their donors to ensure that the money keeps coming

so that they can continue misleading their constituents, thus ensuring a cycle of poor

representation. Alternatively, in a highly regulated system that encourages issue discourse,

provides public funding and emphasizes reporting, disclosure, and prosecution of violators, a

new influx of voters may improve the quality of representation by offsetting the money bias.

Representatives need not pick the interests of contributors over the interests of voters if

contributions are not as essential to electoral success. When money is highly regulated in

campaigns, the opportunity cost in campaign dollars of voting for the preferences of constituents

instead of contributors might be lower, relegating special interest preferences to a secondary

concern.

15
 On the other hand, the strictness of campaign finance regulation may have no significant

effect on compulsory voting’s influence on quality of representation. Nearly universal turnout

may replace money in importance. As President Obama suggests, maybe simply mandating

attendance at the polls counteracts the big-money bias because candidates must represent all of

their voters, not just their donors and voluntary voters. Scholars like Birch, Lijphart and

Brockerton have speculated that voters become more informed when they know they have to

submit a ballot; therefore, they may become less susceptible to manipulative political

propaganda. Lijphart and Galston have also suggested that the simple presence of compulsory

voting may make high levels of election spending irrelevant—money is not needed to ‘get out

the vote.’ Representatives facing a trade-off between satisfying their voters or their contributors

in a country with compulsory voting might systematically prefer the preferences of their voters

regardless of how election spending is regulated.

Do strict campaign finance rules correlate with better representation regardless of the

presence of compulsory voting? Maybe compulsory voting does not significantly change

electoral outcomes, as Brennan suggested, and regulation of election spending is alone correlated

with quality of representation. Contribution and expenditure limits, programs for public funding

and free or subsidized media access, disclosure and sanctions all may control the big-money bias

and allow representatives to focus on pleasing their voting constituents and not just their

contributors. If compulsory voting and strict campaign finance regulation are positively

correlated, some past research on compulsory voting may be erroneous—effects attributed to

compulsory voting may in fact be a product of strict campaign finance rules.

Finally, perhaps neither variable improves the quality of representation. Maybe too many

other contextual and structural variables interfere and each country must go about creating

16
unique regimes to reflect the individual problems that that country faces. Transparency may

matter in Country A but not Country B; nearly universal turnout may matter in Country X but

not Country Y. Representatives and constituents see their roles and relationships differently. In

this scenario, countries wishing to implement either campaign finance reform, compulsory

voting or both to improve quality of representation must carefully design legislation that is

appropriate for its unique system.

Looking Ahead

Mandatory voting and voter turnout have been studied for many years, but there is a gap

in the research when it comes to understanding how the policy and phenomenon interact with a

closely related theme: election spending. Weak campaign finance regulation allows large sums

of money to pour into the political process. Money introduces a socioeconomic bias into the

political process that appears to disproportionately hurt (or at the very least, not help) the poor

and other marginalized groups—groups that often do not vote. The strongest fix to the turnout

imbalance is the institutional one: compulsory voting. Perhaps if everyone voted, election money

would matter less and our representatives would be equally responsive to both the most affluent

and to the groups who need the most help. Political equality, after all, is essential to a well-

functioning government of, by, and for the people.

The research to date suggests that counties with compulsory voting enact more

redistributive policies favorable to minority groups and poorer constituents. The theory that most

authors use to explain this finding is that the delegate-style representatives voting to pass these

types of redistributive policies do so because either they are more responsive to all of their voting

constituents, not just the wealthy ones, or they are responsive to the new median voter who is

17
less wealthy than the median voluntary voter. This is because to win reelection, they must also

win the vote of their poorer constituents who are also guaranteed to show up at the polls. The

quality of representation as defined by constituent-representative congruency on policy

preferences and legislative responsiveness to the median voter in countries with mandatory

voting would therefore be different than in countries with chronically low turnout. Strictness of

campaign finance laws may affect the magnitude of the effect that compulsory voting has on

quality of representation because its absence or weakness may allow big-money interests to

unduly sway policy decisions and votes in their favor. That is to say that the representative’s

trade-off may still be relevant, only to a different degree.

In the remainder of this thesis, I examine the effect of compulsory voting and campaign

finance regulation on quality of representation in various combinations of the two institutions to

determine if either are worth pursuing singularly or jointly in the quest for better representation.

Compulsory voting may be desirable for a variety of reasons; however, if it only achieves the

desired result of higher quality of representation in countries that also strictly regulate campaign

finance, we can no longer view it as a singular potential remedy to the socioeconomic bias in

representation here in the U.S. Alternatively, if compulsory voting consistently improves

government responsiveness to the average voter regardless of campaign finance regulation, we

may wish to give the policy more consideration.

I answer this question by first conducting a large-n statistical analysis to see if the

interaction of compulsory voting with different levels of campaign finance regulation improves

representation as approximated by various related dependent variables. I then conduct case

studies of two countries with compulsory voting and different levels of campaign finance:

Australia and Brazil. These countries allow me to look more precisely at how the policies work

18
(or do not work) to align constituent preferences with policy outcomes. Within each case study, I

examine the electoral system and systems of accountability, the compulsory voting and

campaign finance policies in detail, and whether legislation passed and not passed during a year

directly following an election more closely reflect donor or voter preferences.

I find no consistent statistical relationship between compulsory voting, campaign finance

regulation or their interaction on my dependent variables. Even though we cannot be sure of

causality, nor that the variables are significant with regards to the representative component of

each of the dependent variables, certain individual models do show some meaningful

significance. For example, average turnout is associated with a positive increase in the score of

one of my dependent variables (the Voice and Accountability Index), as is the interaction of

enforced compulsory voting with regular party campaign finance reporting. The interaction

between enforced compulsory voting and candidate contribution limits is significant regarding

Voice and Accountability and Control of Corruption, though candidate contribution limits are

also significant in their interaction with mean turnout on Control of Corruption. Various other

models show some statistical significance, and these are explained further in Chapter 3.

My case studies examine more closely how these policies act and interact. I find that in

Australia, voter preferences prevail when they exist, though in the absence of strong public

opinion, interest groups who donated (and also likely spent money on lobbying efforts) tend to

get what they want. Compulsory voting may matter in aligning public preferences with outcomes,

and nearly unregulated campaign money did not seem to impact this effect. This example

suggests to us that perhaps all you need for high quality representation is some basic

transparency paired with nearly universal turnout. It would follow that any further attempts to

regulate money in campaigns are unnecessary—nearly universal turnout alone makes

19
government highly responsive to the public will.

In Brazil, we see a very different result. The electoral system creates a political chaos that

is not compatible with any meaningful accountability despite nearly universal turnout. The sheer

number of parties and representatives who do not necessarily vote with their party along with a

media that is not particularly interested on reporting roll-call votes make it impossible for

average voter to be fully informed and punish representatives who betray him or her at the polls.

This system likely nullifies any effects that compulsory voting may have otherwise had. Big

moneyed interests typically get the legislation they want or block the efforts that they do not,

though this influence may be entering the system through campaign finance, lobbying efforts, or

other illicit means like bribery.

These results tell us that compulsory voting is not a ‘magic bullet’ solution that

automatically makes all systems more representative. Other contextual factors do matter, such as

the transmission of relevant information to voters and the wider electoral system and political

culture. The prevalence of corruption matters, as do lobbying efforts. Even in a system that

appears highly responsive to its electorate such as Australia, we find that money can matter if an

issue is not salient in the eyes of the public.

Taken together, these conclusions present only a weak link between compulsory voting

and better quality representation that is mediated by other institutions. We therefore cannot say

with certainty if the introduction of compulsory voting would make U.S. policy outcomes better

reflect the opinions of the general public; however, we can speculate given the proximity of our

institutions to those of Brazil and Australia.

20
 Chapter 2: Data and Methods

For my statistical analysis, I employ a “nested analysis” design as my framework: “[this

type of design] combines the statistical analysis of a large sample of cases with the in-depth

investigation of one or more of the cases contained within the large sample” (Lieberman 2005,

435-436). As Lieberman explains, the advantage of a nested analysis is that the strategies of case

selection can be guided by the results of the large-n statistical tests. In accordance with his

methodology, I first run many large-n regression models to see if compulsory voting and

campaign finance have any independent or joint effect on quality of representation (or

perceptions of quality or representation) in democracies. In some models, I also substitute a

simple measure of turnout for the compulsory voting variable.

Representation and government responsiveness cannot be perfectly measured statistically,

nor does a dataset exist including a widely accepted ‘representation’ or ‘responsiveness’ index.

Instead, I attempt to measure these concepts by using many related dependent variables. I use

two indexes (Voice and Accountability and Control of Corruption) and four averaged survey

results (Favoritism, Confidence in Parliament, How well people feel they are represented in

elections and If there is a leader that represents people). If my independent variables—campaign

finance regulation, compulsory voting and turnout—and their interaction effects have consistent,

statistically significant effects across various models, I can be confident in the robustness of the

results.

This chapter explains in detail the dependent and independent variables I use in my

analysis and contains justification for each of my control variables.

Measuring Representation: Dependent Variables

To first determine whether compulsory voting, campaign finance regulation, or the

interaction of the two together have a clear effect on quality of representation in democracies

around the world,10 I run large-n regression models with various dependent variables that

approximate quality of representation. No perfect empirical measure of good representation

exists, though some existing variables measured by scholars and international organizations

contain components that speak to the quality of the link between constituent and representative. I

focus on six different indicators: four are based on survey data and two are aggregate indexes

that contain survey results as components. If my independent variables have a consistent positive

effect on most of the dependent variables, we can conclude that the policies may be speaking to

what the dependent variables all have in common: representation. If the independent variables

are not statistically significant or are significant sporadically, or if they relate to only one or two

of the dependent variables, they might perhaps be related to other components of the dependent

variables that are not representativeness or responsiveness.

First, I use the “Voice and Accountability” index of the World Bank’s Worldwide

Governance Indicator dataset. The variable is defined as one that “captures perceptions of the

extent to which a country's citizens are able to participate in selecting their government, as well

as freedom of expression, freedom of association, and a free media” (Kaufmann and Kraay 2014).

While many of the indicators are objective, a significant number come from surveys to reflect

how citizens of each country feel their country is performing. These include questions on

satisfaction with democracy, trust in parliament, whether the system provides for accountability

and public voice, whether there is dialogue between government and rural organizations and the

transparency of government policy. Each country is given a composite score on a continuous

scale between -2.5 and 2.5 (a higher numeric score reflects better performance on all of the

10
I am only including in the analysis countries that are classified as either a liberal democracy or electoral
democracy by Cheibub, Gandhi and Vreeland in their “Democracy and Dictatorship Revisited” data set.

22
indicators). If representatives frequently and completely ignore their constituents, a country

would perform poorly on this index. For example, the mean Voice and Accountability score for

autocracies, many of which do not have or claim to have any representative institution, is -1.2584.

The mean score for electoral democracies with significantly flawed representative institutions is -

0.5154286. Liberal democracies that emphasize participation and government for and by the

people score a mean of 0.82633. The highest score is Norway with a 1.76; the lowest is Somalia

with a -2.19.11 Australia scored 1.44 and Brazil, 0.37.

I also use the World Bank’s Worldwide Governance Indicators “Control of Corruption”

index. This variable is a composite index that “captures perceptions of the extent to which public

power is exercised for private gain, including both petty and grand forms of corruption, as well

as "capture" of the state by elites and private interests” (Kaufmann and Kraay 2014). The index

is not a perfect measure of corruption in representative bodies; it includes corruption at the

bureaucratic level and corruption in other public institutions, such as the state police. Similar to

Voice and Accountability, each country is given a score between -2.5 and 2.5 (a higher score

reflects better control of corruption) and the score is a mix of objective empirical data and survey

data that include the public’s perception of the prevalence of corruption in its government.

Corruption is one way by which the constituent-representative link is disrupted; corrupt

representatives fail to do what their constituents want or need in favor of special interests for

personal gain (for example, bribery and patronage). While a lack of corruption does not

guarantee in and of itself high quality representation, it is a necessary condition. System-wide

11
Before 1991, the State of Somalia was officially recognized as the Somali Democratic Republic. In 1991, a coup
overthrew the democratic government and initiated a prolonged civil war (United Nations). Many categorized
Somalia as a failed state. In 2012, Somalia became a (fragile) federation with a provisional constitution; however,
the data I am using does not yet reflect this transition and still treats Somalia as an autocracy without representative
bodies.

23
corruption plays a distorting role.12 A country that scores poorly cannot be highly responsive to

the people, though a country that scores well is not necessarily highly responsive. However, if

certain independent variables are associated with a positive increase on this index and also

significant regarding some of the other dependent variables, we may conclude that representation

has improved. Denmark scores the highest with a score of 2.41 and Equatorial Guinea has the

lowest score with -1.61. Australia scores 1.76 and Brazil, -0.12.

Similar to the Control of Corruption Index, I look at the World Economic Forum’s

“Favoritism in decisions of government officials” measure, contained in the Global

Competitiveness Report of 2015. Survey participants were over 14,000 business leaders in 144

countries. They responded to the question “In your country, to what extent do government

officials show favoritism to well-connected firms and individuals when deciding upon policies

and contracts? [1 = show favoritism to a great extent; 7 = do not show favoritism at all]” (World

Economic Forum 2015). Each country has been given a score that is the mean of Respondents’

answers. The lower the value, the less representative government officials are perceived of being

of the average constituent (or median voter) in favor of special interests with money and

influence when making policies and giving work contracts. A high score on this index may be a

better indicator of a high quality of representation than the Control of Corruption index because

it may denote more equal representation instead of only the lack of a political defect. For this

variable, Venezuela (coded as an electoral democracy) scores the lowest with 1.480 and Qatar

(an autocracy) scores the highest with a 5.582. While Qatar is an unexpected leader in this

12
An individual representative may accept bribes, but not actually influence outcomes. Or, the individual bribe may
be about a policy issue on which the public has no strong opinion; therefore, actions related to the bribe are not
necessarily showing unresponsiveness to the public. However, this index speaks to the overall systematic control of
corruption. System-wide corruption (as indicated by a poor score on this measure) likely does not only influence
obscure, concentrated policies but also matters that affect many. For example, Brazil’s Petrobras scandal involves a
widely corrupt political class and the systematic misuse of public funds. (See more in Chapter 7).

24
category, on average the liberal democracies do score slightly better (3.256 for autocracies and

3.358 for liberal democracies). Electoral democracies score the worst, with an average of 2.916.

This suggests an interesting phenomenon; perhaps when citizens do not have an expectation of

equal voice or when they receive highly biased government controlled media (as in autocracies),

they are less likely to identify favoritism. Perhaps some business leaders are afraid of negative

consequences for speaking the truth, or perhaps business leaders interpret favoritism differently

(for example, a business that operates in a favorable policy climate might not view that climate

as a product of their own efforts, or believe that that climate is fair to all). Since I am excluding

autocratic countries from my models (the premise of representation as it relates to turnout rests

upon a citizens ability to vote), these data points will not influence my results; however, one

should look at the test results using this dependent variable with a grain of salt. Australia scores

3.94; Brazil scores 2.58.

Next, I employ the “Confidence in Parliament” measure of The Electoral Integrity Project

(“Perceptions of Electoral Integrity”). This data comes from survey results asking respondents to

“rate how much confidence you have in those organizations in this country, from a 1 (no

confidence at all) to 10 (a great deal of confidence)?” (Norris, Martinez and Gromping 2015).

The survey asked about various institutions including the Courts, the Government as a whole, the

armed Forces, the Police, The Press, etc. Because the legislature is theoretically the most

representative body in government, I have chosen to only focus on the “Parliament” measure.

One would assume that survey respondents in countries with poor representation that do not

seem to enact policies that benefit the majority of their citizens would score lower on this index

than those with quality representation. Autocracies and electoral democracies average answers of

3.712 and 3.984 respectively; liberal democracies average a score of 5.815. Bahrain scores the

25
lowest (1.500); Denmark, the highest (8.667). Australia scores a 7.375, and Brazil, 5. However,

confidence in the entire legislative body is not necessarily the same as confidence in one’s own

representative, especially for countries with First-Past-the-Post/Single-member districts where up

to 49.9% of the voters in a district may lose. In addition, in the United States, it is well

acknowledged that constituents rate their own individual representative highly but have a poor

opinion of Congress (Fenno 1978). A low rating on this variable would not necessarily speak to

the individual representation (or perception of representation) that each citizen is receiving. If all

representatives were representing their own constituencies well, on the aggregate a system would

be highly representative. Yet, it still might receive a low rating. This variable may be only

remotely related to quality of representation.

The last two measures of representation I use are from the Comparative Study of

Electoral Systems (CSES), Module 2 (2001-2006) dataset. Survey Respondents from 35

countries (1 autocracy: Russian Federation; 3 electoral democracies: Kyrgyzstan, Albania and

the Philippines, 31 liberal democracies) answered the questions: “Thinking about how elections

in [country] work in practice, how well do elections ensure that the views of voters are

represented by MPs [Members of Parliament]: very well, quite well, not very well, or not well at

all?” and “Regardless of how you feel about the parties, would you say that any of the individual

party leaders/presidential candidates at the last election represents your views reasonably well?”

(Sapiro and Shively 2007). The first question is coded on a scale of 1 (best) to 4 (worst) and the

second is coded as 1 (yes) or 2 (no). The data in the dataset comes by survey respondent for each

country (there are about 2000 responses per country). Excluding answers such as “Don’t Know,”

“Refused” or “Missing,” I found the mean answer for each country to use in my models. Brazil

scores the highest mean value on whether voters were represented in elections (therefore,

26
 Brazilians have the lowest perception of representation in elections of the survey countries) and

Denmark scores the lowest (Danish respondents had the most confidence that they were being

well-represented in elections). On the question of whether there was an individual leader in the

last election that represented respondents’ views, Russia scored the highest (the most no’s) and

New Zealand scored the lowest (the most yes’s). While these two questions directly touch upon

the people’s perception of the quality of their own representation, the small country sample size

makes the findings regarding these dependent variables somewhat limited, as does the vagueness

of the question—people may have interpreted representativeness in different ways.

All six of these dependent variable are summarized in Table 2.1.

Table 2.1: # in Scale Mean Standard Minimum Maximum Australia Brazil

Summary of Sample Score Deviation
Dependent
Variables
Voice and 195 -2.5 – -0.065 1.00 -2.19 1.76 1.44 0.37
Accountability 2.5 (Somalia) (Norway)
(best)
Control of 193 -2.5 – -0.078 0.99 -1.61 2.41 1.76 -0.12
Corruption 2.5 (Equatorial (Denmark)
(best) Guinea)
Confidence in 125 1-10 4.78 1.63 1.5 8.67 7.38 5.00
Parliament (best) (Bahrain) (Denmark)
Favoritism 142 1-7 3.20 0.88 1.48 5.58 3.94 2.58
(best) (Venezuela) (Qatar)
How well are 35 1-4 2.56 0.18 2.09 2.94 2.44 2.94
voters (worst) (Denmark) (Brazil)
represented?
Is there a leader 35 1 (yes) 1.42 0.17 1.17 1.81 (New 1.21 1.36
that represents 2 (no) (Russia) Zealand)
you?

The significance of my independent variables on these six dependent variables paint an

overall picture about how well governments are performing in their representative functions or if

any patterns exist. All of these variables are correlated enough to be significant; however, none

27
are so closely related that they are in essence measuring the same thing (see Table 2.2 below).

This was my intention—quality of representation is complex, multi-dimensional and difficult to

measure empirically. Public perception of the level of representation is an essential component;

what is representation if not how well a constituent feels his or her views are defended in

government? Yet, public perception, especially as measured by survey data, is subject to various

biases. Respondents may not be well informed about how their representatives are voting or may

be unaware of corruption; respondents may hold opinions contrary to what is actually in their

own interests (and thereby may mistakenly support or oppose a representative); respondents may

interpret the survey questions differently or answer dishonestly; respondents may interpret the

scales differently. These dependent variables all touch on different facets of representation and

have been collected and calculated differently—their variety hopefully controls for some of these

biases. Nevertheless, the story behind representation cannot be told entirely through numbers.

My in-depth case studies should expose some of the biases and illustrate more closely what may

actually be happening in the examined countries.

Table 2.2: Is there a

Correlation Between Confidence How well are leader that
Dependent Variables Voice and in voters represents Control of
Accountability Favoritism Parliament represented? you? Corruption
Voice and Accountability 1
Favoritism 0.53585058 1
Confidence in Parliament 0.746293507 0.579925 1
How well are voters
represented? -0.43567337 -0.37018 -0.373889 1
Is there a leader that
represents you? -0.45997292 -0.47161 -0.454702 0.3338379 1
Control of Corruption 0.812165666 0.798501 0.6957405 -0.4713664 -0.4979414 1

Measuring Compulsory Voting

Compulsory voting is one of my primary independent variables. It is binary; a country

28
either has it or does not (though a lack of enforcement may make the policy null in effect). I

created a second variable that represents the presence of enforced compulsory voting only and

edited the data according to the up-to-date International IDEA database and the Electoral

Commission Report of 2006. For this variable, I classified countries for which unenforced

compulsory voting seems to have no distinguishable effect on turnout as countries without

compulsory voting. I did so because turnout is important as an intervening variable to explain

how compulsory voting affects representation. I ran regression models with both the original data

that does not take enforcement into account and my altered data to see if simply having the

policy has a statistical effect or if enforcement matters. The initial classifications come from

Pippa Norris’s “Democracy Cross-national Data Set,” though I made various changes to the

unenforced compulsory voting category as well to reflect current conditions.13 Of the 191

countries in Pippa Norris’s Data set, 21 have compulsory voting written into law and only 16

enforce the policy.

Table 2.3 below summarizes the changes I have made:

TABLE 2.3: Changes to the dataset for accuracy & enforcement of compulsory voting
Countries that have Unenforced Unenforced Counties coded as
since abolished the compulsory voting compulsory voting not having
practice/are no with no notable effect with a notable effect compulsory voting
longer a democracy on turnout (recoded on turnout (left as is that do (recoded in
(recoded from having in the enforced in the dataset and in the dataset)
compulsory voting to compulsory voting the enforced
not having category from having compulsory voting
compulsory voting compulsory voting to variable)
nor enforced not having
compulsory voting) compulsory voting)
Egypt Costa Rica Singapore Turkey
Chile Dominican republic Greece Nauru
Venezuela Honduras
Guatemala Thailand
Italy

13
Explanations of all of the changes I made can be found in Appendix 1.

29
 While my main focus is on the effects of the policy of compulsory voting, the underlying

question is whether high turnout (the ultimate end of enforced compulsory voting) influences

representation. Because there are so few countries with compulsory voting, I also look at turnout

data in all of my models. If higher turnout has a significant effect on representation but

compulsory voting itself does not, we may conclude that other factors are at play that bias the

results of the compulsory voting countries. For example, Brazil has a chronic corruption

problem; perhaps many of the other countries with compulsory voting also have the same issue.

If compulsory voting countries are not correlated with high quality representation but high

turnout is, any policy that increases turnout may still be worth pursuing to gain better

responsiveness of a government to its constituents.

Measuring Campaign Finance Regulation

Data on campaign finance laws by country is also readily available on International

IDEA, though it is broken down by type of policy (i.e. contribution limits to candidates,

contribution limits to parties, access to subsidized media for candidates, ban on vote buying,

regular reporting). A variable aggregating various measures of party regulations regarding

campaign spending already exists in Pippa Norris’s Democracy Cross-national Data Set. This

variable constitutes a general metric of the stringency of the campaign finance regime but is not a

good measure of how well the regulations are enforced. I have run regression models to see if the

aggregate variable or any of the individual components have any effect on my dependent

variables. Specifically, I have looked more closely at 1) Is there a ban on corporate donations to

candidates?14; 2) Is there a limit on the amount a donor can contribute to a political party over a

14
The ban on corporate donations to candidates showed no statistical significance in any of my models, through
interaction effects or independently. For this reason, it is not discussed further in the next chapter.

30
time period (not election specific)?; 3) Is there a limit on the amount a donor can contribute to a

candidate?; 4) Are there provisions for direct public funding of political parties? 15; 5) Are there

provisions for free or subsidized access to media for political parties?; 6) Is there a ban on vote

buying?; 7) Are there limits on the amount a candidate can spend?; 7) Do political parties have to

report regularly on their finances in relation to election campaigns?; and, 8) Is information in

reports from political parties and/or candidates to be made public? Data is available for 180

countries and most questions are coded 0 or 1; 0 reflects a lack of the regulation and 1 reflects its

existence.

Interaction Effects

In each model, I include an interaction variable comprised of the campaign finance

variable and the compulsory voting variable relevant to that particular model. This is designed to

test whether the strength of the effect of mandatory voting or turnout depends on the type of the

campaign finance regulation that exists in that country. However, the direction of the effect of

campaign finance on quality of representation may be evident when looking only at compulsory

voting countries. These models help guide my case-study selection for the second part of my

analysis. If a higher quality of representation is dependent on the existence of both compulsory

voting and certain or many campaign finance regulations, they must be pursued jointly.

15
Direct public funding is a common and popular reform introduced in many countries and U.S. States. It turns out
that of countries with enforced compulsory voting, 13 of the 16 have some type of public funding for political
parties (see Figure 3.3). The three that did not were Nauru, Bolivia and Singapore. In my models including enforced
compulsory voting and public funding, the interaction effect was collinear with some other variable in the model.
Public funding was also collinear with another variable in models testing its interaction with average turnout.
Unfortunately, because of the sample limitations we cannot come to any conclusions about how this variable
interactions and influences representation; therefore, it is not discussed further in the next chapter.

31
Control Variables

Compulsory voting countries fall mostly within two major categories: Consolidated

western democracies that have had the policy for many years and Latin American countries with

unstable and young democracies. Because these countries are not necessarily like other

democracies, I control for factors that may influence these two categories of countries

specifically. In addition, representation is a phenomenon influenced by many factors, both

contextual and systemic. I control for variables that may have an independent effect on

representation to try to isolate the effects of compulsory voting and turnout. Twelve of my

nineteen control variables can be found in Pippa Norris’s Democracy Cross-National Dataset as

well as from their original source, as listed in Table 2.4.

Representation is influenced by the electoral system a country chooses. In a single-

member district system with first-past-the-post elections, two all-encompassing parties generally

emerge (Duverger 1972). The two major parties have broad platforms and it is likely that voters

do not align perfectly with either major party. Both of the two parties may represent their median

voter, but the median voter may also be further in ideology from many of the other voters within

the same party, depending on their distribution across the spectrum. In addition, if a voter is on

the losing side, their district/state representatives are only of the opposite party, potentially

leaving voters without a representative of their party.16 On the other hand, proportional

representation systems usually breed multiple small parties, giving voters more diverse platforms

from which to choose. As long as they win above the country’s threshold for seats in parliament,

the party will win a number of seats proportional to the number of votes it received. More voters

can feel like they are being represented. For this reason, the type of electoral system (as

16
If 49.5% of voters in every district across the country voted for Party A and 50.5% voted for Party B, Party B
would win 100% of the seats, leaving 49.5% of the population without representation from their preferred party.

32
measured in the Proportional Representation variable) may influence perception of accurate

representation.

TABLE 2.4: Independent variables and their sources

Variable Source
*also found in Pippa Norris’s Democracy Cross-
National Dataset
Proportional Representation World Bank Database of Political
Institutions*
Number of Effective Agora 2000*
Parliamentary Parties
Age of the Current Regime CHEIBUB 2009*
GDP per capita World Bank World Development Indicators
2014*
(in my models, I use the log of this variable)
GINI coefficient of inequality World Bank World Development
Indicators, various years
Ethnic Fractionalization Alesina 2002*
(combined linguistic and racial)
Religious Fractionalization Alesina 2002*
Internal Conflict Institutional Profiles Database 2012
(ethnic/religious/regional)
Population World Bank World Development Indicators
2014*
(in my models, I use the log of this variable)
Average Population Size per Pippa Norris’s Democracy Cross-National
Representative Dataset
Public Participation at the Institutional Profiles Database 2012
National Level (excluding elections)
Public Trust in Politicians Quality of Governance
Social Trust World Values Survey*
Respect for the Rights of Institutional Profiles Database 2012
Minorities
Genuine Media Pluralism Institutional Profiles Database 2012
Transparency of Economic Policy Institutional Profiles Database 2012
Globalization Index KOF 2005*
Federalism Index Geering-Thacker*
Government Effectiveness World Bank Worldwide Governance
Indicators 2011*

33
 Similarly, the effective number of parliamentary parties matters in that it gives voters

fewer or greater numbers of platforms from which to choose when going to the polls and (when

combined with a proportional representation system) a higher possibility that a member of the

party of the voter’s choosing will win a seat.

New democracies are notoriously fragile, and democratic institutions (like a

representative legislative body) are often not yet consolidated. Also, older democracies have had

more time to reform and improve problems such as corruption. They are frequently wealthier and

more stable—citizens have accepted democracy as the only legitimate system and are more

willing to work within the parameters of democracy than to overthrow it. The age of the current

regime may influence the public’s confidence in representation and democracy overall.

GDP per capita likely has some effect on representation and perception of representation.

Richer countries are typically more stable democracies that have institutionalized and accepted

the practice of representation. Furthermore, if a government’s economy is large, a country is able

to do more for its citizens, influencing their perceptions and confidence in the system. If a

representative can bring money home to a district, his or her constituents may be more satisfied

with the way he or she is doing the job. I have chosen to use the log of GDP because when

graphed against my various dependent variables, there seems to be a curvilinear effect (see

Figure 2.1). Because my regression models are linear, a transformed linear variable would

improve the model’s accuracy.

34
Figure 2.1: Relationship between GDP and Dependent Variables

Alongside the importance of GDP in explaining country differences is how the country’s

wealth is distributed. The GINI index measures income inequality. Citizens of countries with

large disparities between the wealthy and the poor may be less satisfied with their

representatives, who can either pass policies that benefit one socioeconomic group or the other

(for example, tax cuts would satisfy the wealthy but lead to a cut in welfare for those in poverty).

It is much easier to be responsive to the interests of everyone if everyone is similar, be it

socioeconomically or ethnically, religiously, linguistically, etc. For this reason, I also include

ethnic (linguistic and racial) and religious fractionalization (an index measuring how divided the

35
country is between different groups) as measures that influence how constituents perceive their

representation and government responsiveness. Internal conflict rates the intensity of ethnic,

religious or regional conflicts and speaks the magnitude of some of the salient and divisive

issues. A government of a country with high levels of internal conflict would struggle to satisfy

all of its population and therefore may suffer from low levels of perceived representation

regardless of the actions it takes.

Total population may influence representation, especially as the number of resources

must be divided by a greater number of citizens and the population has a greater variety of

interests. Similar to the GDP variable, one sees a curvilinear effect when total population is

graphed against the dependent variables (see Figure 2.2). For this reason, using the log of total

population should improve the accuracy of my linear regression models.

Average population size per representative may influence the amount of contact a

representative has with his or her constituents and the amount of casework (constituency service)

a representative can do. A representative may have more favorable ratings if he or she has more

personal interactions with his or her constituents and can provide more individualized attention

to citizens’ needs. In addition, a representative could theoretically hear the opinions of every one

of his or her constituents in a 100-member district; however, in a district of 500,000, such an

understanding and aggregation of every individual’s wants and needs is impossible.

Public participation at the national level (excluding elections) is an important measure of

how well constituents are communicating their wants and needs to their representatives.

Especially in large districts, representatives must somehow gauge the preferences of their voters

beyond the politician’s most recent campaign platform. This variable ascribes a number to each

country by looking at the amount of participation. Participation includes filling out surveys,

36
 contributing to commissions and attending public debates. Theoretically, a representative with a

very active constituency would have an easier time understanding and later responding to the

district.

Figure 2.2: Population and Dependent Variables

A country’s culture may determine some of its political attitudes. Some societies are

naturally more trusting, hierarchical or individualistic, for example. Citizens of countries with a

high distrust of politicians may assume that all representatives are corrupt and favor special

interests over the average voter. The institution of representation itself rests upon a good-faith

37
belief that elected officials will act with the interest of their constituents in mind. Without this

type of trust, perception of quality of representation may be undeservingly low. 17

While social trust may not seem directly related to representation, Robert Putnam made

the case for such a connection in Making Democracy Work (1993). He argued that social trust is

important for building social capital, which in turn is necessary for good, functioning

governance:

[S]ocial trust has long been a key ingredient in the ethos that has sustained economic
dynamism and government performance. Cooperation is often required—between
legislature and executive, between workers and managers, among political parties,
between the government and private groups, among small firms, and so on…Trust
lubricates cooperation. The greater the level of trust within a community, the greater the
likelihood of cooperation. (Putnam 1993, 170-171)

According to eighteenth century Italian economist Antonio Genovesi, without social trust: “there

can be no certainty in contracts and hence no force to the laws” (Putnam 1993, 170). A society

without trust is a society of every man for himself and a society of self-interest. Democracies

derive their power and legitimacy from the people; a society that does not trust its neighbors to

uphold agreements or its government to work on its behalf is unlikely to buy into the system. A

government may therefore fail to fulfill its basic duties or provide basic services. Citizens are

likely to perceive poor representation in government if government is ineffective or widely

inefficient or if they simply do not trust each other or the system, as in the absence of social trust.

The respect for the rights of minorities is important to ensure that all citizens are getting

fair and equal representation. If elected representatives ignore and alienate certain insular groups

17
This variable is highly correlated with three of my dependent variables: control of corruption, confidence in
parliament and favoritism. I ran all of my models with and without this control variable and found that models
without the trust in politicians variable that had shown significance with the trust in politicians variable also showed
significance to similar levels of confidence and with similar coefficients. Some additional models regarding
confidence in parliament and favoritism did show significance for my key independent variables when excluding
trust in politicians; however, the models made little theoretical sense. See Appendix 2. Regarding control of
corruption, models including candidate expenditure limits showed statistical significance in the interaction, but were
not theoretically sound.

38
from the process, the quality of representation and responsiveness may not be high. Respect for

the basic rights of minorities may not be sufficient, but it may be necessary foundation to ensure

that everyone’s voice can be heard. That is to say, if a small population such as an indigenous

community is completely excluded from voting, for example, measures of representative

responsiveness to voters may appear high despite being entirely unresponsive to a group of

citizens. The majority may have a voice in government and confidence in parliament, but certain

communities that are too small to be numerically significant or are not reached by survey

measures may have completely different experiences.18

Genuine media pluralism is important to ensure that citizens are getting news from

various sources not controlled by the same outlet. In a society with free, pluralistic media,

citizens are likely to learn more accurate information about their representative’s behavior,

including voting records, opinions on salient public issues and scandals that include corrupt

behavior. Media pluralism is also important for general political knowledge so people can make

informed decisions and communicate those preferences to their representatives, who in turn can

be responsive. Similarly, if citizens are going to accurately identify their quality of

representation, policies (both passed and rejected) must be transparent. Complex economic

policies tend to favor either the wealthy or the poor (i.e. tax cuts or redistributive programs), and

a society cannot have (or perceive to have) high quality representation if one group is unaware

that certain policies either help or hurt their socioeconomic wellbeing.

