Journal of Politics and International Affairs Fall 2015

8/16/2019 Journal of Politics and International Affairs Fall 2015

1/75

Te Journal o Politicsand International AffairsVolume IX, Issue IIFall 2015Te Ohio State University

Wesley SwansonManaging Editor for Content

Gabriel GiddensExternal Communications

Caroline KinnenInternal Communications

Rachel WilliamsLayout Editor

JPIA Editorial Staff

Jeffrey Glazner Alex LappertCara Schaefer Kasey Powers

Marcus AndrewsAndrea StanicKatlyn Perani

Catey StanleyErik Wisniewski

Adrienne MichelsonMegan Simmons

Ryan MooreChris Delaney

Editors in Chief

Cormac Bloomfeld David Nield


2/75

A special thanks to our faculty advisors, Dr. Paul A. Beck and Ms. Christina Murphy, and TheOhio State Department of Political Science Chair, Dr. Richard Herrmann, for guiding us and

making this journal possible.


3/75

The Journal of Politics

JPIAand International Af arsVolume IX | Issue II | Fall 2015 | Print Edition

Contents

The Effect of Party Af liation on Views Regarding a WomanServing as PresidentChristopher L. Pulliam,Syracuse University

The Effects of Water Access On Government Health CareSpending WorldwideDillon R. Roseen, Daniel Furman & Mala Morjaria,Georgia Institute of Technology

Towards the Contradiction of Legal Supremacy in theEuropean UnionCameron Mailhot,University of Minnesota, Twin Cities

French Aid for a French FranceMatthew Cullom & Andrea Moneton,Georgetown University

8

24

40

56


4/75


5/755

Editorial Foreword

Editors in Chief Cormac Bloom eld David Nield

Welcome,

It is our pleasure to present the Fall 2015 edition of the Journal of Politics & International Affairs. This issfeatures four papers chosen from over one hundred submissions from across the nation, making this issue most competitive in the history of the journal. Together, these papers showcase the best of undergraduate political science and economics in a variety of sub elds including: international relations, American politic

comparative politics, political theory, and environmental politics.This is our second volume as editors in chief of the Journal, and we could not be more proud of the progrethat has been made. From doubling of an editorial staff to multiplying submission rates, the Journal hasgrown rapidly in success. Humbled by our success, we are bent on towards even greater achievements andaccomplishments for the Journal going forward, and are excited for the work we have ahead.

This issue would not be in your hands or on your screens if not for the support of the Ohio State UniversityPolitical Science department. We would like to especially thank Dr. Richard Herrmann for his vision and in the Journal, Dr. Paul Beck for his guidance and advice, and Ms. Christina Murphy, whose assistance andsupport is invaluable.

We would also like to thank most importantly you, the reader, for taking the time to read the Journal. We hthat in reading this journal you will learn about a new topic, a new perspective, or simply enjoy fascinatingarticles on a wide-range of topics. Our goal is to provide an appreciation for the quality of social scienceresearch that can be, and is being, done at the undergraduate level nationwide. It is through you that theresearch contained within these pages nds meaning.

Cormac Bloom eld and David Nield

Editors in Chief


6/75

The Journal of Politics & International Affairs

6

The Effect of Party Af liation on Views Regarding aWoman Serving as President

Christopher L. Pulliam1

The 2016 presidential election could potentially result in a woman serving as the President of theUnited States. Women have largely been ignored in presidential politics and much of the publicopinion literature has attempted to answer why. While the contributions of previous scholarsis certainly noteworthy and valuable, past work is outdated and does not directly examine therelationship between party identi cation and views on female candidates. Today’s politicalclimate is uniquely hyper-partisan and it is all but certain that Hillary Clinton, a woman, willbe the presidential nominee for the Democratic Party. This paper addresses the shortcomingsof the previous literature by addressing the relationship between party identi cation and views

regarding having a woman serve as president within today’s hyper-partisan political climate.This paper suggests a model where party culture in uences the views of its members. I hypothe-

size that among individuals in the United States those who identify with the Republican Party aremore likely to have negative feelings towards having a woman serve as president of the U.S. thanthose who identify with the Democratic Party. The ndings of my cross-tabulation con rm myhypothesis and show that people who identify with the Republican Party are more likely to havenegative views regarding a woman serving as president than those who identify with the Demo-cratic Party. This nding remains, even when controlling for gender.

With the 2016 Presidential election less than one year away, excitement is growing aroundwho the potential nominees will be. For the GOP, Donald Trump, Ben Carson, Ted Cruz,Jeb Bush, and Marco Rubio dominate the conversation. Across the aisle, the Democrat contendersinclude Martin O’Malley, Bernie Sanders, and Hillary Clinton, with Quinnipiac University polling

data showing that Clinton is far ahead of the pack (Langer 2015).

Political scientists in the eld of public opinion have been fascinated with the idea of hav-

ing a woman in the Oval Of ce for years. Many scholars have attempted to answer central ques-

tions regarding how public opinion in uences the possibility of having a female president (e.g.Welch and Sigelman 1982; Lawless 2004; Falk and Kenski 2006). Previous literature has largely

1 Christopher Pulliam is a 3rd year undergraduate studying political science and economics at Syra-cuse University. His research interests include public opinion, political behavior, ideology, gender, right-wing politics, and American conservatism. He would like to thank Professor Shana Kushner Gadarian,Dr. Jonathan Hanson, and Professor Spencer Piston for their invaluable input, advice, and guidance.


7/757

Effect of Party Af liation Regarding a Woman President

operated in a political climate where women are almost non-existent in presidential politics,

outside the role of wife and/or First Lady. All forty-four of the United States’ presidents have

been male. Sarah Palin and Geraldine Ferraro are the only two womenhas largely operated in a

political climate where women are almost non-existent in presidential politics, outside the role

of wife and/or First Lady. All forty-four of the United States’ presidents have been male. Sarah

Palin and Geraldine Ferraro are the only two women who have been Vice Presidential nominees,

and there has never been a female presidential nominee for either major party.

Considering the male-dominant presidential political landscape, it is exciting and note-

worthy to see a female candidate at the forefront of the presidential race. With President Barrack

Obama becoming the rst African American president in 2008, and the possibility of Hillary

Clinton becoming the rst female president in 2016, it is clear that the political landscape has

shifted away from one traditional requirement of electability: being a white male.

This paper will provide insight on the relationship between party identi cation and

whether one sees the potential for a female president as positive or sees the potential for a female

president as negative in today’s political landscape.

Women In a Different Political Landscape

The question of what in uences public opinion towards having a woman as president has been

pondered by political scientists for decades. Some suggest that opinion changes when going be-

tween different demographic groups (Welch and Sigelman 1982; Fox and Smith 1998; Sigelman

and Welch 1984), while others suggest that issue salience may play an important role (Falk and

Kenski 2006). Still others suggest that gender stereotypes play a signi cant role (Sanbomatsu

and Dolan 2009; Lawless 2004; Alexander and Anderson 1993). Debate exists whether polling

data is inherently unreliable due to respondents not being truthful and the potential aw of the

woman-president question or if traditional polling actually overestimates sexism.

Many political scientists have suggested that support for women changes between dif-

ferent demographic groups. Welch and Sigelman found that the core support for a woman to

become president was found among people who had more years of education, were younger, and

attended church less. The core opposition for a woman to become president was found among

people who had less education, were older, and attended church more often. These results sug-


8/75


8

gest that conservatives may be more likely to disapprove of a female president than liberals; yet,

the data did not strongly support this conclusion (Welch and Sigelman 1982).

Similarly, a 1998 study suggests that differences in support for a female candidate may be

the result of regional and cultural differences. Fox and Smith polled students at the University of

Wyoming and University California Santa Barbara. They found that despite similar demograph-

ics in the cases (young and well-educated people), the group in California was more likely to

approve of a female presidential candidate (Fox and Smith 1998).

