19"

A'A‘ l.

$33

,3) ,«gvfinnawm ”1%.

f.)

 

sf! )1... 1:11,;
W _ us!

 

..15(vb.171<.i.oni..~o. . . ‘ .I.....;...vvhfim>. was...” v0: :1... [lion—nzblahu»
é ‘ . . wig? ”a . .fiwrww

.. 3%...m 2%: “and

I

.v 5 v0 J‘IY‘I

Egg?

 

 

m ‘ LIBRARY
1 (, Michigan State
1° 0 University

__f

This is to certify that the
dissertation entitled

THREE ESSAYS ON THE ECONOMICS OF RELIGION AND
POLITICS

. presented by

 

a, _r Ayman Reda

has been accepted towards fulfillment
of the requirements for the

degree m Economics

 

 

2/; we;

Major Professor’s Signature
05/03/05

Date

MSU is an Affirmative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES retum on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

APR 1 5? :‘UiU
101 2 11

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“ I
2/05 c:/ClRC/Dat00ln.ifdd<p.15

THREE ESSAYS ON THE ECONOMICS OF RELIGION AND POLITICS
By

Ayman Reda

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Economics

2005

ABSTRACT
THREE ESSAYS ON THE ECONOMICS OF RELIGION AND POLITICS
By

Ayman Reda

This dissertation contains three essays on the economics of religion and politics.

In the first essay, we model both the religious competition among denominations and
the political competition among parties. The model’s propositions provide the first formal
economic explanation for the ‘Culture Wars’ thesis. In particular, the model seeks to
explain the reason why liberal or less strict denominations tend to favor higher
government taxes, higher government spending and ‘bigger’ governments, while
conservative or stricter denominations tend to favor lower government taxes, lower
government spending and ‘smaller’ governments. This is due to the fact that conservative
denominations demand high levels of commitment from their adherents in the form of
monetary and time contributions and as such their adherents will be less willing to
allocate resources to the government in the form of taxes. Liberal denominations on the
other hand demand lower levels of commitment from their adherents which means that
their adherents are more willing to allocate resources to the government. We Show that
any change in the objectives of the denominations will have an impact on their religious
teachings, which in turn will affect the behavior of their adherents in the political and
economic arenas.

In the second essay, we formally examine the relationship between religious nonprofits
and the public sector with the primary aim of studying the impact of this relationship on

religion in society. The model introduces the first formal examination of the relationship

between religious nonprofits and the government. In the model, we provide a rationale
for why the government may choose to award the funds to a religious charity. If a
religious charity is awarded the exclusive right to provide the social service, it will seek
to utilize this opportunity in order to proselytize its particular religious doctrine to non-
adherents through several means. The model postulates that the utility of the religious
clergy of the denomination(s) awarded the funds will increase, since they can now preach
to a larger group of individuals. Based on the concept of compensating wage
differentials, the clergy are therefore willing to reduce their supply price in order to get
that extra utility. This active proselytizing alters the religious preferences of believers in
the population and leads to a change in the relative powers of the different denominations
and the religious nature of the society as a whole. The paper discusses the implications of
these results with regards to recent policies such as the Faith-Based and Community
Initiatives.

In the third essay, we empirically test the relationship between religious and economic
preferences that we developed in the rational choice model of our first essay. We test
using US state data, whether changes in states’ religious composition over time
influences states’ tax rates. We use church membership rates and religious contributions
by households as alternative measures of a state’s religiosity level. We employ both a
first-differencing and a fixed effects approach. Our paper does not report any significant
relationship between government tax rates and the religiosity of the population. However,
the results do not reject our hypothesis, but only fail to support it. As such, we attempt to
explain the causes of our findings and suggest possible fiiture methods to reexamine the

issue more deeply and extend the analysis further.

ACKNOWLEDGMENTS

At the end of a long and difficult journey, one cannot but remember how it all started.
A high school teacher’s passion in teaching economics led to a student’s love for the
subject, and a determination to go the full distance. In the final and most demanding
stage, I witnessed the same passion in Professor Lawrence Martin’s guidance and
support. As the chairperson of my dissertation committee, Professor Martin gave me help
and encouragement more than I could ever describe within this limited space. Afier long
hours of intellectual pondering in his office, and while generously allowing me to explain
my thoughts, he was carefully discovering my potential and evaluating my interests. Then
he would carefully propose a research agenda and the possible methods of achieving it.
The experience of the past few years is witness that following his advice has been, and
will continue to be, the wise thing to do.

I owe much gratitude to Professor Jeff Biddle for his help and encouragement. His
generous willingness to always listen to my concerns inspired me to maintain my
momentum during the final and exhausting stages of my research. His lucid and engaging
approach stimulated within me an interest in applied economics. 1 am also grateful to
Professor Thomas Jeitschko for his continuous support and guidance. Professor Jeitschko
was always ready to help me in the most difficult of times, and when all other paths
seemed unwelcoming. Furthermore, his advice allowed my early manuscript to evolve
into a much improved dissertation. I am much grateful to Professor Brooks Hull
(University of Michigan-Dearbom) for his help and support since my undergraduate
years. His work in the field of economics of religion motivated my interest in this area of

inquiry. And his advice helped me improve my research perspective and methodology. I

iv

owe a debt of gratitude to Professor Ross Emmett for his keen interest and insightful
advice. His constructive critique of my dissertation expanded my perspective into a wider
and more intricate examination of the issues involved. I was then able to reevaluate my
research in a more complete and informative manner. I am also grateful to several other
professors for their advice and support at different stages of this journey. In particular, I
would like to thank Professor John Wilson and Professor Warren Samuels for giving me
the honor of learning from them. I would also like to thank Professor Michael Twomey
(University of Michigan-Dearbom) for his valuable advice and encouragement.

I also give special thanks to all my friends at MSU who shared the burden of such a
journey. In particular, I am thankful to Olena Nizalova for being a true friend. I am also
thankful to Vladimir Hlasny, Lebohang Lijane, Keoka Greyson-Miller and Paula Kazi.
Special thanks go to David Kamara who is a friend since early high school years, and was
kind enough to read my dissertation and offer valuable insights.

The most difficult of thanks are those intended for one’s parents. Because then you feel
that all words have failed you, and gratitude is but a feeble attempt to mask our inability
to give as they do. This Ph.D. is only a humble attempt at saying, “thank you”. This
degree has been their dream as much as it has been mine, and seeing them proud of it is a
moment I have long dreamed of. I also thank my sisters and brother for their
encouragement and support throughout these years and wish them the best of dreams.

To my eternal friend and loving wife, Malak, this achievement would not have been
possible without you. After our marriage, this journey took a whole new meaning...a
meaning that only we can understand. The pages of this dissertation bear witness to this
reality. But alas, what can I say more to fully express my love and appreciation, other

than to state the fact that the peak of expression is silence.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. viii
LIST OF FIGURES ................................................................................. ix
INTRODUCTION .................................................................................... 1

CHAPTER 1: An Economic Explanation for the ‘Culture Wars’ Thesis

1 .1 Introduction .................................................................................. 3
1.2 Literature Review ........................................................................... 4
1.3 The Model .................................................................................... 9
1.4 Utility Function ............................................................................ 17
1.5 The Scenarios .............................................................................. 19

1.5a. Benevolent Denominations and Election-Seeking Parties .................... 19

1.5b. Benevolent Denominations and Policy-Seeking or Ideological Parties. . ...30

1.5c. ‘Dictatorial’ Denominations and Election-Seeking Parties ................... 35
1.5d. ‘Dictatorial’ Denominations and Policy-Seeking Parties ..................... 38
1.5e. Legitimacy and Endorsement ..................................................... 42
1.6 Empirical Evidence .......................................................................... 51
1.7 Conclusion ..................................................................................... 57

CHAPTER 2: Religious Charities and Government Funding

2. 1 Introduction ................................................................................. 59

2.2 Literature Review ......................................................................... 61

2.3 The Model .................................................................................. 66
2.3a. Stage 1: Religious Competition/Original Preferences ........................ 72

vi

2.3b. Stage 2: Religious Provision of Social Service ................................. 75

2.3c. Stage 3: Religious Competition/Revised Preferences. . ...........78

2.4 Results ....................................................................................... 79
2.4a. Funds to a Single Denomination ................................................. 86
2.4b Funds to Both Denominations .................................................... 9O
2.4c. Believers and Secularists ......................................................... 92
2.5 Conclusion ................................................................................. 98
CHAPTER 3: Religious and Economic Preferences: An Empirical Analysis of State
Tax Rates
3.1 Introduction ............................................................................... 100
3.2 Religious and Economic Preferences ................................................. 105
3.3 Theoretical Model ........................................................................ 110
3.4 Empirical Model and Results ........................................................... 117
3.4a. Church Membership ............................................................... 126
3.4b. Religious Contributions ......................................................... 129
3.5 Conclusion ......................................................................................................... 130
CONCLUSION .................................................................................... 133
APPENDIX A: Religiosity ..................................................................... 136
APPENDIX B: Data Sources .................................................................... 138
BILBLIOGRAPHY ............................................................................... 141

LIST OF TABLES

Table 3.1: Percentage of the adherent population belonging to the ‘conservative’
denominations ...................................................................................... 1 2 1

Table 3.2: Percentage of the adherent population belonging to the ‘conservative’

denominations and the Catholic Church ........................................................ 121
Table 3.3: Summary Statistics ................................................................... 124
Table 3.4: Estimation Results using Church Membership Data-First Differencing
........................................................................................................ 126
Table 3.5: Estimation Results using Church Membership Data-Fixed Effects
........................................................................................................ 128
Table 3.6: Estimation Results using Religious Contributions Data ......................... 129

viii

LIST OF FIGURES

Figure 1.1: Stages of the Game ................................................................... 16
Figure 1.2: Religious Equilibrium ............................................................... 26
Figure 1.3: Policy-Seeking or Ideological Parties ............................................. 34
Figure 1.4: Religious Endorsement of Parties (a) .............................................. 45
Figure 1.5: Religious Endorsement of Parties (b) .............................................. 47
Figure 2.1: Stages of the Game .................................................................... 71
Figure 2.2: Religious Equilibrium ................................................................. 81
Figure 2.3A: Proselytizing by ‘liberal’ denomination .......................................... 89
Figure 2.3B: Proselytizing by ‘conservative’ denomination ................................. 89
Figure 2.3C: Proselytizing by both denominations ............................................ 90
Figure 2.4: Believers and Secularists ............................................................. 92
Figure 2.5: Proselytizing that includes Secularists .............................................. 94
Figure 3.1: Religious Equilibrium .............................................................. 116

ix

INTRODUCTION

Although economics of religion is a relatively new field of inquiry within the
economics discipline, economic analysis of religion can be traced back to Adam Smith in
his seminal work, Wealth of Nations. According to Anderson (1988), Smith explained
religious behavior using a supply and demand framework, and based on the assumption
that religious agents are rational self-interested individuals. Smith also discussed the
history of the Catholic Church from an economic perspective, labeling it as a type of
‘corporate organization’, assuming a monopoly role in the “market for religion”
(Anderson, 1988). As such, he compared such a monopoly structure to one where
religious competition prevails, arguing that it is in the interest of a state to encourage
religious pluralism in order to loosen the moral and political grip of religious bodies on
state affairs. Furthermore, this pluralism will lead to improved toleration between
religious sects, as “the teachers of every little sect, finding themselves almost alone,
would be obliged to respect those of almost every other sect” (Smith, 2000, pp.852).

One can infer from Smith’s analysis of religion that he utilizes economics concepts and
theories to explain the objectives of religious agents and the outcomes from interactions
in religious markets. Smith portrays the objectives and actions of religious agents in
similar fashion to economic actors in commodity markets. He also depicts religious
institutions as corporations or businesses that compete against each other on political,
economic and religious grounds. Furthermore, he evaluates this competition from a
classical economics perspective, by arguing in favor of religious pluralism (or perfect
competition in religious markets) and separation of church and state (or no government

intervention in religious markets). In a competitive environment, each sect is “little” and

“alone” that it exerts only minor influence on the religious nature of society, and the
secular operations of government. This effort by Smith to explain religion using
economic theory as a set of analytical tools set the foundation for the emergence of
economics of religion as a promising and ‘legitimate’ field in the economics discipline.

The essays in this dissertation continue in this tradition. The first essay models the
relationship between religious denominations and political parties in order to explain the
reason why liberal denominations favor higher government taxes and higher government
spending, while conservative denominations favor lower government taxes and lower
government spending. The model explains these outcomes as a result of a rational choice
problem faced by believers when making normal budgetary decisions, and the
corresponding competition between denominations and parties.

The second essay studies the relationship between religious nonprofits and the
government, in order to give us an insight into the difi‘erent motives of these institutions
and the outcomes that arise as a result of their interaction. We study this relationship by
investigating the impact of government funding on religious charities and believers in
general. The paper offers important insights on government faith-initiatives.

In the final essay, we empirically test the hypotheses developed in the first essay.
Specifically, we test using US state data, whether changes in states’ religious composition
over time, measured by church membership rates and religious contributions, influences
states’ tax rates. Using first-differencing and fixed effects methods, the results do not
reject our hypothesis, but only fail to support it. As such, we attempt to explain the causes
for our findings by reassessing the model’s assumptions. We also suggest alternative

methods to reexamine the issue more deeply and extend the analysis further.

Chapter 1

An Economic Explanation for the ‘Culture Wars’
Thesis

1.1 Introduction

In this paper, we explore the relationship between religion and politics in the
frameWork of rational choice models. The paper models the relationship between
religious denominations and political parties. We model both the intra-religious
competition among denominations and the intra-political competition among parties.

The model seeks to explain the reason why liberal or more secular denominations tend
to favor higher government taxes, higher government spending and ‘bigger’
governments, while conservative or more religious denominations tend to favor lower
government taxes, lower government spending and ‘smaller’ governments. The model
explains these real world tendencies as a result of a rational choice problem faced by
believers when making normal budgetary decisions. Denominations who sincerely
represent the religious interests of their adherents will seek to map these preferences into
corresponding political and economic preferences. In particular, liberal denominations
will induce their adherents to favor more government intervention and ‘involvement’ in
the economic arena, while conservative denominations will induce their adherents to
favor less government ‘involvement’. Any change in the objectives of the denominations
will have an impact on their religious teachings, which in turn will affect the behavior of
their adherents in the political and economic arenas. Significant differences emerge in the
outcome of the ‘game’ when we change the objectives of the parties or the

denominations. Therefore, the issue of objectives and intentions will have a major role in

the results of the model. This is imperative since motives have played and still play a
significant role in the overall performance of these two institutions, especially the
religious institution.

Section 1.2 provides a review of the literature in the field of economic of religion that
is related to the topic of this paper. In section 1.3 we discuss the structure of the model. In
section 1.4, we examine the features of the utility function used. Section 1.5 develops the
propositions of the model under different sets of assumptions. In section 1.6, we survey
some of the empirical work done on religion and politics in order to observe any
empirical evidence to support (or reject) our propositions and results. We conclude in

section 1.7.
1.2 Literature Review

The field of economics of religion is a relatively new field in economics. The
beginning is generally attributed to Azzi and Ehrenberg (1975) in their paper about how
households allocate resources between religious and ‘secular’ consumption'. Ianaccone
(1998) defines economics of religion as “the line of research that interprets religious
behavior from an economic perspective, applying microeconomic theory and techniques
to explain patterns of religious behavior among individuals, groups, and cultures.”
Therefore, the attempt made by economists in this field is to extend the tools of
neoclassical economics to the religious sphere. It is important to note that even such an

attempt is not expected to gain wide approval by both economists and theologians.

 

1 Our model can be viewed as one where rational adherents allocate resources between ‘religious
investrnent’ and taxes to the government. However, the denominations and parties make the actual or final
choice by taking the preferences of all adherents as given and operating in a competitive environment
according to specified objectives.

Although this can be observed in any new intellectual ‘adventure’ that economists decide
to embark upon, it would arguably raise the most controversy in the field of religion.

Barros and Garoupa (2002) classify the economic theories of religion developed so far
into two groups, ‘demand-side’ and ‘supply-side’ theories. They refer to the works of
Iannaccone (1990, 1995) and Durkin and Greeley (1991) as examples of ‘demand-side’
theories. Barros and Garoupa (2002) state that both works above have utilized Becker’s
theory of human capital to explain religious consumption patterns as an addictive
behavior. On the ‘supply-side’, they discuss one approach used by Anderson et al.
(1992), Davidson and Ekelund (1997) and Ekelund et al. (1989, 1992, 1996) that treat
religious denominations or sects2 as firms that aim to maximize their ‘profit’ as they
compete against each other. Another ‘supply-side’ approach discussed by Barros and
Garoupa (2002) is one where religious authorities seek to maximize the welfare of their
followers. They mention Iannaccone (1988, 1992), Montgomery (1996) and Zaleski and
Zech (1992) as examples of this approach, where the religious denomination is treated as
a club that employs agents and aims to maximize group welfare subject to various
constraints.

In the area of ‘supply-side’ theories of the religious economy, Zaleski and Zech (1995)
emphasize the pioneering works of F inke and Stark, especially their seminal work, “The
Churching of America, 1776-1990: Winners and Losers in the Religious Economy”
(1993). They note the importance that Finke and Stark give to the dynamics of supply-

side forces over the relative stability of demand-side forces.

 

2 The use of sect here is to refer to any religious denomination within a particular religion. Thus, it is
different fi'om the normal usage in economic theory where it is distinguished from a church. The church-
sect classification is not relevant in this paper.

In his important survey of the literature, Iannaccone (1998) argues that the old notion
that religion is gradually disappearing from the social, economic and political life of
people has been proved false. He states that, “as survey, census, and historical data have
piled up, the continuing vitality of religion has become apparent ...” (Iannaccone, 1998).
In this work, Ianaccone surveys all the important empirical work that has been done in
this field. Many of the empirical work in the field of economics of religion have been to
test the propositions of the developed theories and to test old and new beliefs about
religion.

A substantial portion of the work done in the field of economics of religion has taken
the form of an axiomatic approach. The methodology has utilized the concepts of
economic theory to derive propositions and predictions. But these studies, although of
great value, fall short of employing formal economic modeling. The standard and often
rigorous techniques of economic modeling were not used. In a promising work that is
forthcoming, titled “The Political Origins of Religious Liberty”, Anthony Gill utilizes
this axiomatic approach to study the dynamics of religious liberty in the framework of
religious and political competition. Gill studies the concept of religious liberty in the
framework of a rational choice problem in which religion and politics interact in a
competitive environment. Specifically, Gill models religious liberty using the concept of
opportunity cost. He defines religious liberty as the “degree a government regulates the
religious marketplace”. Gill argues that in their effort to secure political legitimacy,
governments weigh the benefits and costs of regulating religion and this choice problem
determines the regulatory decisions implemented. In our paper, we are not studying the

notion of religious liberty specifically. The model we develop is aimed at explaining the

dynamic relationship between religious and political institutions in all areas of the
political and religious spheres that pose potential conflict or cooperation between these
institutions. We formally derive results and propositions that describe the religious and
political decisions of denominations and parties. The paper specifically explains the
motives behind specific religious positions taken by denominations and specific political-
economic decisions taken by parties. Furthermore, it explains the origins of political-
economic decisions taken by religious authorities.