The globalization index measures how connected a country is, economically, socially and

politically. A highly globalized society may face significant international ramifications for the

political decisions it makes; therefore, a representative may have more to consider and more

18
In case this variable is also too much a part of high quality representation (instead of a control variable), I also
tested various models without it. Results did not change in any meaningful way.

39
interests to balance when making policy decisions. He or she may sometimes vote against his or

her own constituents because of the global consequences of the proposed policy.

The federalism index speaks to the number of opportunities citizens have to engage with

the political process (the more levels of government, the more access points, the closer a

representative can be and the smaller the district size). Unitary states with one single government

would perhaps have a different level of representation (or perceived representation) than a strong

federal state. Moreover, citizens may feel more represented at the local level but not at the

national level. When asked if there is a leader in government that represents the respondent

without specifying which government, answers may vary depending on the respondent’s

interpretation of the question.

Lastly, government effectiveness is an important measure of a state’s capacity to carry

out policy and reform. The index looks at quality of bureaucracy and public administration, the

amount of ‘red tape’ slowing the system down, and how governments fare in carrying out what

many would consider basic state functions (electricity, basic health services and sanitation,

education, infrastructure and public transportation, etc.). Government effectiveness matters

because a government is incapable of providing basic goods and services to its people, it is

unable in capacity to be responsive to their constituency. If a government is not effective, people

are likely to look unfavorably upon the institutions and systems they perceive as incompetent,

thereby affecting the dependent variables that contain survey components. Perceptions of

representation may be affected and actual representation may be poor if a representative cannot

provide for his or her constituents basic needs.

40
 Chapter 3: Data Analysis

My initial linear regression models testing the interaction between enforced compulsory

voting and the campaign finance party regulation aggregate index on each of my six dependent

variables show no statistically significant results for my key variables. P-values are above 0.05 for

enforced compulsory voting, the party regulation index and the interaction in all models;

therefore, we cannot be sure that these variables have an effect beyond pure chance. Other

independent variables (frequently trust in politicians) seem to be explaining the variation with

more certainty. Because of the limiting nature of the campaign finance party regulation index

variable, sample sizes range from 25 countries to 40. Table 3.1 shows the best model of the group.

When I tested the same model using compulsory voting regardless of enforcement, I found nearly

identical results. See Table 3.2.

I next looked at whether turnout had any effect when interacting with the campaign finance

party regulation index. Once again, neither independent variable nor the interaction has any

statistically significant effect on any of the six dependent variables. Table 3.3 shows the model

with the most explanatory power.

These models suggest that we cannot say with certainty whether the amount of party

regulation or the presence of compulsory voting (or high voter turnout) have either an independent

effect on quality of representation or an effect contingent upon the presence or strength of the other

variable. From these models, one cannot conclude that either turnout or campaign finance

regulation have any consistent effect on representation. That is not to say that they do not have any

effect ever; instead, their effects may be dependent on individual country-level variables that could

not be measured in the dataset. I will further explore such variables in my case studies. The models

do suggest that people’s perception of the quality of their representation may have more to do with
41
 trust in politicians, which shows up statistically significant to a 95% confidence level in 11 of the

18 models I ran that included the campaign finance party regulation index.

These results may in fact clarify a broader truth about representation (which is that

representation and peoples’ perception of their representation is too complex to be explained by

the amount of campaign finance party regulation or by average voter turnout during elections). To

visualize the data in an alternative way, I plotted graphs of countries’ values for the party

Participation at the Nat’l Level -0.016

Table 3.1: Explaining DV: Control of
(0.18)
Corruption Index
Representation, Model 1
Genuine Media Pluralism 0.122
Key IVs: Enforced CV & Party (1.35)
Regulation
Internal Conflict 0.082
Interaction 0.116 (1.27)
(1.15) Transparency of Econ. Policy 0.079
Enforced Compulsory Voting -0.407 (0.73)
(0.79) GDP per capita (log) 0.301
Party Regulation Aggregate -0.029 (1.73)
(1.22) Total Population (log) 0.067
Proportional Representation -0.110 (1.09)
(0.60) Ethnic Fractionalization -0.643
Age of Current Regime -0.002 (2.19)*
(0.99) Religious Fractionalization -0.040
Effective Number of Parties -0.041 (0.18)
(1.49) Federalism Index 0.033
GINI Index -0.003 (0.57)
(0.36) Globalization Index -0.009
Social Trust -0.638 (1.30)
(0.88) Government Effectiveness 0.601
Trust in Politicians 0.406 (3.85)**
(4.96)** _cons -4.581
Avg. Population Size/Rep -0.000 (1.95)
(1.19) R2 0.98
Rights of Minorities 0.092 N 40
(1.14)
* p<0.05; ** p<0.01

42
Table 3.2: Explaining DV: Voice and
Accountability
Representation, Model 2
Key IVs: Compulsory Voting &
Party regulation
Interaction 0.016 Internal Conflict -0.033
(0.21) (0.66)
Compulsory voting -0.022 Transparency of Econ. Policy 0.061
(0.06) (0.71)
Party Regulation Aggregate -0.020 GDP per capita (log) 0.125
(1.07) (0.93)
Proportional Representation -0.120 Total Population (log) -0.078
(0.85) (1.67)
Age of Current Regime -0.002 Ethnic Fractionalization -0.312
(0.95) (1.35)
Effective Number of Parties -0.013 Religious Fractionalization -0.020
(0.62) (0.12)
GINI Index -0.004 Federalism Index -0.057
(0.68) (1.25)
Trust in Politicians 0.126 Globalization Index 0.006
(2.14)* (1.22)
Social Trust -0.420 Government Effectiveness 0.236
(0.78) (1.89)
Avg. Population Size/Rep. 0.000 _cons 0.014
(1.11) (0.01)
Rights of Minorities 0.089 R2 0.96
(1.39) N 40
Participation at the Nat’l Level -0.020 * p<0.05; ** p<0.01
(0.28)
Genuine Media Pluralism 0.111
(1.62)

regulation index against different measures of representation. These graphs of all electoral and

liberal democracies show that there are no clear patterns between any of my dependent variables

and the campaign finance party regulation index (See Figure 3.1). If this index alone explained

the dependent variables in a way predicted by theory (the more regulation, the better the

representation/perception of representation), the data points would form a recognizable, upward

sloping pattern for the normally coded variables (Voice and Accountability, Favoritism,

Confidence in Parliament and Control of Corruption) and a downward sloping pattern for the

inversely coded variables (Voters Represented and Leader that Represent).

43
Table 3.3: Explaining DV: Is there a Leader who
Represents You?
Representation, Model 3
Key IVs: Mean turnout & Party
Regulation
Interaction -0.001
(0.66)
Mean Turnout 0.003
(0.35) Internal Conflict -0.136
Party Regulation Aggregate -0.026 (4.19)
(0.29) Transparency of Econ. Policy -0.148
Proportional Representation 0.016 (2.35)
(0.14) GDP per capita (log) 0.875
Age of Current Regime 0.000 (2.61)
(0.33) Total Population (log) 0.185
Effective Number of Parties 0.026 (2.63)
(1.80) Ethnic Fractionalization 0.324
GINI Index 0.012 (2.04)
(1.26) Religious Fractionalization -0.055
Trust in Politicians -0.183 (0.38)
(2.05) Federalism Index 0.109
Social Trust -0.477 (2.38)
(0.83) Globalization Index -0.001
Avg. Population Size/Rep. -0.000 (0.39)
(2.02) Government Effectiveness 0.049
Rights of Minorities -0.116 (0.70)
(2.62) _cons -8.677
Participation at the Nat’l Level 0.007 (1.92)
(0.08) R2 0.98
Genuine Media Pluralism -0.267 N 25
(2.83) * p<0.05; ** p<0.01

Within countries with high voter turnout because of the enforced mandatory voting

policy, the amount of campaign finance party regulation is also randomly associated with levels

of representation. This confirms that the interaction of the two independent variables (enforced

compulsory voting and campaign finance party regulation) has no distinguishable effect. The

presence of enforced compulsory voting does not make the amount of campaign finance party

regulation significant. See Figure 3.2.

44
Figure 3.1

45
Figure 3.2

Before concluding that campaign finance party regulation, turnout or compulsory voting

do not matter, we must acknowledge that the party regulation index may be flawed. Given the

variation in enforcement of campaign finance regulation and the small sample size in the

regression models and graphs (simply because the index was only calculated for 53 countries,

further limited by other variables), we cannot be sure that these countries are representative of all

democracies. Perhaps regulations regarding candidates are more important. It may also be true

that countries that began with lower levels of representation are more likely to pass more

46
campaign finance party regulation. Even if the regulation helps improve representation

somewhat within the country, it may not help enough to show up as significant when compared

with countries with better representation initially that did not introduce the policies. My case

studies may help uncover any biases that stem from the contextual problems of the countries that

do have compulsory voting. The index may also not be measuring the most important aspects of

regulation for each country.

Because the party regulation index has weaknesses and because party regulation is not

the only type of meaningful campaign finance regulation, I tested models with many of the

individual campaign finance variables that regulate parties and candidates with both enforced

compulsory voting and turnout to see if any particular type of regulation had any effect when

combined with high turnout.

Disclosure Related Variables

The variable V36 of the International IDEA political finance database measures whether

political parties have to report regularly on their finances in relation to election campaigns. If a

country has this provision, it receives a 1. If a country does not, it receives a 0. I excluded

countries with no data from my model (as with all of the individual campaign finance variables).

When interacting with the presence of enforced compulsory voting, two of the six models

showed statistical significance. See Table 3.4 Models 4 and 5.

In Table 3.4 Model 4, with Voice and Accountability as the dependent variable, the

interaction between party finance reporting is statistically significant to a 95% confidence level.

Because the p-value is 0.049 (less than 0.05), we can reject the null hypothesis that the

interaction has no effect and accept the alternate hypothesis that the interaction does influence

47
the dependent variable. When political parties are required to regularly report regarding election

finance and people are obligated to show up at the polls on election day, countries see an

increase of 0.4483 in the Voice and Accountability index. This is 44.83% of a standard

deviation, and therefore represents a noteworthy increase. For example, Turkey has neither

policy and scores a -.26 on the index scale. Argentina has both policies and scores a +.24 (a

difference of .5). The mean of countries with neither policy is -.131 and the mean of countries

with both is .473 (a difference of 0.60). Because the index’s scale is small (-2.5 to 2.5), the

increase is noteworthy.

Table 3.4 Model 5 looks at how regular reporting of party finances regarding campaigns

interactions with enforced compulsory voting on citizens’ confidence in their parliament. The

interaction effect and presence of enforced compulsory voting show no statistical significance;

we cannot be sure that their effects are not due to chance. However, the campaign finance

variable is statistically significant; there is less than a 5% possibility that we would see these

results by chance. Regardless of whether a country has enforced compulsory voting or not,

countries with party reporting perform worse on the confidence in parliament barometer by

0.7474 all else being held constant. This result is counterintuitive—in theory additional

campaign finance variables should ‘clean up’ the election process and citizens should have more

faith in their representative bodies and representatives. One possible explanation is simple

correlation; perhaps countries passed the disclosure regulation as an attempt to improve poor

representation compared with countries that found the policy unnecessary because their quality

of representation was already high. Another explanation is that the provision is not enforced in

many countries and therefore is not having the effect that it may if enforced. People are aware

48
Table 3.4: Explaining Model 4 Model 5
Representation, Model 4 and 5 Voice and Confidence in
Key IVs: Enforced CV and Party Reporting Accountability Parliament
Interaction 0.448 1.689
(2.05)* (1.86)
Enforced Compulsory Voting -0.249 -0.863
(1.35) (1.20)
Regular Party Reporting/election -0.076 -0.747
(0.94) (2.17)*
Proportional Representation -0.242 -0.872
(1.97) (1.64)
Age of the Current Regime -0.002 -0.013
(1.43) (1.81)
Effective Number of Parties -0.004 0.111
(0.25) (1.43)
GINI Index 0.001 -0.016
(0.18) (0.66)
Trust in Politicians 0.028 0.746
(0.49) (3.24)**
Social Trust 0.602 1.247
(1.34) (0.57)
Avg. Population Size/Rep 0.000 0.000
(2.49)* (1.41)
Rights of Minorities 0.121 -0.004
(1.90) (0.01)
Participation at the Nat’l Level -0.007 0.405
(0.11) (1.43)
Genuine Media Pluralism 0.014 -0.194
(0.20) (0.62)
Internal Conflict -0.043 -0.256
(1.06) (1.39)
Transparency of Econ Policy -0.082 -0.312
(1.20) (0.92)
GDP per capita (log) 0.140 0.455
(1.14) (0.87)
Total Population (log) -0.108 -0.407
(3.01)** (2.83)**
Ethnic Fractionalization -0.384 0.572
(1.92) (0.65)
Religious Fractionalization 0.236 -0.169
(1.28) (0.21)
Federalism Index -0.050 -0.318
(1.38) (2.11)*
Globalization Index 0.012 0.023
(2.37)* (1.11)
Government Effectiveness 0.274 -0.313
(2.32)* (0.64)
_cons 0.273 7.786
(0.21) (1.39)
R2 0.95 0.87
N 54 44
* p<0.05; ** p<0.01

49
that the rules are not working and extrapolate that there must be something wrong with their

legislature because of the unenforced regulation.

Average turnout and its interaction with party reporting do not yield any statistically

significant results, likely because there are many countries with high turnout without compulsory

voting or party reporting that score well on all of the six dependent variables such as Denmark,

Norway, Sweden and the Netherlands.

Next, I looked at V38: Is information in reports from political parties and/or candidates to

be made public? While 59 liberal and electoral democracies make parties regularly report on

finances and make the reports public, 17 liberal and electoral democracies neither require parties

to report nor make collected reports (presumably from candidates) public. Only one interaction

model regarding public disclosure shows statistical significance, seen below in Table 3.5.

This model suggests that the interaction effect is associated with the Voice and

Accountability index and we are over 95% confident that this is not due only to chance. All else

being equal, when candidate and party disclosure reports are made public, an increase in average

turnout by 1% is associated with an increase in Voice and Accountability of 0.0142. This change

is only 1.4% of a standard deviation. Though the effect is minimal, it exists. While public

disclosure itself is statistically significant, its effect cannot be interpreted independently of

turnout because turnout is a continuous variable that is never 0%. When there is public

disclosure, the interaction term will never be 0; therefore, the interaction will always influence

the final index score.

50
 Table 3.5: Explaining DV: Voice and
Representation, Model 6 Accountability
Key IVs: Turnout and Public
Disclosure Genuine Media Pluralism 0.021
(0.34)
Interaction 0.014
Internal Conflict -0.062
(2.40)*
(1.63)
Mean Turnout -0.009
Transparency of Econ. Policy -0.099
(1.45)
(1.48)
Public Disclosure Reports -1.033
GDP per capita (log) 0.286
(2.62)*
(2.41)*
Proportional Representation -0.196
Total Population (log) -0.094
(1.93)
(2.73)*
Age of Current Regime -0.003
Ethnic Fractionalization -0.271
(1.95)
(1.44)
Effective Number of Parties -0.006
Religious Fractionalization 0.158
(0.37)
(0.94)
GINI Index 0.002
Federalism Index -0.017
(0.40)
(0.49)
Trust in Politicians 0.075
Globalization Index 0.013
(1.41)
(2.85)**
Social Trust 0.102
Government Effectiveness 0.175
(0.27)
(1.58)
Avg. Population Size/Rep 0.000
_cons -0.934
(3.59)**
(0.72)
Rights of Minorities 0.069
R2 0.96
(1.15)
N 51
Participation at the Nat’l Level 0.035
(0.59) * p<0.05; ** p<0.01

Ban on Vote Buying

Only one regression model that includes a ban on vote buying shows statistical

significance (see Table 3.6). The interaction between the ban and enforced compulsory voting is

not significant; however, the individual campaign finance variable is to a 95% confidence level.

Regardless of whether a country has enforced compulsory voting or not, the ban is associated

with a decrease in the value of how well voters feel represented in elections by 0.4496. Because

the dependent variable is coded from 1 to 4, 1 being ‘very well’ and 4 being ‘not well at all,’ this

numeric decrease actually signifies an increase in voter confidence that their views are

represented in elections. Moreover, because the scale is small (1-4), a change by 0.4496 (almost

half a full point) is substantively meaningful. This change represents two and half standard
51
 deviations of the variable. Unfortunately, because of the limited nature of the dependent variable,

the sample size in this model is only 29 countries. Those 29 countries are mostly European, with

a handful of exceptions from Latin America and the Philippians, Iceland, New Zealand, Taiwan

and the U.S. The below results may not be generalizable to all democracies, though they do

suggest that banning voting buying may have a real positive effect.

Table 3.6: Explaining DV: Did you feel

Representation: Model 7 represented in the last
election?
Key IVs: Turnout and Vote
Buying Genuine Media Pluralism 0.135
(1.50)
Interaction 0.274
Internal Conflict -0.038
(1.00)
(1.08)
Mean Turnout -0.159
Transparency of Econ. Policy 0.030
(0.68)
(0.35)
Ban on Vote Buying -0.450
GDP per capita (log) 0.310
(2.56)*
(2.10)
Proportional Representation 0.042
Total Population (log) 0.061
(0.46)
(1.91)
Age of Current Regime -0.003
Ethnic Fractionalization -0.206
(2.12)
(1.20)
Effective Number of Parties 0.002
Religious Fractionalization 0.110
(0.10)
(0.71)
GINI Index -0.005
Federalism Index 0.026
(0.98)
(0.82)
Trust in Politicians 0.180
Globalization Index -0.014
(2.26)
(3.96)**
Social Trust -1.768
Government Effectiveness 0.078
(3.46)*
(0.69)
Avg. Population Size/Rep. -0.000
_cons -0.425
(0.75)
(0.27)
Rights of Minorities 0.096
R2 0.95
(1.42)
N 29
Participation at the Nat’l Level -0.174
(1.96) * p<0.05; ** p<0.01

Contribution Limits

V17 speaks to whether there is a limit on the amount donors can contribute to candidates.

Interactions between candidate contribution limits and both enforced compulsory voting and

52
average turnout are statistically significant in various models. As Model 8 of Table 3.7 shows,

the interaction between enforced compulsory voting and contribution limits to candidates is

statistically significant. The model suggests that when both enforced compulsory voting and

candidate contribution limits are present, the Control of Corruption index increases by 0.6699.

Because the index ranges from -2.5 to 2.5, the over half a point increase is noteworthy (67% of a

standard deviation). The same interaction has a statistically significant effect of similar

magnitude on the Voice and Accountability Index (coded on the same scale). See Model 9 of

Table 3.7. The presence of both policies is correlated with a Voice and Accountability score

0.6576 higher than that score in the absence of both policies (66% of a standard deviation). The

consistency between these two models (Table 3.7) suggests a real effect.

Yet, when tested against favoritism, we see a different effect. Enforced compulsory

voting is significant to a 95% confidence level and the interaction term has a p-value of 0.064; as

such, we can interpret it as statistically significant to a slightly lower degree of confidence. See

Table 3.7 Model 10. When enforced compulsory voting is present alone, the coefficient is

negative, suggesting that the presence of enforced compulsory voting (regardless of the presence

of contribution limits to candidates) is associated with a lower favoritism score by 0.8761

(almost a full standard deviation). There is no theoretical reason for enforced compulsory voting

to by itself cause higher favoritism in government; these results may be simple correlation

because of the non-representative sample of enforced compulsory voting countries. When

candidate contribution limits are present, the interaction of the two policies contributes to an

increase in the favoritism index by 0.0989, an effect much smaller in magnitude than the effect

of these variables on Voice and Accountability and the Control of Corruption index (only

11.23% of a standard deviation).

53
Table 3.7: Explaining Model 8 Model 9 Model 10
Representation, Models 8, 9 and 10 Voice and
Control of Corruption Favoritism
Key IVs: Enforced Compulsory Voting Accountability
and Candidate Contribution Limits
Interaction 0.670 0.658 0.975
(2.21)* (2.91)** (1.93)
Enforced Compulsory Voting -0.348 -0.327 -0.876
(1.47) (1.85) (2.22)*
Candidate Contribution Limits -0.054 -0.108 -0.027
(0.54) (1.45) (0.16)
Proportional Representation -0.240 -0.210 0.351
(1.68) (1.97) (1.48)
Age of Current Regime -0.001 -0.002 -0.000
(0.38) (1.43) (0.07)
Effective Number of Parties -0.008 -0.013 -0.045
(0.35) (0.80) (1.22)
GINI Index 0.003 -0.000 -0.012
(0.48) (0.10) (1.08)
Social Trust 0.301 0.476 -0.198
(0.52) (1.10) (0.20)
Trust in Politicians 0.278 0.040 0.594
(3.77)** (0.72) (4.82)**
Avg. Population Size/Rep -0.000 0.000 -0.000
(0.67) (2.31)* (0.43)
Rights of Minorities 0.086 0.128 0.064
(1.03) (2.05)* (0.46)
Participation at the Nat’l Level 0.036 -0.008 0.152
(0.42) (0.13) (1.06)
Genuine Media Pluralism 0.012 0.018 -0.060
(0.13) (0.27) (0.41)
Internal Conflict 0.065 -0.001 0.051
(1.16) (0.04) (0.55)
Transparency of Econ. Policy 0.095 -0.065 0.052
(1.09) (0.99) (0.36)
GDP per capita (log) 0.141 0.199 0.000
(0.86) (1.62) (0.00)
Total Population (log) -0.045 -0.100 0.161
(1.02) (3.00)** (2.15)*
Ethnic Fractionalization -0.618 -0.453 -0.078
(2.30)* (2.25)* (0.17)
Religious Fractionalization 0.211 0.338 0.250
(0.84) (1.80) (0.59)
Federalism Index -0.023 -0.044 0.088
(0.51) (1.28) (1.14)
Globalization Index -0.003 0.006 -0.021
(0.47) (1.09) (1.72)
Government Effectiveness 0.664 0.330 0.503
(4.44)** (2.96)** (2.01)
_cons -1.633 -0.193 -0.523
(0.96) (0.15) (0.18)
R2 0.97 0.96 0.88
N 52 52 52
* p<0.05; ** p<0.01

54
Table 3.8: Explaining Model 11 Model 12
Representation, Model 11 and 12
Control of Corruption Favoritism
Key IVs: Turnout and Candidate
Contribution Limits
Interaction 0.018 0.026
(3.37)** (2.76)**
Mean Turnout -0.005 -0.018
(1.40) (2.68)*
Candidate Contribution Limits -1.177 -1.731
(3.16)** (2.64)*
Proportional Representation -0.175 0.346
(1.39) (1.56)
Age of Current Regime -0.002 -0.003
(0.92) (0.85)
Effective Number of Parties -0.022 -0.045
(1.07) (1.26)
GINI Index 0.004 -0.004
(0.61) (0.37)
Social Trust 0.292 0.455
(0.61) (0.54)
Trust in Politicians 0.304 0.619
(4.69)** (5.42)**
Avg. Population/Rep. -0.000 -0.000
(0.85) (0.45)
Rights of Minorities 0.073 0.042
(0.99) (0.33)
Participation at Nat’l Level 0.100 0.252
(1.27) (1.82)
Genuine Media Pluralism 0.009 -0.064
(0.11) (0.46)
Internal Conflict 0.024 -0.022
(0.52) (0.28)
Transparency of Econ. Policy 0.068 0.000
(0.88) (0.00)
GDP per capita (log) 0.207 0.083
(1.39) (0.32)
Total Population (log) -0.011 0.209
(0.26) (2.85)**
Ethnic Fractionalization -0.480 -0.022
(2.04) (0.05)
Religious Fractionalization 0.135 0.239
(0.64) (0.64)
Federalism Index -0.017 0.113
(0.42) (1.56)
Globalization Index 0.000 -0.015
(0.03) (1.40)
Government Effectiveness 0.525 0.337
(4.00)** (1.46)
_cons -2.718 -1.726
(1.75) (0.63)
R2 0.98 0.90
N 52 52
* p<0.05; ** p<0.01

55
 Average turnout too seemed to have an effect when interaction with contribution limits to

candidates on various dependent variables. See Table 3.8. Model 11 suggests that the interaction

between average turnout and candidate contribution limits is statistically significant. When

candidate contribution limits are present, an increase by 1 percentage point in average turnout

results in an increase in control of corruption by 0.01834 (only a marginal increase; 1.8% of a

standard deviation). The effect of turnout independent of candidate contribution limits has a p-

value above 0.05; we cannot say with certainty that its effect on control of corruption cannot be

attributed to chance.

As Model 12 of Table 3.8 shows, all three key independent variables are statistically

significant when looking at the favoritism dependent variable. The model shows that when there

are contribution limits to candidates, and increase of average turnout by 1% is associated with an

increase in the favoritism index by 0.008124, indicating less favoritism, though the effect is

minimal (0.92% of a standard deviation). When candidate contribution limits are not present, an

increase in voter turnout actually is associated with a decrease in the favoritism index, suggesting

that more turnout coincides with more favoritism.

The variable V13 measures if there are contribution limits to parties over a time period

that is not election specific, and it was not statistically significant independently or through an

interaction term, but one of the models predicting Voice and Accountability showed that average

turnout was significant regardless of its presence. An increase of 1% in average turnout related to

a minimal increase of 0.008176 in the Voice and Accountability index (0.82% of a standard

deviation). See Table 3.9.

To confirm this independent effect of turnout, I ran a regression model including average

turnout and all of my individual campaign finance variables without the interaction term. See
56
Table 3.10. As reported in Table 3.10 Model 14, average turnout is again statistically significant

and has a similar (but minimal) effect on the Voice and Accountability index. For a 1% increase

in the average, Voice and Accountability increases by 0.008489 (0.85% of a standard deviation).

Lastly, I ran similar models without including the interaction term to see if turnout,

enforced compulsory voting or any of the individual campaign finance variables had independent

effects on any of my other five dependent variables. There were no statistically significant results

related to Favoritism, Confidence in Parliament, whether respondents were well represented in

elections or whether there was a leader that represents respondents. However, V23 (provisions

Table 3.9: Explaining DV: Voice and

Participation at Nat’l Level -0.001
Representation, Model 13 Accountability
(0.01)
Key IVs: Turnout and Party Genuine Media Pluralism 0.073
Contribution Limits (1.01)
Conflict -0.037
Interaction -0.006
(0.97)
(1.04)
Transparency of Econ. Policy -0.099
Mean Turnout 0.008
(1.48)
(2.14)*
GDP per capita (log) 0.242
Party Contribution Limits 0.296
(1.90)
(0.82)
Total Population (log) -0.096
Proportional Representation -0.213
(2.56)*
(2.02)
Ethnic Fractionalization -0.115
Age of Current Regime -0.002
(0.54)
(1.25)
Religious Fractionalization -0.082
Effective Number of Parties -0.022
(0.40)
(1.19)
Federalism Index -0.029
GINI Index -0.004
(0.72)
(0.71)
Globalization Index 0.009
Trust in Politicians 0.049
(1.73)
(0.88)
Government Effectiveness 0.269
Social Trust -0.033
(2.39)*
(0.08)
_cons -0.967
Avg. Population Size/Rep. 0.000
(0.67)
(2.83)**
R2 0.95
Rights of Minorities 0.061
N 53
(0.95)
* p<0.05; ** p<0.01

57
Table 3.10: Explaining Representation, Model 14 Model 15
Model 14 and 15 Voice and Control of Corruption
Key IVs: Turnout and Campaign Finance Accountability
Variables Independently
Mean Turnout 0.008 0.005
(2.86)* (1.36)
Ban on Corporate Donations -0.118 -0.012
(0.83) (0.07)
Party Contribution Limits 0.008 0.219
(0.08) (1.91)
Candidate Contribution Limits 0.113 -0.011
(1.01) (0.08)
Public Funding for Parties -0.039 -0.030
(0.20) (0.13)
Free/Subsidized Media for Parties 0.209 0.364
(1.84) (2.66)*
Ban on Vote Buying 0.017 0.029
(0.09) (0.13)
Candidate Expenditure Limits -0.054 -0.223
(0.62) (2.12)(*)
Party Election Spending Reporting -0.213 -0.007
(1.70) (0.05)
Public Disclosure Reports -0.243 -0.132
(2.02) (0.91)
Proportional Representation -0.259 -0.268
(2.10) (1.79)
Age of Current Regime -0.002 -0.001
(1.46) (0.57)
Effective Number of Parties -0.023 0.000
(1.18) (0.01)
Government Effectiveness 0.290 0.667
(1.98) (3.78)**
GINI Index -0.002 -0.004
(0.30) (0.45)
Trust in Politicians 0.028 0.324
(0.40) (3.88)**
Social Trust -0.277 -0.198
(0.48) (0.29)
Avg. Population/Rep. 0.000 -0.000
(3.39)** (0.50)
Rights of Minorities 0.236 0.196
(2.70)* (1.86)
Participation at the Nat’l Level -0.064 -0.019
(0.76) (0.19)
Genuine Media Pluralism -0.034 0.026
(0.42) (0.26)
Internal Conflict 0.028 0.126
(0.55) (2.08)
Transparency of Econ. Policy -0.143 0.054
(1.52) (0.48)
GDP per capita (log) 0.415 0.260
(2.08) (1.08)
Total Population (log) -0.059 -0.033
(1.17) (0.54)

58
 Ethnic Fractionalization -0.251 -0.706
(0.84) (1.96)
Religious Fractionalization 0.291 0.169
(1.25) (0.60)
Federalism Index 0.026 -0.068
(0.49) (1.08)
Globalization Index 0.007 -0.008
(1.14) (0.99)
_cons -3.174 -2.902
(1.64) (1.24)
R2 0.97 0.98
N 46 46
* p<0.05; (*)=0.05; ** p<0.01

for free or subsidized media access for political parties) and V33 (candidate expenditure limits)

did seem to have an effect on the Control of Corruption index (see Table 3.10 Model 15 above).

The coefficient of free or subsidized media access is 0.364, which means that its presence is

associated with a corruption control index score that is higher by 0.364 units (36.4% of a

standard deviation). However, the coefficient of V33 is -0.223; the presence of expenditure limits

is associated with a lower corruption control index. It is unlikely that expenditure limits cause

higher perceptions of corruption; instead, it is likely that this result is due to some unique

features of the 59 countries that do have expenditure limits.19

Of course, the same could be said for most of my models, especially those including

enforced compulsory voting. Compulsory voting countries are a unique (and small) batch and not

evenly spread throughout the world. The small number of enforced compulsory voting countries

limits our ability to identify significant interaction effects. Models that look at interaction effects

between average turnout and campaign finance variables may be more dependable.

19
When excluding trust in politicians (one of my control variables), the interaction between candidate expenditure
limits with both mean turnout and enforced compulsory voting was significant. However, the interaction coefficients
were negative, meaning that when a country had expenditure limits and enforce compulsory voting or saw an
increase in average turnout by 1% point, control of corruption would be slightly worse. This does not make
theoretical sense and is likely the result of some biases in the countries included in the models.

59
Conclusion

The models that showed statistical significance were all predicting the data well; all of

the regression lines explained over 85% of the variation (the R-squared values were above 0.85

and frequently above 0.95). Further, the models that made theoretical sense and showed

significance of my independent variables generally showed that these key independent variables

were statistically significant to a 95% confidence level or higher.

Even though one can conclude that the effects of the independent variables and

interaction terms on the dependent variables are not due merely to chance in the models shown in

this chapter, no major patterns emerge. Average turnout does seem to be positively associated

with Voice and Accountability, as does the interaction between enforced compulsory voting and

regular party reporting with respect to an election. The interaction between public disclosure

reports and mean turnout also had a small but significant effect on Voice and Accountability.

Enforced compulsory voting in the presence of candidate contribution limits is also

correlated with better scores on both Voice and Accountability and Control of Corruption;

however, only five countries in the model have both policies compared with two that have

enforced compulsory voting without contribution limits, 19 countries that have contribution

limits but no compulsory voting and 26 countries that have neither policy. To conclude that the

two policies matter when present together is premature; the tiny sample includes four South

American countries and two European countries whose experiences are not necessarily

generalizable to a broader range of countries.

The interaction between candidate contribution limits and average turnout also has a

significant effect; a 1% increase in turnout results in a slightly better score in the presence of

these limits on the Control of Corruption index and the Favoritism variable (though the effects

60
were substantively minimal). These two models may prove more robust than the models looking

at candidate contribution limits and enforced compulsory voting in that the variance of countries

with regard to turnout and the campaign finance policy may be greater. Yet at the same time, the

Favoritism model is suspect because it showed that in the absence of candidate contribution

limits, increased turnout results in a decrease in the favoritism value. This is a negative change

associated with increasing turnout for which there is no theoretical justification.

That average turnout shows up independently significant with respect to Voice and

Accountability in various models may suggest that there is a real positive correlation between the

two. Of course, this does not necessarily mean that increasing turnout causes an improvement in

the Voice and Accountability index. Further though the Voice and Accountability index speaks

to quality of representation and perceptions of quality of representation, it itself is not a perfect

measure of representation or responsiveness (nor are any of the other dependent variables).

Higher turnout is associated with positive outcomes, though we cannot be sure that this means

higher responsiveness.

Free or subsidized media access to parties is the only campaign finance variable that

alone (in the absence of the interaction effect) appears to be associated with a better score on any

of my dependent variables, and it shows up significant only with respect to the Control of

Corruption Index.

There seems to be no conclusive evidence that compulsory voting (enforced or

unenforced) is associated with better scores on a variety of indicators. The few models that

showed statistical significance for my key independent variables may have been speaking to a

sample bias or a unique component of that particular dependent variable unrelated to

representativeness or responsiveness. Still, the models are all flawed because of sample biases

61
and the imprecise nature of all of my empirical dependent variables. I do not take these results

conclusively to say that compulsory voting, high turnout or campaign finance regulations cannot

or do not improve outcomes, only that the practices are likely too diverse, complex and

sometimes scarce to be explained here.