Other explanations have involved differences among gender and race. A 1984 study

conducted by Sigelman and Welch found gender may not in uence an individual’s preference

for a female candidate. Women were found to not be consistently more supportive of female

candidates than men. However, among African Americans, women were more likely than men

to support a female candidate. Few gender differences were found among Caucasians (Sigelman

and Welch 1984).

Despite results that show demographic groups have different degrees of approval for a

potential female presidential candidate, others suggest that issue salience may be an important

factor. A 2006 study showed that individuals who said that terrorism, homeland security, and/or

U.S. involvement in Iraq was the most important issue facing the United States were more likely

to feel that a male would do a better job handling the problem as president. The conclusion of

Falk and Kenski was that issue saliency affects gender preferences in presidential candidates

much more than demographic and party identi cation variables (Falk and Kenski 2006).

Several scholars have suggested that holding gender stereotypes play an important role

with varying degrees of effect. A 2009 study claims that the public perceives gender differences

within both parties. They further state that a candidate’s gender plays a role even when the party

of the candidate is known. Additionally, stereotypes involving gender have different implications

for Democratic women and Republican women (Sanbomatsu and Dolan 2009).

Lawless asserts that women have a harder time wooing public opinion during times of

war (i.e. after September 11, 2001). This research challenges the claim that winning elections

have nothing to do with sex based off of the possibility that women fare poorer in elections when

“men’s issues” are dominant (Lawless 2004).

In contrast, a 1993 study that takes a more information-centric approach nds that when


9/759


information about candidates is sparse gender role attitudes are important in initial evaluation of

candidates. Speci cally, people tend to attribute certain leadership abilities and skills in handling

problems based on sex when considering a hypothetical candidate. Further, voters who have

equal gender role attitudes are more likely to favorably evaluate an actual female candidate. Fi-

nally, all incumbents were rated better on “male” and “female” qualities than challengers (Alex-

ander and Andersen 1993).

Outside of possible explanations for why women are almost non-existent in presidential

politics, much of the literature has examined measurement concerns. Some scholars question

how effective polling can be when asking about sensitive issues, especially those regarding race

and gender. Streb, Burrell, Frederick, and Genovese suggest that people may give false answers

when asked if they would vote for a quali ed female presidential candidate. A list experiment

was implemented to see if people were lying on polling questions. This experiment involved

giving the control group a list of four statements. They were asked to identify how many of the

statements made them “angry or upset.” The four statements read: “1. The way gasoline prices

keep going up, 2. Professional athletes getting million-dollar plus salaries, 3. Requiring seatbelts

to be used when driving, and 4. Large corporations polluting the environment.” The test group

was given these four statements with the addition of a fth statement and were asked to do the

same thing. The fth statement read: “A woman serving as president.” The key to this exper -

iment is that respondents had to identify how many statements made them “angry or upset” not

which ones. Since respondents did not have to identify which statements they picked, the re-

searchers saw this method as a way to evaluate the public’s true feeling about a female president

without the potential for lying. They found that despite the growing trend of support for a female

president over the years, about 25% of the American public would be angry about a woman serv-

ing as president. This result was constant among demographic groups (Streb et al. 2008).

Similar to Streb et al., a 1998 study suggests that the woman-president question is in-

herently awed. The researchers state that recent studies at the time found that the candidate’s

gender did not have an effect on voter discrimination based on sex. The study involved polling

students at two different universities (the details of this study were stated on page two of this

paper). The results suggest that bias against women candidates change signi cantly among

regions. The researchers conclude that a woman-president question may be awed and should


10/75


10

not be taken at face value because people may be unaware of their prejudices or are willing to lie

about them (Fox and Smith 1998).

In contrast, Falk and Kenski suggest that traditional polling methods actually overesti-

mate sexism and underestimate the role that individual characteristics play, such as party iden-

ti cation, in deciding if one will vote for a woman for president. They further claim that many

people would in fact vote for a woman if given a realistic matchup (Falk and Kenski 2006).

A problem that is shared with all past literature is that we are now in a political climate

where the possibility of having a woman president is very likely, if not inevitable. Hillary Clin-

ton is the front-runner for the Democratic nomination, and the Republicans have not yet found

a candidate who is guaranteed to challenge her properly. In the past, asking individuals about

having a woman serve as president have been largely hypothetical. Now that the potential for a

woman to sit in the Oval Of ce seems very possible, the American public’s attitudes about this

question may have changed signi cantly.

Additionally, much of the present literature interacts with the relationship between party

identi cation and views on having a woman serve as president; however, none directly explore

this relationship. When the new hyper-partisan political climate is considered, this omission is

even more important to examine. This paper is an attempt to properly examine the relationship

between party identi cation and views regarding having a woman serve as president in the mod-

ern political climate where the potential for having a female presidential candidate is very real

and strong partisanship divides the country.

Party Culture In uences Views on Women

When looking at how the American public views the possibility of having a female president, it

is natural to consider divisions along party lines. Due to the hyper-partisanship present in today’s

political spectrum, differences between Republicans and Democrats are expected with almost ev-

erything. In examining why there may be differences between the two major parties in feelings

towards a female president, it is necessary to consider who makes up each party, how this make-

up could in uence party culture, and who is seen as a presidential front runner for either party.

Republicans tend to be made up of conservatives, including traditionalist conservatives.

An emphasis of traditionalist conservatism is the preservation of traditional home life. This view


11/75


12/75


12

would have more favorable views in regard to having a woman president. This concrete policy

position may in uence the thinking of party members to be pro-women in focus, and thus, more

positive to the idea of having a woman serve as president.

Perhaps most importantly, party identi cation may in uence individuals’ views on having

a woman president due to the fact that the Republican eld of potential presidential candidates is

almost all male, while the top Democratic contender is female. Respondents may remember this

fact in the back of their minds, speci cally that Hillary Clinton will most likely be the Democrat-

ic nominee. Considering this partisan spilt, respondents who identify as Republican may rate

having a female president lower than respondents who identify as a Democrat.

Essentially, my model suggests that individuals are drawn to a political party for a reason

or a set of reasons. Once one identi es with that party, they are in uenced by the culture of that

party. This party culture could be the result of the membership of the party, the policy positions

of that party, and who makes up the “big names” of the party. The culture of the party in uences

and may change the views of the individual. A concrete example of the above will be given for

clarity.

Consider Tom. Tom has very favorable views towards free-market economics. Tom sees

the Republican Party as being favorable towards free-markets and decides to become a member

of the GOP. Now that he identi es as Republican, Tom seeks out what other Republicans think.

He reads blogs, online magazines, and news articles that explain and defend the views of various

wings of the GOP. He keeps up to date on various policy issues and which side the Republican

Party takes. He is aware of current polls and who is being talked about for the GOP presidential

nomination. When faced with the question of if he would vote for a female presidential candi-

date, Tom may recall the piece he read that eloquently defended traditionalist conservatism. He

may also recall how Republicans are against the Paycheck Fairness Act. Finally, he may recall

that potential presidential candidates for the GOP include almost exclusively males and that the

leading Democratic candidate is Hillary Clinton, a female. Tom may want to be like his fellow

Republicans— a legitimate member of the group. So, Tom considers the above and determines

that a true Republican would not vote for a female presidential candidate, or be neutral to the

idea.

Therefore, in order to more effectively understand how the U.S. population views the


13/7513


possibility of having a woman president, party identi cation needs to be considered. When the

above are put into consideration, it seems that party culture may have a signi cant in uence

on views regarding a woman serving as president. A comparison of the two major parties also

suggests that Republicans will have a more negative view than Democrats. The hypothesis for

this paper is that among individuals in the United States, those who identify with the Republican

Party are more likely to have negative feelings towards having a woman serve as president of the

U.S., than those who identify with the Democratic Party.

Research Design

In order to test this hypothesis, the variables need to be measured properly. The independent

variable is party af liation. The dependent variable is feelings towards having a woman serve as

president of the United States. The unit of analysis is individuals.