A model developed by Roemer (1998) is also worth considering because of its relation
to the topic of this paper. Roemer (1998) studies a model of two parties, Left and Right in
a two-dimensional structure of uncertainty with citizens having preferences on tax rate
and religious position of government. Roemer (1998) examines the political competition
between the two parties on the issue of the tax rate and argues that if religion is of no
significance in the society, the parties take extreme positions on the tax rate spectrum.
However, if religion is of significance in the society, then the positions of the parties on
the tax rate will differ. For example, the Left party will propose a tax rate far less than the
extreme (which is 1). Roemer (1998) argues that the portion of citizens that have
religious preferences close to the median constitute the set of swing voters. Thus, any
policy by parties on the tax rate issue is critically related to the tax rate preferences of this
portion of the population. If this group of voters has high average wealth, they naturally
prefer low taxes and this induces both parties to be biased in favor of low taxes to secure
winning the elections. The interesting issue here is thus the tendency of both parties of
opposing views to favor one view because of religious preferences in the population.

Roemer (1998) concludes that given this result, we may argue that parties can

intentionally inflate the importance of religion in order to ensure that a particular
economic policy is enacted.

In terms of the assumptions used, our paper is technically using a one-dimensional
approach in an environment of full information. Roemer (1998) specifically assumes
“that wealth and religious views are not independently distributed”. Therefore, the model
implicitly assumes the relationship between wealth and religion and goes on to study how
this relationship influences the decisions of parties. The preferences are therefore
exogenous. In our model, we derive the political and economic preferences from the
religious preferences of individuals. In other words, the politico-economic preferences
are endogenous. We assume that wealth is the same for all believers, and therefore a
person’s preferences regarding the tax rate are independent of his income level. It is
determined primarily by their religious preferences and their denominational affiliation.
Religion shapes a believer ’s political perspectives. A believer, who prefers a religious
life of low commitment, flexible rules, and minimal spirituality and favors a materialistic
way of life, will likely be in favor of a policy of more taxes or at least not opposed to it.
In contrast, a believer who prefers strict rules, high commitment and religious piety, will
likely favor more spending on religion and thus less taxes. On the issue of income and
religiosity, Iannaccone (1994) comments tha , “inCome correlates weakly with most
dimensions of religious commitment”. Regardless of the presence of any correlation, our
model takes a different approach altogether. Also, the religious preferences in our model
are regarding ‘religious investment’ or religiosity while Roemer (1998) defines religious

preferences as a function of the religious stance of the government.

Furthermore, our model examines a situation with denominations and parties engaged
in competition. The dependence and independence between these ‘grand’ institutions is
studied. The model describes the reasons behind the different positions they take under
different scenarios with different assumptions. The issue of objectives and intentions

plays a significant role.
1.3 The Model

In this model, we use a structure that is characteristic of contemporary politics and
religion. The political structure is one of political parties that are competing against each
other in an election. The ‘secular’ parties are at some stage declaring their platform or
positions on the various issues that interest the voters. The voters respond to this political
competition by voting for their preferred candidate. Thus, this structure may not be
applicable to many periods in history, where the forms of political activity differ
substantially. Also, we construct the religious world as one where denominations take the
religious preferences of all the individuals in the society as given and given some specific
objective, locate optimally on that spectrum of religious preferences. We will also
examine cases where the denominations ‘ignore’ the preferences of believers and instead
seek specific policy objectives driven by a commitment to a doctrine or scripture. The
religious authorities may have their own religious agenda that is independent of their
adherents’ preferences.

As is the case with reductionistic methods of analysis, we are compressing the real
world into a far less complicated structure and substituting harmony for chaos and
consistency for unpredictability. In reality, preferences are not given to denominations to

act upon and denominations may not have consistent objectives over time and may even

have multiple objectives at a specific point in time. Nevertheless, we believe that the
structure that is employed in this paper does lead to some useful insights that will help
explain the behavior of parties and denominations as they engage in religious and
political competition.

The main contribution that this paper presents is its effort to explain the behavior of
political parties and religious denominations explicitly. As evident in the review of
literature on the economics of religion, the behaviors of parties and denominations have
not been jointly studied in a framework that explains both their independence and their
interdependence. The model explains the behavior of religious denominations in the
religious domain and the behavior of political parties in the political-economic domain.
There is an element of independence between these domains and also an element of
interdependence or dependence. Previous work did not study both denominations and
parties in a unified structure. This paper tries to fill that gap.

Building on that previous point, the paper seeks to explain the reasons that induce
denominations and parties to act the way they do. In particular, the paper gives an
explanation for the religious decisions taken by denominations and also an explanation
for the political and economic positions that they adopt. The model evolves as a situation
where religious decisions of denominations directly or indirectly influence the political
and economic decisions of adherents and as such, affect government policies through the
political competition process. The paper explains the religious and political decisions of
denominations as driven mainly by two forces. The economic and political reality of the
world forces the denominations to make decisions in conformity with this reality. The

denominations however are also driven by intentions or objectives. The interplay of these

10

two forces affects the ultimate behavior of denominations. In other words, denominations
act in both a pragmatic and an idealist3manner with varying degrees of emphasis under
different circumstances.

A unique feature of the model is the timeline of events it presents. In the pre-game
stage, the preferences of individuals are given. Denominations or parties cannot influence
these. In the game stages, we model the denominations as the first players. Their
decisions precede those of the-parties. We believe that this formulation is both realistic
and insightful. To a great extent, religion arrives at home earlier than politics. Young
individuals are exposed to religion at an earlier stage in their life than they are to politics.
Religion arrives via the parents, relatives, and fiiends and more importantly for us, the
church. Smith et a1 (1998) explain the influence of parents on their children in the area of
religion as an “investment in a child’s religious human capital”. They argue that this
investment by parents serves as the driving force for continuing interest in religion in the
later stages of a child’s life (Smith et al 1998). As such, our structure assumes that the
religious institution in general influences individuals before they become exposed to the
world of politics. Religious institutions fuel this perpetuating influence over generations
of new children through their impact on parents, who represent the main vehicles of
religious learning and investment.

With their own preferences given beforehand, the denominations compete and attract
adherents and then impose conformity to a standard that they choose. These standards
that represent the teachings of the denomination, will direct the political and economic

decisions of the adherents. In other words, the religious teachings will lead to political

 

3 This means in accordance with religious principles.

11

and economic orientations. This is the critical issue in the model. That religious doctrine
directs political-economic perspectives and actions forms the crucial concept that
underlines the model in general. In more concrete terms, the model presents a structure
where political and economic decisions are endogenized and this endogeneity is due to
religious doctrine. The model is thus presenting religion as a central force in the lives of
individuals, both on the private and public levels. It is our opinion that religion is the
main driving force for individuals in their political and economic decision-making
processes, and this paper is an attempt to study this driving force in the framework of
rational choice models.

We categorize the methods by which religion (via denominations) affects the political
and economic aspects of society into two types, direct and indirect. The indirect method
is when denominations aim to achieve a defined religious objective and fulfill that
objective either by sincerely representing the preferences of their adherents in the
religious competition stage and this representation is translated into a religious position
and a corresponding politico-economic position, or by imposing a specific pre-defined
religious standard on their adherents. The objective of the denomination is not a specific
economic or political policy, but a specific religious policy. In contrast, the direct method
comprises of some explicit political or economic policy objective for the denomination.
We can probably say that in the indirect method, religion drives politics while in the
direct method politics drives religion. There is a significant difference between these
objectives or intentions.

The methodology we employ in this model is based on concepts in the fields of

economic theory and political science. The paper makes use of the framework of spatial

12

location models. We follow, to some extent, the procedure of Barros and Garoupa (2002)
in which they use a spatial location model to formulate a theory of church strictness. In
particular, they use a modified Hotelling model. The games of strategic interaction in this
model utilize the tools of game theory and in particular, the concept of ‘backward
induction’, to derive the sub-game perfect equilibrium results. Similar to Barros and
Garoupa (2002), we study a situation where believers locate on the ‘religious spectrum’
and choose that denomination that maximizes their utility. This follows the Hotelling
model of consumers grouping near different sellers, according to their preferences. The
denominations provide a local public good4. Believers each have an ideal position on the
‘religious’ spectrum.

From political science, we adopt the methodology used by Grossman and Helpman
(2001) in their book, Special Interest Politics, where they examine the political economy
of special interest groups. Specifically, we use their methodology on voting, ‘pliable’
policies and ‘fixed’ policies.

Hence, we employ a 2-party and 2-denomination structure. Although in the real world
there are numerous denominations, we can identify two main types of religious
denominations, the ‘liberal’ and the ‘conservative’ denominations. Also in the US, there
are two main parties, Republican and Democratic. Even in other contexts, we can
categorize them as ‘ruling’ party and ‘opposition’ party. The structure of the model is
divided up into 5 stages. The stages are as follows: -
$22.13 - All individuals in the population have an ideal preference regarding the

‘amount of religious investment’ they are willing to undertake in their religious life. This

 

4 We do not include the religious good in the model since it does not contribute any significant insight or
result.

13

religious investment may refer to the preferred amount of money each individual is
willing to contribute to religious institutions, or the amount of time the person prefers to
contribute in assisting in the social and religious activities of the religious institution
he/she belongs tos. This parameter describes the level of religion preferred by individuals
in terms of measurable variables such as monetary contributions or voluntary assistance
in church activities. Thus, we can safely assume that these variables serve as proxies for
the religiosity of individuals". We are therefore using economic variables to represent
religious preferences. Iannaccone (1994) lists religious inputs as consisting of time used
in involvement in religious practices such as praying, meditating, reading scriptures, and
religious goods as financial contributions, offerings and travel expenditures. We will
regard all these types of activities as inputs into the religious production process and as
representing ‘religious investrnent’ in general. These preferences are given at the
beginning of the ‘game’. We assume the preferences to be located on a ‘religious
investment’ spectrum that includes all members of the population7.
The ideal preferences of an individual i is given by: a,-_

S2123: - In this stage that we call the religious competition stage, the denominations
take the religious preferences of individuals as given and compete against each other. We
initially regard the denominations as benevolent players whose objective is to maximize

the expected welfare of their adherents by choosing an optimal position on the ‘religious

 

5 The ‘value’ of time may be in the form of the opportunity cost in terms of profitable secular activities
such as working, studying, etc.

6 See appendix A, pp. 136.

7 Atheists and agnostics can be located on the extreme left of the spectrum or to the left of the zero point.
To avoid the problem associated with individuals who have no religious preferences, we can simply restrict
the spectrum to those that do have religious preferences regarding investment in religion. This is also due to
the fact that we want to normalize the spectrum to include all values between 0 and 1, since atheists and
‘secular fundamentalists’ may have negative preferences regarding ‘religious investment’.

14

investment’ spectrums. Therefore, the denomination’s objective in this case follows the
indirect method. Denominations do not have a prior position on ‘religious investment’.
Instead, they sincerely represent the preferences of their adherents. Individuals
automatically join the denomination that is closest to their ideal position on the spectrum.
All members of the population must join a denomination.

Stage 3: - After locating in a specific denomination, believers seek to maximize their
politico-economic utility. In this transition stage, each individual belongs to a particular
denomination and as such, is obligated to conform to the teachings and rules of that
denomination. Initially, each individual had his/her ideal position on religious
investment. But in stage 3, every individual who joins denomination 1 is restricted to
behave according to the optimal position chosen by denomination 1 in stage 2.

The same is the case for all adherents of denomination 2. Thus all members of
denomination I have to invest the amount specified by denomination I . This happens
regardless of the nature of the denomination’s objectives or intentions. Whether the
denomination is benevolent or dictatorial, the individual joins the denomination that is
closest to his/her ideal position on ‘religious investment’. This is a clear illustration of
how religion is ‘produced’ by religious institutions. A key issue in religious doctrine is
conformity and submission to the teachings and standards of the religious scriptures. In
the case of benevolence, the preference of a single individual will be considered equally
like all others and therefore the influence of one particular individual is confined to his
weight in the total of adherents. All adherents carry equal weight. This illustrates a main

teaching present in many religious traditions, namely, equality in the ‘eyes’ of the

 

8 We will examine other objectives in the paper also. These differences in objectives form a key role in
the results and insights of the model.

15

Creator. It reflects the justice of the Creator as practiced or implemented by religious
authorities that ‘represent’ Himg. Also, the fact that one optimal position among this
spectrum of adherents’ positions will be chosen and adopted reflects the communal
nature of religions where all individuals regard the group’s welfare as superior to their
own private welfare. As such, adherents are willing to sacrifice their interests for the
good of the whole and in obeying the doctrine that they have ‘willingly’ chosen to abide
by.

Therefore, in this stage, given the fact that they now have to behave in conjunction
with the denomination’s teachings, each individual pursues his interests in the politico-
economic sphere.

M: - In this stage that we call the political competition stage, the political parties
have as given, the positions of the denominations in stage 2 and the ideal politico-
econornic positions of the adherents of both denominations from stage 3. Given the fact
that the religious preference is characterized as ‘religious investment’, we define the
politico-economic parameter to be the ‘tax rate’, in particular, a head tax”). This means
that with their final positions on religious investment decided by denominations,
adherents now develop their politico-economic positions or their preferences regarding
the tax rate in stage 3. Parties then seek to maximize their probability of winning a
majority of votes by choosing a position (tax rate) among this ‘new’ distribution of ideal

tax rate positionsl 1.

 

9 The choice of gender here is in specific reference to the three main monotheistic religions, Islam,
Christianity and Judaism.

'0 This is analogous to the ‘religious investment’ preference in the sense that the tax rate represents the
preferred ‘contribution’ of each adherent to the government.

N This is one of several objectives of parties that will be examined.

16

Stage 5: - In this stage, which we term the voting stage, voters decide which party to vote
for based on the platforms of both parties from stage 4. They vote for the party that gives

them the highest politico-economic utility.

 

O 4

Preferences Religious Competition Transition Stage Political Competition Voting Stage
Given Stage Stage

Figure 1.1: Stages of the Game
Before proceeding to solving the stages in detail, it is usefiil to examine the structure of
the utility function that we will use in this model as it forms the foundation for the model
in general.

1.4 Utility Function

The first part of the utility function represents the religious utility while the second part
represents the politico-economic utility.
i) Religious Utility

The religious utility is given by:

fltR- s (xj—aa 2}

In the function above, R is the religious ‘club’ good that is exogenously provided. fl

represents the weight placed on religion in the population in general. It is the same for all
individuals”. We can regard it as an indicator of the ‘power’ of religious institutions or
religion in the society, as evidenced by its effect on the utility function of all individuals.
This power may represent the prominence of religious institutions or their history. This

parameter is not related to the individual-specific preference on ‘religious investment’.

 

'2 We can have [3 vary by individual, although this will not affect the results of the model. However, a
value of 0 may indicate no preference for religious utility which may indicate that the individual is an
atheist or an agnostic. This could be a way of structuring the preferences of these groups in the population.

17

The variable 0,- represents the ideal position of individual i on ‘religious investrnent’I3

and xj corresponds to the level of religious investment chosen by denomination j in stage
2. We denote the transportation cost incurred by individuals as a result of their obligation
to conform to the teachings or rules of the denomination, by S. It is the ‘price’ of

sacrifice. This value is the same for all individuals in the society.

From the form of this utility, we observe that the individual derives utility from his/her
input into the religious good production as given by {-— S (xj — ai) 2}”. The individual
loses utility as a result of confonnism to the denomination’s rules, but also gets utility
from giving his/her preferred investment level. As such, the total effect from this is given
in the quadratic form {— S (39— aJ 2}.

ii) Politico-economic Utility

(1- flh’ Uijk (Z. Ci) + Vik}

In the politico-economic utility above, Z is the public good produced by the party that
assumes office after the elections. (1- ,3) this refers to the weight that society places on
political and economic issues in general. Also, it may refer to the ‘power’ of these
political and economic institutions in the society. 6,- denotes the private consumption of
individual i after investing in religion and paying taxes. Whatever is left after religious

investment and government taxes is used for private consumption. Also, v”, is the fixed

 

'3 In practice, a,- will usually have a range of small values such as [0, 0.2], since contributions to
denominations do not constitute high percentages of individuals’ incomes. Similarly is the case for the tax
rate. Private consumption is usually the largest component of income spending. The nature of the model
and its primary interest in the religious and political components makes them seem to command very big
portions of income.

'4 This is reasonable because in the religious domain, individuals feel spiritually satisfied by the act of
giving to the denomination as well as receiving from it.

18

position of individual i towards party k. This position is independent of the tax rate or the
religious preference. It refers to the bias of individual i towards party k regardless of the

positions of party k in the elections, thus the term ‘fixed’. This variable could be positive
as well as negative. A positive vik means that the individual has a positive bias towards
party k.

Hence, Urjk( Z, c,- ) represents the politico-economic utility of individual i. The
individual derives utility from the public good provided by the government (winning
party) and from private consumption. This function is maximized in the transition stage.

The composite utility function is therefore:

Uij = 13.{R- Sij—atlz} + (1'16” Uy-IJZ. Ci) + Vik} (1)

We now examine the workings of the model under different assumptions about the

religious and political environments.

1.5 The Scenarios

1.5a. Benevolent Denominations and Election-Seeking Parties

In this case, we consider a situation of two benevolent denominations whose objective
is to maximize the sum of expected welfare of their adherents and two parties whose sole
objective is winning the elections. The parties have no policy- related objective. It is an
instance of the indirect approach to politics by religious institutions, since the
denominations care exclusively about the interests of their adherents. Their influence on
politics and economics is a by-product of their religious teachings.

We proceed to solve the problem using the method of backward induction. This means
that each player in a particular stage will choose his decision optimally based on the

outcome of the following stage. Even though we use this procedure, we will observe that

19

in some cases the decision of one stage doesn’t explicitly depend on that of the one
coming after. This is due to the structure of the assumptions in that particular case. With
other assumptions or cases, the outcome shows impact of one stage on another. For
example, the assumption of benevolence of the denominations is ‘fixed’ and this means
that the decision of parties in the last stage does not influence the decisions of
denominations in the first stage. This is an assumption that we are imposing on the model
so as to illustrate the impact of intentions and doctrine. As we said before, this case is one
where religion drives politics and not vice versa. The nature of the denomination’s
objectives ‘forces’ it to regard the interests of its adherents as superior to all other
considerations. As such, the denomination assumes a democratic, egalitarian and populist
mode.