The major trends (or lack thereof) of the macro-analysis are informative. According to

Lieberman’s nested analysis framework, if regression models are robust, researchers ought to

pick case studies that exemplify the trend to “model-test” in order to further verify the theory and

examine the mechanisms by which the independent variables act upon the dependent variable. If

the theory is not supported (or only weakly supported) by the large-n regression analysis,

researchers ought to pick cases that the model could not explain. These case studies serve to

determine the mechanisms by which compulsory voting and campaign finance interact to affect

quality of representation (if they do at all, and if not, what other institutions seem to matter). The

cases studies may perhaps suggest a new, more robust theory or model. Because no general

trends emerged throughout my regression models and because none of the dependent variables

precisely measure representation, I follow Lieberman’s latter approach. My country case studies

more exactly illustrate how responsive representatives are to their constituencies in various

countries and are more able to examine the interactions between and causal effects of campaign

finance and turnout on representation.

62
 Chapter 4: Design of Case Studies

My regression models all explain the variance in my dependent variables well; however,

results differ across the various dependent variables. Accordingly, I follow Lieberman’s model-

building approach. I have chosen two countries to study in depth from the regression models: one

performs very well on most indicators and the other performs poorly. This section explains how

exactly I pick each case study, what elements of each country I examine and why.

Picking Case Studies

According to the nested analysis framework, I must first determine whether my large-N

regression models produced robust and satisfactory results. If any single dependent variable had

captured representation exactly as I define it, I would be able to look at one set of models and

draw satisfactory conclusions about the effects of enforced compulsory voting, average turnout

and campaign finance regulations. However, all six of my dependent variables are imperfect—

they all capture some aspect of representation but may be also explained by many other

phenomena. Because none of the variables stands out as the best measure of representation, I

look for patterns across all of the models that may perhaps speak to the representative element

that they have in common. While most of the individual models are able to explain most of the

variation in all of my dependent variables (likely because of the many controls I included), no

independent variable was consistently significant. Thus, while my regression models were

frequently robust, the results were neither satisfactory nor conclusive given the various

weaknesses of my dependent variables and the lack of any consistent explanatory effect.

Despite their weaknesses, mean turnout did have an independent, consistent (however

small), statistically significant effect on the dependent variable Voice and Accountability—this

effect is best exemplified in Table 3.9. To further examine this model, Figure 4.1 below shows
the predicted values for each country in this model plotted against their actual values. The closer

the data points are to the line y = x, the better the model was able to predict values. I have drawn

in the reference line. Figure 4.2 shows the linear prediction of the regression model shown in

Table 3.7 looking at the interaction effect between enforced compulsory voting and candidate

contribution limits on the Control of Corruption index. Once again, I have drawn in the y = x

line.

Figure 4.1 Figure 4.2

Had these two models been representative of all of my findings and most of my

dependent variables responsive to a similar magnitude (adjusted for the scale of each individual

variable), I would follow Lieberman’s “Model-testing Small-N Analysis” strategy. I would pick

countries at various points along the line to study in more depth. The studies would clarify,

confirm or reject the original model. However, my models across all dependent variables showed

inconsistencies and I discerned no consistent explanatory effect beyond turnout’s independent

effect on Voice and Accountability. This tells us that the statistical significance of mean turnout

may speak to some unique component of the index, and not necessarily the component that

touches upon quality of representation or representative responsiveness.

64
Figure 4.3

For unsatisfactory results, Lieberman recommends the “Model-building Small N-

Analysis” (2005). When using this strategy, a researcher picks at least one case not well

explained by the model and perhaps one well explained by the model for comparison. Figure 3.2

of Chapter 3 is reproduced here as Figure 4.3. These graphs of only enforced compulsory voting

countries confirm the lack of any recognizable pattern between the amount of campaign finance

party regulation and the indicators of representation in each country,20 which suggests that

quality of representation in countries with enforced compulsory voting is not influenced by the

20
My other two dependent variables are excluded from these figures because they contained even fewer countries.

65
amount of campaign finance regulation. If this were a generalizable pattern beyond the six

countries included in the graphs and turnout (or enforced compulsory voting) was positively

correlated with indicators of good representation, enforced compulsory voting could be seen as a

potentially effective alternative policy to address representation despite any level of campaign

finance regulation.

If not for Brazil and Argentina, there would perhaps be an upward-sloping relationship

between additional party regulation and all four of the above indicators of representation. I look

at Brazil as an example of a country with a significant number of campaign finance reforms on

the books and high turnout, but that scores poorly on all of the above dependent variables (Voice

and Accountability, Control of Corruption, Favoritism and Confidence in Parliament). Brazil

serves as an interesting case study because of its unique combination of policies regarding

compulsory voting and campaign finance and its prima facie poor quality of representation.

In contrast, I study Australia, which very few campaign finance regulations but scores

well on all of the indicators. Both Brazil and Australia defy the expectations of theory—that

campaign finance should help countries achieve ‘cleaner’ governments. While neither of these

countries were actual outliers in my robust but non-satisfactory regression models, they would

both fall under the category of ‘off-the-line’ examples had a model been able to perfectly capture

the theory (in opposite directions—Brazil falls below expectations and Australia exceeds

expectations). These two countries are sufficiently different to perhaps expose some of the flaws

of the large-n analysis and to illuminate different contextual factors that influence representation

beyond campaign finance and voter turnout.

For these two countries, I examine the facet of representation defined as government

responsiveness to its constituents by looking at important legislation that was either passed or

66
rejected during a legislative year immediately following an election. By comparing public

opinion with major donors’ interests, I can conclude for each country whether the national

legislative branch is passing policies that are consistent with what the people want or with what

politically powerful donors want.

However, to fully understand a country’s responsiveness within its specific context, we

must first understand the mechanisms of accountability that shape how representatives do their

jobs. Second we must understand how the campaign finance system regulates money in

elections, and third we must understand more completely how compulsory voting is working in

each country. Only then can we look at the workings of the legislative branch and determine for

whom it is working.

Electoral Systems and Representation

In order to best explain my key independent variables (compulsory voting and campaign

finance regulation), we first need to understand the electoral rules that shape how these elements

exist and interact in each country. Electoral rules are necessary for determining the main players

and rules of the democratic game. While no electoral system alone can guarantee good

representation, each electoral system influences representation by shaping the mechanisms of

accountability. Representatives need not be responsive to their constituents if there is no

mechanism for accountability.

One cannot compare accountability and representation directly between different

countries without taking a more detailed look at the technical structure of elections and political

parties. For example, countries that choose proportional representation often see a greater

number of parties that offer voters more diverse choices of platforms; however, representatives

67
do not inherently represent their constituents better because of the proportional allocation of

seats. If too many small parties are unable to form coalitions that allow them to pass meaningful

legislation, they are in effect unable to represent their constituents well, despite any rhetoric or

ideology. For example, Italy has a system of proportional representation and nine effective

parties. This gives Italian voters a great deal of choice on Election Day, yet, Italy scores poorly

on indicators of corruption and favoritism. It is also plagued by instability and indecisiveness—

its many parties have been unable to form the strong, capable coalitions necessary for effective

governance (The Times 2006).

Additional rules matter in how the overall electoral system functions. Electoral rules

change the incentive structure for politicians—these incentives alter to whom each elected

representative feels most accountable. An electoral system with weak incentives for

representatives to stay accountable to their constituents may show poor responsiveness to its

constituents despite high levels of voter turnout and regardless of any campaign finance

regulation. John Carey of Dartmouth College writes of two types of accountability: individual-

level accountability and collective accountability. Carey defines individual-level accountability

as when “legislators are most responsive to citizens’ demands” (Carey 2008, 8). Individual-level

accountability is possible:

when each [legislator] is responsible for cultivating her or his own support constituency,
which in turn can reward and punish its representative directly at the polls…[such a
system] deter[s] the betrayal of the demands of particular voters who picked an individual
legislator as their representative. (Carey 2008, 8)

The emphasis is on the individual candidate or legislator, and it is his or her duty to communicate

effectively with the voters. He or she is less reliant on the party label for votes and may therefore

be less loyal to the party when its interests conflict with those of his or her voters. Individual-

68
level accountability is often promoted by two-party systems or open-lists that encourage intra-

party competition and cross-party cooperation in government.

On the other hand, collective accountability occurs when representatives “are primarily

concerned with the ideological and policy content of party labels” (Carey 2008, 8). Elected

representatives are directly accountable to party leaders instead of constituents and most often

vote in a block with the party, fearing exclusion from resources or future party tickets. Parties

serve as a mechanism to clearly communicate messages to voters and to punish legislators who

stray from the positions on which they were elected. Often, collective accountability is

associated with proportional representation systems with closed party lists on the ballot.

Both types of accountability work best when legislators are only accountable to one

principal (either the party or the constituency). The greater the number of principals in the

system, the more divided legislators’ loyalty must necessarily be. The pulls in various directions

towards different principals can cause legislators to deviate from their party or from their

constituents’ demands (Carey 2008). The more insulated a legislator can be from these additional

principals, the more he or she can stay truly accountable and responsive to either the party or the

constituency. In strong presidential systems, presidents control important resources and set

legislative agendas. They can pressure legislators in a way that prime ministers within the

legislature cannot, and are therefore an extra principal. Governors in some countries (such as

Brazil) can exert influence because they control important appointments. An interest group that

donates large sums of money to support a candidate’s campaign may also tug at a legislator’s

loyalty. Electoral systems are important in that some designs create more access points for

additional principals that obscure the direct representation link.21

21
Though additional principals in the system do compete for a representative’s attention, they also introduce checks
and balances widely considered to be beneficial to the overall functioning of the system. Separation of powers and

69
 There seems to be little “academic consensus on the relative merits of individualistic

versus collective representation” (Carey 2008, 14). Both systems ultimately strive to keep elected

representatives accountable; collective accountability just keeps representatives accountable to

the people through a political organization. Both have the potential to work well if a country’s

culture and existing institutions guide the choice. Problems arise when a country has neither.

Alternatively, the literature of the field has tended to conclude, “PR [proportional

representation] elections are normatively superior to SMD [single-member district] elections”

(Carey 2008, 13). However, this preference assumes that “the political parties are fundamental

units of legislative representation” (Carey 2008, 13). Especially (but not only) in democracies

that favor collective accountability, political parties are essential to high quality representation

because they “provide the linkages between elites and masses, or between the state and civil

society, which are so critical to citizens’ assessment of the political order and therefore to

democratic legitimacy” (Power 1995, 102). Some scholars find parties even more essential to all

representative democracies without distinguishing between different types of accountability:

Representative democracy cannot exist without political parties. Because of the

information asymmetries that exist between those who govern and those who are
governed, there must be incentives for representatives to build reputation. And there must
be tools for voters to punish and reward authorities. Only when representatives run as
members of political parties, voters can use information shortcuts and can reward
or punish their past performance. Without parties, voters find it too costly to cast
informed and meaningful votes in representative democracy. (Nadia and Walker 2010,
251)

Parties encourage predictable behavior and accountability; they provide critical information

about issues to voters and they serve as access points for citizens to engage in the political

process either through employment, volunteerism or candidacy. Systems with too many parties

resources between different branches and different levels of government is intended to protect individuals from
tyranny, despite introducing competing incentives for elected representatives.

70
may become confusing for voters who must discern from many persuasive messages what is

truly in their best interest. Systems with too little party control may mean that party labels are

empty and voters must follow individual candidates if they want to make informed decisions.

At the beginning of each country case study, I examine the electoral system of that

country. This will allow me to interpret results more holistically. To properly understand the

quality of a country’s representation, accountability and responsiveness, we must first understand

1) what kind of accountability each country prefers; 2) if the system and culture is designed to

ensure that type of accountability; and 3) what additional principals are distracting legislators

from their constituencies. With a better understanding of why political actors act in the ways they

do within their specific political contexts, we can more accurately assess the motives of the

actions they do or do not take. In the case of Brazil, the electoral system more directly causes

problems in representation and impedes responsiveness.

Campaign Finance and Representation

After looking at each electoral system, I describe in detail the campaign finance regime

of each country. Campaigns cost a great deal of money. How they should be financed is highly

contested everywhere and democracies around the world have come up with very different

systems to control the flow of money in elections. Money influences politics in a variety of

ways—on the one hand it allows candidates to spread their platforms and interest groups to

engage in persuasion and debate, on the other, it can be used for personal gain at the expense of

equality and fairness. While most countries with modern systems of political financing have

explicitly banned direct vote buying (the most obvious form of corruption), money can influence

politicians and policy outcomes in a variety of more subtle ways. Those who donate do so for

many different reasons. Some might believe firmly in the same ideology as a candidate. Others

71
(big interest groups and some wealthy individuals) look to ‘buy influence.’ They donate to shape

relevant policy outcomes in a way that is more favorable to their personal or corporate interests.

Campaign finance is important to understand in the context of an electoral system. For

example, more money may be donated to parties in closed-list systems, making party regulation

more important. In open-list systems that encourage intra-party competition, candidates

themselves perhaps should be regulated to better control the flow of money in campaigns. In

addition, detailed campaign finance rules are important to study when assessing the interaction

between campaign finance and compulsory voting to begin to understand the root causes of

outcomes. For example, Brazil bans political donations from trade unions but not corporations,

which may put the two (often antagonistic) actors on different playing fields when vying for

influence. An understanding of Australia’s and Brazil’s campaign finance regimes is important

in assessing the effects mandatory voting may have on government responsiveness.

Different electoral systems are vulnerable in different ways, but countries everywhere

continue to struggle with balancing freedom of speech (which includes funded political

discourse) and corruption, be it bribery, clientelism and patronage or undue influence. Because

campaign finance regulations can perhaps interfere the constituent-representative link, they are

important to study in detail when looking at responsiveness.

Compulsory Voting and Representation

Each country enforces compulsory voting in a different way, and each citizenry has

reacted to the policy in different ways. Not all compulsory voting is created equal and turnout

statistics can be misleading or hide demographic groups that are still underrepresented at the

72
polls. At the beginning of each case study, I explain the details of the policy and its actual effect

on voter turnout.

Performing the Case Studies

After understanding the contextual information, I look to determine more precisely the

responsiveness of each government’s national legislative body to the average citizen. To do this,

I pick one legislative year directly following an election and look equally at legislation that did

pass and did not pass.22 I choose to do so because elected legislators are straight off of the

campaign trail, and campaign donations and voter preferences are likely to take center stage.

While the year before an election might also be appropriate, politicians may pander to the donors

they want instead of reacting to the donors they already have. While studies that look at the

influence of campaign contributions on legislative outcomes use a variety of methods, such as

focusing on one policy (Frendreis and Waterman 1985) or focusing on multiple legislative

sessions (Wawro 2001), other scholars have also used the first year of a legislative session to

look at bills (Fleisher 1993).

I look through significant campaign donations to parties from the relevant election to

determine who were the biggest donors, their ideological leanings and what legislation would

benefit them most, to whom they donated and what, if anything, they claim to expect in return.23

I then look through all of the legislation both rejected and passed during the year. From the many

22
It is also important to remember that money likely influences policy beyond just outcomes; it may affect agenda
setting, policy writing and committee voting (Peoples 2013). Some actually contend that money has more influence
in these early stages of legislation than in final outcomes (Powell 2012). Still, such an analysis would likely require
finding detailed legislative histories and speaking directly with contributors, members of committees and policy
writers. Due to the constraints of this project, I was limited to focusing on macro-outcomes instead of subtle biases
throughout the legislative process.
23
Because the political parties of some countries exercise strong control over their members and members almost
always vote as cohesive blocks, it makes sense to focus on these donations instead of donations to individual
candidates. In countries where parties are not the main actors, candidate contributions may be more telling; however,
I am somewhat limited by the format and type of data that each country makes public.

73
bills, I pick a few examples of major legislation that included either a redistributive component

or represent a collision between corporate and public interests. I then read bill digests, news

articles, party positions and public opinion polls to try to determine to whom the legislature was

responding most by its actions. I look at responsiveness on a national level as opposed to looking

at individual representatives’ actions (for example, roll call votes and their constituents’

preferences)—I do so in part because of the limited data available. Roll call votes are not

available for all bills, nor is public opinion data on most issues broken down to the state or

district level. In addition, while it is important for voters to know to whom their representatives

are responsive, one good or bad representative does not dictate policy singlehandedly. One

member corrupted by campaign donations does not make the system corrupt if money does not

influence how all of the other members make their decisions. Instead, the system’s overall

responsiveness (the aggregate of all elected members of the legislative branch) affects real

outcomes that influence peoples’ everyday lives.

After looking at all of the separate pieces—the legislation (who it hurts, who it helps,

who wanted it passed), relevant campaign donations and public opinion—I assess whether

representatives appear to have been working for their donors or for the public. By looking at a

handful of bills that illustrate conflicts between corporations or the wealthy individuals who gave

money and the average voter, I draw overall conclusions about who is most influential in each

country’s system and why.

Lastly, I compare my findings in each country to assess the policy of compulsory voting,

the effects of nearly universal turnout, and the importance of campaign finance in determining

representative responsiveness on a broader level. These conclusions carry theoretical

implications that can be applied generally to other countries.

74
 Chapter 5: Australia

Australia scores well on all six of the dependent variables I used in my regression

models. Table 5.1 contains a summary of the scores. Australia also has notably less campaign

finance regulation than many other countries. The data interpreted alone suggests that perhaps

compulsory voting helps improve outcomes regardless of weak campaign finance.

Table 5.1: Voice and Control of Confidence Favoritism Are Voters Is there a
Summary Accountability Corruption in (1-7 best) Represented Leader
of Scores (-2.5-2.5 best) (-2.5-2.5 Parliament in Who
on best) (1-10 best) Elections? Represents
Dependent (1-4 worst) You?
Variables (1 yes; 2
no)
Australia 1.44 1.76 7.38 3.94 2.44 1.2

In this case study, I first explore Australia’s electoral system to understand the

mechanisms of accountability. Because Australia’s system is based on collective accountability

and has long-standing, highly disciplined parties, I focus on contributions made to the parties

instead of individual candidates and party platforms instead of candidate platforms. I give an

overview of Australia’s approach to campaign finance and mandatory voting, and then focus on

the voting patters of and donations made during the 2013 elections. Next, I look at legislative

outcomes during 2014 and conduct a detailed analysis of five bills (two that passed and three that

did not).

I find that when the public has strong opinions on an issue, outcomes reflect the majority

opinion despite where the money falls. Regarding less salient issues with concentrated interests,

money (that may come from either campaign donations or lobbying efforts) does seem to guide

outcomes.
Australia’s Electoral System

Between Australia and Brazil, Australia has the older and more institutionalized tradition

of democracy. There have been no interruptions in their democracy, which over time has been

subject to various reforms. Some territories began using the secret ballot (not coincidentally also

known as the ‘Australian Ballot’) as early as the 1850s, and the country achieved universal adult

suffrage by 1903 (Farrell and McAllister 2006). The current constitution has existed since 1901.

The federal system is parliamentarian with a bicameral legislature. The Senate (the upper house)

has 76 members in total; there are 12 representatives for each of the six states and two for both of

the two territories. Senators are elected for six years. The House of Representatives (the lower

house) has 150 members that each represents his or her own district electorate. Each electorate is

approximately the same size. Members have terms of three years. The Prime Minister is selected

from the majority party or majority coalition to fulfill executive functions. Elections are

scheduled for every three years, but they are often more frequent because the House can dissolve

itself with a motion of no confidence in government. The opposition can also call for a motion of

no confidence in the prime minister requiring his resignation and the resignation of his entire

Ministry.

Both houses are elected using preference voting, which is to say voters number their

candidate choices on the ballot. In the upper house, which uses proportional representation,

preference voting takes the form of the Single-Transferrable Vote. Each state elects half of its

senators at a time, and a candidate must receive 1/6th of the vote to win a seat. If a candidate

receives votes beyond this amount, excess votes may be distributed to the voters’ next numbered

choices. House members are elected in single-member electorates and a candidate must receive

over 50% of the vote. If no candidate reaches a majority in the first round, the candidate with the

76
fewest votes is removed from the race and votes for that candidate are distributed to the voters’

marked second choice (what is called the Alternative Vote). This continues until a majority is

reached (Farrell and McAllister 2006).

Voting is compulsory in Australia and enforced with fines (International IDEA). Beyond

simple compulsion, valid votes must have preferences marked for all of the candidates on the

ballot (except one, which vote counters assume is the voter’s last choice). Combined with the

frequency of elections (nearly every 12 to 18 months—one of the highest among comparable

advanced democracies24) and a number of immigrant citizens from non-democratic countries

with limited English proficiency, this is cumbersome for many voters, as it demands a basic

understanding of every candidate’s platform in every election (Farrell and McAllister 2006). To

discourage voter alienation, which leads to invalid or spoiled ballots or donkey votes (the

numbering of preferences in order down the ballot), Australian parties give out ‘how to vote’

cards listing party preferred preference orders. Even easier is the option to vote ‘above the line’

for a single party, introduced in 1983. On the ticket, the party has preference over the order of its

candidates. Most voters tend to use ticket voting—in 2004, the rate of ticket voting was 95.8%

(Farrell and McAllister 2006, 129). Even so, the high level of compulsion of voting in

Australia—for both casting a ballot and ordering a mandated number of preferences—has led to

a high number of invalid ballots, though the number of invalid ballots has decrease since the

introduction of ticket voting in 1983 (Farrell and McAllister 2006).

Parties in Australia are strong, disciplined and dominant in the legislature. As a result,

most representatives vote along party lines. There are almost no “breaches in voting unity”

24
Switzerland likely has the most frequent elections of advanced democracies—there are generally three or four
national votes a year in addition to any cantonal elections (Lane 2015). Typically, the frequency of elections is
negatively correlated with voter turnout rates (Lijphart 1997). True to the theory, Switzerland has low turnout rates
now that voting is voluntary—it was compulsory until 1974. One canton still makes voting mandatory today
(International IDEA).

77
(Carey 2008, 108). Representatives risk sanction or exclusion off party lists in the next election if

they deviate too greatly from the party platform. Parties are also are the best way for those with

political ambition to get involved in politics—candidates who have devoted a great deal of time

to supporting their party fare best in elections. In 2004, Senate candidates spent on average over

22 hours a month on party activities, especially canvassing (Farrell and McAllister 2006, 108).

Even though systems of proportional representation have on average more parties represented in

the legislature (they have an average of 4.57 effective parties, as classified by AGORA in 2000),

Australia historically has seen three major parties: the Labor Party on the left and the National

and Liberal Parties which form a stable right-wing coalition.

While Australia gives its citizens a great deal of choice through the open-list preferential

system, most voters opt to partake in a close-list election. The system is designed with two

competing currents—preferential voting favors individual voter choice, yet ticket voting and

party control over ballot orders favor collective party cohesiveness. The two strains are

definitively not of equal weight: “The evidence clearly leads to the conclusion that among

Australia’s elected representatives (or, at least, the parliamentary candidates), the party emphasis

predominates” (Farrell and McAllister 2006, 13). Despite Australia’s attempts to foster

individualism in campaigning, Carey’s model of collective accountability fits Australia quite

well. Parties serve as effective vehicles for representation, carry reliable and meaningful labels

and serve as access points into the political process. The electoral system itself does not

guarantee good representation, but the design is certainly conducive to it.

Because parties maintain such strong control over their members, representatives rarely

deviate from their party’s position. Therefore, when evaluating Australia’s representation it

makes sense to focus on the party as the main political actor. An individual representative is

78
likely unable to represent his or her constituents when their opinions deviate from the party’s

official stance. Individual roll-call votes, while available online, are less important in

campaigning and holding representatives accountable to their word because voting behavior is

dictated and made public by the party organization. The leadership successfully holds

representatives accountable to the party leadership and can issue sanctions. As such, the question

of Australia’s responsiveness becomes whether parties adopt platforms and fight for legislation

that its voters want, not whether individual representatives fight for the preferences of their

constituencies. Australia is a clear model of Carey’s collective accountability.

Australian Campaign Finance Regulation

The Australian Electoral Commission runs elections and regulates money in campaigns,

though the organization has a markedly uncomplicated law to enforce. Australia does not ban

foreign donations or corporate donations.25 State contractors and trade unions may contribute to

campaigns. Donations may be anonymous as long as they are under the current disclosure

threshold of $12,400 Australian dollars26 (above which the donor must also file his or her own

report to the Commission). Because the party organizations exist independently on the state

level, individual donors can strategically contribute just under the disclosure threshold to all of

the nine state party organizations, thereby funneling a great deal of money into one party while

avoiding all disclosure triggers. There are no overall contribution or expenditure limits on either

parties or candidates and legal definitions are imprecise; the Green party has called monetary

donations from parliamentarians “tithes” to avoid reporting them as contributions (Orr 2007, 78).

25
In fact, British Lord Michael Ashcroft donated $1 million Australian Dollars to the Liberal Party in 2006 (Orr
2007); Details on regulations come from the IDEA political finance database.
26
As of March 8th, 2016, this threshold was equivalent to $9,248.55 US Dollars.

79
There are no media subsidies for political ads. The only ban that exists is on state resources being

used to support parties or candidates.

Public funding is available—an amount is retroactively paid to parties after the election

per first-preference vote received as long as the party wins above 4% of first-preference votes

nationally (Miragliotta 2010, 176). The more private money a party spends effectively to win

votes, the more likely it is that it will receive a higher amount of compensation after the election.

This rule may potentially drive up the cost of elections and the need to collect private donations.

The system is not contingent upon stricter limits on the collection and usage of private money, so

parties end up using a mix of public funds and private donations. For example, over 80% of

campaign money spent by both the Liberal party (on the right) and the Labor Party (on the left)

in the 2006/07 election cycle came from private sources (Miragliotta 2010, 177).

In the absence of most other forms of regulation, the Australian people rely on disclosure

requirements to regulate money in campaigns; parties must report annually (though not in

relation to a specific campaign) and candidates and large donors must submit their own reports.

These reports are made public; however, they do not always contain much information because

the threshold to disclose anonymous donations is so high. The disclosure reports are further

useless in informing vote choices because they must be filed within a few months after the

election. (Orr 2007).27

The site Money, Politics and Transparency scores countries on a scale of 1-100 (best) for

their laws and for the practice of those laws. The “‘in law’ indicator provides an objective

assessment of whether certain legal codes, regulations, and mechanisms exist… the ‘in practice’

indicator addresses de facto issues of implementation, enforcement, effectiveness, and

27
The few citizens that care about the source of candidate or party campaign money and election opponents who
wish to highlight presumed biases cannot use pre-election disclosure reports to inform voter choice. In Australia,
sanction at the polls happens an election cycle later, if it happens at all.

80
accessibility” (Money, Politics and Transparency 2014). Australia scores a 41 ‘in law’ and a 67

‘in practice.’ The ‘in law’ score is the lower of my two case studies, though the ‘in practice’

score is higher, perhaps reflecting state capacity or a greater willingness to enforce the few laws

they do have (Money, Politics Transparency 2014).28 Bolivia also scores the same ‘in law’ score,

yet performs quite poorly on indicators such as Voice and Accountability and Control of

Corruption (worse than even Brazil, which has much more campaign finance regulation in law).

Australia seems to have much better representation than one would predict if looking only at

campaign finance laws, thereby reinforcing the idea that there are other variables at play and

perhaps demonstrating in another way that campaign finance regulations (or lack thereof) are not

always themselves associated with better (or worse) representation outcomes.

Though the ‘in practice’ score is the highest between the two countries I examine in this

paper, it does not represent good enforcement per se. The penalty for violating the law is a fine,

albeit a relatively low one. The maximum penalty for an intentionally false disclosure report is

A$10,000 for parties, but most fines are much less and are not levied proportionally to the size of

the violation. Maximum fines for candidates are much lower. Technically, those who falsify

information during investigations may be imprisoned. In reality, few are investigated and even

fewer are sanctioned in any significant way. The Australian Electoral Commission, in charge of

campaign finance administration, is “under-resourced” and has prioritized what it does have

towards its role in actually administering elections (Orr 2007, 80). Further, the law that it must

enforce makes actual violations quite difficult to prosecute:

The hurdles to prosecution are formidable: proof beyond reasonable doubt, including
proof of dishonesty in the case of misleading returns. Strict liability only applies to
omissions, such as failure to lodge a timely return or the lodging of an incomplete return.
That approximately 40 annual returns are amended each year, but no prosecutions

28
For comparison, the United States scores a 71 ‘in law’ and a 67 ‘in practice’ as well (Money, Politics and
Transparency 2014).

81
 launched, suggests that prosecution is at best used as a threat to extract a late or amender
return, rather than a penalty or deterrent. (Orr 2007, 80)

In general, Australia has a fairly relaxed campaign finance regime—“Australia is decidedly less

regulated than not just the U.S., but other comparable, common law democracies” (Orr 2007,

74). Scholars best qualify the laws and enforcement as “‘lackadaisical’ because of the ease with

which many of its strictures can be ignored or circumvented” (McMenamin 2015, 134). The

Commission has been characterized by its “natural desire to avoid political controversy and

conflict” (Orr 2007, 80) and parties, candidates and donors take advantage of the system. This is

not entirely the fault of the Commission. The culture among political actors, the media and the

electorate encourage this attitude. Politicians and parties know that they are highly unlikely to be

investigated, fined or imprisoned. The media does not choose to spend much time investigating

and reporting on the information in campaign finance filings, partially due to what Orr calls a

“combination of journalistic indifference and lack of skill” (2007, 86) when it comes to political

finance matters; this exacerbates the unimportance of any reports filed with the Commission,

done correctly or not.

Because Australia’s regime has not been strengthened over the years and receives so little

attention, one might assume that money has not traditionally played a large role in influencing

Australian politics; yet, this is not necessarily the case. While some donations are made to

express preference for an ideology or a specific platform, large donors “expect their donations to

favorably influence government policy” and switch their support when they do not see results

(Orr 2007, 81). Business interests tend to donate more to the right-wing coalition (the Liberal

and National parties) due to ideological considerations; however, they do still donate to the

Labor Party if it seems like the Labor Party will win, consistent with the influence motive. What

this suggests is a certain relationship between these donors and recipients:

82
 The pragmatic relationship between business donors and the parties principally rests on
two types of exchanges: discrete and reciprocal. Discrete exchanges are usually the sale
of donors of access to politicians. Reciprocal exchanges are neither clear nor
simultaneous. Rather, they establish an obligation toward the donor that is likely to
increase the donor’s chances of access and influence in the future. (McMenamin 2015,
130)

Australians had high hopes that their disclosure regime would keep their politicians clean and

accountable to the right principals (in Australia’s case of collective accountability, the parties).

Unfortunately, in the absence of other limits, the system has not performed as intended. Scholar

Joo-Cheong Tham claims that the disclosure regime “merely lead to a ‘normalization of large

scale corporate donations’” (Orr 2007, 81) by businesses whose motives are not purely

ideological.

That is not to say that a robust disclosure regime cannot be successful in other systems. In

Australia, media and voter indifference exacerbate the ineffectiveness of the few campaign

finance laws on the books:

electors everywhere may be unswayed by matters concerned with political procedures

and ethics, such as donations, influence, and disclosure. But with compulsory voting,
Australia may be more prone to this than countries with voluntary voting, as the voices of
the apathetic and cynical are no less heard than the voices of the committed and
concerned. (Orr 2007, 85)

In systems of voluntary voting, voters who show up at the polls are more likely to be informed

than voluntary non-voters. These are the people most likely to follow election news and study

candidate and party platforms. These voters may be wary of certain types of donations (for

example, a large donation from a foreigner). To them, campaign finance may be a more salient

issue than to voluntary non-voters. Even though studies have shown that compulsory voting has

increased the level of political knowledge among voluntary non-voters, campaign finance may

be too dull and too technical of an issue for the politically apathetic. Alternatively, people may

take it as a given and not give it much thought.

83
 This is certainly true in Australia. Partly because of Australia’s “absence of earth-

shattering scandals” (Orr 2007, 87), campaign finance has not grabbed the public’s attention.29

Still, the lack of a major scandal does not suggest that the regime is working properly to secure

that the constituent-representative link remains uninterrupted by special and wealthy interests.

Australia’s campaign finance regime is weak in law and lax in enforcement, creating an

environment that legally (or practically) allows for a great deal of money meant to influence to

flow into politics. Responsiveness to the people, to the extent that it exists in Australia, will exist

in spite of Australia’s campaign finance regime, not because of it.

Australian Compulsory Voting

Compulsory voting policy varies greatly from country to country in its execution and in

its penalties for non-compliance. Voters in various countries have different incentives for

compliance. Some consider Australia to be the best example of functioning compulsory voting.

The system rests on fines for failure to vote (currently at 20 Australian Dollars30), but allows

citizens to request waivers and give explanations for not showing up at the polls. Satisfactory

explanations include religious duty to abstain and

physical obstruction, whether of sickness or outside prevention, or of natural events, or

accident of any kind…One might also imagine cases where an intending voter on his way

29
A recent scandal has had prominent figures discussing the possibility of strengthening the campaign finance
regulation that currently exists. A known mafia criminal in Italy now living in Australia, Frank Madafferi, has
donated tens of thousands of dollars to the Liberal party through a fundraising entity. Nothing he has done is
explicitly illegal, yet it has many concerned. A report on the matter found that “‘loopholes’ in the oversight system
meant it was ‘difficult to identify any bribery in the form of political donations’” (Edwards 2015). Labor leader Bill
Shorten, with the support of the Greens, was open to looking at reforming donation regulation. Greens leader
Richard Di Natale said, “We have a political donation system that encourages large corporations to make donations,
and when they make those donations they expect to get a benefit” (Edwards 2015). Professor Ian Ramsey of the
Centre for Corporate Law at Melbourne University agreed that Australia’s system had some “significant
limitations…In particular, there are important issues about the transparency, how much is disclosed, and how
quickly that information is disclosed” (Edwards 2015). Others, like Social Services Minister Scott Morrison was
satisfied with the transparency regime Australia had created and did not think reform was necessary to the
functioning of the system (Edwards 2015).
30
$14.92 US Dollars on March 8, 2016.

84
 to the poll was diverted to save life, or to prevent crime, or to assist at some great
disaster, such as a fire. (Australian Supreme Court in Judd v. McKeon (1926 38 CLR
380)).

Those who do not receive waivers and do not pay the fine risk imprisonment. In 1993, around

half a million Australians did not vote, but 94% justified their abstention with an acceptable

excuse. 23,320 voters paid the A$20 fine and 4,412 voters went to court to resolve the matter. 41

were jailed (Swenson 2007). What makes compulsory voting successful in Australia, though, is

not the fear of sanction—so few non-voters face consequences (approximately 5% pay the small

fine, 0.88% go to court and 0.0082% are sent to jail).31 Instead, “people comply…out of respect

for the law itself and a belief that it is a reasonable one” (Electoral Commission 2006, 22). In

1996, 74% of the population was in favor of compulsory voting in federal elections (Electoral

Commission 2006, 22).