Data from the American National Election Study (ANES) of 2012 will be used to measure the

independent and dependent variables. This study will be used because it is the most recent study

available, and it focuses on investigating national election outcomes. Given that the president is

chosen through a national election, the ANES is ideal.

As mentioned earlier, there is debate in the literature regarding the use of traditional poll-

ing methods when inquiring about certain issues, in this case gender. Considering that one camp

thinks traditional polling overestimates negative views towards women, while the other camp

thinks traditional polling underestimates negative views towards women, this paper will assume

that the ANES is a good way to measure the attitudes of voters. While using the ANES is cer-

tainly not perfect, it has been long held that traditional polling methods are good measures of

respondents’ feelings. Since the goal of this paper is not to resolve the above debate, I will yield

to the wisdom of using traditional polling methods.

The variable “WPRES_GDBD” will be used to measure individuals’ feelings towards

having a woman serve as president. The survey question asks, “Would it be good, bad, or neither

good nor bad if the United States has a woman President in the next 20 years?” The answers

were recorded in three categories: “Good if the United States has a woman President”, “Bad if

the United States has a woman President”, and “Neither good nor bad if the United States has a

woman President.” This question reliably measures the dependent variable because it consistent-


14/75


14

ly measures respondents’ feelings about having a female president. This question is also valid

because it contains no systematic error. When examining the question at face value, asking if

having a woman president in 20 years is good, bad, or neither should be a valid measure of re-

spondents’ feelings towards having a woman serve as president.

The variable “PID_SELF” will be used to measure individuals’ party identi cation. The

survey question asks, “Generally speaking, do you usually think of yourself as a [Democrat, a

Republican, / a Republican, a Democrat], and independent, or what?” The answers were record-

ed in three categories: “Democrat”, “Republican”, and “Independent.” The question is a reliable

measure of the dependent variable because it measures individuals’ party identi cation consis-

tently. The question is also valid because it does not suffer from systematic error. At face val-

ue, asking someone how they identify themselves in terms of political parties should be a valid

measure of respondents’ party af liation.

I will be performing a cross-tabulation between the independent and the dependent vari-

ables. Considering that the independent variable is nominal and the dependent variable is ordi-

nal, a cross-tabulation would be an appropriate tool.

In addition to nding the zero-order relationship between the independent and dependent

variables, I will also be nding a partial-effect relationship by controlling for gender. Con-

trolling for gender is important because women may show a more favorable bias towards having

a female president than men. Also, women tend to identify with Democrats more than men do,

and men tend to identify with Republicans more than women do.

I will be using the variable “GENDER_RESPONDENT” to measure the control variable.

The variable combines the variables for gender face-to-face and web modes of administering the

survey. The question may run into reliability issues because the interviewer observes the individ-

ual’s gender for administering the survey face-to-face. Considering that the survey was admin-

istered by several different interviewers, their categorization of individuals may differ slightly.

Additionally, this question runs into issues of validity. The question relies on the interviewer

correctly identifying if the respondent is male or female. The interviewer may identify individ-

uals incorrectly. Despite these potential issues, it is still unlikely that the dataset is signi cantly

inaccurate. Since there are no other gender variables in the ANES, this variable is the best option

by default.


15/7515


Analysis and Assessment

The ndings of the cross-tabulations support my hypothesis by showing that individuals in the

United States who identify with the Republican Party are more likely to have negative feelings

towards having a woman serve as president than those who identify with the Democratic Party.

The charts below display the ndings of this study. A majority of Democrats (56.2%) see

having a woman serve as president in the next 20 years as good, while only 23.8% of Republi-

cans feel the same way. Additionally, 9.6% of Republicans feel that having a woman serve as

president is bad, while only 3.6% of Democrats feel the same way. Finally, 66.7% of Republi-

cans say that having a woman serve as president in the next 20 years as neither good nor bad,

which is much more in comparison to only 40.2% of Democrats.

Chart 1

Feelings On Having a Woman PresidentWomen Presi-dent in next 20

years?

Democrat Republican Independent Total

Good 56.2%(1226)

23.8%(306)

37.6%(642)

42%(2174)

Bad 3.6%

(79)

9.6%

(123)

4.5%

(77)

5.4%

(279)

Neither goodnor bad

40.2%(878)

66.7%(858)

57.9%(989)

52.6%(2725)

Total 100%(2183)

100%(1287)

100%(1708)

100%(5178)

These three ndings are signi cant at the 0.05 level. Pearson’s chi-square value, for four

degrees of freedom is 391.739. The signi cance value is 0.000, which passes the 0.05 thresholdof signi cance. The null hypothesis is rejected with 95% con dence. Signi cant differences

exist between how Democrats and Republicans feel about having a woman serve as president

in the next 20 years. Democrats tend to have more favorable feelings towards having a woman

president than Republicans. Republicans tend to have more negative or neutral feelings towards


16/75


16

having a woman president than Democrats.

Further, this study nds that 49.7% of male Democrats and 18.6% of male Republicans

think that having a woman president in the next 20 years would be good. Also, 3.6% of male

Democrats and 9.1% of male Republicans say that having a woman president in the next 20

years would be bad. Of male Democrats, 46.7% feel that having a woman president in the next

20 years would be neither good nor bad, while 72.3% of male republicans feel the same way.

Chart 2Feelings on Having a Woman President

(Controlling for Gender)

Male Female

WomanPresi-

dent innext 20years?

Demo-crat

Republi-can

Indepen-dent

Total Demo-crat

Republi-can

Indepen-dent

Total

Good 49.7%(461)

18.6%(124)

32.1%(302)

35%(887)

60.9%(765)

29.4%(182)

44.3%(340)

Bad 3.6%(33)

9.1%(61)

5.1%(48)

5.6%(142)

3.7%(46)

10%(62)

3.8%(29)

5.2%(137)

Neithergood

nor bad

46.7%(433)

72.3%(482)

62.8%(590)

35.4%(445)

60.6%(376)

52%(399)

Total 100%(927)

100%(667)

100%(940)

100%(1256)

100%(620)

100%(768)

The above ndings are signi cant at the 0.05 level. Pearson’s chi-square value for males

is 178.755 with four degrees of freedom. The signi cance value is 0.000, which passes the 0.05

threshold of signi cance. The null hypothesis is rejected with 95% con dence. Signi cant

differences exist between how male Democrats and male Republicans feel about having a wom-


17/7517


an serve as president in the next 20 years. Male Democrats tend to have more positive views

towards having a female president than male Republicans. Male Republicans tend to have more

negative or neutral views towards having a female president than male Democrats.

Among women, 60.9% of Democrats feel that having a female president in the next 20

years would be good, while only 29.4% of Republicans feel the same. Further, 3.7% of female

Democrats and 10% of female Republicans think that having a woman president in the next 20

years would be bad. Finally, 35.4% of female Democrats say that having a woman president in

the next 20 years would be neither good nor bad, which is much less than the 60.6% of Republi-

can women who say the same thing.

The above ndings are signi cant at the 0.05 level. Pearson’s chi-square value for fe-

males, with four degrees of freedom, is 190.637. The signi cance value is 0.000, which passes

the 0.05 signi cance threshold. With 95% con dence, the null hypothesis is rejected. Signi-

cant differences exist between how female Democrats and female Republicans feel about having

a woman serve as president in the next 20 years. Female Democrats tend to have more positive

feelings towards seeing a woman in the Oval Of ce than female Republicans. Female Republi-

cans tend to have more negative or neutral feelings towards seeing a woman in the Oval Of ce

than female Democrats.

When comparing males and females who feel that having a woman serve as president is

good, there is an additive relationship. The change between the percentage of male Democrats

who see it as good and the percentage of male Republicans who see it as good is -31.1%. The

change between the percentage of female Democrats who see it as good and the percentage of

female Republicans who see it as good is -31.5%. These values are not close to zero and yet are

close together, indicating a non-spurious, additive relationship.