But before we start solving the problem, it is useful to examine the decision that is
made in stage 3, the transition stage. As we stated previously, individuals maximize their
politico-economic utility, given that they now belong to a particular denomination. In

general, each individual solves the following problem :-

Max (Ii/712,, Ci) S.t Y-xJ-T, = c,- , Z,- = T,N , where 7}: tax rate”.
{T,-} N: population.
Y: income.
Substituting in the constraints, we have :

Max ujk(]iM Y-xth)
{II}

F. 0. C:- (dz/3,, /dZ,-) (dz,- /d 7,.) + ((111,. /dc,-) (dc/d TJ=0

We can then derive the optimal tax rate for i, (TX) fi'om the above result.

 

‘5 We obtain similar results with an income tax instead of a flat tax.

20

Example: Let us consider the following utility function for simplicity and in order to get

results that can be readily interpreted.
Um, (Z, c) = ban + (I-b)lnc,~
ujtrz. c) = bln(T.-N) +(1-b)ln(Y-19-Tt)
F. 0. C":-
b / T,- + (l- b) /(Y-xJ-TJ.-1 = 0
T.’ = th-x.) (2)
This result describes the optimal politico-economic position of the believers or

adherents. It states the optimal tax rate of each follower of a denomination. Thus the

indirect utility of each believer would be as follows:-

V(T,) = bln(b(Y-xJ)N) + (I-b)ln (Y-x,- b(Y-x,-))

 

'6 In the case of heterogeneous income Y,, we have T; = b (Y,-xj). This means that as income
increases, the preferred tax rate increases and decreases as income decreases. We should note that due to
the model’s construction, religious preferences are not influenced by income levels but primarily by
‘exogenous’ doctrine and denominational competition. However, we can observe that if income levels
increase, say for conservatives, this would induce them to favor more taxes than before, and thus become
moderate on economic issues. Also, if incomes fall for religious liberals, this would induce them to favor
less taxes than before and thus moderate on economic issues such as taxes and government spending.
However this does not provide us with the complete picture, because one should instead consider that
adherents that are on the receiving-side of government welfare may behave differently than if they were on
the giving-side. In other words, religiously conservative believers may favor more taxes and government
spending if they are net receivers of government assistance and less taxes and government spending if they
are net givers. Similarly, religiously liberal believers may favor less taxes and government spending if they
are net givers. The empirical observation by Manza and Brooks (1997) that Catholics have become more
Republican over time may be explained by the fact that their income levels has increased over time and
assuming that they are religiously conservative, they have gradually become net givers instead of net
receivers. In other words, the Catholics are moving towards their ‘religious-economic’ equilibrium.
Furthermore, this might explain Leege’s (1993) observation that liberal Protestants have become more in
favor of the Democratic Party over time. Leege (1993) states that they have become less affluent than
before. As a result, since they are now potential net receivers, they would tend to be in favor of more taxes
and government spending. Their previous economic affluence relative to other denominations induced them
to favor Republican policies, even though they are religiously liberal. As they have become more equated
with other denominations on the income scale, they have become more economically liberal as a result.

21

In (2), Y is the same for all individuals in the economy, as well as b. The variable )9-

will take two values corresponding to the two denominations 1 and 2. The problem facing

parties in the political competition stage, is to choose the tax rate that maximizes their

chances of winning the elections and assuming office. In (2), it is clear that in every

#
denomination, T,- is the same for all members in that denomination, and that the tax rate

preferred by members of denomination I will differ from that preferred by members of
denomination 2 and this difference is mainly due to the positions chosen by the
denominations on the ‘religious investment’ spectrum. This means that since there is
conformity to a uniform religious standard for all members of a denomination, likewise,
they will conform to a uniform standard on political and economic matters. The
denomination is therefore imposing a religious doctrine that in turn imposes a
corresponding doctrine of rules and directives on political and economic issues. The
adherents were differentiated on the issue of ‘religious investment’ (we assumed this to
mean religiosity), and this differentiation introduced further differentiation on the
political and economic issues. The parties when deciding their election platforms will
take all this into account. We now solve this case of 1.5a stage by stage.
The Voting Stag

In this stage, voters will decide which parties to vote for after the parties have
announced their platforms on the tax rate issue. Each voter in this stage belongs to a
particular denomination.

Specifically, voter i of denomination I will vote for party A if and only if”:-

 

n This methodology is developed by Grossman and Helpman (2001) in Special Interest Politics.

22

Vt/(TA) + VIM > Vil(TB) + vilB (3)

Note:- v,“ & vi“; are the ‘fixed’ preferences of i for each party, A & B”.

We have v,- = v,-B - viA, following a uniform distribution given by:-
U,- ~U {(-1 +2c)/2fi (I +2c)/2f}

From (3), we have: Vt: ViB" ViA < VII(TA) — VI‘I(TB) (3’)

From above, (3 ') is the voting rule for individual i.

Given (3 '), the share of votes for party A among voters with preferences v,- is given by:
Si! = 1/2 — C +f[Vu(TA) — Vil(TB)]

Summing across all different S,- 1, and weighing by the number of votes for each
particular v,-,

S,=1/2—c +f[V,(TA)- V/(TB)]

The Political Competition Stage
From the distribution stated above, the share of votes for party A among members of

denominations I and 2 are as follows:

S, = 1/2 — c +f[V,(TA) — V,(TB)] , V,(TA): average of V(TA) in denomination 1.
V; (T B): average of V(TB) in denomination 1.
S; = 1/2 — c +f[V2(TA) — V2(T3)] , V2(TA): average of V(TA) in denomination 2.

V2(T3): average of V( T 3) in denomination 2.

 

'8 We assume that the distribution of positions on the ‘frxed’ policy issues is identical in each
denomination. Thus, (qu- vi“) = (Veg - v.34) follow the same distribution.

23

Total share of votes :-
S = 1/2 — c +f[V7(TA) — V7(TB)], V7(TA): average of V(TA) in population.
Vr(T3): average of V( T 3) in population.
The objective of party A is to maximize the probability of winning (to get a majority) :-
i.e. to solve : MaxS > 1/2
0r 6 <f[V7{TA) - V7019]
Thus A maximizes V7(TA), and B maximizes V;( T B).
To solve this more explicitly, we use the form of the utility fiinction introduced above.
(1,1 2, c ) = bln(T,-N) + (I-b)ln(Y-xj-T,-)
PartyA will seek to maximize: Max {bln(T,4N) + (1-b)ln(Y- -TA)}, (4)
{ T A}
where: X = cox, + (I- w)x2, (4)
a) = share of population that are adherents of denomination I. This value of a) will be
obtained from the religious competition stage.
The solution to this problem is: T A’. = b(Y-)O (5)
Similarly, the solution to B’s problem is : T B. = b(Y-A9 (6)
Politico-economic equilibrium: In this stage, both parties will converge to the same

position in equilibrium. This should be expected since all the parties care about is
winning. The equilibrium tax rate will depend on the average of the optimal positions on
‘religious investment’ chosen by the denominations. Also, it will depend on the share of

the adherent population that each denomination holds. The final winner of the

competition will depend on the distribution of the fixed positions v,-. If party A commands

24

more popularity on the ‘fixed’ policy issue, then it will win the elections and similarly the
case for party B. If both have same popularity, then a coin is flipped to determine the
winner. Given this equilibrium result, we now turn to the preceding stages to derive the
strategic decisions made there.
The Transition Stage

We have seen that in this stage, believers take their denomination’s policy as given and
derive their optimal tax rate. Their optimal or preferred tax rates are influenced primarily
by the denomination’s positions on the ‘religious investment’ issue. This is the tax rate
that they would prefer to see implemented and it is driven by their religious convictions.
This will form the main idea of the model and we will elaborate on it in more detail in
working out the details of the religious competition stage.
The Religious Competition Stage

As previously stated, the denominations act as benevolent players who sincerely seek
to maximize the sum of the ‘religious welfare’ of their adherents. They also compete in a
Nash environment. Each denomination chooses the best response strategy to its
competitor”.

The denominations seek to solve the following problem (for all i in their

denomination):-

Max {R— 509- — af}
{361-}

a,: this represents the preferences of individual i on ‘religious investment’ and

follows a uniform distribution, g( a).

 

'9 This methodology is similar to the application of Hotelling’s Model by Barros and Garoupa (2002).

25

t
Denomination 1: Max /{R—- s(x, —a,~)2}g(a)da (7)
{x1} 0
t

F.0.C: —23/(x,—a,) da = 0
0

We assume g( a) to follow a uniform distribution of density20 1.

 

The solution is given by: x]. = t/2 (7a)
1
Denomination 2: Max ,/(R— S(x2—ai)2}g(a)da (8)
{x,} t
I
F.0.C.- —2s /(x2—a,-).da = 0
I
The solution is given by: x; = 1/2 + t/2 (8a)

Rengflrs Equilibrium: Both denominations will locate equidistant from the end points.

 

The denomination locating on the left will be labeled the liberal denomination while that
on the right labeled the conservative denomination“. This differentiation is partly caused

by the benevolent or democratic nature of the denomination’s objectives.

 

‘ I I I '
x, X;
(liberal) (conservative)

Figure 1.2: Religious Equilibrium

 

2° As discussed in footnote 13, the range of a,- over which people have preferences is usually that of small
percentages of income. Our integral from t to 1 is just for simplicity. It could be normalized over a set of
smaller values, say 0 to 0.1.

21 See Empirical Evidence, (i), pp. 51.

26

Proposition 1:

In equilibrium, the liberal denomination will induce its members to favor a high tax
rate while the conservative denomination will induce its members to favor a low tax rate.
m

The proof is straightforward, since we just need to restate some of our results formally.
From the religious competition stage, we find that the religious equilibrium will have
both denominations locating equidistant from the end points, under some conditions. As a
result, we have a liberal and a conservative denomination. The denomination locating on
the left favors low religious contributions and voluntary services. This means that its
members favor low religiosity levels and thus it is liberal. The one on the right favors
high contributions and voluntary services, or high religiosity. Thus, it is the conservative
denomination. The main reason for this result is because the denominations sincerely
represent the preferences of their adherents and thus act benevolently. This creates the
differentiation in final positions between the two denominations. By choosing these
positions, the denominations have imposed on their adherents a uniform level of religious
investment that all adherents must conform to. This forces the adherents to adjust their
budget choices accordingly. By the religious differentiation between denominations, they
have imposed on adherents a politico-economic differentiation. Liberal adherents now
favor higher tax rates while conservative adherents favor low taxes.

This is given by the following equations:

T1. = b(Y-x,), for all i belonging to denomination I.
T 2* = b(Y-x2), for all i belonging to denomination 2.

It follows that since x, <x2, then T ,* >T2t.

27

Furthermore, Y-xl-TA =6] and Y-xz- T ,4 =02. (Note: T A: T 3 in equilibrium).
This means that in the final outcome of the game, adherents of denomination I will

exercise more private consumption than those of denomination 2, (C I > C 2)

However, although denominations do induce their members to be ‘liberal ' or

‘conservative’ regarding the political and economic issues, in the fiamework of our
model, they end up paying the same tax rate T A = T 322, which is the case in real life since

governments impose taxes unifome (at least the head tax) on all adherents of
denominations. Also, by law, governments cannot discriminate among denominations in
the imposition and collection of taxes. The question may still arise, that if the final tax
rate is the same regardless of the winning party, why induce members to favor different
politico-economic positions? We have noted that the objective of the denomination in this
case directs their behavior. The party positions have no influence on the religious
decisions of denominations. Also, although the tax rate is the same, it is however
influenced by the religious positions of both denominations as given by the following

result:
TB’=T,,* = b(Y-)O whereX = tax, + (1- aux,
Thus, the final tax rate is influenced by both x I and x2. Furthermore, it is influenced by

a). Any deviation by a denomination from their Nash equilibrium position will cause the

 

22 We should note that for this result to hold in the framework of the example above, a few assumptions
should be made regarding the parameters. In denominations I & 2, after the religious competition stage is
over, the adherents are left with [Y-xJ and [Y-le respectively. After paying the same taxes, what remain
fiom their incomes are [Y- Xl-(Y-ID b] and [Y- xg-(Y-AO b] respectively. For the budget constraints of both
types of adherents to be satisfied, we must have: [Y- xJ-(Y-AO b]30 for j=l, 2. This means that we have to
make some assumptions regarding the range of (ar), (refer to footnote 13, 20), or range of values that (b)
can have. But these assumptions do not change the results or propositions that the paper produces.

28

other denomination to gather more adherents and also have their optimal policy be more
influential in the final outcome of the game.

We can now state a further result obtained from the proof above.
Proposition 1’

Liberal adherents are more lavish in private consumption than conservative adherents.

Also, given that T 1* >T2*, from Z,- = TiN, it follows that Z ,‘> 2;.

Corollary

Liberal denominations favor more government spending than conservative
denominations”. Thus, we can reasonably say that liberal denominations favor ‘bigger'
governments and conservative denominations favor ‘smaller ' governments”.

This corollary is insightful, since in real politics, we observe that liberal denominations
are more closely affiliated with the Democratic Party and are thus proponents of ‘bigger’
governments, while the conservative denominations are more closely affiliated with the

Republican Party and are thus proponents of ‘smaller’ governments.

These hold: T A' = T 3* (convergence)
T2*< T].
x, *< x 2‘ (equidistant)
Cl.) 02*

Z, *> 22*

 

23 See footnote 29, pp. 52.
24 See Empirical Evidence (ii), pp.52.

29

1.5b. Benevolent Denominations and Policy-Seeking or Ideological

Parties

In this section, we alter one main assumption of the previous case and that is election-
seeking parties. In this case, we model the parties as having a specific policy preference.
In other words, the parties have a particular tax rate or govermnent spending level that
they prefer prior to the start of the game. Ideology, party interests or candidate interests
may drive this policy preference. This situation is of important historical and
contemporary value, since we do witness that some political parties do have specific
ideological positions on many political and economic issues. A political party may be
driven by an ideology very different from that of the denominations. We can have an
instance where the party’s ideology, supports conservative economic policies or ‘small’
governments and this position may happen to coincide with the interests of religious
denominations on those same issues”. Also, we may have situations where some
business interests influence the preferences of the politicians and cause them to favor
particular policies at given periods in time. We are thus departing from a case of neutral
policy preference towards a specific policy bias. Hence, our assumption in this case is
that parties have prior policy preferences for some unspecified reason and they hope to
adopt that policy as best as they can. But the society is still controlled by the religious
institution in the first stages, in the same way as we had in the previous case. In other
words, even with the ‘ideology’ of the parties, the citizens still belong to one of the two

denominations and conform to the teachings of that denomination.

 

25 Parties may have unique ideological orientations, such as Marxist, Socialist, Labor, Social Democrat
or otherwise, and they happen to coincide with the preferences of religious denominations on specific
issues.

30

Voting Stage

Voters vote for the platform that maximizes their politico-economic utility, according
to:
Vi: viB'viA < VII(TA)_VII(TB) (3,)
Political Competition Stage
This stage is where the new assumption affects the structure of the model. Each party

now chooses its policy in order to maximize the following objective:-

Max —(T,,- Tr“)?
{Tit}

where T [I ideal or ideological position of party. (We note here that the values of

possible T 1: are assumed in such a way that the budget constraint is satisfied in general).

The parties are ‘dictatorial’ in the sense that they choose their ideological positions
regardless of the behavior of all other players in the game. Roemer (2001 , 81) labels such
a behavior as “dogmatic”. The parties do not care about winning and they would “brook
no compromise” (Roemer 2001, 81). Whether their policy is chosen through elections
does not affect their decision-making. This may seem strange, but we are making such

assumptions in order to study a wide variety of scenarios.

Therefore from 17.0. C, T A. = T AA
TBa = TBA
Politico-economic equilibrium: Each party chooses its ideal politico-economic position.

No convergence occurs.

Transition Stage: This stage stays the same as before.

31

Religioug Competition Stage
Since denominations are benevolent, and Since we are considering an indirect,
religion-driven political position, the result is similar as in the previous case.
The denominations locate equidistantly as follows :
Religious Equilibrium
x1. = t/2
x; = 1/2 + r/2

The final tax rates are thus equal to the prior ideological positions of the parties.
Depending on the positions of these tax rates, the winner of the election will be
determined. Adherents of denomination I (liberal) will vote for the party with the higher
‘ideal’ tax rate and the adherents of denomination 2 (conservative) will vote for the party
with the lower ‘ideal’ tax rate. But the voting process is also influenced by the
distribution of the ‘fixed’ policy preferences, vi: v,-B - v“. This means that if
coincidentally, the positions of the parties are located equidistantly in the same way as
that of the denomination, then this ‘bias’ distribution will determine the winner of the
elections. If party A wins the elections (assuming T A ‘< T33, then denomination 2
(conservative) is better off since the tax rate implemented by A is closer to the preferred
tax rate for conservative believers than liberal ones. The liberal denomination is better off
if B wins the elections.

In this scenario, we observe that the ‘dictatorial’ mode of political rule adopted by

parties (or government), restricts the indirect influence that denominations had in case 5a.

In that case, we had as final policies:

TA‘ 2 [70,40

32

TI; 2 b(Y-X)

where, X = cox]. + (1- co)x2‘.

The denominations positions were represented in the decisions of the parties. In this
case, the parties exclude the positions of denominations in choosing the policy level. We
should note that the parties’ ideological positions are determined exogenously. The
denominations therefore have no part in ‘constructing’ government policies. The
denominations however can still influence the result of the election through their
adherents’ voting. The denominations can affect the probability of the winning party

position. If party A’s position is to the far right while B is closer to the middle, then more

will vote for B and thus B wins. If T ,4 I, T 3* are located equidistantly, then denominations

have even less influence since the winner is now determined by v,. Knowing that they did

not play a role in forming the party positions, the denominations can only hope to ask
their adherents to vote for the policy that is closer to their ideal. This also depends on the
fact that the denominations are constrained to being benevolent. If they had more flexible
objectives or ‘doctrine’, then they may be in a better position to exercise influence over
government policies. In short, denominations have little influence over government
policy positions and they are ‘constrained’ by the obligation to be benevolent.
Proposition ;:

If parties become ‘ideologically dictatorial ', the ability of benevolent religious
denominations to influence the political and economic policies (indirectly) is greatly

reduced. The political and economic interests of the denominations is to a great extent

determined by forces beyond their control.

33

This doesn’t mean that denominations will necessarily be worse off under this scenario
than the previous one. It may well be that due to the ideological positions of parties and
the ‘popularity’ distributions for the parties, a denomination may end up better off than in
the previous scenario. But the main issue is that the likelihood of this happening is to a
great extent beyond the realm of the denomination’s influence.

For example, let us have a situation as in figure 1.3: -

 

. I . I . I I . .
T ,4 T 2 M T ,3 T,
Figure 1.3 : Policy-Seeking or Ideological Parties

In the situation in figure 1.3, we have the two denominations locating equidistantly on
the religious spectrum with the corresponding positions on the tax rate spectrum. The
parties locate according to their ideological positions. The conservative denomination is
closer to party A’s position and therefore all the adherents of denomination 2 will vote for

A because of the conformity requirement. Individuals do have an ideal tax preference
given by T," = b(Y-a,-) , for all i. But the assumption of conformity requires that each

adherent ‘replace’ his/her political preferences with that of the denomination. As a result,

they choose their voting option based on the denomination’s preference.