Some of this support may be because compulsory voting is minimally intrusive in

Australia. It allows absentee voting by mail, pre-polling day voting, and voting in a different

electoral division (but still within one’s own state or territory), and provisional votes when a

voters’ status is uncertain (AEC Electoral Handbook). Elections are held on Saturdays and the

government sends balloting supplies to isolated areas (Hill 2014). A partial explanation for the

high levels of compliance may be due to the fact that voters have become used to the policy and

adopted universal voting into their social and civic culture—the policy has existed for 92 years

as of 2016.

Turnout in Australia is just shy of universal. Of those registered, turnout is consistently

between 93 and 96%; however, when calculated as a percentage of the voting age population,

turnout is frequently in the low 80s (International IDEA). Most attribute the gap to young

people—up to a quarter of people below age 25 are not registered to vote and 51% of Australians
31
Calculated from the 1993 statistics using a denominator of the approximate number of non-voters: 500,000.

85
between 18 and 24 years old said they had “not much” or “no interest” in politics. Many claim

they would not vote if it were not required (Electoral Commission 2006, 34).

Compulsory voting in Australia is largely effective despite small penalties for non-

compliance. Most Australians have embraced the policy and their duty to participate in the

political process. While there are asymmetries in turnout and young adults are content to risk the

small fine for not voting, the policy is largely cited as one of the best examples of compulsory

voting and is likely not going anywhere soon.

Case Study

I focus my study of Australia on the most recent election in 2013.32 Correspondingly, I

look at legislative measures from the 2013-2014 parliamentary session. I focus on only one

parliamentary session because of the time restraints of this research project, though a similar

study might choose to look at the entire session or compare sessions with different majority

parties.

In the 2013 election, there were 17,406,251 eligible voters out of a population of

22,262,501. Of those registered, 93.23% voted. Of the total voting age population, only 78.86%

voted, the only percentage below 80% since the 1955 election (International IDEA). The

difference is largely due to the youth vote. The Australian Electoral Commission has found that a

third of all eligible voters who are not registered are between 18 and 24. If a party could turn out

the group of voters who choose to abstain, they could affect political outcomes; therefore,

mobilizing these voters continues to be a political strategy to win office (Beck 2013). What the

simple turnout statistics also do not reflect is the rising number of invalid votes that cannot be

counted for any reason. At 5.91% of all cast ballots, this number was the highest it has been
32
The next election is set to take place in the fall of 2016.

86
since 1984 (International IDEA). This high number may signal voter disengagement or

alienation from the current Australian politics. Still, Australian turnout far exceeds that of many

of its peer countries with voluntary voting.

It is not obvious which parties benefit the most from compulsory voting—both the

Australian Labor Party on the left and the Coalition (Liberal Party and National Party) on the

right supported its implementation in 1924 and both have considered abolishing it at one time or

another. However, most agree that among those who benefit are the socioeconomically

disadvantaged and marginalized groups in Australia whose voices are, by mandate, included in

the political process. It is likely the Labor Party on the left that reaps more of the benefits of

increased turnout, but the right-wing parties have not made any serious reform attempts.

Any advantage the Labor Party may have does not always translate into solid legislative

majorities (or if it had initially, the parties may have shifted over time to reflect the new median-

voter). Currently, the right-wing Coalition, comprised of the Liberal and the National Parties,

holds the majority in the House of Representatives with 90 of the 150 seats. Labor has 55 seats

and 5 seats are crossbench (independents or members of minor parties who do not belong to the

Coalition). In the Senate, the Coalition has a plurality of seats (33 out of 76), but not a majority.

Labor has 25 seats, and 18 seats belong to members of other parties (notably, 10 seats belong to

the Australian Greens).

Campaign Financing of the 2013 Election

In order to determine the interests of the major donors so I could later determine if they

matched closely with policy outcomes, I looked individually at every donation above A$100,000

made to the three major parties (Labor, Liberal and National) at both the national and state

87
levels.33 For every donor above A$100,000, I also looked for other separate donations made to

the opposing parties.34 Because of Australia’s model of collective accountability and its voting

system that operates most of the time as a closed-list system that advantage parties over

individuals, it makes sense for donors to contribute to party organizations instead of individual

candidates. Statistics on donations to candidates are not available on Australia’s disclosure

webpage, nor would they carry much weight if they were available.

While most donors are never required to explain their donation motives, those that offer

explanations maintain that donations to parties are made on ideological grounds alone. The

statistics call this into question, as does the nature of Australian donations and the previous

reporting on the subject. Whereas individual candidates in different parties may hold similar

positions on certain issues, parties typically hold ideologically distinct positions. I found that of

the 56 major donors in my sample, at least 14 had donated money to both sides (though typically

favoring one party over the other). Two of these bipartisan donors were in the oil/energy sector,

one was in mining, four were in real estate/property management, one was in media, three were

banks, three fit into an ‘other’ category and included an association of clubs, a cab-fare credit

card company and an agribusiness organization. Most individuals who donated (as opposed to as

a corporation) donated only to one party, though I was not able to check whether individual

donors were in any way connected to any of the companies (at least some likely are).

Bipartisan donations notwithstanding, from the sample of large donations I reviewed, it

appeared as if the energy and mining sectors preferred the Liberal Party while the unions and

33
Though contributions made below this cutoff are likely still influential, after this mark I began seeing many repeat
donors and individuals (who are much more difficult to track). Many of the largest players had donated above this
mark to at least one party. Ideally I would have looked at donations to a lower cutoff, but given the time restrains of
the project, I could not.
34
Contributions made of $100,000.00 or above in installments to different party organizations will not have
appeared in this search.

88
international trade corporations preferred the Labor Party. Overall, the Liberal Party received

more money than the Labor Party, with the national branch totals at approximately A$43 million

and A$40 million respectively. The Liberal party state branches collected A$54 million and the

Labor Party state branches collected A$38 million. The National Party collected almost A$2

million at the national level and just over A$10 at the state level—this is money that ultimately

helps the Liberal Party form majority coalitions in parliament and push their desired agenda.35

The Liberal Party’s donation totals taken alone would underestimate the strength that the

conservative ideology had in fundraising over the Labor Party.

Some big donations came from mysterious individuals who were not easy to find online.

The Australian media only knows that some of the individuals have unique ties to foreign entities

and governments. This includes the largest donor to the Australian Labor Party, Zi Chun Wang,

who listed only his email, a Chinese phone number and a Chinese address. He donated

A$850,000.00 and has ties to a Chinese-American property development company called the

Ever Bright Group. The company has insisted that Mr. Wang made his donations in his “personal

capacity and was not associated with the company” (Farrell and Evershed 2015). When The

Guardian tried to contact him, the paper did not get a response. Mr. Lawrence Kung, a

Taiwanese-born businessman and longtime friend of the former Taiwanese president, also

donated A$200,000.00 to the Labor Party. He sponsored former Prime Minister Kevin Rudd’s

trip to Taiwan to meet with then-president Chen who was later indicted for the embezzlement of

funds for secret diplomacy.36

35
These numbers (and all future Australian campaign donation numbers in this chapter) come from the Australian
Electoral Commission’s disclosure website, Summary of all Party Returns 2013-2014:
http://periodicdisclosures.aec.gov.au/SummaryParty.aspx
36
The Sydney Morning Herald reports, “The secret fund to which Mr Kung has been linked was, according to
former president Chen, used to support Taiwan’s efforts to secure diplomatic recognition in the South Pacific and
gain influence in Australia” (Baker and Dorling 2009).

89
 The Liberal Party too receives donations that are cause for suspicion. P. Marks

Investment Pty Ltd donated A$431,361.00. Paul Marks, born in Israel as Paul Miszkowski, is the

director of a complex web of companies connected to a mining corporation (Nimrod Resources,

who also donated A$500,000.00 to the Liberal Party). Australian news sources also reveal that

he ran a brothel and that Mr. Marks is close with former Prime Minister Tony Abbott. Though

Mr. Marks maintains that he supports the Liberal Party “because its philosophy, principles and

economic policies will ensure the growth of the Australian economy and improve the lives of all

Australians,” many are still skeptical of his motives, as his support has come at a time when his

business interests have been expanding (Drill, Whinnett and Minear 2015). Another company,

Hong Kong Kingson Investments, has donated A$340,000.00 to the Liberal Party, A$100,000.00

to the National Party (part of the right-wing Coalition with the Liberals), and A$80,000.00 to the

Labor Party. The man behind the donations, Chau Chak Wing, is a Chinese-born Australian

citizen, but the money typically comes from a network of his overseas companies. Wing told the

Sydney Morning Herald, “When I make those donations, I do not put any conditions on the

contributions.” Former Prime Minister John Howard of the Liberal Party also told the press, “He

always struck me as a person interested in a genuine way in building relations between China

and Australia…I never discussed donations with him and the access he had was not so frequent

as to even justify that question” (Snow, Christensen and Garnet 2009).

Australian campaign finance regulation does not forbid any of these donations and only

requires that both the party and the donor disclose them. If money could predict what policies a

government would pass, we would expect the current Liberal/National coalition to pass policies

that favor the energy and mining sectors. These would be policies that do not overburden oil or

gas extraction and do not attempt to fight climate change by promoting clean energy and carbon

90
emission reduction. Banks somewhat preferred donating to the Liberal Party, probably because

the party is seen as more finance and business friendly. These entities would want deregulation

or to obstruct new regulation. Other industries that donated to the Liberal Party, like healthcare

corporations and pharmaceutical companies would want the same—less regulation.

Major Legislation Passed:

I began by looking at the major legislation passed by the Australian Parliament and

subsequently signed by British Queen Elizabeth in the first year of the 44th session (2013-2014).

I looked for controversial legislation for which public opinion could deviate from the opinions of

major donors and legislation that was redistributive in nature (Is the government responsive to

the public or to the corporate interest? Is it responsive to the rich or the poor?). To see if a bill fit

either of these categories, I did a preliminary search to see if the issue had been polled and if

news sources had written about the legislation. From the articles, it became clear what policies

were controversial and where different interest groups stood on each issue.

I looked at two of the 177 Acts that were passed during the session and three bills that did

not pass: the Carbon Tax Repeal Act, the Fair and Sustainable Pensions Bill, the Building and

Construction Industry (Improving Productivity) Bill, the Landholders’ Right to Refuse (Gas and

Coal) Bill and the Live Animal Export Prohibition (Ending Cruelty) Bill.37 I do not claim that

these five bills are representative of all bills that come before the national Australian parliament;

however, they reflect different sectors of the economy and distinct interests where compulsory

voting and/or campaign finance may have influenced outcomes.

37
For an additional bill, the High Speed Rail Planning Authority bill, see Appendix 4. I originally looked at this
piece of proposed legislation because it would potentially involve lucrative construction contracts and strong public
opinion; however, actual construction would be so far in the future (beginning in 2027 at the earliest) that neither
money nor public opinion has gotten involved. The bill’s ultimate defeat does not tell us anything; however, I have
included it in an appendix because it is interesting.

91
Carbon Tax Repeal

My first focus was on the Carbon Tax Repeal Act. This legislation affected the powerful

energy and coal sector and was one in which perhaps big interest groups could potentially use

campaign finance to overrule public opinion.

In 2012, the Labor Party majority parliament passed legislation that charged the 348

highest polluters A$23 ($22.60 USD) per ton of greenhouse gases emitted (BBC News 2014). It

was highly controversial legislation with interest groups and parties aligning in predictable ways.

At the time, surveys found that 36% of Australians believed that global warming is a serious and

pressing problem and the government should be taking immediate steps even if it involve

significant costs. Another 45% of the public believed that the problem of global warming should

be addressed, but its effects will be gradual so we can deal with it by taking small steps that are

low in cost. 18% of Australians responded that until we are sure that global warming is a real

problem we should not take any steps with an economic cost (Lowy Institute Poll 2015). The

CPM (Carbon Pricing Mechanism) legislation was not originally on the Labor Party’s 2012

agenda, but Prime Minister Julia Gillard agreed to create a Multi-Party Climate Change

Committee to form a coalition with the Australian Greens and three independent parliament

members. The committee recommended this legislation, and the Labor government followed its

advice (Talberg et al. 2013). After its passage, the growing consensus seemed to be that the 2012

legislation was more costly than the environment was worth— reports found that two-thirds of

Australians said they were against legislation that would fix a price on carbon, and 45% of those

opposed claimed they were ‘strongly against the policy.’ 57% said they were in favor of a future

Coalition government repealing the legislation if they won the next election (Hanson 2012).

92
 Public opposition to the CPM was still high in 2013—58% of polled Australians opposed

the legislation, though that number was falling (Oliver 2013). In 2014 the number of Australians

who though climate change was a serious threat that needed to be addressed now had risen to

45% (from 36% in 2012). The number who though gradual, low-cost solutions were sufficient

decreased from 45% in 2012 to 38%, and the number who say that we should do nothing because

we are unsure global warming is really a problem fell to 15% (Lowy Institute Poll 2015). A

majority of the public thought the government should be doing something, and this actually

included the Coalition government, which still supported a target of greenhouse gas emission

reductions by 5% relative to 2000 levels by 2020 even though they disliked the current policy

(BBC News 2014).

The Liberal Party made its repeal one of its six main campaign policies, claimed that it

had a mandate to rescind the policy once in office and insisted that the act “increased costs to

businesses and households, has not reduced greenhouse gas emissions” (Talberg et al. 2013).

They cited, as consequences of the legislation, a significant rise in electricity and gas prices after

the CPM’s implementation in 2012 (though not all of the increase can be attributed to the CPM).

Businesses had endured a cost increase of around 14%, though only some of the prices were

passed along to consumers. Some high-profile manufacturing companies, like the Kurri Kurri

aluminum smelter, closed, and blamed the CPM. Coal mining, gas and coal electricity industries

lost production. The Coalition claimed that “all that the Gillard Government has done today [by

passing the CPM legislation] is export jobs and emissions overseas” (Talberg et al. 2013).

Coalition voters did not necessarily view the repeal as such a priority. Exit polls during

the 2013 election revealed that 31% of Coalition voters said their most important issue was

‘economy and jobs.’ Only 3% answered ‘repealing the carbon tax’ as their top policy priority of

93
the Coalitions six specific campaign policies.38 Another poll found that of 13 options, having

“better policies on things like the environment and climate change” was ranked 10th in reasons

that voters chose to support the Coalition, behind options such as ‘no reason’ and ‘don’t know.’

Only 1% of Coalition voters ranked climate change policies within their top three reasons for

voting for the party. However, it is also possible that respondents may have considered the repeal

and economic rather than a climate change issue. Regardless, “the Coalition was open about its

intentions to repeal the CPM upon taking government, it cannot be argued that Coalition voters

expected anything different” (Talberg et al. 2013). Voters seemed to agree with the Coalition’s

policy position and overall favored the repeal, but the importance of the issue to most voters was

much less clear.

The repeal legislation was introduced in November 2013, immediately after the election.

The legislation was complex, as had been the original mechanisms to regulate carbon and cut

down on greenhouse gas emissions.39 Its full ramifications were unknown, but Coalition

members estimate that the CPI should fall by about 0.7% from what it was estimated to be in

2014-15 before the repeal, and average families should spend around A$10.50 per week (or

A$550 per year) less on energy bills than they would have before the repeal. Businesses were

expected to save in administrative costs and substantive costs, and compliance costs were

expected to fall by A$85.3 million per year (Hunt and Hockey 2014).

Both houses passed the bill in July 2014 with the support of the Coalition, the Liberal

Democratic Party and the Family First Party. Clive Palmer of the Palmer United Party abstained

38
This particular exit poll was conducted by JWS Research and commissioned by The Climate Institute.
39
Along with the repeal of the main provision of 2012 bill, the legislation included provisions such as the removal of
a 15% tax offset for the conservation tillage, assistance to the steel industry in relation to the carbon pricing, a
reduction in funds to the Australian Renewable Energy Agency (ARENA), an abolition of the Climate Change
Authority (CCA) (a government agency that reviewed climate change policies and made recommendations to the
Parliament) and a transfer of some of its functions to the Department of Energy and the abolition of the Clean
Energy Finance Corporation. It also gave new powers to the Australian Competition and Consumer Commission
(ACCC) to ensure that savings from the repeal are passed on to consumers (Hunt and Hockey 2014).

94
because of a conflict of interest—he was heavily invested in the coal, iron and nickel industry

(Talberg et al.2013). The Labor Party opposed a repeal, though they supported looking for better

alternatives to the CPM legislation. The Australian Greens strongly opposed the repeal, but also

opposed the Labor Party’s proposals. So too did the environmental community. Though the

business community was not strongly in favor of the CPM legislation, they were also not

vehemently opposed to it, as they all agreed they would have to take some hits in the fight

against climate change. Most of the relevant interest groups were strongly against the CPM, such

as the Australian Chamber of Commerce and Industry, the Australian Retailers Association, and

the Minerals Council of Australia. Others, such as Matthew Warren of the Energy Supply

Association of Australia, were wary of any changes:

It is not in this nation’s broader interest to legislate and repeal and legislate and repeal
energy policy. At some point, the major parties will need to sit down and agree on a
sensible, workable direction for Australia’s energy future. It would be better for everyone
if that day came sooner rather than later. (Talberg et al. 2013)

The repeal legislation was not introduced with any specific alternative in mind, which worried

some of those affected like the Clean Energy Council and the Waste Management Association of

Australia. The Climate Institute was concerned with how Australia would meet the emission

reduction targets to which it had agreed given the lack of alternatives. The Australian Council of

Trade Unions, the Construction, Forestry, Mining and Energy Union (CFMEU) and the

Australian Worker’s Union had accepted the CPM in 2012 and now opposed the repeal (Talberg

et al. 2013).40

It would be unfair to conclude that the Coalition made the repeal of the CPM a priority

because of the 2013 campaign donations from any of the affected industries—while the entire

party’s platform may have been shaped over the long term by its core constituency’s (and

40
The CFMEU and the Australian Worker’s Union were both big donors to the Labor Party during the 2013
campaign (Talberg et al. 2013).

95
donor’s) interests, one might reasonably expect that the energy sector and mining corporations

decided to donate large sums to the 2013 election because they supported the Coalition’s position

on an issue that was relevant to their own interests.41

In countries with voluntary voting, general polls often survey those who they think are

likely voters. Due to various biases in self-reporting and the inexactitude of guesswork, the polls

are not always successful; however, in an attempt to survey only likely voters, the opinions of

likely non-voters may be excluded. These polls may not necessarily reflect the views of the

entire citizenry in a proportional manner. Because of compulsory voting, well-done professional

polling data in Australia should accurately reflect the electorate. Given that the majority of

Australians, regardless of party affiliation, were in favor of the CPM repeal, the Coalition

government did not ignore their interests in favor of big donors in the energy and mining sector.

In fact, big donors and Australians tended to agree when it came to this particular climate change

policy.

Nothing about this bill’s passage is problematic from the perspective of representation.

The Coalition opposed the legislation when the Labor Party passed it in 2012 and has not

changed its stance. Big money came down on both sides of the fight, but party positions were

clearly advertised. Donors may have given money to the side that they supported with the intent

to influence election results instead of the intent to change the positions of elected officials

(though of course this motive was still possible as well, just unproven beyond pure speculation).

Though certain corporations’ large donations to both the Liberal Party and the Labor Party

certainly looks suspicious, the passed legislation did not seem to undermine the public interest or

41
No donor is likely to say directly that his or her donations were made with the intent to influence legislative
behavior or gain privilege and access. When looking to determine true motives, we must rely somewhat on optics.
For this reason, I will follow the ‘innocent until proven guilty’ rule of thumb when deciding whether donations are
made to subvert the public interest.

96
will according to public opinion surveys. The loosely regulated campaign finance system did not

matter. The voters understood the CPM repeal as one of the Coalition’s main goals for their

majority tenure in Parliament and could not have reasonably been surprised or felt betrayed by

the legislation. The fight fell mostly along party lines—collective accountability seemed to work

as intended in this instance. The outcome on this policy issue reflects both donor preferences and

majority public preferences; therefore, from this example, the government cannot be said to be

more responsive to one or the other.

Fair and Sustainable Pensions Bill

Because there were no major redistributive policies considered in the 2013-14 that were

significant and controversial, I looked for bills passed during the 2014-15 session. The Social

Services Legislation Amendment (Fair and Sustainable Pensions) Bill was contested and

controversial, and passed both houses on July 17th, 2014.

The Act’s stated aim is to “have a positive impact on more vulnerable people with lower

levels of assets, as well as improving the sustainability of the pension system into the future…to

provide an adequate level of support to those in need over the long term” (Morrison 2015). A

certain calculus based on an “assets test free area” determines pension amounts. Up to a certain

amount of assets, pensions are not cut beyond the full amount. After the threshold, for every

extra $1,000 in assessable assets, one’s pension is reduced by $1.5 per fortnight. This legislation

increases the thresholds on various social security pensions such as the age pension, disability

support pension, wife pension, care-giver payments, bereavement allowance, widow B pensions

and certain Veterans’ Affairs pensions, but also increases the “taper rate by which a pension is

reduced once the free areas are exceeded” (Morrison 2015). So, despite becoming more generous

97
in certain respects, the payouts fall more quickly in amount once the threshold is reached.42 The

legislation does include provisions to protect the pensions of those who would lose the entirety

of their social security pension under these new rules. The government estimates that it will save

A$2,434 million in the future (Klapdor 2015), indicating that on the whole, people would lose

some of their benefits.

The Labor Party opposition does not agree that the current pension system is

unsustainable. In response to this claim, Shadow Minister for Families and Payments Jenny

Macklin stated:

[The government] would have you believe that the government cannot afford to look
after low- and middle-income retirees. But it is simply not true, and Labor knows that the
pension is sustainable. Australia is considered to have one of the most sustainable
pensions in the world. The Allianz Pension Sustainability Index last year found that
Australia's pension system is the most sustainable in the world. According to the OECD,
Australia spends just 3½ per cent of GDP on the age pension, compared with the OECD
average of 7.8 per cent. Ensuring the continued sustainability of the retirement income
system is important, but it has to be done in a fair and equitable way. (Klapdor 2015)

Opposition believes the true unsustainable problem of the retirement income system in Australia

is actually the superannuation tax concessions.

The Department of Social Services estimates that 88% of over four million social security

and veterans’ affairs pensioners will not be affected by this amendment. Four percent (around

171,500 people) will benefit by about A$750 per year, and eight percent (327,300 people) will be

hurt by this act: 236,000 will lose on average A$3,380 per year and 91,300 will lose about

$4,940 per year. Some people could lose up to A$14,467. Around 6,500 more people will qualify

for social security allowance payments because of the increased ‘free area’ zone (Klapdor 2015).

Models show that those hurt by this amendment will grow to about 40% of retirees by 2055 and

those most affected will be those with average or below average incomes and single women,

42
A table of the changes in detail can be found in Appendix 3.

98
though some argue that the whole pension system (not just this amendment) is set up to

perpetuate this problem (Klapdor 2015).

Modeling by the Labor Party shows more dramatic results: 700,000 people who retire in

the next decade will be affected by the cuts and some will lose almost a quarter of their annual

income. Notwithstanding this claim, Labor’s frequent partner, the Australian Greens, came out in

favor of the legislation claiming that it will make the pension system more fair (Kapdor 2015).

This disagreement within the left about the reform’s outcomes demonstrates the great complexity

of this issue and that support does not split neatly between parties along ideological grounds.

The legislation has some support from the welfare sector, but it has not been

overwhelming. The Australian Council of Social Service and the National Welfare Rights

Network (both large national networks of relevant interest groups regarding social services and

welfare) “have offered qualified support” for the legislation, but COTA Australia (Council on the

Aging, one of Australia’s largest elderly peoples association) does not think this change is what

pension reform needs most. National Seniors and Industry Super Australia have come out firmly

against the amendment (Kapdor 2015). National Seniors is another of Australia’s largest elderly

peoples associations in Australia and Industry Super helps Australians manage and maximize

their retirement funds.

Australian citizens generally like their pensions. A 1992 study found that 56.8% of

Australians wanted their old-age pension to increase and 39.2% wanted it to stay the same. Only

3.6% wanted cuts (Grant 2004). While I could not find a more recent study asking the same

question, other poll results have shown that these figures likely have not changed drastically.

Many are concerned about their retirement finances. Only 55% of non-retired people think that

they will have sufficient money to live comfortably in retirement, and 36% say that they will

99
probably or definitely not have enough money to live comfortably in retirement (Bray and Gray

2016). A survey in early 2014 (before the passage of this legislation) found that 64% of

Australians thought the age pension was too low, and 25% thought it was about right. 47%

thought the disability support pension was too low and 31% thought it was about right. Only 2%

thought the age pension was too high and 8% thought the disability support pension was too

high. For those between 45 and 64, 76% thought the age pension was too low (Essential Reports

2014, Pension and benefits). 75% of those surveyed were opposed to making the eligibility

requirements tougher for the age pension, and 55% were against making requirements tougher

for the disability support pension (Essential Reports 2014, Eligibility). In addition, 64% of

Australians disapproved of including the value of the family home in the assets test for eligibility

for the age pension, while only 26% approved.43 Disapproval rates of this proposition were

similar amongst all party groups listed (Labor, Liberal/National, Greens, Other). Those more

likely to approve were under 35 and had a university education. While the passed legislation

does not add the value of the family home in the assets test, the poll displays the sentiment of

around 2/3s of Australians—they oppose measures that reduce pensions payouts by adjusting the

asset calculation (Essential Report 2014, Pension assets).

Of course, positive public opinion does not prevent the program from being unsustainable

and a government may choose to act regardless with the long-term interests of its citizens in

mind. In addition, governments must constantly balance the amount of taxes citizens want to pay

and the amount of services they demand—the two are rarely in equilibrium. When asked whether

they pay more than their fair share of taxes, 40% of Australians regardless of party thought they

paid too much, 36% thought they paid about the right amount and 11% thought they paid less

43
Including a house in the assets test would theoretically decrease payments to those who are already well-off (those
who have highly valued homes).

100
than their fair share. Surprisingly, when broken down by party, those who identify with the

Labor Party were the most likely to think they paid more than their fair share. When broken

down into four income categories, the richest group was the most likely to think they paid too

much (Essential Report 2015, Share of Tax).

Right wing parties often run on platforms of fiscal responsibility, cutting taxes,

balancing the budget and getting spending under control. Prime Minister Abbott of the Liberal

Party was no exception, though he made a campaign promise not to change pensions. When

questioned in office about his commitment to leaving the pension system intact, Abbott replied

“I am confident that pensioners will be better off because under this government they will lose

the carbon tax, but keep the compensation” (Maher 2014). Coalition voters may have reasonably

expected their pensions not to change under the Abbott administration; 60% had opposed a

toughening of the assets test (Essential Reports 2014, Pension assets). In this way, the Abbott

administration was not very responsive to its constituency or to any other voting group.

Or perhaps the government had become more responsive to a different constituency—the

young. Young people have reason to be more in favor of social security reform because they pay

a great deal of money into the system and are unsure they will receive any back. This holds true

in the United States and in Australia where those under 35 were more likely to support including

house assets testing in the determination of one’s pension amount. These young people are those

who are underrepresented at the polls in voluntary voting systems. One could argue that the

U.S.’s social security system is under no immediate pressure to change because the elderly like

the benefits they receive and the young are politically apathetic. They do not organize; they do

not show up to vote and do not push their concerns onto the national agenda. The elderly have

formed formidable interest groups that fight for and protect their interests and they turn out in

101
high numbers. Because of compulsory voting in Australia, the disparity between turnout of the

elderly and the young is less striking than in the U.S. Young voters are still voting at lower rates,

but to a less drastic degree.

I was unable to find campaign donations during the 2013 election from any old age or

disability interest groups. This does not mean they did not donate money, rather if they did that

they avoided disclosure by donating under the anonymity threshold. Given the power of the

AARP in the United States, I was surprised that these organizations were not among the principal

major political players as determined by campaign donations.44 However, upon closer

investigation, AARP is not a major campaign donor either. In the 2012 election cycle,

OpenSecrets.org reveals that the organization contributed only $29,826 (the sum of its

contributions to candidates, leadership PACs, parties, 527 committees and outside spending

groups). This money was further dispersed amongst candidates—Barack Obama receiving the

most with only $9,500 and most other House and Senate candidates receiving around $1,000.

AARP preferred to spend its money on lobbying—it ranked 33 of 4,375 groups in lobbying

spending in 2012 by spending $9,900,000 (OpenSecrets.org 2012). Lobbying is another legal

channel for interest groups to spend political money to influence policy and political outcomes

entirely separate from the campaign donation process. In addition to lobbying, AARP’s great

weapon is its sheer number of members and its ability to mobilize these members to vote. This is

an advantage that retired peoples associations do not have in Australia because of compulsory

voting, and this lack of an advantage perhaps contributed to the pension reform outcome.

Because it appears as if there were no major campaign donations that relate to this policy

outcome (which is unfavorable amongst most Australians thought slightly less so among the

44
I checked for donations from various elderly and retired peoples organizations, including the Association of
Independent Retirees, the Active Over 50s (ARPA) and COTA. I also looked for donations from the Psychology of
Intellectual Disability Group, the Autism Interest Group and People With Disability Australia.

102
young), we can conclude from this example that in the absence of money in the form of

campaign donations, compulsory voting may have affected outcomes because the preferences of

the young are (almost) equally voiced as those of the elderly. Pension reform continues to be an

ongoing challenge for Australia, and citizens may see more changes in the coming years. The

Coalition government has demonstrated by passing this Fair and Sustainable Pension Act that

they are unafraid of making incremental but mostly unpopular changes with the longevity of the

entire system in mind. By passing this legislation, the Coalition has gone against the clear

preferences of the voters, though it has not betrayed them for campaign contributors. Instead,

other factors might be affecting the accuracy of representation, such as the party’s pure ideology

in favor of a smaller welfare state or a decision to engage in trustee-style representation looking

after the wellbeing of the nation over the long-term. Lobbying groups in favor of pension reform

may be funneling more money and efforts into the cause than elderly associations. This policy

outcome, while it does not reflect the majority of voters nor any major campaign donors, may

reflect these lobbyists instead. Neither campaign finance regulation nor nearly universal voter

turnout can prevent this type of influence, and if that is the true problem, we must look at

alternative solutions, perhaps that target lobbying.

Bills Not Passed

I next looked through major legislation that was introduced during the 2013-2014

session and ultimately was not passed. I again chose bills that had the potential to pit special

interests against the public will or socioeconomic classes against one another.

103
The Building and Construction Industry (Improving Productivity) Bill

The Liberal Party introduced the Building and Construction Industry (Improving

Productivity) Bill in 2013. The bill provided for the re-establishment of the Australian Building

and Construction Commission and gave the Commissioner various powers and functions. It also

would have amended certain laws regarding self-incrimination, liabilities, admissible documents

and the jurisdiction of courts. It would have allowed the Commission “to separate legitimate

protests from unlawful and organized picketing aimed at disrupting building and construction

work” (TimeBase 2013, Building). When asked about the Abbott Government’s commitment to

the Australian Building and Construction Commission in 2013, polls found 29% overall support

for the government’s plan and 22% opposition. 48% either did not know or did not care. Among

Abbott’s own party, 52% of respondents supported his goal and only 5% opposed it. Labor

voters were 13% in favor and 39% opposed. Support was slightly higher among men than

women, at 35% and 23% respectively (Essential Report 2013, Australian Building). The

legislation was contested bitterly in parliament, but nearly half of the Australian public did not

offer an opinion.

The Housing Industry Association Industrial Relations favored this bill as a way to

“[boost] Australia’s productivity and [ensure] law and order prevail at our nations’ building

sites” (TimeBase 2013, Building). The Labor party has also been interested in reforming this

area of law, but has come up with a different solution (the Fair Work Industry Inspectorate).

They portrayed this proposed legislation as highly antagonistic to labor unions, especially the

Construction, Forestry, Mining and Energy Union (CFMEU) who are long-term supporters (and

big donors) of the Labor party. The Secretary of the CFMEU put out a statement about the bill:

“The Federal Government apparently believes that discriminating against blue collar workers in

104
one industry is acceptable” (TimeBase 2013, Building). These allegations of discrimination

ultimately led to the bill’s defeat during the 2013-2014 session.

In general, Australians hold fairly favorable views of unions. In one poll, 45% of

Australians answered that workers would be better off if Unions were made stronger, and 26%

answered they would be worse off. 30% thought either that it would make no difference or they

did not know. Within the Labor party, 68% of voters thought they would be better off and 13%

thought they would be worse off; within the coalition, 28% thought better and 41% thought

worse (Essential Report 2015, Better or worse).

The bill has been re-introduced in 2016 without any changes. During the interim, the

Heydon Royal Commission “found a culture of ‘widespread and deep-seated’ misconduct by

union officials, and detailed a long list of officials whom it said may be guilty of criminal

activity including blackmail, bribes and threats of violence” (TimeBase 2016). These allegations

may have soured the public’s opinion on unions, creating a unique political opening for the

Coalition to push the legislation through.

Given the lack of popular interest in this bill and the general division of support along

party lines, one might have assumed this legislation would pass in 2013, but the Unions were

able to successfully spin the media in their favor to paint the legislation as discriminatory. This

special interest influence did not need to come in through campaign donations—it came from

lobbying and campaigning during the legislative session, just as in the pension reform act

mentioned above. We cannot assume that the big campaign donation of A$266,000,000 that

CFMEU gave to the Labor party was intending to garner unwarranted influence in the policy

process. They did not donate any money to the Coalition parties and the Labor party has always

been the pro-union party. The current head of the Labor party, Minister Bill Shorten, was also

105
the National Secretary of the Australian Workers Union from 2001-2007. It is very possible that

this donation was made to bolster Labor’s chance of winning a majority (or plurality) in

parliament again. The fact that this bill did not pass in 2013 does not necessarily demonstrate

undue influence that stems from campaign donations, but instead may exemplify other ways that

powerful, organized interests can influence the policy process.

In the United States, labor union issues are also highly contested. Even though voting is

voluntary, those heavily involved in unions or who feel strongly for or against them are more

likely to vote (The American National Elections Studies 2015). Studies have also shown that

unions themselves foster democratic participation, most notably through voter turnout. In the

2014 U.S. midterm elections, union members were 13% more likely to vote than non-union

members (McElwee 2015a).