Further, when comparing males and females who feel that having a woman serve as pres-

ident is bad, there is also an additive relationship. The change between the percentage of male

Democrats and male Republicans who see it as bad is 5.5%. The change between the percentage

of female Democrats and male Democrats who see it as bad is 6.3%. Again, these values are not

close to zero and yet are close together, indicating a non-spurious, additive relationship.

In this comparison of males and females who feel that having a female serve as president

is neither good nor bad, the relationship is once again additive. The change between the per-


18/75


19/7519


woman serve as president would be good, there should be few issues of gender bias in the pri-

maries. So far, the race seems to re ect the data of this study. Clinton is considered the far and

away favorite. She leads Senator Sanders and Governor O’Malley in the polls and in the en-

dorsement race by large margins (Langer 2015, Bycoffe 2015).

Once the primaries are over, the story only changes slightly. The research shows that

the total number of individuals who have negative feelings towards having a woman in the Oval

Of ce is 5.4%. Assuming that all respondents have been honest, the level for possible gender

bias among voters is minimal. However, if the 2016 election is similar to recent presidential

elections, even the slightest disadvantage could decide the election. There is a very good chance

that this slight gender bias could be highlighted because she will most likely be running against

a male candidate since the frontrunners in the GOP nomination race are all men (Langer 2015).

Even then, the level of gender bias may be minimal because many individuals who see having a

woman serve as president as a negative would end up voting for the Republican candidate re-

gardless. More work needs to be done after the election in order to properly determine if gender

bias hurt Clinton. Similar work has been done on how President Obama won in 2008 despite

racial prejudice (Piston 2010). This same kind of research should be done in the future in order

to determine how gender bias in uenced Americans’ vote choice in 2016.

Assuming that the results of this study are true, it seems that the United States could have

its rst female president. This suggestion is not surprising considering that the U.S. elected its

rst African American president in 2008, and again in 2012. While issues of racism and sexism

still exist, it seems that they do not signi cantly hinder the presidential election process. Fortu-

nately, the 2016 election will come down to the quality of the candidates, not whether they are

male or female.


20/75


21/7521



22/75


22

The Effects of Water Access On Government Health CareSpending Worldwide

Dillon R. Roseen1

, Daniel Furman2

& Mala Morjaria3

Water access worldwide affects numerous areas of personal health. This paper attempts to an-alyze the effects that water access has on total government health care expenditure in differentcountries. By utilizing a cross-sectional analysis from over 80 countries, we provide a uniqueview of how lack of water access may burden health care. This paper grounds its hypothesis in

Pareto economic welfare policy, which helps explain the relationship between these variables. In addition to water access, we utilized additional independent variables such as private healthcare spending and tobacco use and a dummy variable adjusting for differences in developedand non-developed countries. These variables enabled us to gain a holistic view of the effects of

water access on health care spending.

According to a report from the World Health Organization and UNICEF released in 2014,around 748 million people worldwide lack access to clean, drinkable water (UNICEF2014). 3.4 million people die every year from preventable water-related diseases, such as diarrhea

(Prüss-Üstün 2008), of which 90% can be attributed to unsafe drinking water. Diarrhea was the

second leading cause of death in the world in 2012 among children under ve, causing at least one

child death every minute (World Health Organization 2014). Evidence for the causal relationship between unsafe drinking water and diarrheal deaths is overwhelming (Fewtrell 2004). Water borne

illnesses and deaths disproportionately affect people in poor regions of the world who lack access

to proper health care facilities. In developing countries, access to health care services is especially

scarce given the low number of health workers in proportion to the total population (Peters 2008).

This gap in service raises health care costs, which greatly affects people of lower economic status

1 Dillon Roseen earned his B.S. in Economics & International Affairs from Georgia Tech in De-

cember 2015. His research interests include international distributive justice, development studies, andlaw. After a few more adventures, he intends to start work in the national security sector in Washington,D.C.2 Daniel Furman is a fourth-year Business Administration student at Georgia Tech. His research in-terests include nancial econometrics, revenue management, and social welfare. Daniel is set to graduatein May 2016. Upon graduation, he will be joining The Boston Consulting Group in their Atlanta of ce.3 Mala Morjaria graduated from Georgia Tech in December 2015 with her B.S. in Industrial andSystems Engineering. Her research interests include sports economics and econometrics.Faculty guidance was provided by Dr. Shatakshee Dhongde.


23/7523

Effects of Water Access on Government Health Care Spending

who must pay a larger share of their total income to receive proper care. The WHO reports that

approximately 100 million people every year fall into poverty as a result of the costs associat-

ed with illness (Etienne 2010). A large share of the health care burden in developing countries

comes as a result of completely preventable diseases, such as those caused by unsanitary water.

If health care facilities in developing countries are freed from the burden of treating preventable

water-related diseases, resources could be redirected to provide better quality and more effective

care. Therefore, our analysis seeks to identify the relationship between water access and govern-

mental health care expenditure.

Hypothesis

Our main hypothesis states that if a country has better access to improved potable drinking

sources4, then that country will spend less money on health-related costs5 as a percentage of total

government expenditures. Our hypothesis is supported by the rationale that as fewer individuals

become sick from water-borne illnesses, there will be less of a nancial burden on the health-re-

lated spending of a speci c country. This hypothesis corresponds to the foundation of the social

welfare economic theory. The social welfare economic theory analyzes the relationship between

national policies on the overall well being of a community.

Following Adam Smith’s teachings of the Invisible Hand, social welfare economics

states that the allocation of sums tends to bene t the whole community. Our hypothesis also

makes use of Pareto’s concept of social welfare economics. Pareto found that there is a point

(the Pareto Optimal/Ef cient point) where afterward, any gains in social welfare ef ciencies are

contrasted by subsequent losses by another party (University of Toronto 2014). This can be seen

by the large differences in health care systems and water infrastructure between developed and

developing countries, which corresponds with the points before and after the Pareto Ef ciency

Point. Our hypothesis concludes that successfully improving water access in both developed and

developing countries will result in a reduction of health care spending.

4 Drinking water is water used for domestic purposes, drinking, cooking and personal hygiene(WHO).Safe drinking water comes from a household connection; public standpipe; borehole; protecteddug well; protected spring; rainwater and is free from any microbial, chemical and physical characteristicsthat meet WHO guidelines or national standards on drinking water quality (WHO).5 Total expenditure on health (THE) is measured as the sum of all nancing agents managing fundsto purchase health goods and services.


24/75


24

Literature Review

The existing literature suggests that there is, in fact, a global water crisis, particularly in devel-

oping nations (Kirpich 2014, Frederiksen 2005, Walter 2014). One millennium development

goal (MDG) is to halve the proportion of people in the world without access to safe drinking

water. While this goal was met in 2010 (World Health Organization 2014), a crisis remains in

some countries. In many developing nations, a vast number of people still live without access to

improved water sources.

The seminal work in the area of water infrastructure and health comes from Annette

Prüss-Üstün, Robert Bos, Fiona Gore, and Jamie Bartram of the World Health Organization. In

their work titled “Safer Water, Better Health” published in 2008, they concluded that one tenth

of the global disease burden could be “prevented by improving water supply, sanitation, hygiene

and management of water resources” (Prüss-Üstün 2008). The study was initiated as a progress

report on the MDG to improve environmental sustainability, speci cally Target 10, which is

described above.

A different study performed by Yongsi (2010) nds that in Cameroon a statistically high

number of cases of diarrhea were caused by poor water quality. The overall burden of these

diseases led to an increase in the burden on health care providers in Cameroon. The results of

this study can be applied to many other developing nations. The causes of variability in diarrheal

levels between neighborhoods in Cameroon are particularly interesting. While Yongsi’s report

illustrated the correlation between water quality and the health care burden, it failed to explain

regional variations in diarrheal cases throughout Cameroon. Once these differences are iden-

ti ed, policy makers can learn how to intervene in communities to lessen the total burden on

health care.