Corollary
If the parties' positions are dijferent, we would now have a ‘ liberal’ and a
‘conservative ’ party. Liberal denominations tend to support ‘Iiberal ' parties and

. . . . . . . . 26
conservative denominations wrll tend to support conservative parties .

 

26 See Empirical Evidence (iii), pp. 56.

34

Regardless of the degree of conservatism of the party, the members of the conservative
denomination always vote for that party. This argument also holds for the liberal party
and denomination.

We should note that in this scenario, parties have equal number of votes corresponding
to the locations of the denominations. If one party has more ‘popularity’ than another, it
wins the elections. If for example A wins, the conservative denomination is better off,
since the position that will be implemented is closer to the denomination’s preferred level
than it was in the first scenario, where the parties converged to the midpoint. If however
B wins, then the conservative denomination is worse off.

These hold: TA. = TAA
T; = T,“
TAA< TBA , TAe< TI;
T2r< Tia
x ,*< x; (equidistant)
Cl.) 02*
Z, a) 22.

1.5c. ‘Dictatorial’ Denominations and Election-Seeking Parties

In this scenario, we change the objectives and intentions of denominations from
benevolence or democratic religious representation, to an ‘authoritarian’, ‘dictatorial’ or
‘dogmatic’ religious authority. In other words, the denomination now has a specific, prior
religious position that it seeks to implement. This position is determined exogenously

from an unspecified doctrine. The denomination ‘bypasses’ the preferences of believers

35

and seeks a specific religious investment level. It chooses the position on the ‘religious
investment ' spectrum that would correspond to its preferred level. It is thus ‘dictating’ its
preferred doctrine to its members. It is also an example of indoctrination by the
denominational authorities. Parties seek winning the elections only.
Voting Stage: Same as in case 1.5a.
Political Competition SLagg
Politico-economic equilibrium :

We follow the same procedure as in 1.5a and the result is as follows :

TA I = b(Y-A9

TB. = b(Y')O

The parties’ political positions converge.
Transition Stagg

Same as before.
Religious Competition Stage

Denominations seek to :
A 2
Max —(x,- x, )
{)9}
The denomination seeks to implement xf. To achieve this objective, it would choose x,

to maximize the above objective.

F.0.C. x,’= x,“

t A

3‘72 =x2

36

Religious @uilibrium

Denominations choose the positions on the ‘religious investment’ spectrum
corresponding to their given ideal positions.
Proposition 3:

If denominations are ‘dictatorial ', they will impose on adherents a level of ‘religious

investment ' corresponding to the preferred level by the religious authority.

Corollag:

If denominations are dictatorial, the religious authorities do not sincerely represent the
preferences of adherents. A liberal denomination and a conservative denomination would
not emerge out of a sincere representation of religious preferences, but out of probable
diflerences between the doctrines and preferences of religious authorities.

Therefore, in this scenario, any apparent difference between adherents of different
denominations in terms of preferred tax rate, government spending or size of government,
will be entirely due to the specific preferences of the religious authorities. Adherents
‘blindly’ follow the teachings of the denomination. Some adherents will find their ideal
position close to the denomination’s ideal position, but in terms of the total welfare of
adherents, it will generally be less than if the denominations were benevolent.
Proposition 3’:

Under a ‘dictatorial’ religious authority, the religious welfare of adherents is
suboptimal.

By being benevolent, the denomination is minimizing the total religious transportation
costs of all adherents and that gives the Pareto optimal result. Any other objective is

suboptimal. From standard linear city models, we know that the locations that achieve a

37

social optimum correspond to ‘A and 3A on a scale of O to 1, as in the case with benevolent
denominations. Any other set of locations will be suboptimal.
This may not be the case if the denomination’s preferences coincide with the optimal

level. But that is of small probability.

These hold: T A. = T 3* (convergence)
T2s< T,’
xIA< sz, x,'< x;
Cla> 62‘
21..) 22..
1.5d. ‘Dictatorial’ Denominations and Policy-Seeking Parties

This scenario differs from the previous one with regard to the party objectives. We now
have denominations that seek to impose their doctrine, and parties that seek to dictate
their ideology or interests. This is probably the worst situation that may face the believers
or citizens in a society, from their point of view. On the one hand, the citizens’ religious
views are not taken into consideration, and on the other hand, they are forced or coerced
into accepting government policies that they may not agree with. Some believers may
find themselves fortunate to have similar preferences to that of the religious authority and
we do find this in authoritarian societies, but the majority of believers will feel alienated
or disregarded. Also, citizens would feel that parties ignore their political and economic

preferences and instead try to impose a single view on them.

38

The Voting Slagg

Citizens ‘vote’ for the party platform that is closer to the policy preferred by their
religious authorities”. When the citizens are now part of a specific denomination, the
religious teachings of that denomination will induce them to favor particular economic
and political policies.
Political Competition Stage

Since parties are ‘dictatorial’, they seek to impose their preferred policy. This is similar
to case 1.5b.

Max —(T,, -T,,")2
{Tit}

The solution is : T AI. = T AA

C A

T ,3 = T B
Politico-economic equilibrium: Each party chooses its ideal politico-economic position.
No convergence occurs.
Transition Stage
This stage stays the same as before.
Religious Competition Stggg

Denominations seek to :

Max —(x,-- x,)2

{xi}

The denomination seeks to implement xjA. To achieve this objective, it would choose x]-

to maximize the above objective.

 

27 The word vote here seems strange because the citizens are not choosing their preferences but that of
others.

39

F..0.C x,*= x,A

t A

x2 = x2
From our example, the tax rates that the adherents are directed to follow are:
T1. = WY" 351’)
T; = b(Y- 3‘2).
Proposition 4

With ‘dictatorial' religious authorities, the ‘ religion ’ that the adherents subscribe to is

primarily determined by the religious preferences of religious authorities.
Proposition 4’

Denominations with strict objectives and intentions are less capable of increasing their
number of followers than denominations with more flexible objectives and intentions.

A benevolent denomination can continuously change the number of followers it has as
circumstances change, while a ‘dogmatic’ denomination is always restricted to a strict

doctrine.

Corollag

The decisions of the religious authorities and the party officials are completely
independent.

In addition to the fact that the outcome is not socially optimal, any distinctively liberal
or conservative denomination is completely due to the religious preferences of the
religious authorities. Likewise, any distinctively liberal or conservative party is fully due

to the politico-economic preferences of the respective parties.

40

We can therefore say that the adherents would ‘vote’ for the tax rate that their religious
authority prefers. If the religious authority prefers high government spending, they would
choose high government spending and if the religious authority favors ‘small’
governments, they would side with ‘small’ governments. This kind of ‘blind’ following
of religious authority is not uncommon in the history of world religions. And even when
they submit to these teachings, they may end up facing different political and economic
policies that are driven by another authority with its own particular interests. However,
one should note that the religious and party authorities are also strictly following given
aims and objectives. A liberal denomination may find out that its given doctrine is
unpopular among the adherent population. But the structure of its objectives forces it to
maintain that position even if it ends up gathering fewer followers. There are many
instances in history where denominations or sects have preferred to stick to their
teachings and have refused to alter their rules because they believe it would be a violation
of doctrine. These denominations have remained static, while those whose objectives are
much more flexible can adjust their doctrine over time as circumstances change and thus
assume a more dynamic course. Likewise, parties cannot adjust their preferred policies to
get more votes and unless their ideology changes over time, they may end up losing the

elections continuously.
. at It: It: a a II: at
Given all the above, we can have many values for x, , x2 , T, , T2 , c, . C2 , TA ,
C i #
T3 , Z, and 22 .
These hold: TA: = T;
t A
TB = TB

TAA< TBA . Tnt< TB.

41

T2*< Tl.
xIA< xza, x,'< x2.
Cl.) 02*
2]..) Z,’

1.5e. Legitimacy and Endorsement

In the cases we have examined so far, the religious authorities have specific religious
Objectives. These objectives can range fiom a sincere representation of believers’
religious preferences to an ‘authoritarian’ implementation of a particular preference held
by the religious authorities. The denomination seeks to implement its religious objective
regardless of the actions of political parties. That is, the position chosen by the
denominations is not influenced by the positions chosen by the parties. These religious
positions chosen by the denominations lead indirectly to political and economic
preferences for the adherents. But this doesn’t mean that the denominations have no
interest in politics or economics. A denomination may choose a level of ‘religious
investment’ because this level may lead to a particular tax rate or government spending
level. But in the structure we used in our cases above, there is a difference between
having a specific religious objective and a specific politico-economic objective. We have
seen that if a denomination has a level of ‘religious investment’ or religiosity that it
prefers, it would implement that level in a dictatorial manner. The denomination moves
before the parties and therefore would just easily implement that particular level of
‘religious investrnent’. But if the denomination has a politico-economic interest or

objective, such as a specific tax rate, it wouldn’t be able to implement it exactly as it is,

42

because parties determine the final tax rate in the political competition stage. This is the
structure we have employed so far in this paper.

In this section, we will consider a slightly different set of assumptions, but with
Significant consequences on the results. Here, we assume that the religious authorities
have a direct interest in politics and economics. This means that the religious authorities
seek a specific type or level of politico-economic policy. This preference may be due to
some ‘secular’ interest on the part of religious authorities in the form of ‘rent-seeking’,
personal agendas or some alliance agreed upon between religious leaders and party
officials. We call it direct simply because the objective will be defined in the form of
politico-economic policy rather than religious doctrine. The main difference in this
scenario is that denominations choose by endorsement. This means that a denomination
supports one of the two positions chosen by the parties. It supports the position closer to
its ideal politico-economic preference. For example, the liberal denomination would
endorse the position of a liberal party and in doing so, directs its adherents to vote for that
specific party. This method of endorsement is common in politics. The endorsement
serves as a form of legitimacy provided by the denomination to the party. This legitimacy
is a certificate of support from the religious authorities to the party and it materializes in
the form of voting. We assume that parties are ideological. In his forthcoming book that
studies religious liberty in the framework of rational choice theory, Gill argues that
legitimacy is a relatively inexpensive method for governments to attain and maintain
office. Gill (1998) remarks that this demand for legitimacy by government creates a

potential for agreement between policymakers and denominations, since, “religion tends

43

to be the primary producer of societal norms and values” (Gill 1998, 52-53). Gill goes on
to examine legitimacy as it plays out in the area of government regulation of religion.

In this section, we seek to formally model legitimacy and endorsement of parties by
denominations and how this structure affects the propositions obtained so far in this
model.

VotingStage

The adherents vote in accordance with the directions or teachings given by the
religious authorities. They vote for the party that is endorsed by their religious leaders.
Political Competition Stage

Since parties are ideological, they solve the same problem as in cases 1.5b and 1.5d.

Max —(T,,- n“)?
{Tit}

Therefore from F. 0. C, T; = T;
T; = TBA (We let TA’<TB')

Politico-economic Equilibrium

The parties choose their ideal ideological positions, given that the budget constraint is
satisfied.
T rarflion Stage

This is the same as before. Adherents take what the denominations teach them and
derive their optimal politico-economic policy. But they end up paying the tax rate of the
winning party.

Using backward induction, the denominations know that the parties choose their ideal

positions in the political competition stage. But the denominations have their own

optimal or ideal position on the politico-economic issue. Thus, they observe the two party
policies and decide to endorse that policy that is closest to their ideal position. We
assume that the parties have the same ‘popularity’ with regards to the ‘fixed’ policy issue.
We can therefore add an endorsement stage.
Endorsement Stage

Denomination j chooses to endorse party A iff:-

- (TX- III-“)2 > - (TI;- If)?
where If corresponds to the preferred tax rate of denomination j.

Suppose we have the following situation below:-

 

T 2 TA T B T [A
Figure 1.4: Religious Endorsement of Parties (a)

In the situation of figure 1.4, the two parties locate according to their ideal positions.
We have a liberal (B) and a conservative (A) party”. The conservative denomination, 2
prefers the position of A to B. It therefore would endorse A. If the conservative
denomination endorses A, it would seek to choose its religious position such that it
maximizes the probability that A wins the elections. After all, the voters are in the first
place believers and take the orders of their religious leaders seriously.

Religious Competition
The objective of the conservative denomination is therefore to maximize the

probability that the conservative party wins. The same goes for the liberal party and

 

28 These positions could be derived from a benevolent, sincere representation of adherent preferences, or
could be due to some ‘dogmatic’ stance by the authorities. The origin of the denomination’s position is not
important in this particular scenario. What is important is that denominations have distinct preferences.

45

denomination. The situation in practical terms is where a liberal party would seek a
coalition with the liberal denomination in which the party gets the votes (legitimacy) it
needs and the denomination has a party that is more attentive to its needs than the other

(conservative) party.

Conservative Denomination, 2: Max (Probability that A wins)
{12}
Liberal Denomination, I .' Max (Probability that B wins)

{x,}
Religious Equilibrium
The positions of both denominations will converge to .the middle (or median) of the
‘religious investment’ spectrum.

Proof

After each denomination has decided which party to endorse or legitimize, they are
now interested as per the ‘coalition agreement’, in gathering as many voters as they can
for that party. The only tool the denominations have at hand is the religious variable. By
altering this position, they can influence the final votes. If the liberal denomination wants
the liberal party to win, it would simply seek to maximize the probability that that party
wins. More specifically, the denomination would seek to maximize the number of
adherents it has. This is because if all adherents are conforming completely to the
teachings and rules of the denomination, by maximizing the number of adherents, they
are essentially maximizing the number of those who would vote for the liberal party. To
maximize the number of adherents, the liberal denomination would try to move its

religious position to the right to get more votes for the liberal denomination. As a

46

rational, strategic response, the conservative denomination would move to the left. The
Nash equilibrium is at the median or midpoint of the religious spectrum.

Therefore, the denominations would converge to the midpoint on the religious
spectrum and both sets of adherents are now essentially facing the same religious rules or
standards. There is a tendency towards increased moderation by both denominations in
the hope of attracting more adherents. They are aiming for a larger audience. But they
face a trade-off. By trying to increase adherents for the sake of a particular politico-
economic objective, they are ‘sacrificing’ religion. They try to achieve their political-
economic objectives at the expense of their religious teachings and doctrine.

Mathematically, the denominations maximize the number of adherents as follows:

Conservative Denomination: Max (co) > 1/2
{x2}
Liberal Denomination: Max (I-co) > 1/2

{x,}
Nash Equilibrium (convergence to the middle): to = 1/2.

Proposition 5

In a framework of endorsement strategies, the liberal denomination will m to
endorse the position of the liberal denomination and the conservative denomination {211$
to endorse the position of the conservative party.

If the denominations are located symmetrically on the tax rate spectrum, then the
liberal denomination will M endorse the liberal party and the conservative
denomination will glw_ay§ endorse the conservative party.

We can have situations were a liberal denomination endorses a party with relatively

opposing viewpoints, as shown below:

47

 

T, T; T; T,
Figure 1.5: Religious Endorsement of Parties (b)

In the situation above, party A is extremely conservative to the extent that it is Optimal
for the conservative denomination, 2, to endorse the liberal party (B) instead. In this case,
party B is certain to win. Since we can have many possibilities, we can restrict ourselves
to cases where the denominations endorse parties of similar viewpoints or ideological
orientations, thus the term ‘tend’.

However, if we allow parties to react strategically to this possibility (which we haven’t
done here because parties are ‘dogrnatic’), then they would move their positions so as to
get at least one endorsement. This means A moves to the right until B cannot get the
endorsement of denomination 2 anymore. B would move to the left and stop at the point
beyond which it loses its endorsement from denomination I .

Also, if we give complete fi'eedom to both sides such that they can react strategically to
any movement by any other side, then we will have both denominations and parties
locating at the midpoint.

Proposition 5’

As denominations become directly interested in political and economic policies, their
positions on the religious spectrum will converge to the median or midpoint.

This is because they are interested primarily in a certain political objective, and to
achieve it through endorsement, they tend to ‘moderate’ their religious positions so as to

increase support for that endorsed policy. This is where politics drives religion.

48

C_0L0_113_1Y

If denominations are interested in political rather than religious objectives, they will
mil to ‘sacrifice’ the religious interests of their adherents to achieve their ‘personal '
political-economic agendas.

Also, the religious diflerentiation between denominations is reduced until eventually
there is no notable difference between them. We will have a liberal and a conservative
party, but there are no longer a liberal and a conservative denomination.

Religion and believers in general, are the ‘losers’. As we have seen, any non-
equidistant location of religious positions is sub-optimal. As a result, the location in this
scenario is sub-optimal.

We have the following situation:

T; = 2:,“

TB" = T,“

TAA< TBA ’ TA*< TB.

T2A< T,A, T,* = T2*[=b(Y-)O (in the example we have)]

t *

x, = x2
01:02
A A t *
Z] > 22 . Z1 =22
This scenario has raised a very important point and that is about objectives and
intentions. We have seen that with strict religious objectives, there is a notable

differentiation between denominations in their religious positions. With politico-

economic objectives, this differentiation gradually disappears. The explanation that the

49

model proposes is that of intentions. If a denomination is Sincerely representing the
preferences of their adherents, it will assume a distinct religious identity. But if the
denominations are instead interested in achieving some political objective through a
coalition with parties, they risk losing this religious identity. The coalition and
endorsement that occurs, makes denomination solely interested in the election outcome
and not the religious stance of its members. Political parties exploit the power that
religious institutions have by asking for their support. And in providing this support, the
denominations lose in religious terms.

This situation is not uncommon in history. We have witnessed instances where
denominations have worked hard for certain parties or candidates to win elections or gain
popular support. Denominations alter religious teachings in order to attract more
adherents who are later ordered to vote for the endorsed politician. One can even say that
politics may have benefited from this strategy in trying to reduce the power that religious
institutions have. Providing material or secular incentives and rewards to religious leaders
to gain their endorsement can do this. If the religious leaders become ‘tempted’ by these
rewards, they will accept to garner support for that politician. They do this by simply
changing their teachings to attract more followers. This can be in the form of more
flexible rules, less spiritual requirements or less demand for charity, if it is a conservative
denomination. A liberal denomination would in contrast increase the strictness of the
rules and require more from their adherents. As a result, this ‘strategic change in religion’

by the two denominations leads to increased similarity between them.