Because this particular union regulation policy does not influence the lives of citizens not

involved in the construction industry, nearly universal turnout may not actually affect outcomes

in any meaningful way. If compulsory voting did influence outcomes and could overcome

special interest lobbying, supporters of the anti-union legislation in Australia had a small

advantage in terms of public opinion and the legislation probably should have passed.

In conclusion, this issue was on the agenda because the Coalition has always been

unhappy with the Labor Party’s solution to construction industry oversight. The Australian

people did not have strong opinions on the issue. I found no major donations from the

construction industry except from the worker’s union. The union was strongly opposed to the

proposed regulation; however, it is highly unlikely that the CFMEU donations changed the Labor

party’s position in any meaningful way—their interests were already aligned. Money in this case

seems to matter, but through a different and entirely legal channel: public discourse about

106
proposed legislation. Such public discourse is nearly impossible to regulate; while it can be seen

as manipulative, it can also be seen as democracy in action; that is, people convincing one

another through persuasive arguments.

Landholders’ Right to Refuse (Gas and Coal) Bill

The Australian Greens introduced the Landholders’ Right to Refuse (Gas and Coal) Bill

in 2013. The bill tries to give landholders the right to refuse coal seam gas operations on lands

that produce food—companies must obtain written authorization to begin their resource

extraction after an independent assessment of current and future risks affecting relevant land and

ground water sources. Courts would have the ability to grant injunctions to protect the

landowners if companies do not comply. Prime Minister Abbott has previously said that he

would support such a measure, but his government had yet to introduce the legislation. In the

meantime, the Coalition government of New South Wales has passed legislation giving permit

holders a right of entry, which is at odds with Abbott’s previous position in support of the

landowners. The measure was defeated in 2013 but was reintroduced by the Greens in 2015 and

is currently before the Senate (TimeBase 2013, Greens).

On this topic, few Australians have strong opinions. 22% in total think that coal seam gas

mining resource extraction should be banned entirely, 32% in total think that it should be

restricted on farming land. 12% think that there is sufficient current regulation of coal seam gas

mining. 34% don’t know (Essential Report 2014, Coal seam gas).45

45
The party breakdowns are as follows: 26% of the Labor vote, 15% of the Coalition vote and 39% of the Greens
vote think that coal seam gas mining should be banned entirely. 34% of the Labor vote, 29% of the Coalition vote,
38% of the Greens vote think that it should be restricted on farmland. 5% of the Labor vote, 22% of the Coalition
vote and 2% of the Greens vote think that current regulation is enough and 34% of the Labor vote, 34% of the
Coalition vote and 22% of the Greens vote do not know (Essential Report 2014, Coal seam gas)

107
 On the other hand, energy and mining corporations are active donors, and the big donors

frequently give significant amounts of money to both the Labor party and the Liberal party

(though they give a good deal more to the Liberal party).46 Unsurprisingly, they do not give to

the Greens. Money may be the reason that the Liberal party, despite Abbott’s professed support,

did not place this legislation on the agenda. It may also be the reason that, after another party

introduced the legislation, Abbott still did not support the bill.

Because this issue does not seem to be particularly salient for voters but it is for certain

energy and coal companies, money in this case may be shaping policy outcomes. Abbott’s party

stands to lose a great deal of campaign money if he passed the policy, but will not likely lose

very many votes if he does not. In addition, his own party at the state/territory level had already

committed to policy giving companies the right instead of landowners. Given these various

incentives, it made little sense to fight for the bill. Even though he was breaking his word to give

landowners the right to refuse, he stood to lose very little. This change in policy stance reflects

Abbott’s values—landowners and their votes were less important to Abbott than his own party’s

support and/or important campaign contributions. It also shows us that money can win in

situations where campaign money is present but voter interest is not.

Live Animal Export Prohibition (Ending Cruelty) Bill

The last significant bill that did not pass during the 2013-2014 parliamentary session was

the Live Animal Export Prohibition (Ending Cruelty) Bill 2014. Andrew Wilkie of the Liberal

46
For example, Beach Energy Limited gave A$50,000.00 to the Labor Party South Australian Brach and A$
110,000.00 to the Liberal party South Australian Branch. Woodside Energy Limited gave A$110,000.00 to the
Labor party national branch and A$1,100.00 to the Labor party Western Australian Branch. It also gave
A$119,500.00 to the Liberal party national branch and $6,600.00 to the Liberal party Western Australian branch.
Trepang Services gave A$10,000.00 to the Labor party Northern Territory branch and A$100,000.00 to the Liberal
party national branch.

108
party sponsored the bill, which would have prohibited the export of livestock for slaughter after

July 1st, 2017 and included provisions to ensure that all livestock were treated humanely before

their slaughter. In total, according to a poll, 25% of Australians believed that the country should

not export live sheep or cattle to any country at all (supporting the total ban provision). 52% of

Australians believed that Australia should only export live sheep and cattle to countries that

guarantee that the animals will be treated humanely. 16% believed that Australia should export

live sheep and cattle to any country that wants them. 7% in total responded ‘don’t know’

(Essential Reports 2013, Live animal). 47

By not passing this bill, parliament rejected the unconditional ban on exporting animals

for slaughter. This seems to be in line with public opinion; therefore, campaign money did not

appear to cause policy outcomes to diverge from public opinion. I found no donations from any

major Australian livestock export corporation or from any animal rights association. If either

interest had tried to influence policy, they would have done so by lobbying lawmakers or

changing the minds of the voters. That this issue was on the agenda in the first place does not

represent popular opinion; it instead represents a special interest. The animal rights lobby may

have put pressure on certain politicians to sponsor and introduce this bill. Yet, the animal rights

lobby was not strong enough to win without public opinion squarely on its side. The rejection of

this bill does not guarantee that the most popular position—that Australia should only export live

sheep and cattle to countries that guarantee they will be treated humanely—will become law. A

party may or may not propose this legislation and while its passage would not be problematic

because it is in line with public opinion, its lack of passage or future lack of introduction may

47
29% of Labor voters, 19% of Coalition voters and 35% of Greens believed that there should be a total ban. 50% of
Labor voters, 56% of Coalition voters and 56% of Greens believed that Australia should only export to countries
that will treat the animals humanely. 14% Labor, 20% Coalition and 8% Green believe that Australia should export
to any country that wants the animals and 7% of Labor voters, 7% of Coalition voters and 1% of Greens did not
have an answer (Essential Reports 2013, Live Animal).

109
signify corporate interests interrupting the constituent-representative link through lobbying, not

through campaign finance.

The lesson we can take from this example is that in the natural absence of big campaign

finance money, outcomes seem to mirror public opinion. At most, lobbying was able to get the

issue on the agenda.48

Conclusions

Table 5.2: Carbon Fair and Building and Landholders’ Live Animal
Summary of Tax Sustainable Construction Right to Export
Bills, Repeal Pensions Industry Refuse (Gas Prohibition
Money, Bill (Improving and Coal) (Ending Cruelty)
Public Productivity Bill Bill
Opinion and Bill)
Outcomes
Money To both No CMFEU to Energy/Coal No
sides ALP to both sides,
preference for
Liberal Party
Public In favor of Against Not strong Not strong In favor of
Opinion repeal pension exporting when
changes guaranteed that
animals will be
treated humanely
Outcome Repealed Pensions Not Passed Not Passed Not Passed
Changed (lobbying
efforts)

Big campaign money in Australia did not seem to show too much influence on policy

outcomes, both in what did or did not pass (summarized in Table 5.2). It may have played a role

in Abbott’s decision not to support a bill protecting landowners from coal and gas companies,

but this was in the absence of strong public opinion on the matter. When public opinion was

48
This may not have even been a result of lobbying efforts—perhaps MP Wilkie is just a big animal lover.

110
strong, it seemed to dictate policy outcomes both when campaign money went to both sides and

when there was no campaign money involved. When public opinion was not strong, campaign

donations and lobbying efforts seemed to influence policy outcomes. What this tells us is that

compulsory voting may have mattered (and campaign finance did not) in influencing policy

outcomes when the general public cared about an issue. Australian representatives are on the

whole responsive to the public interest where it exists. Where it does not, representatives are

responsive to different groups who assert their interests either through campaign donations or

through lobbying. The only exception was the pension reform law, where public opinion did not

predict outcomes, but neither did campaign donations. Lobbying interests likely played a larger

role. The pension issue, however, is a somewhat unique case, not only to Australia but also

throughout the rest of the world—politicians are struggling to find a balance between the public

interest and the long-term sustainability of large social welfare programs.

Given the Australian example, it would seem that compulsory voting (or rather, having

nearly universal turnout) does have a real effect on policy in certain situations despite Australia’s

lax campaign finance regime. Australian representatives are on the whole responsive to the

public will and therefore exhibit quality representation. In a voluntary voting system, perhaps the

pension reform would not have passed because the elderly (who strongly dislike policies that cut

benefits) vote at higher rates than the young. Gilens’s research showed us that in a voluntary

voting system such as the U.S., policy outcomes only reflected the preferences of the wealthy

(Gilens 2012). In this particular compulsory voting system, policy outcomes reflected the will of

the public on the whole when the public had an opinion.

Their system of collective accountability ensures that voters can rely on party labels and

party platforms to communicate their interests, and then they can be confident that the

111
representatives they elect fight for what they promised. The implications of this case study would

suggest that, in a system where accountability is clear and there are reliable mechanisms of

enforcement, we can look to increase turnout to combat donor influence. In the following

chapters, I explore if this conclusion holds true given other political landscapes and different

campaign finance regimes.

112
 Chapter 6: Brazil

Brazil scores poorly on all six of the dependent variables I used in my regression models.

Table 6.1 contains a summary of the scores. Brazil has a considerable amount of campaign

finance regulation, and notably more than Australia. This would suggest that perhaps compulsory

voting does not itself improve quality of representation despite a significant amount of campaign

finance regulation. Because compulsory voting is working in Australia yet outcomes are still

poor in Brazil, it seems likely that Brazil has a unique set of institutions and contextual problems

affecting the influence of both compulsory voting and campaign finance regulation.

Table 6.1: Voice and Control of Confidence Favoritism Are Voters Is there a
Summary Accountability Corruption in (1-7 best) Represented Leader
of Scores (-2.5-2.5 best) (-2.5-2.5 Parliament in Who
on best) (1-10 best) Elections? Represents
Dependent (1-4 worst) You?
Variables (1 yes; 2
no)
Brazil 0.37 -0.12 5 2.58 2.94 1.36

In this case study, I again begin by studying Brazil’s electoral system. The system

deserves special scrutiny in that it creates electoral chaos and does not foster any kind of

meaningful accountability. I then give an overview of Brazil’s approach to campaign finance and

mandatory voting, and then focus on the voting patters of and donations made during the 2014

elections. Next, I look at legislative outcomes during 2015 and conduct a detailed analysis of

three bills (two that passed and one that did not).

I find that in each case, money dictated outcomes, though it was never clear whether

money’s influence was coming in only through campaign finance, through lobbying or through

illicit channels. Voters cannot rely on party labels to inform their decisions, nor is it easy for

them to keep track of their representatives’ behaviors. The media does not effectively inform the
public of individual roll call votes, making it impossible for voters to sanction bad

representatives at the polls. All of these contextual factors perhaps impede the effects that

compulsory voting and/or campaign finance regulation could have had. Taken with the Australia

example, we might conclude that compulsory voting can make a meaningful difference in quality

of representation, but there are necessary intervening institutions, specifically, the effective

transmission of relevant information that voters need to make informed decisions on election

day.

Brazil’s Electoral System49

Electoral systems are never wholly determinative in their influence on representation and

quality of government; however, Brazil’s electoral system deserves special scrutiny. Much of the

available literature on the country concludes that the electoral rules have been responsible for or

significantly exacerbated many of the contemporary governance problems from which Brazil

suffers, including a lack of meaningful representation (Ames and Power 2007). The system

leaves gaps in accountability that politicians and parties have been unable to fill. With few

exceptions, Brazilian parties are “virtually irrelevant as vehicles of political information and

socialization” (Ames and Power 2007, 202). The Brazilian people lack effective channels of

participation in a system designed to encourage endless party and candidate proliferation. Given

the chaos of politics and elections, any real accountability—and therefore, responsiveness—falls

through the cracks.

Brazil’s system design was heavily influenced by the outgoing military regime. The

Constituent Assembly of 1986, comprised 40% by members who were affiliated with the

military regime, drafted the new democratic system and electoral rules through compromise,

49
The research and some of the writing in this subsection comes directly from the final paper I wrote for my Senior
Seminar class on Comparative Democratization with Professor Michael Kraus.

114
conceding a great deal to the powerful military faction. Despite the Assembly’s best efforts to

create a lasting, resilient system that fit the country’s culture and needs, the final product was

“influenced heavily by short-term political considerations…[and] provisions [were] often

substantially out of line with the sentiment of the majority of the membership” (Schneider 1991,

334). The most striking example is the adoption of a presidential system despite clear preference

for a parliamentary one. The landslide victory for presidentialism can be attributed to the “non-

too-veiled threat of possible armed forces intervention” if the assembly voted another way

(Schneider 1991,336).

Nearly 30 years later, much of the system laid out in the 1988 Constitution remains the

same. Brazil has a bicameral legislature; the upper house, the Federal Senate (Senado Federal), is

comprised of 81 members—three members per state and three members for the federal district,

elected by a majority vote. The lower house, the Chamber of Deputies (Câmara dos Deputados),

has 513 members elected through a system of proportional representation.

Many scholars in Brazilian politics have noted the “notorious weakness of Brazilian

parties” and the striking historical “party underdevelopment” (Power 1995, 101). Parties have

very little power and voters and candidates have very little loyalty to party organizations. This

phenomenon has been cited as “a handicap of Brazilian political development” (Power 1995,

101), because it inhibits much-needed political stability. Such underdevelopment is partially the

fault of frequent interruptions of authoritarianism, making their purpose essentially obsolete for

protracted periods of time (Power 1995). The drafting of the new Constitution of 1988 from

scratch gave Brazilians a unique opportunity to design their electoral rules to mold the party

system to fit the country’s needs. In a time of great transition, one would expect the Constituent

Assembly to seek stability and predictability. Yet, “at each branding point, Brazil has chosen the

115
decision-rule that most encourages party proliferation” (Linz and Stepan 1996, 183), a path

antithetical to the much needed party control and development which could serve to better link

constituents to their accountable representatives.50

Why would this be so? On the one hand, the military party (originally ARENA,

rebranded in 1980 as the PDS) was substantially more organized, cohesive and popular than

many of the small opposition parties. Maintaining a fragmented, weak party system would give

the military more relative power. For members of the Constituent Assembly not affiliated with

the military regime, many wanted to preserve their own autonomy from parties (Mainwaring

1991). Their interest in weak electoral rules was not simply out of self-interest, however, as weak

electoral systems often lead to high turnover rates (in Brazil, turnover is sometimes 60% in an

election) (Mainwaring 1991). Many politicians associated strict rules that controlled parties and

disallowed or discouraged party proliferation with authoritarian control. The fear was warranted;

the recent military regime had imposed strict party rules that greatly favored their own. Wanting

to move far away from any policies that could be considered anti-democratic, politicians chose

the policies that guaranteed little party discipline and promoted individualism (Mainwaring

1991).

To begin, Brazil chose a system of proportional representation (Linz and Stepan 1996).

Proportional representation is uncommon in presidential systems; however, this choice reflects

the intention of the Constituent Assembly to eliminate the separate executive branch. Even

though the ex-authoritarian members of Congress imposed a presidential system, the Assembly

left intact electoral rules that correspond to a parliamentary government and allow for more

50
Systems of individual-level accountability do not necessarily demand strong party control, but given the chaos of
the Brazilian system (explained throughout this section), some control could help to clarify the positions of many
candidates. If party labels meant more, they could convey relevant information to voters who are asked to choose
between a large number of candidates.

116
parties in the national congress. In typical parliamentary systems, “Parties are corresponsible for

governing…they must continue to support the policies of the executive or that executive will

fall” (Mainwaring 1991, 31). This encourages party discipline, loyalty and stable coalitions. In a

presidential system, a party member can vote out of line with the chief executive and the chief

executive does not lose power, thus encouraging more independence from one’s party. The

president in Brazil is therefore in a difficult position—his or her party may have fewer seats

because of the sheer number of parties represented (due to proportional representation) and there

is no real consequence if members of the president’s own party vote independently (because the

executive derives its power independently of Congress). Of course, lack of party discipline

presents an opportunity for the president as well; he or she can seek support within many

different opposition parties whose members do not feel beholden to their party line (Mainwaring

1991). In 1991, scholars claimed that the Brazilian presidency was the “most powerful

presidency in the world” (Ames and Power 2007, 188) due to the resources it controls. Because

the independent executive has its own legislative agenda, it can exert pressures on legislators not

usually present in parliamentary systems. In fact, Presidents appear to get most of what they

want from Congress (Ames and Power 2007). This additional principal has the potential to divert

legislators’ attention from their constituents because the presidency is so strong and controls so

many resources.

Brazil’s proportional representation system is further unusual in that it has an extremely

low threshold for representation (Linz and Stepan 1996). A party can win seats in Congress by

acquiring a mere .04% of the national vote (Linz and Stepan 1996, 183-184), thus, there is little

incentive for small parties to merge in order to consolidate votes. The number of effective

political parties in most consolidated parliamentary democracies is typically between three and

117
seven. No consolidated presidential democracy had more than 2.7 effective parties in the

legislature between 1979 and 1989. In 1992, Brazil had 8.5 effective parties, “making [the lower

house] three times more fragmented (or less aggregated) than any long-standing consolidated

presidential democracy in the world” (Linz and Stepan 1996, 181), let alone still being outside of

the normal range of parliamentary democracies. The absolute number of parties today in the

Chamber of Deputies (the lower house) is 28 (Cascione).

Thresholds are considered one of the main ‘barriers to entry’ for a new party. Brazil’s

barriers on the whole are exceptionally low (Mainwaring 1991). Frequently, parties will splinter

into factions because they easily can. Frustrated party members can simply create their own party

instead of being coerced into a vote. If they choose to do so, they are given almost all of the same

congressional privileges as the larger, established parties. For example, all parties are given

space for their leadership, secretarial assistance, telephones and free television airtime during

campaigns (Mainwaring 1991). Each party, no matter how many candidates, receives two hours

of free media time on national television and radio and one hour of free time on statewide

television and radio (Linz and Stepan 1996). All of these automatic benefits are incentives for

candidates, once elected, to form new parties.

A provision called the candidato nato gave any politician who switched parties during his

term a guarantee that he could appear on his new party’s ticket in the next election (Mainwaring

1991). Brazilians celebrated their ability to switch parties at will: “The mandate is granted by the

people [as opposed to the party] and should be taken away only by the people” (Mainwaring

1991, 36). This permissive attitude, combined with the personalistic nature of Brazilian politics,

demonstrates that voters have few qualms about voting for their preferred candidate regardless of

any party changes (Schneider 1991). In the first three legislative sessions of Brazil’s new

118
democracy, about a third of Congressmen changed parties (Power 1995, 109). This was not

discouraged until 2007 when the Supreme Court banned switching parties within a term, though

each case is to be evaluated on an individual basis and candidates can switch parties between

terms and still be guaranteed a spot on the ballot just as easily as before (Soares 2015).

Open lists make Brazilian politics even less party-oriented. Voters vote for candidates

within a party. Parties are first awarded a number of seats proportional to the number of votes

they receive, and then the seats are distributed according to which candidates won the most

votes. This type of system encourages intra-party competition (Power 1995) and “provides a

strong incentive for individualism in campaigns, especially since a politician’s prestige and

power are greatly enhanced by a massive vote total” (Mainwaring 1991, 24). If a candidate is

popular enough, he can win his party various seats; however, he is not dependent on the party for

his mandate and therefore feels no pressure to adhere strictly to the party’s rules; in fact, “the

overwhelming majority of Brazilian elected officials do not believe that they owe their mandate

to their party (Power 1995). The part of the candidato nato provision that still stands after the

Supreme Court ruling in 2007 protects the politician. He can “violate all of the party’s

programmatic concerns, consistently vote against the leadership, and still be guaranteed a place

on the ballot…Parties put up with flagrant violations of party programs and organizational

commitments if a politician brings in a lot of votes” (Mainwaring 1991, 25-28). If a party were to

try to exert its control over a member, the member would simply leave and the party would lose

votes.

Moreover, each party is allowed to have 1.5 candidates multiplied by the number of open

seats in the election. For example, in São Paulo, a party can have up to 90 candidates for deputy

and 126 candidates for state deputy (Mainwaring 1991). One can imagine that a ballot might

119
overflow with names of candidates, all of whom each voter is unlikely to know. Voters must rely

on cues from the party labels to inform their decisions; yet, party labels mean little. Voters know

this—despite the option of simply giving one’s vote to a party, about 90% of voters still choose

to vote for individual candidates (Ames and Power 2007). This stands in direct contrast to the

Australian open-list preferential system where a similar percentage chooses to vote for a party.

Parties are vehicles for effective participation and representation. If a politician’s

behaviors are unregulated and unpredictable, citizens cannot reliably know for whom or for what

they are voting. An electoral system that allows for so many candidates and encourages such a

complex party system might do well to rely on collective accountability. Unfortunately, the

Brazilian system encourages neither collective accountability to a party nor individual-level

accountability to the constituencies: “Mechanisms of accountability are vitiated because the

electorate can not keep track of the performances of all of the deputies and senators, nor can it

infer much about their performances and positions on the basis of party affiliations”

(Mainwaring 1991, 40). Ballots overflowing with candidates make it difficult for even the most

politically active citizen to make informed decisions, especially when the party labels mean so

little.

Given the emptiness of party labels, it would make the most sense to look at individual

roll-call votes when evaluating Brazil’s responsiveness; however, these are not easy to find

online. In my own review of certain bills, I rely on the media to explain who voted for and

against each piece of legislation. Despite the chaos, the media still uses party labels to explain

outcomes and will only list the individual votes of certain prominent politicians. This is, of

course, problematic and illustrates again another hole in accountability that I will explore in

more detail later.

120
 Brazil is infamous for its corrupt politics—most notably, bribery—which is a clear

indicator that something is wrong within its system of accountability. Corruption is a signal that

the link between voter and legislator has been interrupted; the loyalty of a representative has

been pulled too far towards a different principal, be it a big campaign donor or an influential

corporation promising a job. Corruption is a historical problem in Brazil, and one that many have

come to accept as a necessary evil. Brazilians say, “rouba, mas faz,” which means candidates

“steal, but make things happen” (Baquero 2015, 150). Citizens have largely accepted that

corruption is just a natural phenomenon and there is little they could, or should do.

Corruption contributes to “growing alienation, and apathy of citizens regarding

politics…Brazilians believe that governments are unresponsive to their needs” (Baquero 2015,

139). Surveys have found that political parties and Congress suffer from extremely high levels of

distrust, at 81% and 71% respectively (Baquero 2015, 151). Low levels of trust in democratic

institutions makes citizens doubt the representation they are receiving and the viability of the

system as a whole.

Many in the Constituent Assembly of 1986 did want to adopt a recall procedure to hold

politicians more accountable to their constituents. The recall would discourage corruption; if

news of the illegal behavior spread, a constituency could initiate a process to remove that

representative from office, thereby raising the risk of accepting bribes. The process is another

channel by which Brazilians could have secured better representation in politics. The reform did

not pass (Power 1995), and even if it had, its effectiveness presumes fair information is

distributed to all who it affects. Strong parties could have acted as watchdogs to alert citizens of

foul play; however, parties in Brazil may not want to alienate or drive away candidates that bring

121
them many votes. As it stands, the electoral system is permissive of bad behavior and poor

representation.

Beyond explicit corruption, Brazilian politics rely heavily on clientelism and patronage

politics that interfere with quality representation for all. Representation in Brazil has a notable

class-bias, exacerbated by staggering economic inequality. Scholars who study Brazil note,

“representation means clientelism rather than representing clearly defined social groups”

(Mainwaring 1991, 39). The culture runs deep though the country’s history. No regime has been

able to rid the country of this vice, nor has there been a good faith attempt to do so, perhaps

suggesting that the problem is more cultural than institutional. Self-serving politicians benefit

from these networks and leverage access to resources to enhance their own power. With the

return to democracy in the late 1980s, Brazil saw “a newly competitive political system, and an

unprecedented explosion in the number of clients (potential voters)—that combined in a way that

exacerbated older traditions such as patronage politics and the patrimonial state” (Power 1995,

107). Especially because all literate citizens between the ages of 18 and 70 were obliged to show

up at the polls, candidates reverted to old tricks to win over voters. It is possible that in a system

of chronic clientelism and corruption, high levels of turnout may actually exacerbate the

problem.

‘Clients’ with money have “easy access” to representatives and bureaucrats and therefore

do not need parties to articulate or promote their interests. On the contrary, “Popular

interests…are not effectively represented through such informal channels, and other forms of

popular representation (unions, neighborhood associations, peasant organizations) have difficulty

gaining access to the corridors of power” (Mainwaring 1991, 40). Especially for the rural,

isolated poor (some of whom are illiterate), the weak, ever changing party system that “fail[s] in

122
aggregating and articulating the demands of the population” (Baquero 2015, 154) introduces

significant barriers to understanding issues, to effective participation and to quality

representation and government responsiveness to their needs.

The weak party structure has proven inept at correcting this type of asymmetrical

representation. Of course, strong parties and party bosses are too susceptible to developing

pervasive clientele networks; nonetheless, should the country decide to clean up its act, a party

patronage system might be easier to eradicate than attacking the problem individual by

individual in the slow, ineffective, backlogged and sometimes corrupt court system (Schönleitner

1998). A stronger party system may have some inherent advantages in facilitating representation

and controlling this type of behavior without undermining individual-level accountability.

Brazil’s people want representatives to answer directly to their constituencies—

legislators’ mandates come from the voters, not the party. However, electoral rules favor party

chaos and candidate proliferation, making it especially hard for voters to keep up and punish bad

behavior. In a different system, parties could pick up some of the slack by disseminating the

relevant information to the public and punishing rouge or corrupt members; however in Brazil,

parties lack any meaningful control. In addition, the media does not fill this void. The system of

accountability is in great need of reform and does not encourage any kind of meaningful quality

of representation or responsiveness on a population-level. Individual members of the legislature

may offer their constituents good representation, but there are no system-wide incentives to

encourage all members to act as close proxies for their voters.

123
Brazilian Campaign Finance Regulation

Partly because of the specific electoral rules that foster intra-party competition and

individual campaigning, Brazilian elections are among the most expensive in the world

(Mainwaring 1991):

Under the Brazilian electoral system, the need for funds is pressing because of the high
correlation between amounts raised and votes won. The open-list system puts the
responsibility for raising campaign money on the candidate rather than on the party.
Victorious candidates regularly outspend their opponents by large sums. (Casas Zamora
2013, 51-52).

The 2014 presidential, congressional and state elections may have cost more than $3 billion

(Bevins 2014).51

Because of the large sums of money entering Brazil’s system and the system’s notable

vulnerability to corruption, campaign finance regulation is of central importance to maintain the

integrity of elections and governance, yet effective regulation and administration is

understandably difficult. The Superior Electoral Court (Tribunal Superior Eleitoral) regulates

Brazil’s campaign finance. Brazilian campaign finance law bans foreign donations, donations

from corporations with government contracts or partial government ownership, and donations

from trade unions. There is, however, no ban on corporate donations to political parties or

candidates.52 While there is no explicit ban on anonymous donations, there is a “de facto ban

since the mechanism created to make contributions (only through bank or electronic operations)

requires the identification of all contributors” (International IDEA). State resources cannot be

used to support any political party or candidate. Brazil has banned any donations to parties or

51
None of the money spent was spent on Get Out The Vote campaigns—Brazil’s compulsory voting makes it
mandatory that all literate citizens between 18 and 70 show up at the polls.
52
A Supreme Court ruling in September 2015 has actually changed this provision—beginning in the 2016 elections,
corporate donations to political parties and to candidates will be banned. Currently, contracting corporations are the
main source of political campaign funding. It may serve to look at how political campaigns and the following
policies change after the next election cycle, as this new rule is likely to have profound implications on how
campaigns are financed and the incentives that motivate politicians’ behaviors.

124
candidates from other sources as well, such as non-profits that receive foreign funds, religious

and charity entities, sports entities, and any organizations that receive public funds.

Contribution limits to parties and candidates are election-specific and are calculated as

2% of the income or revenue of the ‘juridical persons’ (a corporation) and 10% of the income of

the ‘natural person’ (International IDEA). The larger the corporation or the wealthier the

individual donor, the more they can donate to a candidate: “While such rules do not level the

playing field for private donors influencing the electoral process, they do allow for some control

over the origin of donations from lawful sources” (Casas Zamora 2013, 49). There are no

expenditure limits for either parties or candidates.

Public funding is regularly provided to political parties who are registered with the

Superior Electoral Court. 5% of the funds are allocated equally amongst all registered parties and

95% is proportional to the votes a party won in the most recent Chamber of Deputies election.

Use of public funds is restricted, however. At least 20% must be spent on party research and/or

political education, and no more than 50% can be used to pay salaries. At least 5% must be spent

to encourage female participation (International IDEA). Public funding is an important source of

party campaign money; in 2009, 75% of all political party money came from the public financing

program. On the contrary, around 65% of campaign funds for individuals comes from corporate

donations. The big donors are often banks and construction companies, and may “compete with a

candidates’ relationship with the political party or constituency” (Casas Zamora 2013, 51-52).

Parties also receive free access to radio and television—this access is permanent, though

more time is granted during campaign season. Time is distributed in a similar manner as is public

funding; 1/3 of time is equally distributed and the remaining 2/3 is distributed based on the

proportion of votes received in the last election for the Chamber of Deputies. 10% of this time

125
must be dedicated to promoting female participation. Individual candidates are not entitled to

free television or radio access (International IDEA), though candidates are free to create a new

party entitled to free media access. This provision is equalizing in light of a ban on paid TV

ads—it “help[s] counterbalance the extremely unequal market of private resources, as the state

provides political contenders with airtime depending on their past electoral record rather than on

their fundraising capacity” (Casas Zamora 2013, 50).

Parties must submit disclosure reports monthly during the four months before and two

months after an election. Candidates must also submit income and expense reports through their

finance committee. All reports must include the identity of donors and the amount they

contributed. All of the disclosed information is published online. Of course, disclosure reports

are subject to underreporting or omission—candidates and donors sometimes agree to keep

donations secret, especially when in conjunction with other illegal activity like vote buying and

accepting money from organized criminal groups (Casas Zamora 2013).

The Supreme Electoral Court does not have full audit capacity—it must request technical

support from the Audit Courts. They also work with Revenue Services to investigate illegal and

underreported donations. Candidates who violate the laws may be removed from office, fined,

and lose public funding for up to two years. Parties can be deregistered (International IDEA).

Like the Australian Electoral Commission, the Supreme Electoral Court is rarely strict—when

violations are found, they normally just request a correction.

While regulation and enforcement has been improving, scholars still note the “weak

oversight and impunity in the face of noncompliance” as its biggest handicap (Casas Zamora

2013, 51). Yet, there is reason for optimism in the Brazilian campaign finance regime:

…in recent years the Brazilian press has developed a significant investigative capacity.
Along with non-governmental organizations, the news media can be counted on to police

126
 campaign finance regulations, as they did when corruption allegations surfaced against
the PT [Workers’ Party] government in 2005-6. (Ames and Powers 2007, 209).

This stands in direct contrast with Australia’s combination of a weak regulatory agency and an

indifferent press. Perhaps because explicit corruption is a much more salient issue in Brazil, the

people have developed an interest in and sensitivity to stories of big money playing an

inappropriate role. Naturally, not all corruption happens through the financing of elections, but

political donations are certainly one avenue for money to buy influence.

Brazil receives a high score on its ‘in law’ campaign finance regulation with a 76 out of

100 but a low score ‘in practice’ with a 47. Money, Politics and Transparency cite the unequal

use of non-financial state resources to support candidates despite the ban, the frequent omission

of contributions on disclosure reports, and other violations of regulations such as the vote buying

ban. While the Superior Electoral Court does impose some sanctions, not all violators comply

and the regulation so far has been ineffective at changing political finance behavior (Money,

Politics and Transparency 2014).

Brazilian Compulsory Voting

Brazil’s compulsory voting came with its transition back to democracy in the 1980s

(though as a policy it was first introduced in 1932) (Birch 2009). Voting is obligatory for literate

citizens between 18 and 70 years old. 16-18 year olds can vote on a voluntary basis, as can those

older than 70 years old. Those in the military are banned from voting. (International IDEA).

Citizens who do not vote and do not provide justification (proof they were not in their electoral

district on election day or ill) are fined (The Superior Electoral Court; Cepalani and Hidalgo

2015). The fine is set between 3-10% of the regional minimum wage (around $3.50 Brazilian

Reais or $1.60 USD). Those who do not vote and do not pay the fine are barred from

127
professional exams, enrolling in public universities and receiving wages from the government,

borrowing money from credit institutions and renewing passports (Electoral Commission 2006;

The Economist 2013; the Superior Electoral Court; Cepalani and Hidalgo 2015). Anyone who

does not vote nor justify his or her abstention in three consecutive elections is removed from the

list of registered voters and barred from voting in the future (the Superior Electoral Court).

Because the non-monetary penalties are so severe, some speculate that Brazil’s compulsory

voting disproportionately turns out the middle and upper class (those who are affected most by

the denial of certain state services and benefits). In fact, “the causal effect of compulsory voting

on turnout among the more educated is at least twice the size of the effect among those with less

education” (Cepaluni and Hidalgo 2015), and thereby the policy actually increases political

inequality. However, because compulsory voting in Brazil has existed since the beginning of its

democracy, it is impossible to say if political equality would improve or worsen in the absence of

compulsory voting.53 Regardless of the validity of Cepalani and Hidalgo’s theory, their results do

suggest that voters take their obligation to vote seriously for fear of sanction.

Case Study

My study of Brazil will focus on the campaign finance data from the most recent

parliamentary elections in 2014 and the 2015 legislative year. Voter turnout as a percent of

registered voters in this election was around average for Brazil, at 80.60% of the 140,488,492

registered Brazilians. The turnout of the voting age population was only 75.09% (slightly lower

than the 2006 and 2010 elections but not abnormal)—it bears remembering that in Brazil, voting

53
In a hypothetical voluntary voting Brazil, it is possible that turnout rates would fall into the 50-60% range and the
socioeconomically disadvantaged would vote in lower proportions than the wealthy. Patterns in many other
countries (such as the U.S.) support this prediction. In such a case, Brazil’s enforced compulsory voting as it stands
today does in fact improve political equality. Cepalani and Hidalgo’s study may suggest that Brazil’s political
equality could be further improved if sanctions burdened all socioeconomic groups equally.