Furthermore, as countries spend less on health care they can redirect governmental re-

sources to improve infrastructure to help prevent water-borne illnesses. We can better see the ef-

fects of proper water interventions by looking at Eder, Schooley, Fullerton, and Murguia’s (2012)

research on post-project impacts in Bolivia. Their study looked at the water intervention process

by analyzing the $26 million Development Assistance Program (DAP) that supported water

access projects. After six years, DAP communities had better infrastructure for community water

systems and household water systems than control communities. Also, DAP sponsored commu-


25/7525


nities were around “30% more likely to be rated good to very good or satisfactory for status of

water infrastructure” (Eder 2012). The communities were able to improve their health and living

standards by investing in water infrastructure, thus reducing the overall health burden.

Other existing sources provided us with good modeling techniques to frame our research.

Dinar and Saleth’s (2005) study introduced an interesting modeling technique in which countries

were grouped according to their level of development to help identify the most common con-

straints on water systems. They nd that these constraints stem from poor resource management

and inef cient water use rather than a physical limitation on water access. Therefore, they stress

that improving management infrastructure can solve the water crisis. Dinar and Saleth chose to

use the Water Institution Health Index indicators to evaluate and compare the relative health of

water institutions. In their study, 43 selected countries were arranged “into three broad groups

with countries having good, moderate, or poor performing water institutions” (Dinar 2005). This

grouping helped identify the weaker performing water institutions that need the most help to

better allocate resources and perfect inef ciencies. Our analysis will adopt a similarly grouped

modeling technique in order to better identify how countries in various stages of development

respond when there is improved water access.

Our study brings a unique perspective to prior literature and hopefully can serve as a sup-

plement for past work. We focus on a cross sectional analysis of countries around the world to

get a holistic perspective on health care spending. Currently, there is little analysis on the direct

effects of improved water access on government health expenditures in a country. Many studies

analyze the microeconomic impact of poor water access on an individual or household’s health

care costs. Macroeconomic studies on water access, however, tend to be limited to speci c re-

gions or countries and fail to link water access to a country’s health care spending. The research

in this paper attempts to provide a broad level, holistic perspective that other research lacks.

Data

This study is primarily interested in the impact a country’s access to improved water sources has

on its general government health care expenditure. Our dependent variable in this model is gen-

eral government expenditure on health (GGHE). To measure GGHE, we looked at a country’s to-

tal government health expenditure as a percentage of its general government expenditure (GGE).


26/75


26

We also incorporated several independent variables in our study. Our primary indepen-

dent variable is the proportion of the population using improved water sources as a percentage

of the total population. Access to an improved water source is the primary independent variable.

If our hypothesis proves to be true, a country with better access to improved water sources will

have a signi cantly lower GGHE.

Table 1.1 De ning independent and dependent variable

L ABELS V ARIABLES D ATA S OURCE

water Proportion of the population using Improved WaterSources (% of total pop)

United Nations StatisticsDivision

publichealth General Government Expenditure on Health (% of

Total Government Expenditure)

The World Health Organi-

zation privatehealth Private Share of Total Health Expenditure (%) The World Health Organi-zation

tobacco Smoking Adults (% of pop over age 15) The World Health Organi-zation

mortality Infant Mortality Rate (per 1,000 births) UNICEFdeveloping Dummy variable signifying if a country is develop-

ing (1) or developed (0)Worldbank

Variables Included in Simple Regression:publichealth= β + β water

1. Improved Water Sources - The percentage of the total population who use any of the

following types of water supply for drinking: piped water into dwelling, plot or yard;

public tap/standpipe; borehole/tube well; protected dug well; protected spring; rain-

water collection and bottled water (if a secondary available source is also improved).

It does not include unprotected well, unprotected spring, water provided by carts with

small tanks/drums, tanker truck-provided water and bottled water (if secondary sourceis not an improved source) or surface water taken directly from rivers, ponds, streams,

lakes, dams, or irrigation channels. This is the standard de nition provided by the

United Nations site for MDG indicators.

2. General Government Health Expenditure - GGHE as % of GGE, provided by the

World Health Organization Global Health Expenditure Database.


27/7527


Additional Variables Included in Multivariate Regression:

publichealth= β + β water + β 2privatehealth + β 3mortality + β tobacco

1. Private Share of Total Health Expenditure - Private share of health spending as given

as a percentage of total health expenditure (THE)6. The WHO Global Health Expendi-

ture Database provided this data set. We included this variable because countries with

a larger share of THE funded by private entities (such as individuals, NGOs, etc.)

would have a lower percentage of THE come from a public source.

2. Infant Mortality Rate – As de ned by UNICEF,the probability that a child born in a

speci c year will die before reaching the age of one, expressed as a rate per 1,000 live

births. UNICEF also provided the data for this variable. Infant mortality could affect

THE, which is why we included it in our regression. In a country with a high infant

mortality rate, public health spending would likely be low because there is inadequate

access to proper medical services.

3. Smoking Adults - Prevalence of current tobacco use of both sexes aged 15 and over,

provided by the World Health Organization Statistical Information System. We felt it

was important to include this variable in our research because of the large effect that

tobacco use has on health. For instance, the Centers for Disease Control and Preven-

tion estimate that cigarette smoking causes 1 in every 5 deaths in the United States

(CDC 2014).

Taken as a collective, these independent variables provide the foundation upon which we

constructed our statistical model. Ideally, these variables will give us a clear picture regarding the

factors that in uence total public health spending. For the purpose of the research, we often uti-

lized data compiled by Gapminder but originally available from sources like the WHO, UNICEF,

and the UN Statistics Division.

The addition of a dummy variable,developing, allowed us to control for different types

of countries: both developed and developing nations. Appendix 1.1 shows a breakdown of the

developed nations and developing nations used in this study. Splitting these countries up allowed

us to control for the large differences in health care systems and water infrastructure between

6 Total Health Expenditures includes both public and private sources


28/75


28

developed and non-developed countries. Developed nations tend to be richer than non-developed

nations and therefore spend more money on health care systems. Additionally, water infrastruc-

ture in developed nations far surpasses that in the developing world. The dummy variable gave

us deeper insights into the realities for both types of countries.

Table 2.1 Descriptive Statistics: Breakdown of variables based on 81 surveyed countries’ data.

Data sources are noted above. Includes data from both developed and developing nations.V ARIABLE O BSERVATIONS M EAN S TD . D EV M IN M AX

publichealth (% of GGE) 81 11.14 4.55 .97 21.25water (% of total pop) 81 87.56 16.14 40 100 privatehealth (% of THE) 81 42.28 19.77 .70 89.97

mortality (per 1,000 births) 81 29.46 27.65 1.9 104.5tobacco (% of tobacco users) 81 24.38 10.51 5.5 51.8developing 81 0.38 0.49 0 1

Table 2.2 Descriptive Statistics: Developed Countries OnlyV ARIABLE O BSERVATIONS M EAN S TD . D EV M IN M AX

publichealth (% of GGE) 50 12.62 4.46 4.22 21.25water (% of total pop) 50 96.8 5.18 79 100 privatehealth (% of THE) 50 33.40 15.51 .70 79.71

mortality (per 1,000 births) 50 11.90 10.72 1.9 45.3

tobacco (% of tobacco users) 50 25.51 10.84 5.5 51.8developing 50 0 0 0 0

Table 2.3 Descriptive Statistics: Developing Countries OnlyV ARIABLE O BSERVATIONS M EAN S TD . D EV M IN M AX

publichealth (% of GGE) 31 8.74 3.63 .97 14.20

water (% of total pop) 31 72.65 16.69 40 98 privatehealth (% of THE) 31 56.61 17.47 23.76 89.97

mortality (per 1,000 births) 31 57.77 22.58 16.1 104.5

tobacco (% of tobacco users) 31 22.57 9.83 6.60 42.30developing 31 1 0 1 1


29/7529


Figure 1.1 Water Sources and Government Health Expenditures for 81 countries

Gauss Markov Assumptions:

It is just as important to interpret the results of data as it is to verify the reliability of data. The

data in the study was veri ed using the Gauss Markov Assumptions for Multivariate Regression.