50

1.6 Empirical Evidence

In this section, we survey some of the empirical work that has been done on

denominations’ religious and political behavior in order to assess the practical validity of
our propositions. This section will provide an overview of empirical studies that relate to
the assumptions and results in this paper.
(i)- Relig'mus Preferences: With regards to the assumptions made in the model regarding
religiosity and religious investment, we labeled adherents who prefer low religiosity
levels as liberal while those who prefer high religiosity levels as conservative. In his
survey of the literature on Economics of Religion, Iannaccone (1998) remarks that most
measures of religious involvement such as voluntary contributions or participation in
religious activities are positively related to the level of religiosity in the denomination. He
states, “the members of liberal Protestant denominations contribute a relatively small
proportion of their income to churches (around 1.5 percent), whereas the members of
conservative Protestant denominations, such as the Southern Baptists and the Assemblies
of God, contribute significantly more (between 2 percent and 4 percent), and Mormon
contributions average 6 percent of income” (Iannaccone 1998).

Furthermore, Iannaccone (1998) states that religious involvement in terms of, “rates of
church attendance, follow a similar pattern, with liberal Protestant denominations ranking
lowest, conservative Protestants attending more, and sect members, such as Mormons and
Jehovah’s Witnesses, attending still more (Dean Hoge and Fennggang Yang 1994;
Iannaccone 1992, 1994)” (Iannaccone 1998).

In a statistical survey on the determinants of religious participation, Iannaccone

comments that, “members of conservative and sectarian denominations attend and give

51

much more than members of liberal denominations even after controlling for
socioeconomic differences”. (Iannaccone 1998)

Hunter (1991) studies these two ‘groups’, liberal and conservative, under the titles
progressivism and orthodoxy respectively. He states that, “it nearly goes without saying
that those who embrace the orthodox impulse are almost always cultural conservatives,
while those who embrace progressivist moral assumptions tend toward a liberal or
libertarian social agenda” (Hunter 1991, pp.46).

Therefore, our assumption regarding religious preferences as locating progressively
under two distinct banners, liberal and conservative, is consistent with a substantial part
of the theoretical and empirical literature on religion and politics.

(ii)- Rflion and Economics: In this section, we review some views regarding the
relationship of religion to economics. We recall that the model derives a relationship
between religious preferences and economic preferences. In particular, individuals with
liberal religious preferences will be induced by the denomination to favor liberal
economic policies in the form of higher taxes, more government spending and ‘bigger’
governments. Such adherents will prefer more government spending on programs such as
social security, health care and public education. In contrast, the model postulates that
conservative adherents will favor low taxes, low spending on welfare projects and less
government intervention in their economic lives. 2")We argue that these views are strictly
a result of the budgetary choices faced by believers when trying to allocate their

resources between religious spending and secular obligations. Strict believers want to

 

29 This does not mean that conservative believers are opposed to welfare programs. They are specifically
opposed to government provision of these services using tax revenue. This means they prefer that such
spending be done by their religious institutions using religious donations of adherents.

52

give less to the ‘secular’ government in order that they can give more to their religious
authorities.

But the question that arises is whether we do observe these tendencies in real life. Do
liberal religious views translate into liberal economic views in the real world? Do
conservatives reject socialist policies by governments? Do they favor a laissez-faire
approach by governments? The answer is that there are mixed views and results on these
issues. Some authors support the propositions above and some provide evidence to
disprove them. One should note however that our model is explaining these phenomena
using a single dimension of analysis, namely religiosity. Other authors explain their
different conclusions using alternative dimensions.

Two prominent sociologists of religion, Wuthnow (1988) and Hunter (1991), hold the
position that religious prefer'ences affect the views of believers and denominations on
economic issues. According to Davis and Robinson (1996), who object to the conclusions
of Wuthnow (1988) and Hunter (1991), state that, “according to Wuthnow (pp. 132, 219-
23, 239), religious liberals take politically liberal positions on a wide variety of
contemporary issues, including .. ., and economic justice, while religious conservatives
take politically conservative stances”.

They state that,
Wuthnow (1988, p.114) sees religiously conservative “special purpose
groups” such as the Moral Majority and the Christian Coalition as having
arisen to combat the growing role of government in guaranteeing equal rights
for women and minority groups and in providing welfare and other forms of
assistance to the poor. Wuthnow (p. 248) notes that religiously conservative
leaders, such as evangelical economist George Gilder, Moral Majority founder
Jerry Falwell, and Christian Coalition founder Pat Robertson, have voiced

strong opposition to redistributional efforts by the government. Through
analysis of a sample of Americans interviewed in 1984, Wuthnow (pp. 219-

53

23) finds that religious conservatives are more opposed than religious liberals
to government spending on social programs. (Davis and Robinson, 1996)

Hunter (1991 , pp.111) argues that the notion of freedom among orthodox believers also
refers to freedom in “economic self-determinations” and as a result, these believers are
proponents of “‘free’ enterprise”. He argues that evangelicals “trace the [above]
relationship to the Old Testament land laws that linked private property to the freedom
from state coercion, especially from taxation”. (Hunter 1991, pp.111)

Hunter (1991, pp.96-97) observes in a 1987 study of denominational leaders that
orthodox religious leaders prefer economic growth as a way of reducing poverty and also
tend to reject using redistribution of wealth to solve the problem.

As such, Wuthnow and Hunter are of the view that religious preferences affect the
economic preferences of believers and religious authorities. Our model also arrives at
similar conclusions on these issues. The explanations presented for this pattern are
however different. Wuthnow and Hunter explain these attitudes as a result of particular
religious doctrines held by the different denomination members. Hunter (1991, pp.49)
argues that, “the culture war emerges over fundamentally different conceptions of moral
authority, over different ideas and beliefs about truth, the good, obligation to one
another”. Thus, they explain this behavior in a Weberian framework. Although our model
explains the economic preference as an expression of religious preferences, it specifically
uses the budget constraint and allocation of resources as the key concepts. The reason
that a conservative prefers low government spending or intervention is not because
his/her doctrine praises the laws of capitalism and free market. It is because the level of

commitment or involvement or strictness the believer prefers ‘forces’ the believer to give

54

a substantial amount of his/her resources to the denomination and as such, he/she are
inclined to favor giving less to the ‘secular’ government. This relationship is due to the
impact of the budget constraint on all believers. The economic reality puts certain
restrictions on the economic preferences of individuals. Therefore, both the religious
doctrine and the economic reality facing individuals, determine their political, economic
and religious preferences.

As we noted before, some studies have obtained very different results and observations
regarding the effect of religion on economic attitudes. Davis and Robinson (1996), in a
statistical survey, show that orthodox adherents do not necessarily hold conservative
economic viewpoints. They state that, “the orthodox are more liberal on issues of
economic inequality” (Davis and Robinson, 1996). But the reason for this observation
according to them, is “[that] the tendency of the religiously orthodox to draw adherents
from disadvantaged groups makes them more liberal on economic issues”(Davis and
Robinson, 1996). In our model, income inequality is not present and therefore we can
ignore this effect. Our result holds in a situation of income equality. If a majority of
conservatives are poor, this means that they care about welfare programs and thus they
would support more government intervention. They are unable to provide to the
denomination their preferred levels of religious investments. Also, Davis and Robinson
(1996) do find that orthodox adherents believe the government welfare spending is too
high. As such, even their results don’t support their hypothesis fully.

Furthermore, there are no final conclusions on this matter and different researchers

have differing views.

55

(iii)— Religion and Politics: In this section, we review the empirical work done on the
political positions and affiliations of denominations and adherents in the area of US
politics. In our model, and specifically under the sections that deal with ideological
parties, we argued that liberal denominations tend to support liberal parties, while
conservative denominations tend to support conservative parties. We also showed the
conditions under which a denomination may endorse the position of a party with a
‘conflicting’ ideology (see section 1.5e).

In his study of religion and politics in the US, Layman (1997) states that, “ there is
already a good deal of evidence that political activists arid voters with orthodox beliefs
and affiliations tend to support the Republican Party while the Democratic Party draws its
activists and voters disproportionately from the ranks of religious liberals and secularists
(Green, Guth, and Fraser 1991; Green et al. 1996; Guth and Green 1986, 1987; Kellstedt,
Smidt, and Kellstedt 1991; Miller and Wattenberg 1984; Rozell and Wilcox 1995;
Wilcox 1992)”. Laymen (1997) states that, “as the cultural tensions between America’s
religious and secular populations grow, the political impact of religious commitment also
should grow, with strongly committed individuals becoming more Republican relative to
their less religious counterparts”. Furthermore, Layman (1997) states, “highly religious
individuals have consistently been more likely than their religious counterparts to identify
with the GOP, ...”.

To explain the changing political attitudes of Liberal Protestants from supporting the
Republican Party to supporting the Democratic Party, Leege (1993) cites the decline in
church attendance by liberal Protestants as a possible cause. This means that if attendance

falls, in the fi'amework of our model, the religiosity of adherents may have fallen. The

56

liberal Protestants are more liberal than before, and as such they are more likely to vote
for the Democratic Party than the Republican Party.

These results indicate that liberal denominations do indeed endorse liberal party
positions (Democratic), while conservative denominations endorse conservative party
policies (Republican).

1.7 Conclusion

So far, we have shown that with a budget constraint facing individuals, denominations
that act in the interests of their adherents will differentiate themselves from other
denominations and as a result, will induce their members to favor particular political and
economic policies. The budget constraint facing adherents is also facing the
denomination since all it does is sincerely represent the interests of its members. But if
the denominations are interested in some other objective such as maximizing the number
of adherents, they will converge to the same position on the religious spectrum and
choose the same optimal ‘religious investment’ level. In that case, no differentiation
occurs and therefore both denominations induce their adherents to favor the same
political and economic policies. There is no difference in preferred tax rates, private
consumption and preferred government size. We can deduce from these observations,
that if denominations care less and less about their adherents’ preferences, any notable
differentiation between denominations and among members of different denominations
will vanish gradually. A more ambitious deduction is that with a decreasing benevolence
by denominational authorities, there would be corresponding decreasing differences

between denominations until eventually denominations lose their unique identities and

57

their impact on the domain of politics and economics. With more benevolence come
more differentiation and more power in influencing political and economic policies.

This model has sought to explain the reason why liberal denominations favor higher
tax rates, more government spending and ‘bigger’ governments, while conservative
denominations favor lower tax rates, less government spending and ‘smaller’
governments. We have shown that with adherents facing a budget constraint, if we
approximate religiosity in terms of ‘religious investment’, benevolent denominations who
preach low levels of religious investment and low religiosity, will induce their believers
to favor high taxes, government spending and ‘bigger’ governments. The opposite holds
for conservative denominations. These factors together influence the specific party that a
denomination decides to endorse.

The model has sought to explain a real world phenomenon using simple concepts.
Religious preferences, budget constraints, and denominational objectives explain the

different religious, economic and political decisions taken by all players in the society.

58

Chapter 2

Religious Charities and Government Funding

2.1 Introduction

On 29th January, 2001, the White House Office of Faith-Based and Community
Initiatives was established with the declared objective of creating a partnership with faith-
based agencies in order to achieve common goals in the field of social service provision.
This development raised severe controversy with regards to its legality under the laws of
the Constitution. It also initiated an extensive debate from all interested parties on the
effectiveness of faith-based social service providers in relation to other social service
providers, and whether they possess unique advantages that warrant any special treatment
from the government. This debate will continue and even intensify as the program
progresses. Religious denominations and their affiliated faith-based agencies will be in
continuous competition against each other to obtain government funding in order to
achieve their specific interests. Political parties now have a useful tool that they can refer
to in times of elections when they are in need of legitimacy or endorsements. It seems
therefore that without careful examination, such an initiative may not necessarily be a
productive partnership or lead to the desired common goals that many would expect.

The purpose of this paper is to examine some aspects of the relationship between
religious nonprofits and the government. The paper presents an economic model that
describes this relationship in order to give us an insight into the different motives of these
institutions and the outcomes that arise as a result of their interaction. We study this

relationship by investigating the impact of government funding on religious charities and

59

on religion in the society as a whole. We model a government that awards funds to
religious charities or denominations30 so that these funds are used for the production and
distribution of a social good or service. We provide a rationale for why the government
chooses to award the funds to a religious charity in particular. A unique nature of
religious charities allows them to have an advantage over secular nonprofits in the
production of the social good. Based on this formulation, the government decides which
religious denomination or denominations to award the funds to according to the specified
government objective. We argue that religious charities will intentionally use part of
these funds for proselytizing purposes3 1. As a result, the religious preferences of believers
in the population are altered and this alteration leads to a change in the relative powers of
the different denominations and the religious nature of the society as a whole. The fact
that a particular denomination is given the rights to provide the social service means that
the government is intentionally or unintentionally giving one sect an advantage over
others. This is mainly because the ‘award-winning’ sect will use the funds to pursue its
specific religious causes. Any law passed by the government to ensure the complete
secular nature of the service is bound to fail under instances of imperfect or costly
monitoring. Also, requiring the religious charity to exclude any religious content in the
provision of the social service may mean that the religious charity will have to act as a
secular producer and this nullifies their efficiency advantage which was the reason they

were given the award initially.

 

30 In this model, the terms religious charities, faith-based agencies and denominations will be used
interchangeably.

3' Proselytizing can still occur even if denominations do not consciously pursue it. This is usually in the
form of externality effects arising fi'om the existing religious and cultural environment. The different
methods through which such effects occur will be discussed later in the paper.

60

In instances where the government provides equal grants to all types of denominations,
this is shown to still have far-reaching consequences on religion and believers.
Furthermore, the model examines the instances where particular denominations may
decide to discriminate in the provision of the social service by restricting their services
based on the religiosity or religious affiliation of the social service recipients.

The model derives results that have important policy implications for present and
fiiture programs by the government to fund religious and secular charities. The paper
raises many significant issues that are sometimes overlooked by many parties involved in
the debate. In section 2.2, we discuss the theoretical and empirical literature on nonprofits
and faith-based agencies. We later explain the model in detail in section 2.3 and derive

results in section 2.4 under different assumptions. We then conclude in section 2.5.
2.2 Literature Review

The literature on the economics of nonprofits and altruistic behavior has developed
substantially over the years in the areas of economic theory and empirical research.
According to Rose-Ackerman (1996), economists are increasingly becoming interested in
exploring areas of related interest that were originally thought to be outside the realm of
economic inquiry. Rose— Ackerman (1996) states that many theories have been developed
to explain the motives behind charitable donations, ranging from donations driven by a
sense of commitment as in Sen (1977) and Sugden (1984), to donations driven by the
intention of signaling one’s wealth to others as in Glazer and Konrad (1996). Andreoni
(1990) also presented another explanation for altruistic behavior based on the desire to

gain utility from the act of giving, referred to as the warm-glow effect.

61

Additional literature has focused on the economic behavior of nonprofit institutions in
particular. Rose-Ackerman (1996) presents the theories developed that aim to explain the
functions of nonprofits. One function of such organizations is to provide a channel for
private donations, as many donors may trust nonprofits more than other types of
institutions, such as for-profit agencies. Also, nonprofits may have emerged as a response
to information problems that donors are facing as in the works of James and Rose-
Ackerman (1986), Steinberg and Bradford H. Gray (1993) and Weisbrod (1989).
Furthermore, “nonprofits may provide a more diverse collection of services than is
possible in the public sector” (Rose-Ackerman, 1996).

The empirical economics literature has mainly sought to test the hypotheses proposed
by several theorists. Much of the empirical work has focused on testing the hypothesis of
crowding out of private contributions by public spending with results ranging from
complete crowding out as in Roberts (1984), Sugden (1982) and Warr (1982) to partial
crowding out as in Comes and Sandler (1984, 1994), Steinberg (1987) and Andreoni
(1989, 1990). With regards to the effectiveness of nonprofits, Rose-Ackerman states that
initial studies such as Clarkson and Martin (1980) concluded that for-profits were more
efficient than nonprofits due to the advantage that the profit motive presents. Others such
as Hawes and Phillips (1986), Weisbrod (1988, 1994) and Aaranson et al (1994)
concluded, using data from nursing homes, that the lack of the profit incentive served as
an advantage for nonprofits over competing agencies. Also, Rose-Ackerman (1996)
argues that some studies such as Kagan (1991) and Kisker et a1 (1991) conclude that for-

profits provide services at a lower cost but also at a lower quality. As such, “when

62

nonprofits provide higher quality services, they may also charge higher prices to
compensate, [thus] reducing their advantage” (Rose-Ackerman, 1996).

An essential component of the nonprofit sector is religious charities and faith-based
agencies”. The vast majority of the theoretical literature on the economics of nonprofits
has focused on the economic choices of private donors, the economic functions and
decisions of nonprofit agencies, and the impact of income and wealth redistribution
policies by the public sector on the decisions of donors and nonprofits. There has not
been any formal examination of the impact of government policies on the religious state
of society, both on the individual and group level. In addition, the models have treated all
types of nonprofits identically, while none has sought to examine religious nonprofits in
particular.

In this paper, we formally examine the relationship between religious nonprofits and
the public sector with the primary aim of studying the impact of this relationship on
religion in society. Also, we specifically investigate the religious nature of religious
charities in order to study their uniqueness relative to other types of social service
providers. Reinikka and Svensson (2003) develop a simple model to highlight the
differences between religious nonprofits and for-profit service providers. In their model,
managers of nonprofit agencies will seek to hire altruistic workers that are willing to
work at lower wages and produce the service at a lower cost, thus allowing the agency to
offer the service cheaper than it is offered by for-profit agencies. In our model, instead of
assuming that managers intentionally seek to employ workers who are willing to accept

lower wages, we endogenize workers’ wages, and are able to derive the proposition that

 

32 We do not differentiate between congregations and faith-based organizations as in Wuthnow et al
(2004). All nonprofit agencies that have a distinctly religious identity fit into our model.

63

religious nonprofits are more efficient than secular or non-religious nonprofits in the
production of social services. This essential result will then allow us to study the impact
of government funding of religious or faith-based agencies on all parties involved.

The majority of empirical work on religious nonprofits has been in the fields of
sociology of religion and religious studies. These empirical studies have aimed to study
the unique nature of religious charities and also to present comparative analysis with
secular or for-profit organizations. Smith and Sosin (2001) examine the general
characteristics of faith-based agencies in a series of interviews with faith-based officials.
They classified faith-related agencies in terms of source of funding, link to religious
authority or congregation and the underlying religious culture in which different agencies
exist. Their study provided an insight into the degree of linkage between faith-related
agencies and the doctrine they subscribe to, and also policy implications with regards to
possible funding of such agencies by the government (Smith and Sosin, 2001). Ebaugh et
al (2003) also discuss other authors who have classified the different characteristics of
faith-based agencies, such as Unruh (2001) and J eavons (1998).