128
is not mandatory for those between 16 and 18 years old, those above 70, or the illiterate. Some,

such as those serving the armed forces, are actually forbidden from voting. The 2014 election

saw a relatively high percentage of invalid votes (13.95%) but values fluctuate wildly from year

to year between a low of 8.64% in the 2010 election and 40.10% in the 1990 election

(International IDEA 2016).

The Senate of 2015-2019 is comprised of 17 different parties and four members without a

party. Figure 6.1 shows coalitions built for the 2014 presidential election.54 While easy to

interpret, it does not capture the nature of the ever shifting coalitions or the tendency of party

members to abandon their party position and act of their own accord. Table 6.2 summarizes the

composition in more detail, though it does not fully capture the complexity of the Brazilian party

system or the difficulties in trying to govern within it. Because of the frequent creation,

dissolution and shifting of parties, many do not fit neatly into a two dimensional ideological

spectrum. The table gives us a better sense of how difficult the system is for voters to navigate.

Figure 6.1 Figure 6.2

54
This graphic, and Figure 6.2 come from: http://www.nivela.org/articles/post-election-brazil-new-conditions-old-
models/en

129
Table 6.255 Right-leaning Center Left-leaning
Government and Brazilian Democratic Democratic Labor
Government Support Movement Party Party (PDT) (3)
Bloc (PMDB)* (18) Workers’ Party
(PT)** (11)
Progressive Progressive Party (PP)
Democracy Bloc (6)
Social Democratic
Party (PSD) (4)
Union and Strength Party of the Republic
Bloc (PR) (4)
Brazilian Labor Party
(PTB) (2)
Social Christian Party
(PSC) (1)
Brazilian Republican
Party (PRB) (1)
Christian Labor Party
(PTC) (1)
Opposition Bloc Democrats (DEM) (4) Brazilian Social
Democracy Party
(PSDB) (11)
Green Party (PV) (1)
Socialism and Sustainability Brazilian Socialist
Democracy Bloc Network (REDE) Party (PSB) (7)
(1) Popular Socialist
Party (PPS) (1)
Communist Party of
Brazil (PCdoB)
(1)
Other Members without a Party (4)
*Plurality in Senate and Chamber of Deputies **Presidency

The Chamber of Deputies currently consists of 25 different parties and one member

without a party. Figure 6.2 again shows the coalitions built for the 2014 Presidential race, and

Table 6.3 summarizes the composition of the Chamber by ideology and by coalition.

55
Blocs taken from: https://en.wikipedia.org/wiki/Federal_Senate, but member party composition taken from
http://www25.senado.leg.br/web/senadores/em-exercicio/-/e/por-partido. The left/center/right spectrum in this table
is very simplified—certain parties are much further right or left than others placed in the same category. It is also
worth noting that the left-right spectrum is not the same in Brazil as it is in the U.S.; the whole spectrum appears to
be shifted somewhat to the left.

130
Table 6.356 Right-leaning Center Left-leaning
Bloc 1 Progressive Party (PP) Humanist Party of
(49) Solidarity (PHS)
Brazilian Labor Party (7)
(PTB) (19)
Social Christian Party
(PSC) (9)
Bloc 2 Party of the Republic
(PR) (40)
Social Democratic
Party (PSD) (33)
Republican Party of
the Social Order
(PROS) (5)
Bloc 3 National Ecologic Brazilian Democratic
Party (PEN) (2) Movement Party
(PMDB)* (68)
Bloc 4 Brazilian Republican National Labor Party
Party (PRB) (22) (PTN) (13)
Social Liberal Party Labor Party of Brazil
(PSL) (2) (PTdoB) (3)
Unaffiliated Democratic Labor
Party (PDT) (20)
Workers’ Party
(PT)** (58)
Communist Party of
Brazil (PCdoB)
(12)
Minority Democrats (DEM) Brazilian Social Brazilian Socialist
(28) Democracy Party Party (PSB) (31)
(PSDB) (51) Popular Socialist
Solidarity (SD) (14) Party (PPS) (8)
Green Party (PV) (6) Socialism and Liberty
Sustainability Party (PSOL) (6)
Network (REDE) Party of the Brazilian
(5) Woman (PMB)
(1)
Members without a Party (1)
* Plurality in Senate and Chamber of Deputies **Presidency

56
Bloc and member information taken from http://www2.camara.leg.br/deputados/liderancas-e-
bancadas/bancadas/bancada-atual. Blocs are in no particular order.

131
Campaign Finance of the 2014 Election

Interpreting campaign finance data in Brazil presents quite a challenge, not only because

of the unusually high number of political parties, but also because of the unusually high number

of candidates who do not feel beholden to their party’s platform. Whereas in Australia legislators

can be expected to vote along party lines and therefore to the extent that parties are influenced by

campaign donations, legislators, too, are influenced by the party’s biggest donor, Brazilian

legislators operate differently. Elected representatives may have distinct donors from those of the

party, yet the sheer number of candidates and donations makes potential influence hard to track.

Because there are so many candidates in so many parties running for 592 seats in the National

Congress and because donations are more easily interpreted on the party level on the online

election finance database, the following observations are made without regard to the individual

candidate to whom they were made. Given Brazil’s emphasis on individual campaigning and

quasi-individual-accountability system, this is not ideal; however, that the data is so

overwhelming and difficult to generalize in a meaningful way shows how difficult it is to report

well on campaign finance and uncover influence. A lack of quality, accessible and

comprehensive reporting would mean that voters could not make fully informed decisions or

sanction representatives who exhibit too much influence from outside sources.

Table 6.4 shows the industries that contributed to the top five parties (by number of seats

in the current Brazilian legislative branch) during the 2012 state and municipal election cycle.

The data for the 2014 parliamentary elections is likely different (though highly correlated);

however, it is not available in the same format. The below information is still meaningful in that

it can show the party preferences of certain industries and what industries over the long term may

be shaping party platforms, though it should not be taken to precisely reflect interests on the

132
Table 6.4 PMDB (center) PT (left) PSDB (center) PP (right) PR (right)
Total % Total Total % Total Total % Total Total % Total %
Total Total
Construction 53,509,624.31 11.5 87,897,698.59 16.6 44,491,387.10 11.7 11,269,797.59 7.1 7,632,647.64 6.8

Manufacturing 27,653,009.11 5.9 32,441,880.86 6.1 25,002,860.49 6.6 9,174,133.34 5.8 4,992,561.59 4.5

Trade/ commerce 25,528,276.64 5.5 16,667,329.94 3.1 20,898,228.07 5.5 7,581,376.05 4.8 7,452,604.84 6.7

Financial 9,306,646.65 2.0 13,032,210.79 2.5 10,838,438.32 2.9 2,897,399.58 1.8 334,251.35 0.3
Activity/
Insurance
Administrative 6,600,222.94 1.4 7,909,668.93 1.5 5,336,131.76 1.4 1,323,233.36 0.8 1,349,485.22 1.2
and Customer
Service
Transportation/ 6,327,698.06 1.4 6,895,419.05 1.3 5,774,087.95 1.5 1,253,417.68 0.8 1,165,319.25 1.0
Storage/ Postal
Services
Professional 2,762,686.88 0.6 8,479,482.69 1.6 3,355,734.57 0.9 939,515.69 0.6 912,894.19 0.8
Activity, Science
and Technical
Real Estate 2,271,500.54 0.5 2,326,475.72 0.4 3,442,600.35 0.9 314,212.01 0.2 382,829.00 0.3

Health and 2,199,182.92 0.5 6,623,976.35 1.2 1,318,185.62 0.3 180,233.60 0.1 63,456.05 0.1
Human Services
Extractive 1,971,658.74 0.4 1,705,400.36 0.3 1,784,492.06 0.5 792,248.03 0.5 490,519.75 0.4
Industries
Information and 1,809,593.37 0.4 1,643,831.75 0.3 1,401,829.86 0.4 323,928.48 0.2 243,992.00 0.2
Communication
Water, Sewage, 1,719,285.27 0.4 5,923,138.87 1.1 3,908,126.98 1.0 62,553.53 0.0 555,519.84 0.5
Waste
Management and
Decontamination
Education 1,039,261.13 0.2 1,586,671.02 0.3 522,643.25 0.1 208,485.49 0.1 112,290.17 0.1

Agriculture, 1,037,416.85 0.2 3,344,466.01 0.6 1,989,047.40 0.5 359,980.09 0.2 177,189.34 0.2
Livestock,
Forestry.
Fisheries,
Aquaculture
Food and 650,496.30 0.1 1,050,598.03 0.2 673,784.94 0.2 337,386.85 0.2 390,631.09 0.3
Housing
Electricity and 416,334.52 0.1 1,903,722.24 0.4 748,689.46 0.2 248,686.53 0.2 101,337.67 0.1
Gas
Art, Culture, 332,824.44 0.1 217,157.11 0.0 174,446.21 0.0 105,659.15 0.1 135,056.70 0.1
Sport and
Recreation
Public 187.10 0.0 73,399.97 0.0 4,111.32 0.0 9,355.09 0.0 1,674.07 0.0
Administration,
Defense and
Social Security
Other Service 232,319,234.90 49.9 253,281,753.40 47.8 181,554,076.70 47.9 83,732,867.01 52.6 54,822,418.76 49.1
Activities
No information 88,132,376.31 18.9 77344386.08 14.6 65,992,527.47 17.4 38,024,527.82 23.9 30,330,803.50 27.2

TOTAL 465,587,516.98 100 530,348,667.70 100 379,211,429.90 100 159,138,997.00 100 111,647,482.00 100

133
national level—because of the division of power between levels of government, state and local

elections are likely focused on different policy areas. Data is in USD57.

What is first striking about this data is that the party donor profiles look very similar. No

industry seems to greatly favor one party or one ideology, though there are subtle differences.

The center and left-wing parties each collected between one and four times the amount that the

two right wing parties spent on the state and municipal elections in 2012. Nearly all of the

sectors favored the PMDB, the PT and the PSDB over the PP and PR. The right wing parties

(especially the PR) got a higher percentage of their money from donors for whom the Superior

Electoral Court has no data—perhaps these donations come from individuals, as anonymous

donations are prohibited (not by law, but rather by the process of making a contribution).

Notably, less of the right wing parties’ money came from construction, manufacturing and

financial banks/insurance, whereas slightly more tended to come from trade and commerce

entities.

These patterns are counter-intuitive—typically certain sectors, such as mineral extraction

and the financial industry favor right-wing parties because they stand to lose a great deal of profit

from the additional regulation for which the left often advocates. This may partially be because

the Brazilian spectrum operates differently—there are so many parties that they fall on a

continuum that is markedly different from the one we use in the U.S. This pattern could also

reflect intent to play the game of Brazilian politics and win favor with the parties likely to win,

who are more often left or center-left. Whether they donate to influence policy or to express

ideological agreement may depends on the corporation, country or political party; however, one

57
I used the October 2012 average exchange rate between the Brazilian Real and the US dollar (1 USD is equal to
0.492373 Reais). The data contained in this chart all comes from the TSE Statistics database found at
http://www.tse.jus.br/hotSites/estatistica2012/estatistica-prestacao-contas/doadores.html.

134
may look twice at big corporate donations to a party that favors redistribution and increased

government involvement in the private sector.

Another explanation for this discrepancy in Brazil may be the pervasiveness of corruption

and the center-left majority’s willingness to accept bribes. This is not mere speculation; the

ongoing Petrobras Scandal illustrates clearly how campaign donations fit into corruption

schemes in Brazil. Petrobras (a semi-public corporation that is 51% State owned) coordinated

with other construction companies to bring all those involved a great deal of money at the

expense of the State:

Starting in 2004, according to prosecutors, a small number of top Petrobras officials

colluded with a cartel of companies to overcharge the oil company for construction and
service work. The cartel would decide which of its member companies would win a
contract to, for instance, service an oil rig or build part of a refinery. This fake
competition was overseen by Petrobras confederates who were rewarded with bribes.
They kept some of the money, but shared much of it with political figures. (Segal 2015).

Petrobras officials speculate that bribes totaled nearly $3 billion—The Workers’ Party (PT) alone

received nearly $200 million over the past decade—“money that was supposedly used to finance

political campaigns” (Segal 2015). President Dilma Rousseff herself (of the PT), a former

member of the Petrobras board of directors between 2003 and 2010, denies that she was part of

the scheme and has yet to be indicted or impeached (though such proceedings are likely coming

soon, provoked by other recent scandals and cover-ups). Nearly 50 current politicians are being

investigated, including ex-president Luiz Inácio Lula da Silva, and some have already been

arrested (Douglas 2016).

Unsurprisingly, in the 2014 election, many of the largest donors included construction

corporations who view the government as a source of lucrative contracts. Within the first month

of the campaign, businesses of all types had donated approximately R$224 million (roughly

135
equivalent at the time to $100,800,000 USD58). This total only reflected 14% of the donations to

over 26 thousand candidates for the election season and does not include hidden or illegal

donations, where the Secretary General of the NGO Contas Abertas, Gil Castello Branco,

speculated that a significant amount of money may be.

The JBS group which owns two large food brands (Friboi and Seara) donated R$93

billion (approximately $41,850,000,000 USD) to many different candidates, including federal

candidates of the PP and PR on the right and state candidates, including Gleisi Hoffmann of the

PT on the left for governor of the state of Paraná (Galvez da Silva 2014). That a corporation may

support parties on ideological grounds (as seen by many large donations to various candidates)

but pour money into the campaigns of individuals on the opposite end of the spectrum (like the

JBS group) supports the theory that donors want influence. Donors may see party positions as

only loosely related to legislative votes (in the same way that voters choose to vote for

individuals instead of parties), and therefore give to a diverse set of candidates who they want as

allies in office.

The second, third and fourth largest donors were contracting companies—OAS, Queiroz

Galvão and UTC Engenharia. Fifth was a telecommunications company, Telemont Engenharia

de Telecomunicações that donated R$11.9 million (about $5,360,360 USD) with a notable

preference towards the PMDB, though the company also gave money to the PP, PR, PROS and

PSDB (preferring the center and the right over left-wing parties and candidates) (Galvez da Silva

2014). The PMDB is the party with the largest number of seats in both houses, but the PSDB

stands clearly in their opposition. To donate to both suggests a strategy of garnering favor among

voting blocs that could ultimately lead to a majority willing to pass a bill.

58
To put the donations in USD, I am using the exchange rate of July, 2014 (1USD = 2.22 Reais).

136
 Because Brazil has a separate executive branch of government led by a president who

signs bills into law, interest groups hoping to push a bill through have yet another principal to

appease. Especially in a situation of divided government (like during the 2014 legislative session

when the PT held the presidency but the PMDB coalition held the legislature), it makes sense

that political players looking to influence or bribe would donate heavily to both the center

government coalitions and the left-wing PT. They might also donate to both presidential

candidates as insurance that protects against any outcome.

In fact, during the 2014 presidential election campaign between incumbent Dilma

Rousseff (PT) and Aécio Neves (PSDB), some companies (though not all) did just that. Rousseff

received more money than Neves, though Neves was competitive. Neves collected nearly double

what Rousseff did in individual contributions, reflecting “his ample support among wealthier

Brazilians” (Haynes 2014), whereas Rousseff’s main base is comprised of poorer and working-

class Brazilians (Marcello and Levin 2015). Neves’s main corporate support came from banks,

construction and ethanol companies (Haynes 2014). Those in the ethanol businesses “have

protested for years that they cannot compete against gasoline at the pump because Rousseff’s

government has kept down official fuel prices and scrapped a traditional gas tax to control

inflation” (Haynes 2014). Copersucar, the world’s largest sugar and ethanol producer and trader,

donated a million reais to Neves and did not donate anything to Rousseff (Haynes 2014). While

this may have been purely ideological, it may also have been only practical. Rousseff has been

very oil-friendly (in her policies and her own connections with Petrobras) and may be unlikely to

ever pass a policy favoring ethanol at the petroleum industry’s expense. Copersucar may have

decided that she was not worth the investment. Certain banks also only donated to the Neves

campaign, including Banco BMG and insurance company Porto Seguro SA. Other banks that

137
had benefitted from the Workers’ Party’s regime, such as BNDES (Brazilian Development

Bank), donated heavily to Rousseff (Haynes 2014).

Other groups, like Grupo JBS and Ambev (a brewing company) gave heavily to

Rousseff, but also did donate some money to the Neves campaign (around 1/5th and ½ of the

total donation to Rousseff, respectively). Both sides were heavily financed by the construction

industry:

Big builders and engineering companies, which profit from government-sponsored

infrastructure projects, account for about half of donations to each campaign. Major
construction groups, like donors in most industries, steered about two of every three
dollars to the incumbent…By the start of September, builder Andrade Gutierrez had
given nearly 17 million reais to Rousseff and about 12 million reais59 to Neves. OAS, the
industry’s biggest donor, went even further, giving over 30 million reais60 to the
president, more than five times its contribution to the challenger. (Haynes 2014)

Gutierrez and Ambev have both made public statements about these contributions, claiming “the

companies aimed to support democratic debate by donating to multiple parties” (Haynes 2014).

Yet, it has been recently revealed that Gutierrez made undeclared donations to the Rousseff

campaign by paying over R$10 million ($2.86 million) for opinion polls commissioned by the

Workers’ Party (Update 1 2016). Despite pouring so much money (legally and illegally) into the

left-wing PT and the center PSDB, most of Gutierrez’s donations to district and state level

campaigns and federal deputies went to the PSC on the right to influence elections for the state

of Paraná (which is interestingly not the state of the corporate headquarters—Minas Gerais).

Without insider knowledge, we cannot say with certainty that the intentions of these donations is

not actually to foster democratic debate; however, we can say with certainty that these

corporations have much to gain from having powerful political allies in both the ruling parties

and the opposition.

59
Approximately $7,500,000 USD and $5,300,000 USD respectively, using the exchange rate of August, 2014.
60
Approximately $13,200,000 USD using the exchange rate of August, 2014.

138
Major Legislation Passed

While I tried to pick my cases in a similar manner to the way I picked my Australian

cases, Brazil’s Chamber of Deputies and Senate websites present the information in different

formats and to different degrees. Public opinion polling data was not as readily available. I used

preliminary searches for news articles to see if a significant amount had been written about the

legislation to help me discern which policies passed during the year were important and

controversial. I considered significant an amount that included various different news sources

that went into enough depth as to explain the proposed changes, who was involved and who

would be affected and the presence of blogs and/or international sources covering the topics.

Certain laws would generate a handful of search results, but they all would contain the same or

very similar short content—I judged these laws as too inconsequential to capture the public’s

attention and therefore too inconsequential for interests to diverge enough that a representative

would have to weigh the consequences of losing votes or losing donations. The Brazilian media

has enough capacity and motivation to cover important changes, so I defer to their judgments.

Changes to Unemployment Insurance

While this act did not involve large corporate interests and big campaign donations, it did

involve a policy that affects voters asymmetrically—wealthy individuals may prefer to cut

unemployment insurance programs to save money and cut taxes, whereas lower-income

individuals might prefer to keep or expand it as a safety net. I examine this piece of legislation to

see if the government is more responsive to the wealthy than to the poor.

In December 2013, Brazil’s unemployment rate was at record low of 4.3%. A year later

in 2014, despite minor fluctuations, the rate stayed nearly the same (Ferreira 2016). Hoping to

139
gain control over the country’s growing debt, President Dilma Rousseff passed a temporary

measure (Medida Provisoria 665/14) to cut government spending on unemployment benefits in

December of 2014. These temporary measures come exclusively from the executive branch and

can enter into effect without any participation of the legislature. Unless approved by Congress,

these provisions expire after a certain amount of time. According to the measure, to qualify for

unemployment benefits a recently laid-off worker would have to prove that he or she had worked

18 consecutive months. The previous law only stipulated six months. Given the strong economy

at the time and her recent reelection, Rousseff likely thought of her action as an easy way to cut

government spending without affecting too many people. Unfortunately, beginning in January of

2015, the economy took a significant turn for the worse and unemployment rose steadily.

By mid-June of 2015, unemployment had risen to nearly 7% (Ferreira 2016). Congress

took action. Both houses passed Lei n° 13.134, legislation meant to end Rousseff’s temporary

measure. However, the legislation did not simply nullify Rousseff’s restrictions. It lessened

them, but kept them well above where they had been before December of 2014. The law

stipulated that a worker would have to have worked 12 consecutive months to qualify for the

benefits. The second time a worker applies for the benefits, he or she would have had to have

worked for nine consecutive months; the third time, six consecutive months. The legislation

included a provision for retroactive payments to those who would have qualified for benefits

between February of 2015 until the passage of this legislation but did not because of Rousseff’s

measure (Baldo 2015). The primary purpose of this bill was to save money—Congress hoped to

save R$6.4 billion in 2016 by no longer providing benefits to an estimated 1.6 million workers

(19% of the current total) (Martello 2015). This particular legislation did not change the amount

of benefits any qualifying recipient would receive.

140
 The president of Brazil has the line item veto power subject to an override by Congress.

Rousseff exercised this power to take out two provisions of Lei n° 13.134: 1) a provision that

made access to salary raises more difficult and 2) a provision that made access to these benefits

harder for rural workers than for urban workers (Baldo 2015). Congress did not override either of

these vetoes.

Some public opinion polls found that people were already dissatisfied with the benefits

package before eligibility became even more restricted. A comprehensive study of

unemployment insurance found that 55% of recipients surveyed though the number of benefit

packages was not enough, and 54% thought the amount in each package was not enough. The

same report found that some of the fears that drive welfare reform, such as dependency,

deliberate inactivity and dishonesty, were unfounded—recipients understood that the insurance

was only a short-term solution to help while they look for new employment (Marinho et al.

2010). Another poll by the Brazilian Institute of Public Opinion and Statistics (IBOPE) found

that 38% of respondents thought social security benefits (such as the unemployment benefits)

were discouraging factors when considering whether to start their own business venture (IBOPE

2013). Interpreting these results, over a third of Brazilians would not feel comfortable relying on

the unemployment insurance as a back up plan and therefore would be more reluctant to engage

in entrepreneurship. These results were published in 2013, before access to benefits became even

more restricted. For those who were already skeptical of the safety net before the policy changes,

they would likely feel even more insecure about trying to start their own businesses after the

changes took place.

PSDB, DEM and PPS were the legislation’s main opponents. Rubens Bueno, the leader

of PPS, stated, “We are taking money from the Brazilian worker to cover the fiscal gap of the

141
country” (Passarinho and Alegretti 2015; my translation).61 PMDB and PT led the coalition to

pass the new bill. In the House, the vote was close—252 votes in favor and 227 against. Many

PT voters felt betrayed by their party, calling party leaders traitors62 (54 of the 64 PT Deputies

voted for the bill). In the gallery, unionists from the Força Sindical (A Brazilian trade union)

threw paper bills printed with pictures of PT politicians and the expression “Petro Dólar,”

accusing the party of being bought by the oil industry. They shouted angrily at the PT: “Out,

Workers’ Party” and “The United worker will never be defeated” (Passarinho and Alegreti 2015;

my translations).63 The President of the Chamber, Eduardo Cunha (PMDB) threw the protesters

out and then proceeded with the vote (Passarinho and Alegreti 2015).

Part of the anger towards the PT specifically may have come from the betrayal party

members felt towards the country’s President. In the 2010 election, Rousseff won the vote of the

lower working class people—she received 26% more of the vote of those receiving minimum

wage than her opponent, José Serra of the PSDB. In addition, she won 65% of the black vote and

60% of the mixed-race vote (Estado 2010). Her support was concentrated geographically in the

poorer northern, northeastern and southeastern regions—areas that are likely to benefit more

from more unemployment insurance. Illiteracy is also highest in these regions, especially the

northeast—this is a population for whom voting is still voluntary. Not coincidentally, these are

the regions whose economies are suffering the most in the current recession. Salvador and

Recife, cities in the northeast of Brazil saw unemployment rates up to 12.6% and 10.4%

respectively in February 2016. Rio de Janeiro and Porto Alegre (cities in the South that had

voted for Serra in 2010 and Neves in 2014) had the lowest rates of 5.2% and 6.4%, respectively

61
The actual text of the article reads, “Estamos tirando dinheiro do trabalhador brasileiro para cobrir o rombo fiscal
do país” (Passarinho and Alegretti 2015).
62
The refrain that opponents used was “PT pagou com traição a quem sempre lhe deu a mão,” (Passarinho and
Alegreti 2015) which translates roughly to “the Workers’ Party betrayed those who always gave it support.”
63
“Fora PT,” and “Trabalhador unido jamais sera vencido” (Passarinho and Alegreti 2015).

142
(Ferreira 2016). Rousseff had passed the initial measure making access to unemployment

insurance much more restricted in 2014. Though she stood up for the rural unemployed by

vetoing a section that disadvantaged them more than their urban counterparts, her belated actions

were not enough to placate her party base.

In a country with voluntary voting, one might expect that measures that disadvantage the

poor non-voters would be easier to pass than measures that cut profits for politically engaged

individuals and corporations, perhaps because those affected most do not vote. It seems that in

Brazil, the same holds true despite compulsory voting. Solutions that disadvantage those with the

least economic power are still likely not that politically damaging to the self-interested politician.

This may be true in part because of a particular campaign finance rule—in Brazil,

campaign donations from trade unions to parties or candidates are forbidden. These

organizations were the primary opponents to this type of policy change, yet they were denied a

possible channel of influence. Trade unions may still lobby representatives; however, if a

politician will not lose personal campaign funds nor lose many votes, they may feel more at ease

ignoring lobbyists. Though the Unions did make some noise about this legislation in the

Chamber, their protests were easily ignored. From the Union’s spirited protest, we might think

that the PT will actually lose votes because of this; however, given the many accusations of

corruption, it is impossible to tell if their anger stems uniquely from this legislation or from a

host of different events. Whether the PT were to lose a significant number of votes or not, it

clearly did not care, demonstrating a lack of responsiveness to some of its core voters—in 2014,

those who were less educated (the people more likely affected by this legislation) approved of

the PT’s mandate by double the amount of those who enrolled or completed higher education

(Arantes and Pavanelli 2014). The number is likely to fall and the PT will likely suffer in the

143
upcoming elections. It is impossible to say whether the policy outcome would have been

different had trade unions been allowed to donate to campaigns; however, compulsory voting

certainly was not enough to hold the PT to the interests of a significant chunk of its voter base.

From this example, we can conclude that compulsory voting did not explain the results, whereas

campaign finance laws may have.

Though this policy is not the result of a conflict directly between corporate interests and

labor interests, one could easily imagine a situation where the two would be squarely opposed to

one another. Allowing political donations from corporations but not unions could conceivably

have a real effect despite mandatory voting, though the domination of corporate interests may be

somewhat mitigated by the near universal turnout of the working class (which is far from a

reality in voluntary-voting countries). No labor versus business examples stand out from the

2015 legislative session, though it may be worth examining in another study.64

The larger issue perhaps weakening the effects of compulsory voting is that all affected

voters likely do not know how their representative or party of choice voted. Anger was targeted

at only one party—the PT, which was not alone responsible for the legislation, nor did every

single PT member vote in favor of the bill. Finding the exact roll call votes and party support for

this bill was challenging. No parties seemed to campaign on this reform as a major platform

issue. Even if they had, very few parties voted as cohesive blocks, so individuals would have to

follow their own representative’s performances closely. Even more importantly, a list of those

who specifically voted for and against the bill (aside from a few prominent individuals) did not

appear in any of the news sources reporting on the measure. Given the chaos of the system and

64
While interesting, the practical implications of this study would be limited, as Brazilian campaign finance
regulation has since been modified—beginning the 2016 elections, corporate donations will also be banned, perhaps
placing unions and corporations on more equal footing. A future study might compare the balance of power between
business and labor in influencing policy before and after this change.

144
the already weak system of accountability, if the media is not making it clear who exactly is

responsible for certain measures, we cannot reasonably expect the public to keep track. These

appear to be routine problems in Brazil that stem from the electoral system but are perhaps

perpetuated by the media that fails to convey relevant information to voters.

While this particular legislation only affects a portion of the population, its passage

carries broader implications about the system in practice—universal turnout cannot matter if a

constituency is uninformed about the performance of their representatives. For exactly this

reason, systems of accountability and the media are both crucial to quality representation. We

have already seen how convoluted the system of accountability in Brazil is, and while the media

in Brazil is certainly capable of conducting thorough investigations, they do not seem to focus on

the details of legislative processes. This leaves people unable to make informed decisions about

whether incumbents really are representing their interests and whether they deserve reelection.

The Brazilian public tends to think that all politicians are corrupt (a stereotype perhaps true and

perhaps exacerbated or perpetuated by the media)—maybe public trust in the system would

improve if people knew what their representatives were doing and could elect those who voted

for policies their constituents liked (and perhaps if they were indeed less corrupt).65

65
The day after Congress passed the changes to the unemployment insurance, as part of the same effort to cut
government spending, Congress also passed Lei n° 13.135 to change the qualifications for the death pension (Pensão
Por Morte). Before, there was no minimum period of contribution to the National Social Security Institute or
minimum period of marriage (or civil union) before a widow or widower qualifies for the benefit. Only couples
older than 44 years old are eligible (Baldo 2015).65 Those who are not eligible by those standards will receive a four-
month temporary pension only (Salomão 2015). This legislation affects all Brazilians uniformly. It does not affect
those already receiving the benefits and no person plans for his or her spouse to die—this bill was unlikely to draw
much fierce opposition except possibly from unions, who were focused primarily on the changes in unemployment
insurance object instead of this bill. No major corporations had any direct investment in this issue. While no party
seemed to run explicitly on this reform, most party blocks in the Senate voted in favor. The coalitions lined up
similarly to the law changing unemployment insurance—the PSDB and DEM coalition voted against the measure
(as well as a few other Senators voting out of line with their parties). The legislation had support from both the right
and the left, making it somewhat uncontroversial. I found no public opinion polls on the subject, affirming that this
was of small consequence. News articles did not focus exclusively on this change; instead, they spoke more of the
legislation changing unemployment insurance and briefly mentioned these changes as a secondary concern. Nothing

145
 Social security benefits are redistributive policies—they collect money from all workers

and then give it to those who qualify for the particular program. The tightening of eligibility

requirements, while debatably responsible governance in the long-term, is less redistributive. The

compulsory voting literature has found that in general the policy leads to less inequality through

the passage of more redistributive policies (or that the repeal of compulsory voting precipitates

increased inequality), but here we see an example of pension cuts in a country with compulsory

voting. Nearly universal turnout was not able to prevent retreat from redistributive policies, even

in the absence of corporate interest. While perhaps it prevented more severe cuts, the working

class’s great anger over the policy cuts did not discourage the party (especially the PT) from

betraying one of its voter bases. Brazil is currently suffering from a deep economic recession

marked by high inflation and rising unemployment—the country is in a period where the

government needs to cut back on its spending but individuals also need the most help. In passing

this measure, it is possible that legislators are trying to do what is best for the country as a whole

and acting within their function as trustee-type representatives. It is also possible that votes for or

against this policy are reflections of representatives’ ideologies, and neither money nor votes

could have changed outcomes.

Access to Brazil’s Biogenetic Diversity Act

The next piece of legislation, Lei n° 13.123, illustrates what happens when the interests of

big, moneyed corporations (the pharmaceutical and cosmetic industry) and the poor,

marginalized people of society (the indigenous communities) conflict. According to the Brazilian

Constitution of 1988, the country has the obligation to preserve the integrity and the genetic

about the passage of this legislation reflects either high or low responsiveness to the Brazilian people, any interest
group or any corporation.

146
diversity of the country, which includes between 15-20% of total biodiversity in the world

(Cunha 2015). By law, the Ministry of Science, Technology and Innovation supervises the

pharmaceutical industry and requires authorization to collect materials (Cunha 2015).

Unsupervised, Brazil worries that the industry will overexploit the environment and the

indigenous peoples who live in the Amazon regions. For example, they worry about

deforestation and biopiracy (Cunha 2015). Biopiracy refers to “biological resources and

indigenous knowledge that have been stolen and patented by multinational foreign businesses

without the native communities who created the knowledge sharing in the profits” (Cunha 2015;

my translation).66 In the past, Brazil has seen biopiracy from foreigners who enter under the

guise of religious missionaries, members of NGOS or ecotourists (Cunha 2015).

For many years, the process of obtaining a permit to gain access to these resources has

been inefficient—the Ministry might take two or three years to approve a request. In 2015,

approximately 13,000 patent holders were waiting for access. Some corporations decide that the

wait is not worth it—since 2005, the government has collected R$ 231 million in fines for 581

acts of infringement by both Brazilian and international corporations, including Avon, Natura,

Ambev, Johnson & Johnson, L’Oréal, Unilever, and Pfizer (Schreiber 2015). These companies

think that resources from the Amazon have huge profit potential. In 2013, products using

resources that came from this zone of biodiversity made up less than 2% of the retail market, but

still, herbal medicines alone brought in R$ 964 million in profits (Schreiber 2015).

The new law revises rules regarding research and development using the biogenetic

diversity of the Amazon regions of Brazil by allowing companies to register for authorization

online—a process that will be almost instantaneous (Union for Ethical Biotrade 2015). Though

66
The original text reads: “recursos biológicos e conhecimentos indígenas estavam sendo roubados e patenteados
por empresas multinacionais estrangeiras, sem que as comunidades nativas que geraram os conhecimentos
participassem dos lucros” (Cunha 2015).

147
the government and the private sector are unsure of the amount by which the market will grow

because of the new process, they are very optimistic (Schreiber 2015).