Assumption 1: Linear Parameters

We checked that our regressions were linear in parameters, meaning our model could be written:

Y= β + β X1 + β X2 + β X3 + β X4+ β k Xk + u

Assumption 2: Random Sampling

The data we used included every country that had information for our desired variables. The

equation for a random sample of the population can be shown as:

Yi= β + β Xi 1 + β X i2 + β X i3 + β X i4+ β k X ik + u i

Assumption 3: Zero Conditional Mean

Based on our regression equation, the residuals for all our countries were plotted in the histogram

below. While Figure 2.1 shows a slight rightward skew in our data, it is not enough to violate the

normality assumption.


30/75


31/7531


Results

Table 4.1 Simple Regression: STATA Results, water regressed on dependent variable publichealthI NDEPENDENT V ARIABLE S IMPLE R EGRESSION

water .101***(3.42)

Intercept 2.26(0.86)

Number of Observations 81R-Squared .1292

* denotes signi cance at the 10% level ** denotes signi cance at the 5% level *** denotes signi cance

at the 1% level

In the initial simple regression, the relationship between improved water access and public health

care spending was positively correlated for the 81 surveyed countries. From an economic stand-

point, the coef cient of .101 shows that there is an increase in the value of publichealth by.101

when water increases by one unit. Because this coef cient is signi cant at the 1% level there is

a statistically signi cant correlation between access to improved water sources and public health

care spending and the correlation is economically viable. However, this relationship violates our

hypothesis. Instead of a negative correlation like we initially predicted, there was a positive cor-

relation between improved access to potable water sources and public health care spending. Put

simply, this means that, as a whole, countries with better access to water spend more money on

health care. Our initial explanation for this result is that countries with better access to water are

typically richer. Presumably, rich countries would also spend more public money on health care,

rather than relying on outside, private sources for support. Since this regression includes both

developed and non-developed nations, future regressions that split developed and non-developed

nations may have different results. We ran an additional regression using dummy variables in

Table 6.1 to test this hypothesis. First, though, we looked at a multivariate regression without the

dummy variables to serve as a baseline for the dummy variable regression.


32/75


32

Table 5.1 Multiple Regression: STATA Results without Dummy VariablesI NDEPENDENT V ARIABLES U NRESTRICTED M ULTIPLE R EGRES -

SION

R ESTRICTED M ULTIPLE R EGRESSION

water .023(.64)

privatehealth -.165***(-7.44)

-.172***(-9.20)

mortality -.0001(.04)

tobacco -.077**(-2.15)

-.079**(-2.25)

Intercept 17.86***(4.33)

20.36***(14.92)

Number of Observations 81 81

R-Squared .526 .520

When water is included in a multivariate regression with the other independent variables,

it is no longer statistically signi cant. Instead,tobacco becomes signi cant at the 5% level and

privatehealth becomes signi cant at the 1% level. This result does not t our initial hypothe-

sis. However, the resulting signi cance of privatehealth would logically have a strong effect on

public expenditures when looking at the data as a random sample of 81 countries. Our ndings

suggested that water access is not the most integral component of GGHE. However, this makessense logically as other factors in our analysis overwhelmed the effects ofwater . Certain vari-

ables, like privatehealth and tobacco , were signi cant in both the unrestricted and restricted re-

gressions. The signi cance of these variables could have easily disguised the subtle yet important

effects thatwater has on GGHE.

Furthermore, our variablemortality is not individually signi cant. An F-test was conduct-

ed to determine ifmortality and water were jointly signi cant. The F-statistic of .44 was signi-

cantly lower than the critical value at a 5% signi cance level. Thus, we failed to reject the nullhypothesis, which indicates the two variables are not jointly signi cant.

When thinking logically aboutwater and GGHE, it is important to look at the countries

being analyzed. For this reason, we attempted to split the group of countries into developed and

non-developed countries as de ned by the World Bank.7 We split our sample of 81 countries

7 Developing countries are classi ed as countries with a Gross National Income per capita per year of


33/7533


from the multivariate regression into smaller samples of developed and non-developed countries

as indicated in Appendix 1.1 . To perform a better regression using a dummy variable, we added

the development dummy variable and labeled itdeveloping . Developing countries were marked

(1) and developed countries were marked (0). We also added the slope interaction dummy vari-

able, d 1, to analyze howdeveloping interacts with the variablewater. In this new regression, we

once again foundmortality to be insigni cant and therefore dropped it from our multivariate

regression model.

Table 6.1 Multiple Regression: with Slope Changing Dummy VariablesI NDEPENDENT V ARIABLES M ODEL

Developing 24.98***(2.88)

Water .262***(3.05)d1 (water*developing) -.263***

(-2.82)Privatehealth -.160***

(-7.21)Tobacco -.083**

(-2.46)Intercept -5.25

(-0.62) Number of Observations 81R-Squared .573

The regression with a slope changing dummy variable onwater has quite a different

result than the one with no distinction between the development statuses of countries. For this

model, we foundwater to be statistically signi cant. The effect the variable has on GGHE can

be described by saying there is a (.262 - .263 * developing) unit change in publichealth , when

water increases by 1 unit. That means as a developed countries’ improved water access increas-es by 1%, GGHE (as a proportion of GGE) also increases by .262%. However, as developing

countries’ improved water access increases by 1%, GGHE (as a proportion of GGE) decreases

by only.001%. While this result does not seem like a substantial effect in developing countries,

the remainder of the effect is captured by a shift in the intercept caused by the dummy variable

less than 11,905 USD.


34/75


34

developing. For developing countries, the intercept is shifted upward 24.98 units, showcasing

the difference between developing and developed countries. Both the water and dummy variable

coef cients are signi cant at the 1% level. This statistical signi cance shows us that there is a

correlation between access to improved water sources and public health care spending when a

country’s development status is taken into account. Furthermore, our multivariate regression an-

alyzes other variables, such as private health care spending and tobacco use. These variables are

also statistically signi cantly at a one and ve percent level, respectively.

Conclusions

Put simply, access to improved water sources in developing countries does affect the amount of

public health spending. As water access increases, government expenditure on health decreases.

We veri ed our hypothesis, but only in instances where we looked at developing countries. In

developed nations, the amount of water access also affects public health spending. However, the

relationship is positively correlated, negating our hypothesis. Overall, the impact of water access

on public health spending is rather minimal. It is apparent that there are other, more in uential

factors that determine the level of government expenditure on health.

It is important to note that we also performed an analysis on other variables, such as

GDP, mortality, the number of physicians in a country, private health care spending, and tobacco

use. Our analysis on GDP showed that there was a small, positive correlation between GDP and

health care spending. This correlation was statistically insigni cant. Subsequently, we recognized

the importance of analyzing the number of physicians in a country. However, the data available

would have cut our sample size greatly. As noted above, mortality was also removed from our

regression due to lack of signi cance. Private health care spending and tobacco use showed sig-

ni cance in our model. Therefore, these variables also affect GGHE.

The analysis of water access on GGHE in this study showed us two distinct scenarios.

One in which a developing nation with better water access spent less public money on health.

Another scenario showed a developed nation with better water access and more government

expenditure on health. The evidence demonstrated throughout should be used to guide research

in developing nations on the improvement of water access in order to better the nations overall

health and thus lower GGHE.