In their study of faith-based agencies in Texas, where the concept of the F aith-Based
Initiative was first implemented, Ebaugh et a1 (2003) argue that faith-based agencies do
not differ fi'om secular agencies in the types of services that they provide to their
recipients. However, Ebaugh et al (2003) concluded that faith-based agencies are unique
in that in addition to providing the social service, they also convey a collection of
religious services simultaneously. This led to their conclusion tha , “‘organizational
culture’ significantly differentiates religious from secular agencies, with the former

supporting a culture thoroughly imbued with religious values in terms of staff interaction

64

with clients” (Ebaugh et al, 2003). With regards to specific social programs, many studies
have arrived at a variety of results concerning the effectiveness of faith-based agencies.
Desmond and Maddux (1981) studied the success of religious programs aimed at drug
addicts in San Antonio and concluded that the programs had a positive impact relative to
other programs. Other studies include Berrien, McRoberts and Winship (2000) where
they studied the effect of clergy-police partnership on crime rates in Boston. (Wuthnow et
al, 2004). In his book, “When Sacred and Secular Mix: Religious Nonprofit
Organizations and Public Money” (1996), Stephen Monsma presented a comprehensive
historical and empirical study of the relationship between nonprofits and the public
sector. He examined the history of the relationship between these two sectors and studied
this relationship in the context of different religious denominations. These studies and
others provide useful insights as to the future of the F aith-Based Initiative and other
programs that involve some form of economic alliance between the government and
faith-based agencies.

Although extensive applied research has been done on religious nonprofits, little has
been said about the impact that government funding of these nonprofits has on religion
and believers in general. The literature has mainly dealt with the characteristics of faith-
based agencies and their effectiveness in relation to other service providers. We will
explain the unique nature of faith-based agencies or religious charities and provide an
economic rationale for their efficiency. The paper fills an important gap in the literature
by explaining how denominations and their affiliated faith-based charities compete with
each other for adherents, for public funds and-ultimately-for religious hegemony.

Ultimately, the paper predicts the impact on religion and believers of any relationship that

65

might exist between these agencies and the government sector. Several empirical results

that already exist in the literature will be shown to support our propositions.
2.3 The Model

The paper makes use of the framework of spatial location models. We follow, to some
extent, the procedure of Barros and Garoupa (2002) in which they use a spatial location
model to formulate a theory of church strictness. In particular, they use a modified
Hotelling model (See Hotelling (1929)). Also, denominations or charities compete in a
Nash environment. The games of strategic interaction in this model will utilize the tools
of game theory and in particular, the concept of ‘backward induction’, to derive the sub-
game perfect Nash equilibrium results.

Similar to Barros and Garoupa (2002), we study a situation where believers locate on
the ‘religious spectrum’ and choose that denomination that maximizes their utility. This
follows the Hotelling model of consumers grouping near different sellers, according to
their preferences. Furthermore, we model the choices of individuals in similar fashion to
Becker’s theory of household production. See Becker (1965, 1974), Becker and Murphy
(1988), and Stigler and Becker (1977). In our case, we have individuals’ resources
allocated between religious consumption (in the form of religious investment) and secular
consumption (consumption of private commodities)33 .

The main components of our model include the government, the adherent population
and the religious charities or nonprofits. The government consists of the decision makers

that make choices with regards to public revenue and public expenditure. The religious

 

33 In this model, it is charities and the government that actually choose the equilibrium resource
allocations. Charities or denominations choose a religious investment level according to some given
objective and all adherents conform to this level.

66

charities or denominations comprise the religious authorities that direct and coordinate
the institutions’ plans and services, and the clergy that are essentially the religious
workers performing their duties in accordance with the rules and regulations set by the
religious authorities. The adherent population consists of the total number of believers in
the society.

Government: - The government has hands that it intends to award to a nonprofit agency.

We assume that the funds are obtained exogenously. The purpose of the funds is the

production of a social good or service, Z . This service or good is assumed to be different

from the religious ‘club’ good, R ,that is already produced by religious denominations.

The government may opt to award the funds to a religious nonprofit or a secular
nonprofit or may have to choose between two competing religious nonprofits. We assume
that the government seeks to award the funds to the most efficient nonprofit agency. In
particular, the government awards the funds to the nonprofit agency that can produce the
service with the least cost of production.

Denomination: - Denomination or charity k is a religious institution that cares for the
welfare of a community of adherents. The religious authorities in charge of the
denomination are assumed to have a benevolent nature and as such, seek to maximize the
total welfare of their adherents. For simplicity purposes, we have two denominations in
the model, A andB. In addition to the task of providing its specific religious ‘club’

good to its adherents, a denomination may also be given the responsibility of providing

the social good or service, Z . This happens when the government awards the funds or

social service contract to the denomination.

67

Religious Clergy: - They represent the staff working in the religious organization. In

other words, they are the religious producers or workers. These workers are employed by

the charity to produce the religious ‘club’ good, Rk. But in addition to producing the

religious good, they will also be assigned the task of providing the social service, Z , in
the event that the charity they work for is awarded the contract by the government. The
objective of the religious clergy is to maximize their utility by preaching their doctrine to
a larger community of adherents. Their utility increases as the size of the community
increases. In other words, the amount of utility they derive is directly proportional to the
amount of preaching they can accomplish.
Believers/Adherents: - An individual i is a believer who belongs to a particular
denomination. There are N believers in the population. In an extension to the model, we
consider a set of non-believers or pure secularists, n, in addition to the N believers. The
utility of a believer i in denomination k is given by:

There are two main components of this utility, namely the religious utility and the

secular utility. The religious utility is given by:
2
fl R. - s (r. - a.- )

In the above function, R, is the religious ‘club’ good that is exogenously provided by

denomination/C. The parameter ,8 represents the weight placed on religious

consumption in the population. It is the same for all individuals“. We can regard it as an

 

34 We can have fl vary by individual, although this will not affect the results of the model. However, a
value of 0 may indicate no preference for religious utility which may indicate that the individual is an
atheist or an agnostic. This could be a way of structuring the preferences of these groups in the population.

68

indicator of the ‘power’ of religious institutions or religion in the society, as evidenced by
its effect on the utility function of all individuals. This power may represent the
prominence of religious institutions and their impact on society in general. This

parameter is not related to the individual-specific preference of ‘religious investment’,

given by al.. The religious parameter a, represents the ideal position of individual i on

‘religious investment’. A principle assumption of our model is that preferences can be
represented by a unidimensional variable, in the form of ‘religious investment’. All
individuals in the population have an ideal preference regarding the ‘amount of religious
investment’ they are willing to undertake in their religious life. This religious investment
may refer to the preferred amount of money each individual is willing to contribute to
religious institutions, or the amount of time the person prefers to contribute in assisting in
the social and religious activities of the religious institution he/she belongs to”. This
parameter describes the level of religion preferred by individuals in terms of measurable
variables such as monetary contributions or voluntary assistance in church activities.
Thus, we can safely assume that these variables serve as proxies for the religiosity of
individuals. As such, we can construct a spectrum of religiosity based on ideal religious
investment preferences, starting from 0 (minimum religiosity) to 1(maximum religiosity).

In his survey of the literature on Economics of Religion, Iannaccone (1998) remarks
that most measures of religious involvement such as voluntary contributions or
participation in religious activities are positively related to the level of religiosity in the

denomination. For example, he states, “[that] the members of liberal Protestant

 

35 The ‘value’ of time may be in the form of the opportunity cost in terms of productive secular activities
such as working, studying, etc.

69

denominations contribute a relatively small proportion of their income to churches
(around 1.5 percent), whereas the members of conservative Protestant denominations,
such as the Southern Baptists and the Assemblies of God, contribute significantly more
(between 2 percent and 4 percent), and Mormon contributions average 6 percent of
income” (Iannaccone 1998).

Furthermore, Iannaccone (1998) states that religious involvement in terms of, “rates of
church attendance, follow a similar pattern, with liberal Protestant denominations ranking
lowest, conservative Protestants attending more, and sect members, such as Mormons and
Jehovah’s Witnesses, attending still more (Dean Hoge and Fennggang Yang 1994;
Iannaccone 1992, 1994)” (Iannaccone 1998).

In a statistical survey on the determinants of religious participation, Iannaccone
comments that, “members of conservative and sectarian denominations attend and give
much more than members of liberal denominations even after controlling for
socioeconomic differences”. (Iannaccone 1998) Thus, for expositional purposes, we will
label the two denominations A and B as liberal and conservative respectively. The
conservative or stricter denomination consists of the adherents that prefer high levels of
religious investment while the liberal or less strict denomination consists of those that
prefer low levels of religious investment. This distinction is made to allow us to examine

the policy implications of our model.
The level of religious investment chosen by denomination k isrk. We assume that

denominations can move on the religious spectrum at no cost. We denote the cost
incurred by individuals as a result of their obligation to conform to the teachings or rules

of the denomination, byS. It is the ‘price’ of sacrifice. This value is the same for all

70

individuals in the society. From the form of this utility, we can infer that the individual
loses utility as a result of conformism to the denomination’s rules, but also gets utility

from giving his/her preferred investment level. As such, the total effect is given in the

quadratic form, S (I; -— a, )2.

The secular utility is given by:
(1 _ fl){uik (Zk’Xi)}
In the above function, (I - fl ) refers to the weight that society places on secular

consumption in the population. Z k is the social good or service that may be produced by

the denomination if awarded the funds and X, is the private consumption of individual
i after investing in religion.

The initial budget equation is given by: Y = X I. + al.. The actual budget equation
is Y = X, + I‘k .

Hence, the composite utility function will be as follows:

U, = ,6{Rk-s(rk_a,)2}+(1- e){n,(z,,x,.)} (I)

Where i6 {I,...,N};k E {A,B}.

Stages of the Game: - The sequence of events in the model is as follows: -

 

A \

V I

Religious Competition/ Religious Provision of Social Service Religious Competition/
Original Preferences Revised Preferences

Figure 2.]: Stages of the Game

71

2.3a. Stage 1: Religious Competition/Original Preferences

W

In this stage, the religious preferences of individuals in the population are formed and
denominations compete in a Nash environment on this spectrum of preferences. To study
this gradual formation of religious preferences, we employ an approach similar to the
religious human capital method used by Smith and Sawkins (2003). This formation of
preferences represents a long period of learning by the individual in his/her interactions
in the society. It also represents an accumulation of parental investment decisions over

time, in addition to the cumulative influence of the social and cultural environment. The
outcome of this long process is given by al.. We assume that the individual has been

primarily exposed to the influence of one denomination only. This can be due to the
religion of the parents and the circle of close relatives and friends, or it can be due to the
cultural and political structure of the country where the individual lives. This limited
exposure to the teachings of a particular denomination has shaped the individual’s
preferences over time. The exposure to teachings of other denominations is assumed to be

negligible. We let this relationship be given by the following religious production

function:al. = g [AH/L “13]”, which then becomes: a, = g [APE] since

Ask : The influence of denomination k on individual i. This is due to direct effects

such as preaching, learning, advertising, training, etc. The direct effects of a

 

36 k and —k refer to the two rival denominations in the model. We will later label them as A and B.

72

denomination on an individual’s religious preferences are due to the personal effort of the
individual in the form of time invested in religious consumption, monetary contributions
that provide religious utility and voluntary participation in religious activities. We

construct the direct effect as a production process given by:

AI], = 1k ([1911]) where,
L, : time devoted by individual i to denomination k.

\I" k : all other human capital contributing to religious investment or education.

. . 2
62., when, /6‘I’,,a i, /6Lka‘I’k _>_ 0
These mean that as more time and human capital investment are devoted to

denomination k ’s teachings, the individual becomes more committed to its teachings

over time. We should note that A.

k increases if we are considering the stricter or

conservative denomination since with more time and education, the believer prefers a

higher ideal religious investment level. In contrast, if we are considering the liberal or
less strict denomination, 1k decreases with more time and education. But this increase
or decrease must be constrained because, for instance, it is not plausible to have a
believer become an atheist (a, = 0) if exposed to a substantial level of liberal
teachings. Changes in religious preferences are usually small and slow”. A different but

useful formulation would be to assume that all individuals start with 1k = 0 and increase

 

37 To make the concept more understandable, one can imagine that all individuals start at the midpoint of
the spectrum and move left or right depending on the teachings they are influenced by.

73

depending on the teachings they are primarily influenced by. That is, a person exposed to
less strict teachings would have Ark increase but still be to the left of the Spectrum’s
midpoint.

/ls_k : the influence of denomination —k on individual i .

E : all other factors that influence an individual’s religious preferences. These are the

indirect effects such as learning through parents, fiiends and the community.

To consider the indirect effect, we can safely assume that all the individuals that

influence the religious preferences of iwill experience the same process that i goes

through at some point in their lives. More formally, we can state it as follows:

E = §[A.k,/I_k ,E] and so on and so forth.
BIT" / GE 2 038: indirect effect such as parents, strengthens the commitment and

learning outcome of A7, , given that E was influenced primarily by 21'

Off, / 8E < O: indirect effect such as parents, weakens the influence of ’14. ,

given that E was influenced primarily by lik .

This means that if the surrounding community around i is committed to the doctrine of

k , then i will also be influenced primarily by k ’s doctrine. The doctrine of past
generations passes on to future generations through parental and social investment in an

individual’s religious capital. This is similar to Smith and Sawkin’s (2003) concept of

 

3 . . . . . . . . . .
‘ 8 A, and IL, In these derivatives are those of mdrvrdual l and not hrs/her parents, relatives or friends.

74

social interactions that argues that religious involvement by individuals is proportional to

their interactions with the surrounding religious community.

Therefore, we assume that a, is formed initially by the influence of a single

denomination, k , through direct and indirect means. This is given by:
a, = g [2, , E]

The denominations A and B take the ideal religious preferences of individuals as
given and compete against each other in a Nash environment. We regard the
denominations as benevolent players whose objective is to maximize the expected
welfare of their adherents by locating at an optimal position on the ‘religious investment’
Spectrum. Then, individuals automatically join the denomination that is closest to their
ideal position on the spectrum. We assume that all members of the population voluntarily

join a denomination.
2.3b. Stage 2: Religious Provision of Social Service

In this stage, the government decides which nonprofit agency to award the funds or
grants to. We suppose that the objective of the government is to award the funds to the
nonprofit agency that produces the good or service with the lowest production cost. Much
work on religious nonprofits emphasizes the quality of service that they provide in
comparison to other types of nonprofits and to for-profit service providers. The argument
made is that religious nonprofits produce better quality services due to their long
experience in service delivery and unique methodology and as such, should be
encouraged by the public sector through funding. In our paper, the government funding is

provided for efficiency reasons and not for quality concerns. The main results of the

75

model hold even if we assume the quality-driven objective for the governrnent.39 Thus,
the government seeks the most efficient producer of the social service. In this stage, if a

religious denomination or charity is awarded the funds, that denomination will fulfill the
task of producing the social service, Z k . We will later examine a case where more than

one charity may be awarded funds, as this is the case with the current Faith-Based
Initiative. Also, the denomination(s) awarded the fimds should be a more efficient
producer of the service than the government itself.

The funds are provided for a specific purpose, namely, the production of a social
service. The government, through some objective, has decided that it is in its own interest
to delegate the task of producing the social service to the religious charity. This social
good or service may be for educational purposes, such as building a school or a library. It
may also be for health purposes, such as building a hospital, clinic or a drug rehabilitation
center. We should note that the rules governing grants under the Faith-Based Initiative
state that charities are prohibited from using the government funds for religious worship,
teaching or proselytizing, but should only be used for the provision of non-religious
programs. As such, charities should aim to separate their religious activities from the
non-religious services that are provided using the awarded government funds (Guidance
to Faith-Based and Community Organizations on Partnering with the Federal
Government, White House Faith-Based and Community Initiatives,
http://www.whitehouse.gov/government/fbci/). However, our central argument in this

paper is that, coupled with the task of providing the service, the denomination and clergy

 

39 In addition to being a reasonably sound objective for the government, the efficiency objective also has
the advantage of being more convenient for empirical purposes.

76

are incidentally given an opportunity to proselytize the particular doctrine or religion that
they abide by.

In the course of producing and providing this service, the charity may seek to include
religion via several means. A careful examination of the methods by which religious
charities convey religious messages or doctrine as part of their social service delivery
process is done by Unruh (2004). One method, that she calls religious self-descriptions, is
through printed media such as program descriptions, brochures, pamphlets, etc. Another
method is through religious objects present in the surrounding environment. This may
include wall pictures, architecture of the buildings, artifacts, etc. The religious music
played may also have a spiritual effect on the recipients of social services (Unruh 2004).
In addition to these subtle and indirect ways, more direct methods can be employed such
as prayers for the recipients, reading out loud religious text or quotations and
participation in group worship (Unruh 2004). Also, the staff can also have an effect,
through their dress codes, manner of conduct and their informal conversations with the
consumers or recipients of this social service. Any imposed separation of religion and
social services is arguably an impossible task to achieve. This is mainly due to the fact
that religious charities are, by definition, religious institutions with a religious identity
that is essential to their operation and performance. Any forced disjointing of the
religious element inherent in them and the social services they provide will only succeed
up to a limit beyond which it cannot go further. Religion in both its implicit and explicit
form of expression will continue to leave a noticeable mark on the activities and social
service philosophy of faith-based agencies. Furthermore, government monitoring of

services provided by religious charities is both costly and difficult to implement.

77

The essential idea is that religious charities have numerous methods of utilizing the
government funds for religious purposes. The religious clergy in particular will aim to
preach their religion to the consumers or recipients of the social service that are initially
non-adherents of their denomination. This is because by doing this, they can preach to a
larger audience than before and thus gain more utility as a result. In so doing, they alter
the distribution of preferences on the religious spectrum and shift the average religiosity
level of the population in the direction of their preferred level. We are assuming that all

members of the population, adherents and non-adherents, are consuming this social good
or service that is provided by denomination k 40. The non-adherents are particularly

targeted by denomination k ’s clergy.

We let the outcome of the proselytizing be represented by A._ k. That is, the individual

i is now affected not only by denomination k , but also by denomination -k. The
individual, through the consumption of the social service and its associated religious
messages, is now exposed to the teachings of another denomination, -k . The individual
is thus exposed to the doctrine of more than one denomination and this alters his/her
lifelong beliefs.
2.3c. Stage 3: Religious Competition under Revised Preferences

With the added influence of another denomination’s teachings on the preferences of an
individual, the individual will accordingly revise his/her religious preferences. The

function that now determines the ideal religious preferences of i is given by:

 

40 This is consistent with the result obtained by Wuthnow et al (2004) in which they show that Faith-
Based organizations cater to a varied community of social service recipients.

78

a, = g [ANAHE]
For simplicity, we assume that this can be given by:
a, = g [2,,E] :1: 2L,

Whether A._ k is (+) or (—) will depend on the denomination that is awarded the funds.

We have,
a. = a. i- i, (2)

Denominations will now compete in the same manner as in stage 1 but with revised
religious preferences of individuals.
We now investigate the model in more detail to derive results under different

assumptions.
2.4 Results

Stage 1: Given the original spectrum of religious preferences, the denominations
maximize the welfare of their adherents and locate at the Nash equilibrium positions.
With two denominations, there are two corresponding equilibrium positions on the

spectrum. Each denomination maximizes the following:

Denomingtion lg:

Ma’{nit—shat}+(1—eiiuiktxiiilgtava

VII 0

Subject to Y = X, + r,
g (a) : distribution of ideal preferences.