According to the 2010 census, there are 896,917 indigenous peoples in Brazil, slightly

above 0.4% of the total population. 32.4% (290,601 people) are still illiterate (IBGE 2013), and

therefore are not obligated to vote. These communities have certain rights protected by the

Convention on Biological Diversity of 1992 and the Indigenous and Tribal Peoples Convention

of the International Labor Organization of 1989 (Cunha 2015). These international conventions

guarantee that indigenous peoples will have a say in legal changes that affect them; however, the

indigenous organization in Brazil, the National Indian Foundation (FUNAI) claims that they had

no say in this policy change and their rights had been violated and ignored (Cunha 2015;

Schreiber 2015). The agriculture community, environmental community and scientific

community say they too were left out of the conversation. A lawyer from the Environmental

Institute, Maurício Guetta, says they were in favor of changes to the existing regime, but not the

ones that were made: “the drafting of the Bill within the government and its proceedings in the

Congress were evidence of the inequality in the game of political forces,” exemplified by the

“complete exclusion of the traditional people in the debate.” These are people, who, according to

Guetta, “are invisible people… they do not have any kind of political representation, either in the

executive branch or the legislative branch. There are no Indian Deputies” (Schreiber 2015; my

translations).67 The indigenous communities have tried to get support from a few representatives

of the PSOL, PT and PSB parties, but a majority of the Deputies ignored all of their proposed

amendments (Schreiber 2015). The PT has traditionally been the party that celebrates indigenous

67
The original texts read: “a elaboração do Projeto de Lei dentro do governo e sua tramitação no Congresso
evidenciam a desigualdade no jogo de forças político;" “completa exclusão dos povos tradicionais do debate;" “São
povos invisíveis, claro, porque infelizmente não possuem qualquer representação política, seja no Poder Executivo,
seja no Poder Legislativo. Não existe deputado índio” (Schreiber 2015).

148
rights, however, “the PT’s support for the indigenous cause has so far been more rhetorical than

real,” (Millett et al. 2015) as illustrated in this example. Rousseff herself, representing the PT,

was a strong proponent of the legislation, making futile any attempts to convince a majority of

PT members in the legislature to vote against it.

Unsurprisingly, the pharmaceutical industry is a major donor to political campaigns. I

counted nearly 18 different companies that had each donated over R$200,000.00 to various

parties and candidates for the 2014 election, and around another 60 that donated in smaller

quantities.68 Most parties, if not all, got some money from the industry. Some of the largest

donors included Recofarma Industria do Amazonas LTDA, which gave R$1,750,000.00 to

PROS for the governor race of the State of Amazonas, and R$590,000.00 to the PT, PSD, DEM

and SD for Federal Deputy and Senator races. Cristalia Produtos Químicos Farmaceuticos LTDA

donated R$2,000,000.00 to Rousseff’s campaign (PT) and R$75,000.00 to a PT Federal Deputy

campaign. From the dispersed nature of their donations, it is unclear whether the industry as a

whole has an ideological preference. Of course, without an incriminating statement of an

industry official we cannot conclude with certainty that the donations were meant to influence;

however, that most companies donated to various parties often not in coalitions together suggest

suspicious motives. Whether the intent was present or not, the government seemed to be very

responsive to the pharmaceutical industry’s needs while easily brushing aside concerns of the

less powerful.

Despite compulsory voting (and international conventions), the interests of the

indigenous Brazilians remained invisible. Though the communities are small enough that their

votes may not play any role in changing outcomes, that they were so easily excluded from the

68
These campaign finance statistics have been taken from the database found here:
http://inter01.tse.jus.br/spceweb.consulta.receitasdespesas2014/abrirTelaReceitasCandidato.action.

149
conversation suggests that merely bringing people into the electorate does not guarantee that all

have a voice. The government was not only not responsive to the community in determining the

final outcome, but also not responsive to the community throughout the process. This is a

profound exclusion that demonstrates that minority voices can still be ignored, whether they vote

or not. Their votes are expendable because of their small numbers while the money of the

pharmaceutical industry is not. Compulsory voting, in this case, did not counteract the money

bias in Brazil’s government.

None of this is to say that the pharmaceutical industry in Brazil is all-powerful. During

the same session, the legislature passed two different laws increasing regulation. The first law,

Lei n° 13.236, added regulation regarding the labeling of drugs to prevent errors in

administration. This legislation was introduced by Humberto Costa of the PT in the Senate and

signed by Rousseff without any vetoes. While generally supported, it would certainly cost the

industry to re-label all of their products. The second law, Lei n° 13.235, changed the stringency

of generic drug testing. Similarly named products must be tested for effectiveness, safety and

quality (Docpharma 2016). Testing is cumbersome and costly for companies, and something that

pharmaceutical companies might want to avoid. Pharma in Brazil does not seem to always get

that which is easiest and most profitable out of the laws; however, when its interests came in

competition with the indigenous peoples of Brazil, they certainly got their way.

Bills Not Passed

I searched through 120 bills that were proposed in the Chamber of Deputies during the

2015 legislative session. I chose to focus on the legislature’s attempt to regulate the food and

beverage industry because the industry is one of the largest donors to political campaigns and

150
because various politicians from many parties introduced similar legislation, none of which

passed.

Ban on Trans Fat

The Brazilian government is well aware of the risks associated with trans fat. There are

ample scientific studies that link these fats with significant health risks such as high cholesterol

and clogged arteries, both of which can lead to fatal heart attacks and strokes. Processed food

companies use trans fat to give flavor and texture to certain common products, such as

margarine, lasagnas, frozen pizzas and chips (Rosa 2015). Sodium is also a significant health

risk— in excess, sodium can cause hypertension (a problem that affects a quarter of all

Brazilians) and increase the risk of heart attack, kidney problems and stroke. In 2008, research

by the Household Budget Survey (Pesquisa de Oçamentos Familiares) found that the average

consumption of sodium by Brazilians was 12 grams a day, over double the five-gram

recommendation of the World Health Organization (Rosa 2016). In response to this alarming

statistic, the Brazilian government met with the Brazilian Association of Food Industries

(Associação Brasilia das Industrias da Alimentação—Abia). Abia is an entity whose members

make up about 73% of the businesses in the food sector. They came to an agreement, made

entirely behind closed doors, which stipulated that food companies would voluntarily cut the

amount of trans fat and sodium in their products. Companies were to slowly reach this goal by

2020. Even though the agreement contained no enforcement mechanism, the government found

in 2010 that 93.4% of affected products had achieved the stated goals (Rosa 2015).

While this sounds like fantastic news for Brazilian public health, medical experts,

nutritionists and health organizations all agreed that the goals were set much too low. Health

151
experts called the lack of ambition in the goals stipulated by the government and the industry a

serious obstacle to improving public health (Rosa 2015). Furthermore, the voluntary goal only

applied to companies within the Abia consortium. A study done by the Institute for Consumer

Protection (Instituto Brasileiro de Defensa do Consumidor--Idec) in 2014 found that 11% of the

products they analyzed still contained levels above the goal of the dangerous substances.

In June of 2015, the U.S. Food and Drug Administration announced a new ban on trans

fat—processed food industries have three years to stop using the substance in their products.

This bold action sparked new interest in a ban among some Brazilians, especially those in the

Health Ministry (Ministério da Saúde). Most experts tend to agree that there should be some kind

of enforcement mechanism in place (that is, some penalty for non-compliance). Though the

government has not ruled out the possibility of a law, it has held its position that the voluntary

agreements are the best strategy at the moment (Rosa 2015).

Some legislators disagree. First, on June 17th, Jorge Tadeu Mudalen of the DEM party

introduced a bill prohibiting the use of trans fat in products and the marketing of products that

contain trans fat. On June 24th, Goulart of the PSD introduced a bill to limit the quantity of trans

fat in produced food. Rodrigo Maia of the DEM party introduced legislation to prohibit the use

of hydrogenated vegetable fat in food produced for consumption in Brazil on June 23rd. On July

8th, Rogério Peninha Mendonça of the PMDB introduced legislation that would prohibit the use

of hydrogenated fat in processed food products. On July 9th, Veneziano Vital do Rêgo of the

PMDB tried to ban foods that contain trans fat in school meals. Finally, on August 19th,

Pauderny Avelino of the DEM party introduced legislation to prohibit the production and

commercialization of all food products that contain trans fat. None of these bills passed.

152
 On March 28th, 2016 Idec presented research on the subject of trans fat that it had

conducted throughout the previous months. Idec had surveyed 1600 people and found that 95%

agree that trans fat in food should be prohibited. Brazilians understood that the substance was

harmful and could be substituted for other, healthier alternatives (Idec 2016). While this survey

was conducted and presented after the 2015 legislative session, it is unlikely that opinions have

changed drastically on the matter in less than a year.

Even though most Brazilians would likely be supportive of a ban, it seems from my

search of relevant articles that the issue of trans fat in foods is not a main concern of current

Brazilian politics (an unsurprising fact given the state of the economy and the various corruption

scandals). Though many people do seem to care and would prefer the government to ban the

dangerous substances, those who really care can theoretically be selective about their own food

choices—they do not need government laws to protect themselves. In addition, trans fat is less

costly than healthier fats and some may prefer to have the cheaper option available.

Unfortunately, other bills providing for the disclosure of nutritional content in foods were

also archived: João Derly of the PCdoB introduced legislation to require the food industry to

disclose the amounts of ingredients in foods made in the absence of the consumer. João Daniel of

the PT tried to require that the beverage industry put details about the sugar content in sodas and

juices on its products labels. Felipe Bornier of the PSD brought a bill to require labels that

identify the amount of sugar in processed food products. Hiram Gonçalves of the PMN authored

a bill to require food and beverage producers (as well as pharmaceutical companies) to disclose

on the label whether the products contain lactose. Lastly, a bill by Rodrigo Maia (DEM) tried to

prohibit the use of hydrogenated vegetable fat in food marketed and consumed in the country.

153
 These labeling requirements and bans on trans fat do not directly cost the Brazilian

government anything. They could improve public health, and they enjoy some bipartisan support

(as we see by the variety of members from different parties that have introduced similar

legislation). Especially given the apparent overwhelming support for a ban on trans fat, one

might logically ask why it has not yet been passed.

One explanation might be the issues lack of salience, but the main reason likely lies

within the food industry and its money. In addition to the influx of money from the construction

industry during the first month of the 2014 campaign, the food industry contributed staggering

amounts. The top donor in that first month was the JBS group that controls two major food

brands: Friboi and Seara. JBS gave R$93 billion to federal candidates of the PP and PR, and to

Gleisi Hoffman (PT) for a gubernatorial race in the state of Paraná. Hoffman, who was President

Rousseff’s chief of staff from 2011-2014, did not win the governorship; however, she continues

to be a member of the Federal Senate. Hundreds of different corporations in the food industry

donated to the 2014 campaigns. Similar to the construction industries, they donate to parties all

across the spectrum and to candidates on both federal and state levels. In addition to the

impressive amount of money that JBS (the largest food processing company in Brazil) donated,

the second largest Brazilian food processing company, Brasil Foods, donated R$1,500,000.00.

They donated only to Rousseff. The third largest food producer, Margrig, donated

R$1,850,000.00 to various senate and governor campaigns. The company donated significant

amounts to the PT, PMDB, PDT and PTB.

Campaign money is probably not the only money these corporations spend on politics;

they likely lobby a great deal as well. While we cannot say with certainty that it was the

campaign donations that made most of the current Brazilian representatives resistant to place

154
strong restrictions on the industry, we can say that the industry has enough influence to override

a sensible policy with which an overwhelming majority of Brazilians seem to agree. Perhaps the

outcome would have been different if the public had felt more strongly about the regulation or if

interest groups had organized to make the issue more salient to voters and to politicians;

however, it is clear that nearly universal turnout was not enough.

Conclusions

Table 6.5: Changes to Access to Brazil’s Biogenetic Ban on

Summary of Unemployment Diversity Act Trans Fat
Bills, Outcomes, Insurance
Money and
Public Opinion

Money No campaign Yes—from Pharma to all Food and

donations, but parties Beverage
lobbying money Industry to
all parties
Public Opinion Against changes Indigenous Not a salient
Community/Scientific issue, but
Community/Environmental public in
Community against favor of ban
Outcome Changed, access Restrictions Lifted in favor of Not Passed
restricted Pharma, other voices excluded

The results from my case study of Brazil (summarized in Table 6.5) show significantly

different results from those of Australia. In Australia I found that money (through both campaign

donations and lobbying efforts) only seemed to matter in the absence of strong public opinion. In

Brazil, money (or the lack thereof) seemed to prevail in all three cases—changes in

unemployment insurance, pharmaceutical access to Brazil’s biodiversity and the unsuccessful

ban on trans fat. From this, we can conclude that money does appear to matter and compulsory

voting does not in and of itself guarantee quality representation of the average voter. This may be

155
the case because of the lack of clear channels of accountability and lack of knowledge about how

one’s own representatives are performing. Media that does not make clear and accessible role

call votes in a system where party positions do not dictate member positions is an obstacle to

informed voting. A politician may campaign on attractive rhetoric then consistently vote for

policies that do not benefit his or her constituents. If voters are not informed, they may reelect

the representative and perpetuate their poor representation. Both electoral institutions and

organizations tasked with conveying relevant information to voters seem to be at fault.

Money does seem to matter, though we cannot conclude definitively that campaign

finance is the reason that the Brazilian government is more responsive to corporate interests.

Because of the court decision in 2016 banning corporate donations, a future study could look at

legislation, money and public opinion in the same way as I have here to see if policies are

fundamentally different. Were a study to still find responsiveness to corporations over

individuals, we might conclude that perhaps it is lobbying efforts that matter in winning

influence or that influential money is coming in through other illicit channels.

The results of this case study, when compared with those of Australia, are somewhat

counterintuitive. Australia’s government was responsive to the people on issues that the people

cared about despite their by-law weak campaign finance regime. Corporations and wealthy

individuals donate a great deal, but their interests only prevailed in the absence of public

preference. Brazil has many more campaign finance laws on the books, but corporations have

more power. This assessment does not consider Brazil’s low enforcement capability. In practice,

the system may approximate Australia’s nearly lawless approach (aside from disclosure and

transparency). Regardless of whether campaign finance regulation is similar or stronger in reality

than that of Australia, Brazilian representatives are notably more responsive to money, making it

156
clear that compulsory voting is not a universal solution to a money bias in government. This case

study corroborates the placement of Brazil relative to Australia on my six empirical measures.

Whether Brazil’s poor representation is mostly the fault of its electoral system, its mixed

(or lack of) accountability incentives, its campaign finance, its amount and type of lobbying or

its explicit corruption is subject for further consideration. Here, we can only conclude that the

compulsory voting is not a ‘magic bullet’ solution to the problem of low government

responsiveness to its citizens.

157
 Chapter 7: Conclusions

From the result of a large-n statistical analysis, we first saw that neither compulsory

voting, mean turnout, an aggregate campaign finance index nor any individual campaign finance

variables had consistent positive effects on any of the six dependent variables that each contained

components that spoke to quality of representation and government responsiveness to voters.

Though we did see some statistically significant results with regard to the multi-faceted Voice

and Accountability index (such as a positive relationship with average turnout and the interaction

between enforced compulsory voting and regular party reporting/public disclosure/candidate

contribution limits) and a few with Control of Corruption (the interaction between candidate

contribution limits, enforced compulsory voting and mean turnout), these results may have

reflected a relationship with a different component of the index rather than a relationship with

representation generally. No general patterns emerged that would allow us to conclude that any

one variable or interaction of variables influenced representation in a predictable way. Results

were robust, but not satisfactory, especially since each dependent variable only measured

representation indirectly. Trust in politicians was highly statistically significant in most of the

models regarding control of corruption, confidence in parliament and favoritism, perhaps

suggesting that political culture and attitudes are especially important for securing high quality of

representation (or perceived representation).

To examine more closely the policy of compulsory voting, I chose two case studies of

countries with compulsory voting and different levels of campaign finance regulation. For each

country, I first looked at the electoral system, the existing type of accountability, the campaign

finance system and the compulsory voting policy. Then, in a year directly following a national
election, I looked at significant legislation that passed, the related campaign donations and the

public opinion. Summarized results can be found in table 7.1.

Despite Australia’s lax campaign finance regime, big money did not seem to dominate in

policy areas where the public felt strongly, such as the Carbon Tax repeal. Money showed some

influence only in the absence of issue salience among the voters, as with the failure of the

Landholders’ Right to Refuse bill. Previous research on compulsory voting in Australia found

that the implementation of the policy led to higher pension spending at the national level (Fowler

2013). The Fair and Sustainable Pensions Bill, passed in 2014 seems to move in the opposite

direction; however, changes seemed minimal and were unlikely to bring national pension

spending to pre-compulsory voting levels. Fowler’s study looked at the average increase over

time, so this one subtle policy change does not necessarily challenge his findings. This bill

passed despite public opposition; however, no campaign money seemed to be involved. Though

lobbying efforts were likely substantial, it is unclear if and by how much lobbying efforts really

did shape this outcome. We cannot necessarily conclude from this example that Australian

representatives are unresponsive to their constituents because social security programs often

force politicians to carefully consider important tradeoffs: the short-term preferences of

recipients or the long-term sustainability of the system. Lastly, it is also possible that this

outcome simply reflects the party system and/or the ideological preferences of the Coalition

government.

Overall, my findings lead us to believe that Australia has a high quality of representation

and the federal government is in general responsive to the voters’ interests. Taken alone, the case

study of Australia would suggest that campaign money does not seem to overpower public

opinion and compulsory voting does help match outcomes with the general public sentiment.

159
 In Brazil, we see a different result—in two cases corporate interests took precedent (the

Access to Biodiversity Bill and failure of bills attempting to ban trans fat), and in the third case

where corporations had no stake in the matter, the vulnerable still lost by having welfare

programs cut. The case of cuts in the unemployment insurance system may challenge past

research on compulsory voting that claims that the policy reduces inequality (Chong and Olivera

2008; O’Toole and Strobl 1995) or that its repeal increases inequality (Carey and Hoiuchi 2015).

However, these studies measure overall inequality and do not claim that compulsory voting

protects against all reductions in redistribution as long as they keep inequality levels below those

from before compulsory voting. Inequality after cutting unemployment benefits may still be

substantially less than what it would be if everyone did not vote.69 In addition, it is also possible

that without compulsory voting, these cuts might have been larger. We only know for sure that

compulsory voting in Brazil was not enough to prevent this policy change altogether.

In the two cases involving big money, outcomes were consistent with a system that is

highly responsive to donors and big political actors. The pharmaceutical industry, the cosmetic

industry and the food and beverage industry got the policy outcomes they wanted. It is unclear

whether the corporate influence in the system came exclusively from legal campaign donations,

from illegal campaign donations, lobbying efforts, bribery or some combination of the four.

There is no evidence that compulsory voting in Brazil makes a significant difference in

putting the preferences of voters above or on par with those with economic power. This is likely

due to some combination of unique contextual and systematic factors. The Brazilian political

climate, a function of the bizarre electoral system, is chaotic and not conducive to informed voter

choice. Representatives lack important channels of accountability either to the party or the

69
Because compulsory voting in Brazil was timed with its return to democracy, we cannot easily compare today’s
inequality levels with those before compulsory voting and say that the policy itself has made a difference.

160
people. Civil society interest groups are either poorly organized or disadvantaged—for example,

labor unions have been barred from making political donations, yet corporations still could until

the 2016 election cycle. Issue salience may also influence outcomes—a severe economic and

political crisis have likely shifted the priorities and immediate concerns of average Brazilians

away from certain causes (like banning trans fat from food products) and towards others (like

impeaching the president and combating soaring inflation).

Lastly, and perhaps most importantly, the media does not consistently convey important

and relevant information to voters such as individual roll-call votes on bills that would allow

voters to sanction representatives who betray them at the polls. Interestingly, the control variable

that measured genuine media pluralism did not show up statistically significant in any of my

regression models, which suggests that the mere presence of various media outlets is not enough.

The missing component is the effective transmission of information, which is not by itself

guaranteed by the presence of a free and pluralistic press. If a diligent and effective media could

be quantified, perhaps its interaction with compulsory voting would be significant in improving

measures of representation.

It would be wrong to conclude from this case study that compulsory voting can never

improve government responsiveness to the people, especially given that it likely does have a

positive effect in Australia. Instead, we can conclude that the introduction of compulsory voting

is not guaranteed to improve responsiveness everywhere if other contextual factors such as

mechanisms of accountability and a diligent media are not working in a way that allows for the

new voters to make informed choices at the polls. A certain political culture (for example, an

intolerance of corruption) and institutions that foster meaningful accountability may be necessary

conditions for compulsory voting (or even campaign finance regulation) to have a real effect.

161
 Because Australia and Brazil have different qualities of representation despite both

practicing compulsory voting, we cannot conclude that compulsory voting will always improve

representation. The implication of this finding is that imposing mandatory voting on the United

States without changing anything else may not improve quality of representation. Because the

United States operates within a different electoral system with different campaign finance

regulation, has a different enforcement capacity, encourages a different type of accountability,

has different lobbying rules and has a different media, we cannot say with certainty that

compulsory voting would make the government more responsive to the voters. In this way,

President Obama’s suggestion that mandatory voting would counteract money in campaign

finance is not universally supported by the experiences of other countries that have experimented

with the policy.

We can, however, speculate given the institutions and electoral system that do exist in the

United States. To begin, the U.S. is different than both Brazil and Australia in that it elects

representatives through single-member districts in first-past-the-post elections and employs

individual accountability instead of collective accountability. While this encourages two major

(weak) catchall parties with diverse candidates and fractured constituent bases, it also creates a

new mechanism of accountability. The media is not the only political player that may report roll

call votes or campaign finance information; an incumbent’s challenger has a powerful incentive

to disseminate information about how a representative is performing. Because generally two

nominees face off for one seat, using an incumbent’s record to convince the electorate that the

incumbent has performed in a way that his or her constituency does not like will in theory steal

votes away from the incumbent and help put the challenger in office. In contrast, in an open-list

proportional representation election, for example, six out of twenty candidates win a seat. One

162
candidate calling out the record of an incumbent may steal votes away from that one incumbent,

but those votes may not ultimately be distributed to the challenger who did the investigative

work. It is more efficient for challengers to focus on their own attributes and win over voters

than to put their opponents down.

We see this at play frequently in U.S. elections—candidates will call out their opponents

for votes made years ago. In the 2016 Democratic primary race, Senator Bernie Sanders has

often pointed out that former Senator Hillary Clinton voted for the Iraq war in 2003, a conflict

that most Democrats today regard as a disaster. Clinton, in turn, scolds Sanders for voting against

implementing background checks in the Brady Handgun Violence Prevention Act in the 1990s.

In this way, candidates themselves serve as meaningful sources of information about a

representative’s job performance. This makes the job of the media much easier—part of their

investigative work has been done, and they are often happy to report on issues that cause

controversy.

Negative attack ads, made either by party committees, candidate committees or Super

PACs also employ this technique. Attack ads are on the rise—a study found that 70% of

broadcast advertisements between January 1st and April 22nd of 2012 were negative compared to

only 9.1% of broadcast advertisement during the same period during 2008 (Emery Jr. 2012).

Politicians running for office in the U.S. (at least on the national level) cannot escape their record

and voters who care can easily find information that is relevant to them.

If the research that suggests that voters become more informed in compulsory voting

regimes (Lijphart 1997; Birch 2009; Brockington 2005) is true, voluntary non-voters in the U.S.

could easily find information to make educated decisions on Election Day if they were mandated

to vote. The necessary condition of effective dissemination of information to voters is present in

163
the U.S., making our system more similar to that of Australia than to that of Brazil. Accordingly,

compulsory voting has the potential to successfully improve government responsiveness in the

U.S.

Nevertheless, it is not a given that compulsory voting actually would achieve nearly

universal turnout in the United States. Compulsory voting would have to be adopted with a series

of other policies that make voting easier, like automatic registration. If citizens must register

themselves to be effected by compulsory voting, they may simply choose not to do so. To

accommodate the entire population, we might have to extend voting hours, move Election Day to

a weekend day or make Election Day a national holiday. If we do not, we risk placing too large

of a burden on citizens who ultimately may choose to disobey the law and not vote. If voting is

compulsory yet requires a photo ID, some will have to make a rational choice about whether

obtaining a photo ID or not voting is more costly (in money and opportunity cost). They may

choose not to vote. The effects of this may be to make the non-voting population even more

invisible because people will assume that the entire electorate really is voting when in fact they

are not.

In addition, even successful compulsory voting does not increase turnout symmetrically

across all demographic groups. In Australia, we still see young adults choosing to abstain, either

by excuse or risking sanction. In Brazil, the non-monetary sanctions, such as being denied

enrollment to a public university, public employment, passport renewal and certain bank

transactions disproportionately affect middle and upper class voters who generally use these

services more (Cepaluni and Hidalgo 2015). The article suggests that the type of enforcement

mechanism a country adopts matters in who it ultimately turns out. If the U.S. were to adopt a

similar policy to that of Brazil, we might see a similar effect—certain populations that currently

164
exist under the radar, such as the homeless, may still not vote because certain sanctions do not

reach them. The difference is that in a system of compulsory voting, people may assume they are

voting and outcomes do naturally take their interests into account instead of assuming that they

need someone to stand up for their interests.

It is also unclear if compulsory voting would improve outcomes in the U.S. because of

the conditions of campaign finance and the extensive amounts of lobbying that take place in

Washington. Though we do not have the same culture of corruption as Brazil that has led to the

government greatly favoring certain powerful industries, we do have some very large and

powerful interest groups that spend a great deal of money on campaigns and on lobbying efforts

(for example, AARP, the NRA or the NFIB). Part of these interest groups’ power comes from

their ability to mobilize voters—an advantage that would be lost in a compulsory voting regime.

If most of their power truly lies in voter mobilization, then compulsory voting in the U.S. would

serve to mitigate their influence (as long as we have basic disclosure and transparency in election

spending, according to the Australian model). However, if most of their real influence lies in

campaign donations or lobbying expenditures, it is not clear based on either the Brazilian or

Australian example that compulsory voting would neutralize this advantage. Of course, a more

informed electorate (achieved because of compulsory voting) might be aware of this bias and

perhaps less likely to vote for representatives who primarily serve big interests.

President Obama may very well be right that nearly universal turnout would totally

change our political landscape, though we cannot say with absolute certainty that compulsory

voting would improve representation in the United States despite our ineffective campaign

finance system.

165
Table 7.1: Summary of Australia Brazil
Findings
Electoral System Parliamentary System with Presidential with proportional
proportional representation, representation and open lists,
Bicameral legislature, preference bicameral legislature.
voting.
Straightforward and fairly stable Chaotic and volatile
Type of Accountability Collective Accountability No guarantee of accountability
Campaign Finance Lax—only disclosure and Good regulation, but very
transparency. Some donors donate to expensive elections. Also illegal
both sides. donations.
Compulsory Voting Yes, with monetary fine. Many Yes, monetary and non-monetary
young adults still choosing not to fines. Voluntary for illiterate and
vote. for 16-18, 70+ year olds. Sanctions
for non-compliance may
disproportionality turn out middle
and upper class.
Corruption? No major scandals Many scandals, bribery
Media Does not care too much about Given the chaos of the system,
campaign finance but seems to does not necessarily act as
convey information—system not messenger of information voters
overly complicated or hard for need to make informed decisions
voters to follow. at the polls.
To whom is government The voters, when the issue is Corporations that donate, lobby
more responsive? salient. or bribe.

Future Research

To understand the complex relationship between voter turnout, campaign finance and

representation more completely, future research could repeat this method for longer periods than

one year. It may serve to look at legislation and influence over different sessions over which

different parties preside. One might also include more countries, including some that have high

voter turnout and no compulsory voting.

One may look at Brazil’s legislation before the ban on corporate donations and

afterwards to see if the policy changes in any significant way to better reflect the preferences of

166
the people. This would serve to more directly pinpoint from where the influence is truly coming,

be it campaign finance, lobbying efforts or bribery.

I also recommend doing more research on campaign finance systems that give free or

subsidized media to candidates or political parties, as a potential way to drive election costs

down and eliminate the reliance on large campaign contributions. Given that TV ads are a major

source of campaign costs, incentives to channel money to official candidate and party

committees by subsidizing their TV access may make the independent expenditure by Super

PACs economically inefficient and undesirable.

167
 Appendix 1: Changes to the Data Set

I have recoded Egypt as a country without compulsory voting for both the original

variable and my new variable, given that it is currently a military regime and did not hold 2015

elections. Even before the coup in 2013 no sanctions for nonvoting were ever determined by law.

The 2012 election drew out 54.99% of the voting age population; the 2010 election, only 16.16%.

Regardless of Egypt’s compulsory voting category, it will be excluded from my regression

models because it is not a considered a democracy. I have recoded Venezuela as having no

compulsory voting, since it formally abolished the practice in 1993. Italy abolished compulsory

voting in 1993 and saw a slow decline in turnout from the mid 90% range to only 68.33% in the

most recent 2013 election. Chile is still coded in the data set as having sanctioned compulsory

voting; however, lawmakers abolished the practice in 2012 and turnout in the 2013 election was

only 53.66% of the voting age population. Guatemala abolished compulsory voting in 1990,

though it was never enforced nor did turnout change drastically from the elections before 1990

and those afterwards. I will recode all of these countries listed in the downloaded dataset as

having compulsory voting to not having compulsory voting (nor enforced compulsory voting).

Costa Rica has formal compulsory voting but it does not enforce the practice, exhibiting

average voter turnout rates in the mid 60% range. The Dominican Republic does not enforce its

compulsory voting law and has seen between 62.08% turnout at the highest and 30.56% turnout

at the lowest in the last 25 years. In the last decade, Honduras has had two consecutive elections

drawing only slightly over 50% of the voting age population out to the polls, even though

compulsory voting is formally written into law. Turnout was 39.35% in Thailand’s most recent

2014 election with an unusually high percentage of spoiled ballots. I consider these to be

countries with compulsory voting, but I do not consider them countries with enforced
compulsory voting.

Though Singapore enforces compulsory voting by removing non-voters from the voting

registry (making them ineligible to participate in future elections until they give a valid excuse),

this non-monetary sanction has proven relatively toothless and Singapore’s 2015 election had

only 52.13% turnout of the voting age population. Of those registered, however, Singapore has

consistently seen turnout levels of over 90% at every election since 1976. Because there is some

enforcement mechanism in place, I consider Singapore to have enforced compulsory voting for

the purposes of the regression analysis. Greece removed administrative sanctions for failure to

vote in 2000; however, turnout is still above 70% so I consider it part of the group with enforced

compulsory voting.

Nauru and Turkey are coded as not having compulsory voting in the data set, but Nauru

has had the practice since 1965 and has enjoyed close to universal turnout in the past 20 elections.

Turkey also enforces the practice and has had high turnout since compulsory voting was put into

practice first in the 1983 election. I recoded both countries as having compulsory voting and

enforced compulsory voting. Turnout statistics also come from International IDEA.

169
 Appendix 2

Appendix Table 2.1: DV: Confidence in

Explaining Representation, Parliament
excluding Trust in
Politicians Genuine Media Pluralism 0.063
Key IVs: Mean Turnout and (0.19)
Candidate Expenditure Limits Internal Conflict -0.164
(0.80)
Interaction -0.069 Transparency of Econ. Policy -0.374
(2.65)* (0.96)
Mean Turnout 0.040 GDP per capita (log) 0.244
(2.16)* (0.39)
Candidate Expenditure Limits 3.561 Total Population (log) -0.435
(2.07) (2.56)*
Proportional Representation -1.658 Ethnic Fractionalization 0.467
(2.64)* (0.46)
Age of Current Regime -0.000 Religious Fractionalization -1.931
(0.03) (1.98)
Effective Number of Parties 0.206 Federalism Index -0.239
(2.23)* (1.38)
GINI Index -0.026 Globalization Index 0.000
(0.85) (0.01)
Social Trust -0.669 Government Effectiveness 0.711
(0.28) (1.36)
Avg. Population/Rep. 0.000 _cons 12.155
(1.07) (1.82)
Rights of Minorities -0.271 R2 0.83
(0.89) N 43
Participation at the Nat’l Level 0.308
(0.91) * p<0.05; ** p<0.01

Appendix Table 2.1 shows statistical significance for mean turnout and its interaction

with candidate expenditure limits, which alone are not statistically significant. However, this

model does not make much theoretical sense. When expenditure limits are not present, a 1%

increase in average turnout is associated with a 0.040 increase in the confidence in parliament

indicator (25% of a standard deviation); however, when expenditure limits are present, a 1%

increase in average turnout is associated with a decrease in the confidence in parliament indicator

by 0.029 (1.78% of a standard deviation). Because this negative change is small, it might be

explained by a sample bias of the countries that do have both policies—perhaps they introduce

expenditure limits because they had especially low confidence in parliament in the first place.

170
Appendix Table 2.2: DV: Confidence in
Explaining Parliament
Representation, excluding
Trust in Politicians
Key IVs: Enforced
Compulsory Voting and Party
Regulation Genuine Media Pluralism -0.294
(0.75)
Interaction -0.906 Internal Conflict -0.548
(1.92) (1.75)
Enforced Compulsory Voting 5.433 Transparency of Econ. Policy -0.005
(2.24)* (0.01)
Party Regulation Aggregate 0.031 GDP per capita (log) -0.937
(0.29) (1.22)
Proportional Representation -0.668 Total Population (log) -0.563
(0.78) (1.81)
Age of the Current Regime 0.000 Ethnic Fractionalization 0.627
(0.04) (0.40)
Effective Number of Parties 0.291 Religious Fractionalization -0.962
(2.14) (0.89)
GINI Index -0.042 Federalism Index -0.423
(1.19) (1.53)
Social Trust -0.403 Globalization Index -0.007
(0.12) (0.25)
Avg. Population Size/Rep 0.000 Government Effectiveness 1.225
(0.56) (1.92)
Rights of Minorities 0.071 _cons 26.535
(0.17) (2.42)*
Participation at the Nat’l Level 0.467 R2 0.88
(1.08) N 33
* p<0.05; ** p<0.01

This model shows us that only enforced compulsory voting is statistically significant to a

95% confidence level. When enforced compulsory voting is present, it seems to be associated

with an increase in confidence in parliament by 5.43 points, which is a change of more than 3

standard deviations. Given that Brazil (typically on the low end of most indicators) scores a 5.00

for this measure and Australia (typically on the high end of most indicators) scores only a 7.38 (a

difference of only 2.38), this result makes little sense. This model is likely flawed in some

significant way, so we cannot conclude here that enforced compulsory really is associated with

greater confidence in parliament.