35/7535


Appendix

Appendix 1.1 List of Developing CountriesArmeniaBangladeshBoliviaBurkina FasoCambodiaCameroonCentral African RepublicCongo, Dem. Rep.Cote d’IvoireDjiboutiEthiopia

GambiaGuineaGuinea-BissauHaitiIndonesiaKenyaKyrgyz RepublicLaoLesothoLiberia

MalawiMaliMauritaniaMoldovaMozambiqueMyanmar Niger PakistanPapua New GuineaPhilippines

Appendix 1.2 List of Developed CountriesAlgeria IrelandAndorra IsraelAustralia ItalyAzerbaijan JamaicaBarbados JapanBelize KazakhstanBosnia & Herzegovina KuwaitBotswana LatviaBrazil LebanonCanada LuxembourgChile MalaysiaChina Maldives

Colombia MaltaCroatia Marshall IslandsCyprus MauritiusCzech Republic MontenegroDenmark NamibiaFiji NetherlandsGabon New ZealandGermany NiueGreece NorwayHungary Oman

Iceland RussiaIran South KoreaIraq St. Lucia


36/75


37/7537



38/75


38

Towards the Contradictionof Legal Supremacy in the European Union

Cameron Mailhot 1

Whether in terms of trade, rights, security, or employment, member states of the European Union(EU) have grown increasingly uni ed and interdependent over the years. However, when ex -amining these developments through the lens of the normative and practical values of domesticconstitutional supremacy, a theoretical and legal paradox emerges. In this paper I conduct ananalysis of independently collected legal sources and responses in order to map not only thiscontradiction’s position in the EU but also the ways in which a subset of member states of the

EU have addressed it. In combining an examination of the constitutional structures, positions,and elite responses in three EU member states—Germany, the United Kingdom (UK), and Ro-

mania—, I conclude that there has emerged a general framework of action by which these stateshave attempted to approach this inconsistency by conducting themselves within the boundariesof constitutionality and rule of law so as to t a “legally square” peg into a “normatively round”hole. These ndings not only provide context for understanding the ways in which member statesof the EU have balanced a loyalty to constitutional supremacy with a membership in an increas-ingly supranational order but also lend empirical support for the observed normative value inthe supremacy of national constitutions. Moreover, this paper presents a broader understandingof the ways in which a set of countries have addressed con icts between normative beliefs andconcrete actions in a globalizing world, and it presents a possible qualitative framework for sim-ilar analyses in the EU and beyond.

Nearly 70 years ago, then-Foreign-Minister Robert Schuman of France presented to his West-ern European counterparts a game-changing proposal: the establishment of a supranationalinstitution aimed at reducing interstate threats in Europe by engaging in collective action around

the coal and steel markets. The newly formed European Coal and Steel Community (ECSC) initi-

ated the idea and framework of a united Europe. Though initially serving as a communal approach

to a common concern, over the following half century, the work of this body and its subsequent

institutions culminated in the movement of some of the most politically, militaristically, and eco-

1 Cameron Mailhot is a fourth-year undergraduate studying political science and German (minor)at the University of Minnesota, Twin Cities. His major and minor elds of study are international relationsand comparative politics, respectively, with particular research interests in political conceptualizations ofhuman rights, global norms and governance, inter- and trans- national organizations, and European Union politics. He would like to thank the political science department at the University of Minnesota for theirteaching and support.


39/7539

Towards the Contridiction of Legal Supremacy in the EU

nomically aggressive states toward novel forms of cooperation never before seen on the war-torn

continent (“The History of the European Union”).

Forty-two years later in 1992, the 12 states of the European Communities, which devel-

oped from the ECSC, entered into an historic agreement, the Treaty of Maastricht, to form the

EU (European Community Members). The EU has, since its creation, expanded its direct vertical

in uence into many more area of its member states’ domestic politics and economies. Not only

does it have a say in policies regarding migration, rights, and employment, but internal and ex-

ternal demands for harmonization and integration have led to increasingly binding supranational

Union law and court rulings, oftentimes requiring member states to implement or submit to deci-

sions not necessarily supported by their politicians, of cials, or their public (“EU Law Guide”).

Consequently, a noticeable transition towards legal supremacy of Union law has developed.

A robust discourse has emerged around the ways in which the EU’s legal system relates

to its member states. While some have discussed the positive implications of a supranational or

pluralist style of governing (e.g., Barber 2006, Douglas-Scott 2012, Pliakos 2007), others have

argued against it (e.g., Barents 2009, Borowski 2011). Meanwhile, various scholars have been

building a literature on the causes and effects of a loyalty stretched to two different levels of

government (Cruz 2008, Klamert 2014; Kwiecień 2005, Pernice 2002, Stiernstrom 2005), while

others have dedicated their attention to questions regarding past, present, and future complexities

in establishing a supranational constitutional structure for the Union (Alter 2003, Kumm 2005,

Stone Sweet 2009). Nevertheless, there still exist important considerations pertaining not only

to the recognition of multiple supreme legal systems and to the subsequent discord but also to a

break from norms and standards of legal ordering.

As it appears, this paper will focus on the ways in which member states of the EU have

addressed the theoretical and practical contradiction of constitutional supremacy and the supra-

national nature of the EU. To evaluate this situation, I examine three member states of the EU:

Germany, the United Kingdom, and Romania. While its constitution does not explicitly create a

contradiction of supremacy, Germany is nevertheless a founding member of the ECSC and has

developed a strong legal system based on a written constitution and a sense of rule of law, repre-

senting, in this regard, a model example for analysis. On the other hand, the UK operates under


40/75


40

a longstanding “small-c” constitution2 and therefore provides a unique context for addressing the

con icting notions of the sovereignty of Parliament and the emerging legal supremacy of the EU.

In contrast, contemporary Romania is a relatively new country, and, much like other post-com-

munist states in the EU, it dealt with the duality of seeking EU membership while simultaneously

engaging in state-building. Consequently, it is a timely example of the shared condition faced by

many Eastern European countries still seeking EU membership.

I will begin my analysis with a brief examination of the theoretical contradictions that

have arisen between nations’ constitutional systems and the supranational legal supremacy of the

EU. I will then transition into a discussion of a set of particularly pertinent events in each of the

three member states to highlight the ways in which they have approached this paradox. Through

this process, I conclude that member states havenot , in fact, departed from a sense of loyalty to

the supremacy of national constitutions, as multiple scholars have considered in their discussions

on the EU (see, e.g., Bellamy & Castiglione 1997; Elazar 1998; Moravcsik 2001; Tsebelis &

Garrett 2001). Instead, as I will contend, the norm of constitutional supremacy remains, and the

member states under consideration have taken legal and theoretical steps to maintain this value

while meeting their obligations as members of the EU.

The Emergence of a Contradiction

Within a society ordered under a constitutional system, determining ultimate legal hierarchy isa

priori a relatively simple task. When concerns for legality arise, constitutions generally provide

not only the framework for sorting out the issue but also the ground on which such judgments

are legitimized (Kay 2011, 715). However, an increase in the number and in uential authority of

international institutions in the past half-century has shone light on a newly emerging phenom-

enon: explicit legal pluralism cutting across domestic governments and international politics.

Whether in the domain of economics, war, or sovereignty, states have begun to transfer authority

and decision-making powers to inter- and supranational institutions (“International Organiza-

tions”). The EU is an extreme example of such a condition, but it nonetheless provides a platform

on which discussion of this emerging trend can develop. To fully understand the ways through

2 Small-c” constitutions, as de ned here, signify the intersubjective “body of rules, practices,[norms, customs,] and understandings, written or unwritten, that…[shapes] who holds what kind of pow-er, under what conditions, and subject to what limits” in a given territory (Law 2010, 377).


41/7541


which member states interact with this issue, it is necessary to have a basic understanding of the

current situation.

At its commencement, the ECSC, the rst predecessor to the EU, was nothing more than

a weak institution tasked with both distributing and monitoring resources for rebuilding Western

Europe. Requirements within the ECSC, such as noti cation of the High Authority for any viola-

tions of free competition, were set in an air of goodwill, binding members insofar as they freely

remained party to the community’s treaty. The treaty established a court for adjudication, but its

role was restricted to clari cation and interpretation of the treaties, and individual member states

retained great in uence and autonomy in appointment procedures (European Commission, c).