More specifically, we can write denomination A ’3 problem as follows:

79

Denomingtion A:

Max!{fl{R.-S(rt-a.-)2}+(1-fl){Ur-A(Xr)llg(“)da

VA} 0
st Y = X i + rA
Substituting in the budget equation, we have,

Denomingtion A:

i

Max {,3{RA -S("A ‘0.)2} +(1—fl){UiA(Yi —rA )}}g(a)da

erl 0

Denomination B will have similar equations.

First Order Condition for A:

!

jt—wtr.—4)+(1—e)(aa/a«i)(—1)}da

0

r; =t/2-[(1—p)/(2ps)](ao,, /aX,)

Similarly, for B, we have:

r,‘ =1/2+t/2-[(1—fl)/(2fls)](6v,s WC)

We assume g(a) =1.

These represent equidistant allocations (from the end points) on the religious spectrum,
if we ignore the marginal utilities of private consumption, which are not relevant for our

analysis.

80

 

Figure 2.2: Religious Equilibrium

As discussed earlier, r; would be the optimal position chosen by the liberal

denominationAand I}; is the optimal position chosen by the conservative

denominationB .
542%: In this stage, the government chooses whom to award the funds to. Let the funds
be given by T

Ideally, the funds for the social service should be used to cover mainly the costs of

capital and labor. But the denominations may use part of the funds for proselytizing

purposes.

Therefore, actual expenditure of the firnds is given by:
T = Cost of Capital + Cost of Labor + Proselytizing Costs

TA = Cost of Capital + COAL + A ’s Proselytizing Costs, A ’s expenditure of T.
TB = Cost of Capital + (UBL + B ’s Proselytizing Costs, B ’s expenditure of T.

To = Cost of Capital + (UGL , G ’s expenditure of T.

Where a) is the wage rate paid to the social service workers or producers and L is the
number of hours they are willing to work.
The government will award the funds to a denomination if and only if the government

is less efficient in the production of the service than the two religious charities.

81

Assuming that to be the case, the government then chooses to award the funds to the
denomination that is more efficient or produces the service at lower costs.

If the government cannot observe the proselytizing costs, then it bases its decision on
the factor costs of both denominations. For example, the charity that has more volunteer
workers will get the fimds since its total labor costs will be lower than the labor costs of
the other denomination. We argued that the government chooses to award the funds to a
religious charity because it expects the service to be produced at a lower cost. But instead
of assuming that the cost of production is lower for the denomination, we construct a
process that explains why this is the case. In other words, we endogenize the production
costs of the social service.

The producers of the social service are the persons in charge of the religious
denomination or charity and we denote them by m . They are the religious workers.
Clergy members have a utility function that describes their preferences in the religious

arena. Specifically, this is given by,
Um = Um (Xm,L,,,,P) where,
X m : private consumption by m.
Lm : total number of hours worked by m.

P : volume of proselytizing done by m.
The budget constraint of m is given by: (omLm = X m. That is, total wages equal
total consumption. We assume that the clergy do not consume the social service.

But the volume of proselytizing is given by: P = Lm 3.0 , where,

9 : Proselytizing per person per hour.

82

0' : total number of individuals being preached to.

Thus P =(Total number of hours spent on preaching)x(Preaching per person per
hour)x(tota1 number of individuals preached to).

Furthermore, the total number of individuals preached to is determined by two factors:
r, and Z k . It depends on the number of adherents already in the denomination, and the

number of non-adherents preached by the clergy as they consume the social service.

Without the funds, k is as follows:
P = Lm 19.0(1; )
With the funds,
P' = Lm.6.a(r;,Zk)
The difference is:
P' — P = Lm.o.o(r, ,2, ) — Lm.t9.0'(r, ) = Lm.o.(o(r,,z, ) — o(r;, ))

Clearly, we can observe that the difference is due to the presence of the social service.

We can now solve m ’3 problem:

Max U... = U. <X..L..PI

{L..}

Without the service, we have:

Max Um = Um [a,L,.,L.,L..t9-0('i )]

Lm}

The solution to this maximization problem will give us an indirect utility function

given by: V (a); , P)

83

With the service, we have:

I??? Um = Um [comLm ,1, ,L,.o.o(r, , z, )]

This gives us: V(C(),:,P)

ButO' (r, , Z k ) 2 0' (r, ), because with the presence of the social service, the clergy

are preaching to a larger audience or community, since it also includes all non-adherents.

That is,P' —P2 0.

Using the concept of compensating wage differentials, we equate both indirect utilities:
V(a);,P) = V((o;,P')

Given that P. — P 2 0, thena); < (0;.

In addition to compensating the clergy for the work that they perform providing the
social service, we should also consider the positive attributes of their job in the form of
religious proselytizing“. This extemality resulting from the social service provision
increases the clergy’s utility. Therefore, we are using wage differentials to ‘compensate’
the clergy for the extra utility they gain. This is done by the clergy trading off a higher
wage rate or supply price for a lower wage rate in return for the extra utility from

proselytizing.

 

4‘ The concept of compensating wage differentials was first discussed by Adam Smith in his, Wealth of
Nations. He states that “the whole of the advantages and disadvantages of the different employments of
labour and stock must, in the same neighbourhood, be either perfectly equal or continually tending to
equality” (Smith, 2000, Ch. X, pp. 114). In other words, it is not only wages that should be equalized, but
also the negative or positive attributes associated with the job. For more on compensating wage
differentials, see Rosen (1986).

84

In summary, the award to the denomination increases the utility of the religious leaders
or workers since they can now preach to a larger group of individuals. This increase in
the size of the group receiving the clergy’s preaching, will increase the clergy’s

willingness to ‘give’ more and ‘take’ less. Religious workers are now willing to offer the
social service at a lower supply price (a); < (0;) than before. This preference of

religiously-oriented workers towards working in a religious work environment is
supported in the work by Ebaugh et al (2003) where their survey shows that employees of
religious charities are attracted to their jobs primarily due to the religious nature of the
job.

Proposition 1: The supply price of the social service is inversely related to the amount of
preaching that religious workers perform.

Corollafl: Religious workers are willing to ofl'er the social service at a lower production
cost than non-religious workers. Religious charities are thus more eflicient at the
production of social goods and services than non-religious charities.

This unique characteristic of religious workers gives religious charities an advantage
over non-religious providers of social services. Because they gain utility from providing
to more consumers, they are willing to reduce their supply price in order to get that extra
utility. Ebaugh et a1 (2003) conclude with empirical evidence that faith-based agencies
rely more on volunteer workers than secular agencies. In particular, “volunteers
outnumber paid staff by more than 2:1 in faith-based agencies, while the ratio of
volunteers to paid staff in secular agencies is a little less than one volunteer for every paid

employee (.89:1)” Ebaugh et a1 (2003). As such, this supports our argument that the

85

religious nature of working in a religious charity serves to lower the supply price or
wages that workers demand to the extent that some may decide to work free of charge.

A non-religious provider will have this form of utility function:

MaxUm = Um (X,,1.,)
{LII

According to our compensating wage differentials argument, it is clear that workers

in non-religious settings will demand higher wages, because no proselytizing exists.
Hence,((t);I > a); )and((t);l > a); ).

A government thus awards the funds for a social service to a particular
denomination because of the decrease in costs associated with the denomination’s
production. But we should note that that the portion of the hands spent on proselytizing
should not exceed the difference in costs between the govemment’s provision and the
denomination’s provision of the service. Even with lower costs, the denomination may
use a large portion of the funds such that very little amount of the social service is
produced relative to if the government is the producer.
2.4a. Funds to a Single Denomination

Initially, we examine the case where only one denomination receives funds from the

government. We assume for now that the government has awarded the funds to

denominationA, because(a); > a); ). That is,A ’3 religious producers have a lower

supply price of the social service thanB ’s producers. Since the government cannot
observe the portion of funds spent on proselytizing, the denominations may have to alter

their factor costs to increase their likelihood of obtaining the award. This means that the

86

denominations may be asked to present a summary of their projected factor costs that the

government uses to decide the final recipient of the funds.

Stage3: With A producing the social service, it would seek to influence the preferences
of B ’S adherents. A produces both Z A and AA (the proselytizing outcome).

New preferences are given by:
All of A ’s adherents maintain the same preferences because the proselytizing is
directed only at B ’s adherents.

All B ’s adherents revise their preferences due to the influence of A ’s proselytizing.

The new preferences are given by:

(1;: g'[/IA,A.B,E]

a} =gl/IBIEI #1,.

a}. =aj—AA

The new preferences are reduced by ALA because A is originally on the left part of the

spectrum and seeks to lower the level of ‘religious investment’ or religiosity preferred
byB ’s adherents. A wants B ’s adherents to become more liberal. We assume that the

proselytizing affects all B ’s adherents equally.

The denominations now compete on this revised spectrum and locate optimally. We
now include the social service Z A in the utility function of all adherents as they are now

all recipients of the social service.

87

(I

AzMax {,3{R,—s(r,—o,)}+ +-(1 ,B){o. (2, Y r,)}}g(o)do

IrA} 0

+ :[{fl{RA -S(r,, -a,-)2} +(1-fl){U,A(ZArY-r,)}}g(a)da

B: Max-HER {RB— (r—a) )}+ +—(l ,B){U. BA(Z, Y— rB)}}g(a)da

{7'3}

+]{B{RB— s(rB —aj)}+ (l— ,B){UjB(Z,,Y—r3)}}g(a)da
r, =t/2—/1/2—[(1—,6)/(2,Bs)](ao,, /aX,)
r; =1/2+t/2—A/2—[(1—fl)/(2fls)](dojB/6Xj)

Wenowhave,
t'=t—AA/2,wheret=I/2.
r,“ =r,’ -/i./2A

r8“ =rB'—/l,/2

88

 

 

 

 

.......... . I 3 I a

l.AA AB rBA

 

Figure 2.33: Proselytizing by ‘conservative’ denomination
Promsition 2: If the government awards fimds to a charity with a low average religiosity
level, the average religiosity of the population as a whole will reduce or move to the left

on the spectrum.

Proot: Consider: t' = t — A, / 2
Also, r,“ < r,‘

<r‘

A
r B

B
Implication: If the government awards the funds to the liberal denomination, the
population as a whole will become more liberal or less conservative. Through similar
arguments, if the funds are awarded to a conservative denomination, the level of
religiosity in the population will increase as believers become more conservative. See
figure 2.3B. As shown, this is due to the fact that denominations will utilize this

opportunity given to them to proselytize their own doctrine and alter the preferences of

the adherents of other denominations. The exposure to the teachings of the liberal

89

denomination influenced the religious preferences of the conservative believers, inducing
them to become more liberal. Since they were previously only exposed to conservative
teachings, they happened to belong to a conservative denomination. But with this new
exposure to liberal teachings, they have an opportunity to revise their preferences. They
are now affected by two contrasting religious positions, liberal and conservative. The
funds gave the liberal denomination an advantage over the conservative denomination in
the last stage of competition. The conservative denomination was forced to reduce its
optimal position on the religiosity spectrum as a strategic response to the proselytizing
efforts of the liberal denomination. One should also notice that the funds reduce the
number of extremists in the denomination that didn 't receive funding. There are more

moderates in that denomination.
2.4b Funds to Both Denominations

In this case, both denominations receive funding from the government. We assume that
the government awarded equal funds to both denominations to produce two different
social services that will be consumed by all members of the population”. However, both
social services provide equal proselytizing potential for each denomination. We will have

the following situation:

fie/1L,

......... . L i i I A in...“

 

 

TA 1'3
Figure 2.3C: Proselytizing by both denominations

 

42 Awarding one denomination more funds than another would result in a similar situation to the case
where only one denomination received funding. That is, the average religiosity of the population of
adherents will tilt in the direction of the denomination with more funds. We specifically consider the case
of equal hands in order to derive the important result regarding the number of extremists in the population.
We show that even with equal funds to all denominations, the funding of charities still has significant
consequences on the society’s religious composition and character.

90

With this situation, the population is neither more liberal nor more conservative. This is

because the proselytizing efforts of both charities offset each other, if we assume
thatAtB = 1,. Both social services provide equal potential for proselytizing and as a

result, we assume that the magnitude of the proselytizing outcome by both denominations
will be equal, although the direction of the outcome or the proselytizing effect on
preferences will be different. The denominations’ positions are closer to the center
however. There are less people at the extremes of the spectrum.

Proposition 3: If the government awands equal grants to the denominations, then the
proselytizing eflorts of the denominations have a zero net eflect on the overall religiosity
of the population (see figure 2.3C). The population is neither more liberal nor more
conservative than before. However, there are fewer individuals at the extremes.
Consequently, there are more individuals concentrated at the center of the spectrum.
Both denominations have equal number of adherents.

Corollagy: If the government awards equal grants to denominations, the result is a
decrease in the number of extremists in both denominations as shown by the shrinking of
the spectrum. There are more moderates as a result.

Implication: This result gives us a possible motive for having government funding of
religious charities. We have shown that with equal funds to opposing denominations, an
important result is that there are fewer extremists on both sides and more moderates. By
giving out funds, the government has given both denominations the opportunity to
influence the other. As a result, this interaction between the players in both

denominations encourages openness and toleration. Adherents are given the chance to

91

learn more about the ‘other’ and this learning yields increased understanding and
toleration by both sides”. This may well be an important outcome that influences the
decision a government takes regarding funding of religious charities. This result offers
important insight into the long-term effects of the Faith-Based Initiative in which several
charities of diverse religious doctrines are awarded grants under the new system. It may
well be that such a system will help to bridge the differences among different and often

rival religious groups and promote a more tolerant society.
2.4c. Believers and Secularists

In this case, we add a new section to the population that we label, ‘secularists’ or non-
believers. The unique feature about this portion of the population is that, in addition to
being located at the extreme left of the religiosity spectrum, they do not belong to any
organized community. That is, they are not members of an institution’s congregation.
They are independent individuals who located at the extreme left of the spectrum“.
However, they are also recipients of the social service,Z . But with the social service
being produced exclusively by a religious organization, they are now exposed to
aggressive proselytizing by the denomination producingZ . The charity in question will
seek to attract new members from the ‘secular sect’ in addition to its efforts at attracting
new members from the rival denomination.

Stage 1: The spectrum is as follows: (Figure 2.4: Believers and Secularists)

 

‘ I I I I ‘

fi *

-L 0 TA t 1.B l

 

43 This is similar to Wood and Warren (2002), in which they discuss how faith-based community
organizations act as bridging institutions that enhance social capital between different sects. They link this
to Putnam (2000), where he argues that social capital leads to increased tolerance between groups.

44 Barros and Garoupa (2002) label a similar group as the ‘non-church’.

92

There are L non-believers or secularists in the population.
a, = g, (E), the initial preferences ofl=1,...,L

It is clear that the definition of a non-believer or secularist is an individual who did not
receive any religious teachings from a religious institution. Also, the influence of friends,
relatives and society was insufficient to induce this individual to be committed to any
religious entity“.

In this stage, the denominations have no influence over the secularist portion of the
population. Thus, their location is an equidistant equilibrium as Shown in figure 3.
W: Let A be the award winner. With A producing the social service, it would

seek to influence the preferences of B ’s adherents and the non-believers. A produces
both Z , and 2,.

New preferences are given by:
All of A ’s adherents maintain the same preferences because the proselytizing is
directed only at B ’s adherents and the non-believers.

All B ’s adherents revise their preferences due to the influence of A ’s proselytizing.

All the non-believers revise their preferences by increasing their religiosity or
appreciation for religion. Some of the non-believers would actually commit to a
particular doctrine and join a denomination while some would still decide to maintain
their non-believer stance. Those that decide to join a denomination would choose to join

the liberal denomination, A.

The new preferences of B ’s adherents are given by:

 

45 We can think of non-believers as ‘ultra-liberals’.

93

a; =g'[/i,,lB,E]

aizg [A’B’E] _2'A

a}. =aj-xi,

For the non-believers,

al 2 gr (1A 9E)

a; = a, + xi,

The new preferences for the non-believers are increased by A», because A seeks to

increase the level of ‘religious investment’ or religiosity preferred by the secularists.
Denomination A wantsB ’s adherents to become more liberal and the non-believers to
become ‘more religious’.

Stage 3: The denominations now compete in a Nash environment given this revised

spectrum of religious preferences.

 

 

 

 

 

 

 

 

' I I I I ’
-L O [A It t rB t l
’15. AA A, £8.13
| I I I I
-L 0 TA A t 1.B A 1.B ‘ 1

Figure 2.5 : Proselytizing that includes Secularists

94

A: Max {,6{R,—s(r,—a,)2}+(1-fl){u,,(Z,,Y—r,)}}g(a)da

+ ]{fl{R, -s(r, —aj)2} +(l —B){Uj, (Z,,Y—r, )}}g(a)da

t—Ii
A

+I{fl{RA -s(rA —aj)2}+(1—B){oj, (Z,,Y—r, )}}g(a)da

0

where the last component is due to the entrance of non-believers into A.

j{fl{RB —s(r,, —o,)’} +(1—fl){ujB(Z,,Y —rB)}}g(a)da

B:

E

927‘.

+]{,3{RB -S(rB —aj)2} +(1—,B){UJB(Z,,Y—r8)}}g(a)da

(I

From F.O.C, we have:

r,’ =r/2—i +22/2—[(1—p)/(2ps)](ao, /aX,.)

r,‘ =l/2+t/2—/l/2—[(1—,B)/(2,Bs):|(auj3/6Xj)
The locations are:
¢=n—h+iwr

rB' =r,‘—,I,/2

95

From the above equations, we have that the new position for A is more to the left than
the previous scenario. This is to account for the additional adherents from the ‘secular’

subset of the population“.

The spectrum facing the denominations is the same ranging from(0,1— A).

However, A has more adherents at the leftmost part due to the entrance of non-
believers. Therefore, A has more adherents than B.

By including a non-believer portion of the population, denomination A will have to
proselytize two groups of individuals and in quite different ways. It will seek to lower the
religiosity level of B ’s adherents to attract them into the liberal arena and at the same
time trying to increase the appreciation for religion of the secularists to attract them into
the religious or non-secualr arena. It is doing both of these tasks using the same social
serviceZ .

One should note that denomination A can use the same teachings or methods for
both targeted groups. In fact, all A is doing is teach its own doctrine or interpretation of
doctrine through several means. By teaching the ‘liberal doctrine’, A is exposing B ’s
adherents to a relatively more liberal interpretation of religious doctrine and at the same
time exposing the non-believers to this same liberal interpretation which by nature of the
religious spectrum, is not very different from their original preferences. The main
difference is that the exposure is having distinct effects on both targeted groups.
Promsition 4: In a society that includes non-believers, the award of a government grant

to the ‘less strict’ (liberal) denomination will induce it to attract the members of the

 

46Thisisthecaseif: -A,+A,2/t<—A,/2.Thisholdsfor A, <t/2.