 Appendix 3

Table 6.1: Fair and Before Legislation After Legislation

Sustainable Pensions Bill
Changes

Assets test free areas for A$ 202,000 A$ 250,000

single homeowners
Assets test free areas for A$ 286,500 A$ 375,000
homeowner couples
Assets test free areas for non- A$ 348,500 A$ 450,000
homeowner singles
Assets test free areas for non- A$ 433,000 A$ 575,000
homeowner couples
Taper rate for every A$ 1,000 A$ 1.5 A$ 3.00
in assets/fortnight
 Appendix 4

High Speed Rail Planning Authority

Another bill attempted to introduce a High Speed Rail Planning Authority—a

commission that needs to begin planning soon if the project will ever happen (so that new

projects are not constructed on the corridor where the train tracks would go). The high-speed rail

would connect cities along the east coast. It would be one of the “largest infrastructure projects

ever undertaken in Australia,” and Labor calculated that it would cost A$114 billion. The project

could begin in 2027 and would be completed in 2065. The Labor party is a big proponent of the

project—they claim it will help Australia cope with its growing population (allowing for people

to live farther from their work in cities), travel time will be faster than flying (when including the

airport time) and it will be good for the environment. The Climate Change Think Tank also

supports the proposal (Law 2014). There is some disagreement on whether the primary cost of

the endeavor will fall on taxpayers or whether private entities will finance a large part of the

operation. Either way, proponents lament “Toby Abbott’s complete disinterest in rail

infrastructure” (Law 2014). Labor had allocated A$52 million in the budget during its reign in

the previous government, but the Liberal government took it out. The National party actually

supports the proposal, providing for an interesting coalition split (Law 2014).

While I could not find any reputable polling on public opinion of the high speed train, an

online poll attached to an article about the subject on the Newcastle Herald website reveals that

of 1,445 readers, 93.01% are in favor of the project (McGowan 2016). This poll is likely very

biased—those choosing to read an article about the possibility of a high-speed rail are probably

proponents of the legislation. Still, there is clearly a cohort of Australians who feel very

favorably towards the idea.

173
 Because the idea is only in its initial stages, campaign money likely does not play a huge

role. No construction company has yet been commissioned, nor are there any specific interests

who would clearly benefit. The legislation was merely to establish a planning authority. A lack

of public opinion data on the issue also reveals that this has not yet become an important national

issue that people are seriously considering (perhaps because tangible progress is so many years

away). Because of the issue’s lack of salience among donors and among the public, its defeat

was unsurprising and did not preference on population or interest group over another. The bill’s

lack of passage—while a statement about the Liberal party’s stance on the issue—does not say

much about the responsiveness of the government to its constituency.

174
 Bibliography

2013 Annual Report. 2013. Melbourne, Australia: Institute of Public Affairs.

"A Malign Legacy; Proportional Representation Will Return Italy to Political Instability." The
Times, April 12, 2006, 19.

Abdel-Samad, Mounah. 2009. "Exchanging Favours: The Predominance of Casework in

Legislators' Behaviour in Jordan and Lebanon." The Journal of Legislative Studies 15 (4):
420.

Agostini, Claudio A. 2012. Financiamiento De La Política En Chile: Campañas Electorales

2009-2010: Centro de Estudios Públicos; Corporación de Estudios para Latinoamerica.

Aidt, T. S., Jayasri Dutta, and Elena Loukoianova. 2006. "Democracy Comes to Europe:
Franchise Extension and Fiscal Outcomes 1830-1938." European Economic Review 50: 249.

Ames, Barry and Timothy J. Power. 2007. "Parties and Governability in Brazil." In Party
Politics in New Democracies, edited by Paul Webb and Stephen White. Oxford; New York:
Oxford University Press.

Angell, Alan. 2010. "Democratic Governance in Chile." In Democratic Governance in Latin

America, edited by Scott Mainwaring and Timothy Scully. Stanford, California: Stanford
University Press.

Angell, Alan. 2007. "The Durability of the Party System in Chile." In Party Politics in New
Democracies, edited by Paul Webb and Stephen White. Oxford; New York: Oxford
University Press.

Ansolabehere, Stephen, John M. de Figueiredo, and James M. Snyder. 2003. "Why is there so
Little Money in U.S. Politics?" The Journal of Economic Perspectives 17 (1): 105-130.

Arantes, Larissa and Lucas Pavanelli. "Mais Pobres Votam Em Dilma, Mais Ricos Vão De
Aécio." O Tempo, Aug. 13, 2014.
http://www.otempo.com.br/hotsites/elei%C3%A7%C3%B5es-2014/mais-pobres-votam-em-
dilma-mais-ricos-v%C3%A3o-de-a%C3%A9cio-1.898920.

"Australia Votes to Repeal Carbon Tax." BBC News, 17 July 2014.

"Australian Building and Construction Commission." Essential Report, accessed Apr. 10,
2016, http://www.essentialvision.com.au/australian-building-and-construction-commission;.

Baker, Richard and Philip Dorling. "Free Trip Links Rudd to Ousted Taiwanese Leader." The
Sydney Morning Herald, July 24, 2009.
Baldo, Cáren Cecília. "Entenda o Que Muda Com as Novas Regras do Seguro-
Desemprego." Diário Gaúcho, June 17, 2015.

Baquero, Marcello. 2015. "Corruption, Political Culture and Negative Social Capital in
Brazil." Revista Debates 9 (2).

Bartels, Larry M. 2008. Unequal Democracy: The Political Economy of the New Gilded Age.
Princeton, NJ: Russell Sage Foundation.

Bartels, Larry M. 2005. "Homer Gets a Tax Cut: Inequality and Public Policy in the American
Mind." Perspectives on Politics 3 (1): 15-31.

Beck, Katie. "Australia Election: Why is Voting Compulsory?" BBC News, 27 August 2013.

Bennett, Stephen Earl and David Resnick. 1990. "The Implications of Nonvoting for Democracy
in the United States." American Journal of Political Science 34 (3): 771-802.

Bertho, Fabien. 2013. “Presentation of the Institutional Profiles Database 2012 (IPD 2012)”.
Trésor: Direction Générale. Accessed Jan. 20, 2016.
http://www.cepii.fr/institutions/doc/ipd_2012_cahiers-2013-03_en.pdf.

"Better Or Worse Off with Stronger Unions." Essential Report, accessed Apr. 10,
2016, http://www.essentialvision.com.au/better-or-worse-off-with-stronger-unions-4;.

Bevins, Vincent. "Petrobras Corruption Case Overshadows Brazil Presidential Campaign." Los
Angeles Times, 24 Oct. 2014. http://www.latimes.com/world/brazil/la-fg-ff-brazil-
campaign-finance-20141024-story.html.

Birch, Sarah. 2009. Full Participation: A Comparative Study of Compulsory Voting. New York:
United Nations University Press.

Black, Duncan. 1948. "On the Rationale of Group Decision-Making." Journal of Political
Economy 56 (1): 23-34.

Bray, J. Rob and Matthew Gray. 2016. Ageing and Money: Public Opinion on Pensions,
Inheritance, Financial Wellbeing in Retirement and Caring for Older Australians. Canberra,
Australia: Australian National University College of Arts & Social Sciences.

Brennan, Jason. 2014. "Part I. Medicine Worse than Disease? Against Compulsory Voting."
In Compulsory Voting: For and Against, edited by Jason Brennan and Lisa Hill. New York:
Cambridge University Press.

Brennan, Jason and Lisa Hill. 2014. Compulsory Voting: For and Against. New York:
Cambridge University Press.

176
Brockington, David. 2005. "Compulsory Voting and Political Information: A Cross-National
Examination." paper delivered at the annual meeting of the American Political Science
Association, Washington, DC, 1-4 September, 2005.

"Building and Construction Industry (Improving Productivity) Bill Reintroduced." Timebase,

Feb. 3, 2016.

"Building and Construction Industry (Improving Productivity) Bill: Government Resurrects

ABCC." Timebase, Nov. 20, 2013.

Burnham, Walter Dean. 1987. "The Turnout Problem." Elections American Style 97 (98): 109.

Burstein, Paul. 1979. "Some "Necessary Conditions" for Popular Control of Public Policy: A
Critique." Polity 12 (1): 23-37.

Canes-Wrone, Brandice, David W. Brady , and John F. Cogan. 2002. "Out of Step, Out of
Office: Electoral Accountability and House Members' Voting." American Political Science
Review 96 (1): 127-140.

Carey, John and Yusaku Horiuchi. "Compulsory Voting and Income Inequality." Working Paper,
Dartmouth College.

Carey, John M. 2009; 2008. Legislative Voting and Accountability. Cambridge; New York:
Cambridge University Press.

Casas Zamora, Kevin. 2013. Dangerous Liaisons: Organized Crime and Political Finance in
Latin America and Beyond. Washington, D.C: Brookings Institution Press.

Cascione, Silvio. "Major Parties Lose Ground in Brazil's Congress After Election." Reuters, Oct.
6 2014.

Casellas, Jason P. 2009. "The Institutional and Demographic Determinants of Latino

Representation." Legislative Studies Quarterly 34 (3): 399-426.

Cepaluni, Gabriel and F. Daniel Hidalgo. 2015. "Compulsory Voting can Increase Political
Inequality: Evidence from Brazil."

Cheibub, José Antonio, Jennifer Gandhi, and James Raymond Vreeland. 2010. “Democracy and
Dictatorship Revisited, 2010.” Public Choice, vol. 143, no. 2-1, pp. 67-101.

Chong, Alberto and Mauricio Olivera. 2008. "Does Compulsory Voting Help Equalize
Incomes?" Economics and Politics 20 (3): 391-415.

Clark, Peter B. and James Q. Wilson. 1961. "Incentive Systems: A Theory of

Organizations." Administrative Science Quarterly 6 (2): 129-166.

177
"Climate Change and Energy." The Lowy Institute. The Lowy Institute Poll, accessed March 16,
2016, http://www.lowyinstitute.org/lowyinstitutepollinteractive/climate-change-and-
energy/.

"Coal Seam Gas Mining." Essential Report, accessed Apr. 10,

2016, http://www.essentialvision.com.au/coal-seam-gas-mining-2.

"Company Overview of Harold Mitchell Foundation Ltd." Bloomberg Business. Bloomberg

L.P., accessed Mar. 31, 2016,
http://www.bloomberg.com/research/stocks/private/person.asp?personId=8431879&privcap
Id=54078836.

"Company Overview of Paul Ramsay Holding Pty Ltd." Bloomberg Business. Bloomberg L.P.,
accessed Mar. 17, 2016,
http://www.bloomberg.com/research/stocks/private/snapshot.asp?privcapid=32371021.

Compuslory Voting Around the World. 2006: The Electoral Commission.

"Consulta Aos Doadores e Fornecedores De Campanha De Candidatos - v.: 1.6.2

." SPCE WEB: Prestação de Contas Eleitorais. Tribunal Superior Electoral, accessed Apr.
10, 2016,
http://inter01.tse.jus.br/spceweb.consulta.receitasdespesas2014/abrirTelaReceitasCandidato.
action.

Crain, W. Mark. 1995. "The Right Versus the Obligation to Vote: Rejoinder to Yeret, and
O'Toole and Strobl." Economics and Politics 7 (3): 281.

Crain, W. Mark and Mary L. Leonard. 1993. "The Right Versus the Obligation to Vote: Effects
on Cross-Country Government Growth." Economics and Politics 5 (1): 43.

Cunha, Carolina. "Marco Da Biodiversidade: Novas Regras Para a Pesquisa Científica e

Exploração do Patrimônio Genético no Brasil." UOL Vestibular, Mar. 13, 2015.

Dahlberg, Stefan, Sören Holmberg, Bo Rothstein, Felix Hartmann & Richard Svensson. 2015.
The Quality of Government Basic Dataset, version Jan15. University of Gothenburg: The
Quality of Government Institute, http://www.qog.pol.gu.se.

Douglas, Bruce. "Petrobras Investigation: Portuguese Police Arrest Former Executive." The
Guardian, Mar. 21, 2016.

Downs, Anthony. 1957. "An Economic Theory of Political Action in a Democracy." Journal of
Political Economy 65 (2): 135-150.

Drill, Stephen, Ellen Whinnett, and Tom Minear. "Golf Fanatic and Millionaire Businessman
Paul Marks Became the Liberals’ Best Friend." Herald Sun, Mar. 27, 2015.

178
Duverger, Maurice. 1972. "Factors in a Two-Party and Multiparty System." Party Politics and
Pressure Groups: 23-32.

Edwards, Michael. "Mafia in Australia: Lawyers Call for Reforms to Donation Laws After Links
to Politicians Revealed." Abc, June 30, 2015.

"Eleições2012: Estatísticas De Prestação De Contas - Doações." Estatísticas TSE, last modified

Mar 30, 2016, http://www.tse.jus.br/hotSites/estatistica2012/estatistica-prestacao-
contas/doadores.html.

Emery Jr., C. Eugene. "Charting the Increase in Negative Campaign Ads." The Providence
Journal, Aug. 5, 2012.

"Entre Mais Pobres, Dilma Teve 26 Pontos De Folga." Estado, Nov. 7, 2010.

Estabilidade do Emprego e Benefícios São Motivos Para Não Empreender. 2013: IBOPE
inteligência.

Farrell, Paul and Nick Evershed. Feb. 2 2015. "Mystery $850,000 ALP Donor is Linked to
Chinese-Australian Property Group." The Guardian.

Farrell, David M. and Ian McAllister. 2006. Australian Electoral System: Origins, Variations
and Consequences. Sydney: UNSW Press.

Fenno, Richard F. 1978. Home Style: House Members in their Districts. Boston: Little, Brown.

Ferreira, Joana. "Brazil Unemployment Rate 2001-2016." Trading Economics, last modified
Mar. 23, 2016, accessed Apr. 2, 2016,
http://www.tradingeconomics.com/brazil/unemployment-rate.

Fiorina, Morris P. 1974. Representatives, Roll Calls, and Constituencies. Lexington, MA:
Lexington Books.

Fiorina, Morris P. and Samuel J. Abrams. 2009. Disconnect: The Breakdown of Representation
in American Politics. Norman: University of Oklahoma Press.

Flavin, Patrick. 03. "Campaign Finance Laws, Policy Outcomes, and Political Equality in the
American States." Political Research Quarterly 68 (1): 77.

Fleisher, Richard. 1993. "PAC Contributions and Congressional Voting on National

Defense." Legislative Studies Quarterly 18 (3): 391-409.

Fowler, Anthony. 2013. "Electoral and Policy Consequences of Voter Turnout: Evidence from
Compulsory Voting in Australia." Quarterly Journal of Political Science 8 (2): 159.

179
Frendreis, John P. and Richard W. Waterman. 1985. "PAC Contributions and Legislative
Behavior: Senate Voting on Trucking Dergulation." Social Science Quarterly 66 (2): 401-
412.

Fuentes, Claudio. 2011. Financiación De Partidos Políticos En Chile. Stockholm, Sweden:

International IDEA.

Fukuyama, Francis. 2014. Political Order and Political Decay: From the Industrial Revolution
to the Globalization of Democracy. New York: Farrar, Straus and Giroux.

Gage, Nicola. "Sarah Hanson-Young Received a $300,000 Individual Campaign Donation." Abc,
Feb. 2, 2015.

Galston, William A. "Telling Americans to Vote, Or Else." The New York Times, Nov. 6, 2011,
9.

Galvez da Silva, Paulo. "Os 10 Maiores Doadores De Campanha Nas Eleições

2014." Pragmatismo Politico, Sept. 3, 2014.

Gilens, Martin. 2012. Affluence and Influence: Economic Inequality and Political Power in
America. Princeton, NJ: Princeton University Press.

Grant, Richard. 2004. Less Tax Or More Social Spending: Twenty Years of Opinion Polling:
Parliament of Australia.

"Greens Introduce Landholders’ Right to Refuse (Gas and Coal) Bill 2013 into
Senate." Timebase, Dec. 10, 2013, https://www.timebase.com.au/news/2013/AT848-
article.html.

Griffin, John D. and Brian Newman. 2005. "Are Voters Better Represented?" The Journal of
Politics 67 (4): 1206-1227.

Griffiths-Sussems, Susan. "They'Ll Outspend Us - again." The Greens., last modified April 28,
accessed Mar. 17, 2016, http://greens.org.au/magazine/national/theyll-outspend-us-again.

Hanson, Fergus. "If Government Listened it would Cut Carbon Tax." The Australian, 6 June
2012, 14.

Haynes, Brad. "Brazil's Neves Closing Campaign Cash Gap Thanks to Banks and
Ethanol." Reuters, Oct. 16, 2014.

Hill, Lisa and Jonathan Louth. 2004. "Compulsory Voting Laws and Turnout: Efficacy and
Appropriateness." University of Adelaide, 2004.

Hill, Kim Quaile and Jan E. Leighley. 1992. "The Policy Consequences of Class Bias in State
Electorates." American Journal of Political Science 36 (2): 351-365.

180
Hill, Kim Quaile, Jan E. Leighley, and Angela Hinton-Andersson. 1995. "Lower-Class
Mobilization and Policy Linkage in the U.S. States." American Journal of Political
Science 39 (1): 75-86.

Huneeus, Carlos. 2005. "Chile: A System Frozen by Elite Interests." In Electoral System Design:
The New International IDEA Handbook. Stockholm, Sweden: International IDEA.

Hunt, Albert R. "How Record Spending Will Affect 2016 Election." Bloomberg View, April 26,
2015. http://www.bloombergview.com/articles/2015-04-26/how-record-spending-will-
affect-2016-election.

Hunt, Greg and J. B. Hockey. 2014. Clean Energy Legislation (Carbon Tax Repeal) Bill 2014:
Revised Explanatory Memorandum: Parliament of Australia.

"IBGE Mapeia a População Indígena." IBGE. Instituto Brasileiro de Geografia e Estatística, last
modified Apr. 19, 2013, accessed May 1, 2016,
http://saladeimprensa.ibge.gov.br/noticias.html?view=noticia&id=1&idnoticia=2360&busca
=1&t=ibge-mapeia-populacao-indigena.

"International IDEA: Institute for Democracy and Electoral Assistance," accessed October 18,
2015, http://www.idea.int/.

Kaufmann, Daniel and Aart Kraay. 2014. “Worldwide Governance Indicators, 1996-2014.”
World Bank: http://info.worldbank.org/governance/wgi/index.aspx#home.

Key, V. O. 1949. Southern Politics in State and Nation. 1st ed. New York: A. A. Knopf.

Kuziemko, Ilyana, Michael I. Norton, Emmanuel Saez, and Stefanie Stantcheva. 04. "How
Elastic are Preferences for Redistribution? Evidence from Randomized Survey
Experiments." The American Economic Review 105 (4): 1478-1508.

La Ley General De Elecciones, Public Law 14.852, (1962).

Lane, Charles. "Why Americans should Vote Less often." The Washington Post, June 10, 2015.

Law, James. "High Speed Rail: Australia could Build Network for $30 Billion Less, According
to Beyond Zero Emissions." News.Com.Au, Aug. 19, 2014,
http://www.news.com.au/travel/travel-updates/highspeed-rail-australia-could-build-
network-for-30nbspbillion-less-according-to-beyond-zero-emissions/news-
story/fe9708e26caceeb1ed5cf6bb47a0bf65.

Leahy, Joe. "What is the Petrobras Scandal that is Engulfing Brazil." Financial Times, Mar. 31,
2016.

"Lei 13.127 Desobriga a Constitução De Novo CNPJ a Entidades De Autogestão Em Saúde."

Rede Saúde Filantrópica. CMB, last modified June 10, 2015, accessed Apr. 3, 2016,

181
 http://www.cmb.org.br/redesaude/index.php/noticias/586-lei-13-127-desobriga-a-
constituicao-de-novo-cnpj-a-entidades-de-autogestao-em-saude.

Leighley, Jan E. and Jonathan Nagler. 2014. Who Votes Now?: Demographics, Issues, Inequality
and Turnout in the United States. Princeton, New Jersey: Princeton University Press.

Leviston, Zoe, Jennifer Price, Sarah Malkin, and Rod McCrea. 2014. Fourth Annual Survey of
Australian Attitudes to Climate Change: Interim Report. Perth, Australia: Commonwealth
Scientific and Industrial Research Organisation.

Lieberman, Evan S. 2005. "Nested Analysis as a Mixed-Method Strategy for Comparative

Research." American Political Science Review 99 (3): 435-452.

Lijphart, Arend. 1997. "Unequal Participation: Democracy's Unresolved Dilemma." The

American Political Science Review 91 (1): 1-14.

Lijphart, Arend and Carlos H. Waisman. 1996. Institutional Design in New Democracies:
Eastern Europe and Latin America. Boulder, Colo: Westview Press.

Linz, Juan J. and Alfred Stepan. 1996. Problems of Democratic Transition and Consolidation:
Southern Europe, South America, and Post-Communist Europe. Baltimore: The Johns
Hopkins University Press.

Maher, Sid. "Poll Vow to See Pension Reform Put on Hold." The Australian, April 16, 2014.

Mahler, Vincent A. 2008. "Electoral Turnout and Income Redistribution by the State: A Cross-
National Analysis of the Developed Democracies." European Journal of Political
Research 47 (2): 161-183.

Mainwaring, Scott. 1991. "Politicians, Parties, and Electoral Systems: Brazil in Comparative
Perspective." Comparative Politics 24 (1): 21-43.

Marcello, Maria Carolina and Asher Levine. "Brazil's President Rousseff Sees Popularity Drop
to New Low." Reuters, July 1, 2015.

Marinho, Danilo Nolasco C., Moisés Villamil Balestro, and Walter, Maria Inez M. T. 2010.
Políticas Públicas De Emprego no Brasil: Avaliação Externa do Programa Seguro-
Desemprego. Brasília: Ministério do Trabalho e Emprego.

Martello, Alexandro. "Dilma Sanciona Lei Que Altera Regras Seguro-Desemprego Com
Vetos." Globo, June 17, 2015.

Matsusaka, John G. 2010. "Popular Control of Public Policy: A Quantitative

Approach." Quarterly Journal of Political Science 5 (2): 133-167.

182
McCormick, Robert E. and Richard B. McKenzie. 1979. "The Cost of Voting: Its Fiscal Impact
on Government." Public Choice 34 (3/4): 271-284.

McElwee, Sean. "How Unions Boost Democratic Participation." The American Prospect, Sept.
16, 2015a.

McElwee, Sean. "One Big Reason Congress Ignores the Poor: They Don't Vote." Vox, Sep. 24,
2015b. http://www.vox.com/2015/9/24/9388045/2014-voter-turnout-redistribution.

McGowan, Michael. "China-Backed Company Proposes $24 Billion High Speed Rail Network
from Newcastle to Sydney." Newcastle Herald, Jan. 22 2016.

McMenamin, Iain. 2015. "Partisan Interest and Political Finance Reform in Australia." In The
Deregulatory Moment?: A Comparative Perspective on Changing Campaign Finance Laws,
edited by Robert Boatright: University of Michigan Press.

Miguel Wilson, José. "¿Cuánto Cuesta Una Campaña Parlamentaria En Chile?" La Tercera,
October 21, 2013.

Miller, Warren E. and Donald E. Stokes. 1963. "Constituency Influence in Congress." The
American Political Science Review 57 (1): 45-56.

Millett, Richard L., Jennifer S. Holmes, and Orlando J. Pérez, eds. 2015. Latin American
Democracy: Emerging Reality Or Endangered Species?. 2nd ed. New York: Routledge.

Miragliotta, Narelle, Wayne Errington, and Nicholas Barry. 2010. The Australian Political
System in Action. South Melbourne, Vic: Oxford University Press.

"Money, Politics and Transparency," accessed March 11,

2016, http://data.moneypoliticstransparency.org/.

Morrison, Scott. 2015. Social Services Legislation Amendment (Fair and Sustainable Pensions)
Bill 2015: Revised Explanatory Memorandum: Parliament of Australia.

Navia, Patricio and Ignacio Walker. 2010. "Political Institutions, Populism, and Democracy in
Latin America." In Democratic Governance in Latin America, edited by Scott Mainwaring
and Timothy Scully. Stanford, California: Stanford University Press.

Norén, Laura. "Congressional Demographics." The Society Pages, last modified Oct. 10, 2012,
accessed Jan. 28, 2016,
http://thesocietypages.org/graphicsociology/2012/10/10/congressional-demographics/.

Norris, Pippa; Martinez I. Coma, Ferran; Gromping, Max; Nai, Alessandro. 2015, "Perceptions
of Electoral Integrity, (PEI-3.5)", http://dx.doi.org/10.7910/DVN/UO9ABD, Harvard
Dataverse, V3.

183
Norris, Pippa. 2015. “Democracy Cross-national Data.” John F. Kennedy School of Government,
Harvard University: https://sites.google.com/site/pippanorris3/research/data.

Oliver, Alex. 2013. The Lowy Institute Poll 2013: Australia and the World: The Lowy Institute
for International Policy.

Orr, Graeme. 2007. "Political Disclosure Regulation in Australia: Lackadaisical Law." Election
Law Journal 6 (1): 72-88.

O'Toole, Francis and Eric Strobl. 1995. "Compulsory Voting and Government
Spending." Economics and Politics 7 (3): 271.

Passarinho, Nathalia and Laís Alegretti. "Cámara Aprova Texto-Base De MP Que Muda Regras
do Seguro-Desemprego." Globo, May 6, 2015.

Pearson, Kathryn. 2010. "Demographic Change and the Future of Congress." PS: Political
Science and Politics 43 (2): 235-238.

Peltzman, Sam. 1984. "Constituent Interest and Congressional Voting." Journal of Law and
Economics 27 (1): 181-210.

"Pension and Benefits Payments." Essential Report, accessed Apr. 10,

2016, http://www.essentialvision.com.au/pension-and-benefit-payments.

"Pension Assets Test." Essential Report, accessed Apr. 10,

2016, http://www.essentialvision.com.au/pension-assets-test.

Peoples, Clayton D. 2013. "Campaign Finance and Policymaking: PACs, Campaign

Contributions, and Interest Group Influence in Congress." Sociology Compass 7 (11): 900-
913.

Pitkin, Hanna Fenichel. 1967. The Concept of Representation. Berkely, CA: University of
California Press.

Poole, Keith T. and R. Steven Daniels. 1985. "Ideology, Party, and Voting in the U.S. Congress,
1959-1980." The American Political Science Review 79 (2): 373-399.

"Posicionamento do Idec Para a Audiência Pública Da Anvisa Sobre o Uso De Gordura Trans
Artifical Em Alimentos." idec, accessed Apr. 10, 2016,
http://www.idec.org.br/pdf/posicionamento-idec-audiencia-publica-gordura-trans.pdf.

Powell, Lynda W. 2014. Influence of Campaign Contributions in State Legislatures: The Effects
of Institutions and Politics. Ann Arbor: University of Michigan Press.

Power, Timothy J. 1997. The Political Right in Postauthoritarian Brazil: Elites, Institutions, and
Democratization. University Park, Pennsylvania: The Pennsylvania State University Press.

184
Power, Timothy J. and J. Timmons Roberts. 1995. "Compulsory Voting, Invalid Ballots, and
Abstention in Brazil." Political Research Quarterly 48 (4): 795-826.

"Publicada Lei Que Equipara Similares a Medicamentos Genéricos - Lei 13.235/2015."

Docpharma: Assessoria e Consultoria, last modified Jan. 1, 2016, accessed Apr. 3,
2016,http://www.docpharma.com.br/noticia-inteira.php?cod=47.

Putnam, Robert D., Robert Leonardi, and Raffaella Nanetti. 1993. Making Democracy Work:
Civic Traditions in Modern Italy. Princeton, N.J: Princeton University Press.

Putnam, Robert D. 2000. Bowling Alone: The Collapse and Revival of American Community.
New York: Simon & Schuster.

Rehfeld, Andrew. 2009. "Representation Rethought: On Trustees, Delegates, and Gyroscopes in

the Study of Political Representation and Democracy." The American Political Science
Review 103 (2): 214-230.

Research Institute. 2014. Global Wealth Report 2014: Credit Suisse.

Rosa, Guilherme. "A Lenta Batalha do Brasil Contra o Sal e a Gordura." Brasileiros: Informação
na dose certa para viver melhor. Brasileiros, last modified Aug. 21, 2015, accessed Apr. 5,
2016, http://brasileiros.com.br/2015/08/lenta-batalha-contra-o-sal-e-gordura-no-brasil/.

Roscoe, Douglas D. and Shannon Jenkins. 2005. "A Meta-Analysis of Campaign Contributions'
Impact on Roll Call Voting." Social Science Quarterly 86 (1): 52-68.

Saad, Lydia. "Half in U.S. Support Publicly Financed Federal Campaigns." GALLUP, last
modified June 24, 2013, accessed Dec. 6, 2015, http://www.gallup.com/poll/163208/half-
support-publicly-financed-federal-campaigns.aspx.

Salomão, Lucas. "Senado Aprova MP Que Restringe Accesso a Pensão Por Morte." Globo, May
27, 2015.

Sampaio, Inaldo. "Dilma Sanciona, Sem Vetos, Lei De Autoria De Humberto Costa." Inaldo
Sampaio: A Política com "P" Maiusculo., last modified Dec. 30, 2015, accessed Apr. 3,
2016, http://www.inaldosampaio.com.br/2015/12/dilma-sanciona-sem-vetos-lei-de-autoria-
de-humberto-costa-2/.

"Sancionada Lei Que Beneficia Unafisco Saúde." Sindifisco Nacional, last modified May 28,
2015, accessed Apr. 3, 2016, http://www.sindifisconacional-
rj.org.br/index.php?page=vernoticia.php&id=1047.

Sapiro, Virginia, and W. Philips Shively. 2007.“Comparative Study of Electoral Systems:

Module 2, 2001-2006.” Ann Arbor, MI: Inter-university Consortium for Political and Social
Research, http://doi.org/10.3886/ICPSR03808.v2.

185
Schneider, Ronald M. 1991. "Order and Progress": A Political History of Brazil. Boulder:
Westview Press, Inc.

Schönleitner, G. 1998. "Poverty and Inequality in Brazil: Obstacles to Economic Development

and Democratic Consolidation?" Journal Fur Entwicklungspolitik 14 (3): 277-294.

Schreiber, Mariana. "Farmacêuticas Têm Vitória Com Projeto De Lei Polêmico Sobre
Biodiversidade." BBC News, Feb. 11, 2015.

Segal, David. "Petrobras Oil Scandal Leaves Brazilians Lamenting a Lost Dream." The New
York Times, Aug. 7, 2015.

Shah, Paru. 2014. "It Takes a Black Candidate: A Supply-Side Theory of Minority
Representation." Political Research Quarterly 67 (2): 266-279.

"Share of Tax." Essential Report, accessed Apr. 10,

2015, http://www.essentialvision.com.au/share-of-tax.

Smith, Richard A. 1995. "Interest Group Influence in the US Congress." Legislative Studies
Quarterly: 89-139.

Snow, Deborah, Nic Christensen, and John Garnaut. "Man of Property Donates Millions." The
Sydney Morning Herald, July 4, 2009.

Soares, Eduardo. "Brazil: Federal Supreme Court Decision Changes Electoral Law." The Law
Library of Congress. Library of Congress, last modified Nov. 2, 2007, accessed Dec. 10,
2015, http://www.loc.gov/law/foreign-news/article/brazil-federal-supreme-court-decision-
changes-electoral-law/.

"Somalia - UNOSOM II: Background." United Nations, accessed Jan. 19,

2016, http://www.un.org/en/peacekeeping/missions/past/unosom2backgr1.html.

Stadelmann, David, Marco Portmann, and Reiner Eichenberger. 2014. "Full Transparency of
Politicians' Actions does Not Increase the Quality of Political Representation." Journal of
Experimental Political Science 1 (1): 16-23.

Stepan, Alfred C. 1989. Democratizing Brazil: Problems of Transition and Consolidation. New
York: Oxford University Press.

"Summary of Donations Reported by Donors - by Party 2013-2014." Australian Electoral

Commission, last modified June 18 2015, accessed Mar. 17, 2016,
http://periodicdisclosures.aec.gov.au/SummaryDonor.aspx.

Swenson, Katherine M. 2007. "Sticks, Carrots, Donkey Votes, and True Choice: A Rationale for
Abolishing Compulsory Voting in Australia." Minnesota Journal of International Law 16
(2): 525.

186
Talberg, Anita, Kai Swoboda, Juli Tomaras, and Paula Pyburne. 2013. Bills Digest no. 16 2013-
14: Parliament of Australia.

The Economist. 2013. "Where is it Compulsory to Vote?" The Economist, 19 Sep. 2013.

The Electoral Commission. 2006. Compulsory Voting Around the World.

"The Money Behind the Elections." opensecrets.org. Center for Responsive Politics, accessed
Dec. 6, 2015, https://www.opensecrets.org/bigpicture/.

Thomas, Paul E. J., Peter John Loewen, and Michael K. MacKenzie. 2013. "Fair Isn't always
Equal: Constituency Population and the Quality of Representation in Canada."Canadian
Journal of Political Science 46 (2): 273-293.

Tillmann, Philipp Manuel. 2014. "Entry into Electoral Races and the Quality of Representation."
Ph.D., The University of Chicago.

Toni, Ana and Amorim, Alice. "Sustainability in Post Election Brazil: New Conditions, Old
Models?" Nivela, last modified Mar. 4, 2015, accessed May 5, 2016,
http://www.nivela.org/articles/post-election-brazil-new-conditions-old-models/en.

Tucunduva, Paola. "Brasil Impede Crescimento Econômico Com Alta Carga De Impostos
Corporativos." Programa Alma do Negócio. Alma do Negócio, last modified June 3, 2013,
accessed Apr. 3, 2016, http://www.almadonegocio.tv/brasil-impede-crescimento-
economico-com-alta-carga-de-impostos-corporativos.

"Update 1- Brazil Builder made Undeclared Donations to Rousseff Campaign-Paper." Reuters,

Apr. 15, 2016.

"Voter Turnout 1948-2012." The ANES Guide to Public Opinion and Electoral Behavior. The
American National Election Studies, last modified Oct. 26, 2015, accessed Apr. 28,
2016, http://www.electionstudies.org/nesguide/2ndtable/t6a_2_2.htm.

Wawro, Gregory. 2001. "A Panel Probit Analysis of Campaign Contributions and Roll-Call
Votes." American Journal of Political Science 45 (3): 563-579.

Whinnett, Ellen. "Prime Minister Tony Abbott Mate Paul Marks in Chinese Mining
Deal." Herald Sun, Mar. 24, 2015.

World Economic Forum, and Harvard University. 2015. The global competitiveness report.
Geneva: World Economic Forum.

Yan, Holly. "Obama: Maybe it's Time for Mandatory Voting." CNN.Com, March 19, 2015.

Yeret, Eliyahu. 1995. "Dictators, Democracies, and Voters: A Comment on the Right Versus the
Obligation to Vote." Economics and Politics 7 (3): 263.

187
Zingher, J. N. and B. Farrer. 2014. "The Electoral Effects of the Descriptive Representation of
Ethnic Minority Groups in Australia and the UK." Party Politics.

188