The member states nevertheless moved forward in their coordination, expanding the

domain into energy (European Atomic Energy Community) and trade and investment (the Eu-

ropean Economic Community). This was met by a corresponding transfer of legal supremacy to

international (secondary) lawmaking.3 Quali ed Majority Voting procedure4 expanded into new

sectors, such as immigration, security, and foreign policy, and with the rati cation of the Treaty

of Amsterdam of 1997, the European Court of Justice gained even more authority to produce

legally-binding rulings when EU law was in con ict with domestic law. The result has effec-

tively been a reduction in the legal legitimacy of the legislation, constitutions, and courts under

which member states govern themselves (European Commission, a; European Commission, b;

European Commission, d). While the large portion of legislation in the EU is passed without a

direct contradiction to member states’ domestic laws, there nevertheless remain theoretical and

empirical instances of supranational requests for the change of national law, as highlighted below

(Barker 2001, 42). As a result, it has become more possible for various bodies in the Union to

have indisputable authority over the member states’ constitutional sovereignty.

A few questions may arise at this point. Most noticeably, why does this contradiction

matter? As long as a state is able to maintain relative order within its territory, why is it important

to consider from where a “sovereign” 5 state derives its legal order? The implications are two-3 Wherein “secondary” lawmaking signi es the laws, decisions, and communications formed inthe framework of bodies, institutions, or interstate agreements in the institution outside of the varioustreaties themselves.4 Quali ed Majority Voting, as used here, signi es a form of decision-making wherein legislativesupport is determined by at least two distinct forms of majority approval (such as by a majority of votes by legislators representing at least a majority of the total constituency).5 Note : Though a highly contested concept with its own highly nuanced eld of literature, for


42/75


42

fold. First and foremost, if states submit to internationally binding standards without consistently

and explicitly consenting, the very essence of what is rhetorically understood as the standard for

states’ independence is undermined. Over 97% of member states of the United Nations function

under the supremacy of a written constitution, so a break from this order may signal not only

a break from norms of governing domestically but also an even more serious consideration for

new forms of international political ordering. Moreover, with such a change, states are no longer

(relatively) free to justify their independence in governing and administrating themselves, instead

bowing to a supranational usurping of domestic authority. This essence of sovereignty is essen-

tially removed from where it is constitutionally located. Therefore, the contemporary situation in

the EU is an extreme challenge to the demarcated boundaries of legal sovereignty, and it is in this

sense worthwhile to consider such implications carefully.

Arguably just as dif cult to pinpoint is the second theoretical concern: thenormative

value of the constitution in domestic politics. The debate surrounding the role of constitutions in

a society is not new. Scholars discuss its normative role as a source of legal authority in societ-

ies, such as in establishing rule of law in a territory, stabilizing a transitioning and/or war-torn

society, and creating and protecting some set of shared values (e.g., Ghai & Galli 2006; Epstein,

Knight, and Shvetsova 2001; and Samuels 2006). This organic law is established to provide the

highest source to which citizens, lawmakers, and courts can refer any points of contention or

justify their actions—be it with individual or governmental violations, authoritative overlapping,

or disagreements as to the constitutionality of legislation, to name a few.

Empirical observations support this theoretical argument. As Atanas Slavov (2011)

emphasizes in his study on the emergence of constitutional law in post-communist Bulgaria, the

constitution was formed to stabilize the country through the societal norm of a single source of

higher law. It provided much-needed stability and order for a country and people in transition.

Moreover, insofar as rhetoric has a practical value, direct considerations of the language in vari-

ous constitutions provides further con rmation: Article VI of the U.S. Constitution, for example,

gives itself the authority as the “Supreme Law of the Land,” while the Constitution of Germany

literally translates to “basic [or ground] law” (U.S. Constitution ; Constitution of the Federal

simplicity and respect for the scope of this paper, “sovereignty” is understood here as legitimate, (howev-er conceptualized and consequently determined) “supreme power or authority” and/or “the authority of astate to govern itself…” (“Sovereignty”).


43/7543


Republic of Germany of 1949 ).

It is also necessary to consider the practical implications at the international level. While

it may be that what is happening is simply a sharing of power among domestic and supranational

legal frameworks, this claim does not address the instances in which these two levels direct-

ly con ict with each other and compete for jurisdiction. Civilians, politicians, lawmakers, and

other actors need a clear understanding of the legal system under which they operate, and with

the increasing complexities of supranational institutions’ role in domestic law, this is even more

crucial: the ambiguity of overlapping normative and empirical legal supremacy makes it all the

more dif cult to not only engage in effective lawmaking but to correctly interpret and maneuver

the legal system.

It is therefore evident that the internal relationships of the EU raise concerns as to how

member states are able to balance the role of international law and legal hierarchy with domestic

constitutions. To address this concern, I will now move to an analysis of speci c case studies

of Germany, the UK, and Romania to understand, if possible, how member states in the EU

have approached this issue. In independently gathering pertinent laws, court cases, comments,

and legal reports, I examine each case study from a wide range of viewpoints. Additionally, in

choosing these three countries—each with their own constitutional styles, regional histories, and

positions in the EU—I aim to balance a consideration for a wide variance in modes of approach-

ing this issue with a worthwhile and accurate analysis. While the decisions and courses of ac-

tion highlighted below do not characterize all relationships between national constitutions and

international law, they nevertheless provide a model overview of the features of the relationship

between national authority and supranational law for EU member states.

The German Case: Steadfast yet Flexible

While the German constitution, the Basic Law, provided a new legal source of domestic gover-

nance for the (West) German state following World War II, it is also one of the more experienced

constitutions in regards to interactions with European law, providing multiple sources from

which I am able to draw an analysis. Studies of the legal relationship between Germany and the

EU often discuss two normatively important court cases:Solange I and Solange II . In both cases,

the constitutional court of Germany ruled on the role of the domestic government in the imple-


44/75


44

mentation of EU law, but with the latter overruling the former, a shift in relations between EU

and German governing institutions developed.

In 1962, the European Court of Justice (ECJ) made a landmark decision for the Europe-

an Economic Community (EEC) withCosta v ENEL . In response to recent moves by the Italian

government to nationalize companies, Italian citizen Flaminio Costa objected to his govern-

ment’s overtaking of the electric company ENEL in which he had stock, arguing that consol-

idation violated free market aspects of the 1958 Treaty of Rome and the Constitution of Italy.

Though the Constitutional Court of Italy ruled in support of the new laws, overturning the earlier

ones that contradict the movement toward nationalization, this matter was nevertheless brought

to the ECJ. As not only a ruling on this case but also a general clari cation on legal hierarchy, the

ECJ ruled that member states cannot override any agreements, laws, or decisions made under the

legality of the Treaty of Rome, thus setting EEC law above national law (Costa v ENEL ). The

ECJ further strengthened this norm in the 1970 Internationale Handelsgesellschaft case ruling in

favor of EU supremacy over challenges based on national laws or precedence (Beck).

However, the German courts took a bold step in overturning this international precedent.

The German Federal Constitutional Court (FCC) ruled inSolange I of 1974 that as long as Com-

munity law did not con ict with German law, it would be treated as national law. Nevertheless,

in the case that Community lawwere to infringe on German citizens’ constitutional rights or free-

doms, the court decided that the latter prevailed over the former (Beck). This ruling is therefore

important as it detailed how the national government viewed its relationship with the EEC. At

this time, the German courts appeared to have held a higher regard for the supremacy of domes-

tic law over international law.

Just 12 years later, the FCC overturned its previous ruling. InSolange II of 1986, the

FCC stated that as long as fundamental rights were generally protected, the FCC no longer had

jurisdiction to rule on cases pertaining to secondary law (Beck). Of even greater importance, this

highlights a changing attitude surrounding the supranational institutional order. No longer was

such law seen in the German courts as “secondary” to national legislation. Instead, there was an

acceptance of the shift of ultimate primacy from the constitutionally legitimized courts and legis-

latures of Germany to international institutions created through executive treaties.

Following the rati cation of the Treaty of Lisbon in 2009, the

Journal of Politics and International Affairs Fall 2015

Documents