96

‘stricter’ (conservative) denomination and also non-believers. The liberal denomination
will seek to lower the religiosity or strictness level of the rival denomination 's adherents,
and also increase the religiosity of the non-believers. As a result, there are more liberal
than conservative believers in the population.

M: (See figure 2.5).

As shown in figure 2.5, the non-believers who joined A meant that A has more
adherents then B. Even with the strategic response by B , it could not account for the
fact that A is also targeting non-believers who are located at the far left end of the
spectrum where B has little or no power.

Let the grant be given to B. In this case, B would surely seek to influence A ’s
adherents to induce them to be more conservative. However, can B induce non-believers
to be conservative?

It is clear from observing the spectrum that if B attempts to increase the religiosity
of the non-believers, this would mean that they move to the right and join A.
Proposition 4’: If B (conservative) is awarded the funds, any proselytizing that also
targets non-believers will result in an increase in the number of liberal adherents in the
population. In other words, the population becomes more liberal. Therefore, targeting
non-believers may benefit A (liberal) instead.

Implication: This result implies that if conservative denominations and charities become
interested in the size of their adherent population relative to that of the rival
denomination, then they may seek to discriminate against non-believers in the provision
of the social service. This is also the case if both denominations care about the religious

orientation (conservative or liberal) of the total population of believers.

97

Since the proselytizing is a necessary by—product of the social service provision, then
any service provided to non-believers will also mean that their religous preferences are
altered so that they are more religious and eventually become liberal. This implies that
the conservative charity is actually performing a favor for the liberal denomination at no
cost to the liberal denomination. Therefore, in the event that the conservative
denomination is attentive to its rival ’s relative stance, the likely response of the
conservative denomination would be to restrict the provision of the social service to
individuals who are more likely to accept conservative teachings.

Based on this result, one can conjecture that if part of the objectives of denominations
is maximizing adherence or achieving a religious orientation of the population towards its
doctrine, a liberal denomination will not discriminate against any group in the
population in the provision of the social service, while a conservative denomination is
likely to discriminate against non-believers or any group to the left of the liberal

denomination on the religiosity spectrum.
2.5 Conclusion

In this paper, we have examined the relationship between religious charities and the
government sector. In particular, we have sought to investigate the impact of government
finding on relgious charities and on religion and believers in the society. In order to do
this, we developed an economic rationale for the efficiency advantage of religious
charities over non-religious providers of social service. We have shown that religious
workers or clergy are willing to offer social services at lower costs than their non-
religious counterparts and are thus more cost efficient. This advantage is mainly due to

the proselytizing opportunity that the government funds have offered to the charities. By

98

catering to more social service consumers, the clergy can utilize this opportunity to
preach their particular doctrine to more people and thus gain more religious utility.

This unavoidable by-product of the government funds will give denominations an
advantage over their rival denominations that didn’t receive government funds and this
will eventually change the religious balance of power in the society as a whole. But even
with equal grants to all denominations, we have shown that the government funds will
have the effect of reducing the number of extremists in all denominations and increase
the number of moderates. Furthermore, by including non-believers in the model and
under certain conditions, we argue that strict denominations may seek to actively
discriminate against this group of the population in the provision of the social service.

These results provide important insights into the fiiture of the F aith-Based and
Community Initiatives. In its goal to encourage religious agencies in their social service
missions, the government should be aware of the implications on religion and believers in
society. The government should consider the effect that such programs have on the
interaction between denominations and on the ideological character of society. Religious
authorities should also consider the effect that these programs may have on their
denomination’s position in society relative to other denominations. And finally, believers
and adherents should be aware of the consequences of such programs on their religious

preferences and affiliations.

99

Chapter 3
Religious and Economic Preferences: An Empirical

Analysis of State Tax Rates

3.1 Introduction

Economics of religion is gradually developing to become an important specialty in the
field of economics. Economists, political scientists, sociologists and even religious
scholars are increasingly aware of the important contributions that this growing specialty
can convey to our understanding of human behavior and structure of institutions in
society". One area of concern for economists of religion is to thoroughly study all
aspects of ‘religious markets’, where adherents or believers represent the demand side,
while religious institutions represent the supply side. The analysis is also extended to
study the interaction between players in religious markets and external players, such as
political parties, government, secular nonprofits, media, etc. This area of concern is
therefore an attempt by economists of religion to study religion using economic theory as
a set of analytical tools, and religion as the object of study.

Another agenda for economists of religion, which is to some extent similar to the first,
is to reexamine past theoretical and empirical literature with the aim of introducing
religion as a key variable in the analysis of economic phenomena. In other words,
religion has traditionally been excluded as an explanatory variable in the study of

changes in economic conditions over time. Examples of studies that have sought to fill

 

47 For a comprehensive review of this literature, see Iannaccone (I998).

100

this gap in the literature include Heath et al (1995), Lipford and Tollison (2003), Ewing
(2000) and Paldam (2001). Heath et a1 (1995) attempt to study the impact of religion on
economic performance. They employ U.S statewide data to study the effect of religion on
state per capita income and conclude that fundamentalism has the most significant effect
relative to other denominational groups such as non-fundamentalist Protestants, Catholics
and Jews. Using cross-state data for the US, Lipford and Tollison (2003) study the
relationship between religious participation and per capita personal income. By
employing a simultaneous equation estimation method, they find that church membership
is negatively related to income. On one hand, higher incomes means higher opportunity
costs for involvement in religious activities, and on the other hand, higher participation
levels in religious activities reduce the effort and willingness to seek material wealth. In
studying the effects of specific religious doctrine and cultures, Ewing (2000) examines
the effects of Catholic affiliation on the wages and earnings of individuals relative to
those of non-Catholic affiliations, using the National Longitudinal Surveys of Youth. The
study finds that Catholics receive higher wages than non-Catholics. Ewing (2000) argues
that the Catholic religion adds to an individual’s human capital over time, and that it
sends a signal that such individuals have attractive labor market qualifications such as
discipline, honesty, trustworthiness, enthusiasm and low demand for leisure. Paldam
(2001) studies the impact of religion on corruption, as the latter is believed to have some
significant effect on economic development. Paldam’s (2001) findings state that, “two
groups of religions decrease corruption — Reform Christianity [Protestants and Anglicans]
and Tribal religion — while the others [Catholics, Orthodox and lslarn] increase

corruption in a similar way”. But Paldam (2001) does concede that the Christian

101

implication of the study should be treated cautiously, since Protestantism emerged as a
reaction to the corruption inherent in the Catholic Church at that time, and thus raises the
issue of ‘reverse causality’.

In recent years, much empirical work has centered on studying the impact of religion
on economic development, which Paldam (2001) refers to as the ‘Weber link’48. These
studies include Grier (1997), Barro and McCleary (2003), Guiso et a1 (2003) and
Montalvo and Reynal-Querol (2003). Grier (1997) studies the impact of religion on
economic development for the main British, French and Spanish ex-colonies using data
over a period of 30 years. Grier’s (1997) findings propose that British ex-colonies that are
predominantly Protestant perform significantly better, economically, than French and
Spanish ex-colonies that are mostly Catholic. Grier (1997) argues that the results do
support the Weberian hypothesis that Protestantism exerts a positive influence on
economic development, although this is only one of many factors that affect economic
performance.

Barro and McCleary (2003) also attempt to fill this gap in the literature by studying the
impact of religion on economic grth using a cross-country panel data. Initially, the
authors acknowledge the possibility of reverse causality, i.e. how economic development
affects religion. The secularization hypothesis, ‘which argues that religiosity levels in
societies decrease over time as these societies attain higher levels of economic and

technological progress, has acquired a central position in the sociology of religion

 

48 Many historians, economists and sociologists have criticized Weber’s hypothesis. A notable critic,
R.H. Tawney (1926), argued that “the Protestant ethic, with its insistence on hard work, thrift, etc., had
contributed to the rise of capitalism, but at the same time Protestantism itself was being influenced by an
increasingly capitalistic society." This and other positions argue that Weber’s monocausal relationship
between religion and economic development is restrictive.

102

literature over time”. Barro and McCleary (2003) argue that this theory has been strongly
challenged in recent years by rational choice models that have instead focused on the
market forces of demand and supply in religious markets as the key determinants of
religious behavior and changes in religion over time”. Furthermore, within the rational
choice literature, more focus has been directed towards supply-side explanations of
religious behavior by analyzing the competition between different players in the religious
markets5 1. In order to focus on the effect of religion on economic development, Barro and
McCleary (2003) include instrumental variables in their analysis, by initially analyzing
the determinants of religiosity levels. Their findings suggest that higher religious beliefs
tend to fuel higher economic growth by encouraging growth-augmenting facets of
individuals, while higher church attendance levels lower growth as more resources are
directed towards religious activities (Barro and McCleary, 2003).

Guiso et a1 (2003) utilize the World Values Survey to analyze the impact of religion on
economic attitudes. Their findings support the view that religion in general has a positive
impact on economic growth, with Christianity being more positively linked to factors
promoting economic growth, while Islam is negatively linked. Within Christianity,
Protestants tend to trust others more than Catholics and are more opposed to cheating on

taxes and agreeing to bribes than Catholics. However, Catholics have more favorable

 

49 For detailed discussions about the secularization hypothesis, see Wilson (1966), Berger (1967), Martin
(1978) and Chaves (1994).

50 A more complete elucidation of this new paradigm and its main argument can be sought from Stark
and Bainbridge (1987), Finke and Stark (1992, 2000) and Iannaccone (1997, 1998). For a discussion of
Adam Smith’s impact on the field of economics of religion, see Anderson (1988) and Iannaccone (1991).
Also, Smith dedicates Chapter 1, Article III of “The Wealth of Nations” to a discussion of the market-like
structure and behavior of religious establishments.

5' For a discussion of the supply-side explanations of religion as a methodological approach, see Finke
(1997). For theoretical and empirical applications of this approach, see Iannaccone, Finke and Stark (1997),
Anderson et al (1992), Davidson and Ekelund (1997) and Ekelund et al (1989, 1992, I996), Iannaccone
(1988, 1992), Montgomery (1996) and Zdeski and Zech (1992), and Chapters 1 & 2 of this dissertation.

103

attitudes towards private ownership and competition than their Protestant counterparts.
Even with all these empirical observations, the authors admit that further research is
needed before a statement can be made as to which religious group is more favorable
towards economic grth (Guiso et al, 2003). In another study, Montalvo and Reynal-
Querol (2003) examine the impact of religious diversity on economic development using
religious polarization as a measure of religious diversity and as the key independent
variable explaining grth patterns across countries. Their findings suggests that higher
religious polarization, which implies more potential for religious conflict, has a negative
impact on factors that promote economic growth, such as investment, education and
working-age population grth (Montalvo and Reynal-Querol, 2003).

In a similar approach to the above studies, this paper attempts to fill another gap in the
literature where religion has been ignored as a factor in the analysis. In particular, we aim
to study the impact of religion on government behavior. The paper seeks to reconsider the
question of what affects a govemment’s decision to alter taxes by including in the
analysis the role that religion plays. Another equally important objective of this study is
to empirically assess the propositions projected by the ‘Culture Wars’ thesis and their
political and social implications. Are the govemment’s revenue and expenditure
decisions responsive to the religious preferences and affiliations of believers in society?
Does the government, in the process of formulating its fiscal plans, take into
consideration the demands and prescriptions of religious institutions and religious
doctrine? Are the choices of political leaders and parties influenced by the religious
inclinations of their constituencies? These are the questions that this paper will attempt to

answer, in order to empirically evaluate the arguments proposed by the ‘Culture Wars’

104

thesis and to increase our understanding of the interconnectedness between religious,
political and economic preferences, while at the same time aiming to include religion as a
possible key factor explaining economic changes in society.

The main question therefore that will be investigated in this paper is whether changes
in the religiosity levels of US states affect state tax rates over time. Our paper does not
report any significant relationship between government tax rates and the religiosity of the
population. However, the results do not reject our hypothesis, but only fail to support it.
As such, we attempt to explain the reason for our findings and suggest possible future
methods to reexamine the issue and extend the analysis further.

The next section will discuss in detail some of the arguments in the literature that shed
light on the ‘Culture Wars’ thesis, and we will develop a theoretical model that explains
this thesis and that will serve as our foundation for the empirical work to follow. In the
third section, we present our empirical model, explain the data and methodology used,

and discuss the results. We conclude in the last section.
3.2 Religious and Economic Preferences

The aim of this paper is to empirically test the argument that religious preferences and
affiliations of individuals in a society affect, directly or indirectly, the decisions of
governments regarding public revenue and expenditure. In particular, we aim to test the
hypothesis that the religiosity of an individual is inversely related to the amount of taxes
he or she is willing to pay to the government. This also extends to the hypothesis that the
religiosity of an individual is inversely related to his/her preference on the size of
government, measured by government expenditures. In general terms, the analysis is

essentially Weberian, in that we are interested in how religious preferences, derived from

105

religious doctrine and culture, influences the economic and political preferences of
individuals and hence, society as a whole. Our analysis will be in the context of the US.
religious and political framework. It then follows that we are interested in explaining the
nature of what has been in recent decades called, ‘culture wars’. This refers to the
polarization that is now evident in the religious and political maps of the US. between
the conservative and liberal sides. It has become a common feature of religious and
political discourse in the US. to employ the conservative-liberal dichotomy in analyzing
most religious, economic and political phenomena. Almost any issue, whether economic,
political, social or moral will be explained in terms of opposing conservative and liberal
views. As such, all issues have to go through this polarization process that reduces
complex tastes and preferences into a one-dimensional spectrum extending from most
liberal to most conservative.

Two prominent sociologists of religion, Wuthnow (1988) and Hunter (1991), hold the
position that religious preferences affect the views of believers and denominations on
economic issues. Davis and Robinson (1996), (who object to the conclusions of
Wuthnow ( 1988) and Hunter (1991)), argue that, “according to Wuthnow (pp. 132, 219-
23, 239), religious liberals take politically liberal positions on a wide variety of
contemporary issues, including economic justice, while religious conservatives take
politically conservative stances”.

They state that,
Wuthnow (1988, p.114) sees religiously conservative “special purpose
groups” such as the Moral Majority and the Christian Coalition as having
arisen to combat the growing role of government in guaranteeing equal rights
for women and minority groups and in providing welfare and other forms of

assistance to the poor. Wutlmow (p. 248) notes that religiously conservative
leaders, such as evangelical economist George Gilder, Moral Majority founder

106

Jerry Falwell, and Christian Coalition founder Pat Robertson, have voiced
strong opposition to redistributional efforts by the government. Through
analysis of a sample of Americans interviewed in 1984, Wuthnow (pp. 219-
23) finds that religious conservatives are more opposed than religious liberals
to government spending on social programs. (Davis and Robinson, 1996)

Hunter (1991, pp] 1 l) argues that the notion of freedom among orthodox believers also
refers to freedom in “economic self-determinations” and as a result, these believers are
proponents of “‘free’ enterprise”. He argues that evangelicals “trace the [above]
relationship to the Old Testament land laws that linked private property to the fieedom
from state coercion, especially from taxation”. (Hunter 1991, pp.l 11)

Hunter (1991, pp.96-97) observes in a 1987 study of denominational leaders that
orthodox religious leaders prefer economic growth as a way of reducing poverty and also
tend to reject using redistribution of wealth to solve the problem.

As such, Wuthnow and Hunter are of the view that religious preferences affect the
economic preferences of believers and religious authorities. The explanations presented
for this pattern are similar, although distinct to some extent. Wutlmow and Hunter
explain these attitudes as a result of particular religious doctrines held by the different
denomination members. Hunter (1991, pp.49) argues that, “the culture war emerges over
fundamentally different conceptions of moral authority, over different ideas and beliefs
about truth, the good, obligation to one another”. Thus, they explain this behavior in a
Weberian framework.

However, one must note that this view is not shared by all sociologists of religion, as
some have arrived at results that contradict the ‘culture wars’ thesis, or at least fail to

affirm that it exists. Davis and Robinson (1996), in a statistical survey, Show that

orthodox adherents do not necessarily hold conservative economic viewpoints. They state

107

that, “the orthodox are more liberal on issues of economic inequality” (Davis and
Robinson, 1996). But the reason for this observation according to them, is “[that] the
tendency of the religiously orthodox to draw adherents fi'om disadvantaged groups makes
them more liberal on economic issues” (Davis and Robinson, 1996). However, Davis and
Robinson (1996) do find that orthodox adherents believe that government welfare
spending is too high. As such, their results don’t support their hypothesis fully.

We attempt to explain this polarization in religious and economic preferences and its
manifestation in society. Specifically, we are interested in why individuals with relatively
low levels of religiosity (namely liberal) prefer higher levels of government taxes and
higher government spending, while those with higher levels of religiosity (namely
conservative) prefer lower government taxes and spending”.

It is clearly (and logically) evident that the ‘culture wars’ phenomenon also extends to
the political sphere. People’s religious preferences influence their economic preferences,
which in turn will impact their political choices and affiliations. In his study of religion
and politics in the US, Layman (1997) states that, “ there is already a good deal of
evidence that political activists and voters with orthodox beliefs and affiliations tend to
support the Republican Party while the Democratic Party draws its activists and voters
disproportionately from the ranks of religious liberals and secularists (Green, Guth, and
Fraser 1991; Green et al. 1996; Guth and Green 1986, 1987; Kellstedt, Smidt, and
Kellstedt 1991; Miller and Wattenberg 1984; Rozell and Wilcox 1995; Wilcox 1992)”.

Laymen (1997) states that, “as the cultural tensions [‘wars’] between America’s religious

 

52 In Chapter 1, we argued that conservative opposition to government spending is not a reflection of
opposition towards welfare programs. What it signifies is a preference for religious provision of such
services using religious donations, rather than a secular (government) provision using tax revenues.

108

and secular populations grow, the political impact of religious commitment also should
grow, with strongly committed individuals becoming more Republican relative to their
less religious counterparts”. Furthermore, Layman (1997) states, “highly religious
individuals have consistently been more likely than their religious counterparts to identify
with the GOP ...” As such, the polarization extends to the political sphere with
Republicans being generally labeled as conservatives and Democrats generally labeled as
liberal.

Our aim is therefore to explain these patterns that we observe in society in order to gain
some insight into the effect that religion has on the economic and political aspects of
society. To achieve this, we will develop a theoretical model that utilizes economic
theory and rational choice assumptions. Then, we will aim to test the propositions of our
model using state-level panel data for the US.

In addition to our discussion about the arguments of the ‘culture wars’ thesis and its
political inferences, it is important that we also examine what has been said on the
economics of taxation and how states decide tax changes, so that we can better
understand the contribution of our theory, and the usefulness of including religion in the
analysis of tax changes. In a paper titled, “When Do States Legislate Tax Changes?”
David Merriman surveys the theories that have been proposed to explain the positive and
normative aspects of state tax policy. Merriman discusses five main classes of theories
offered in the literature to explain the determinants of tax policy changes over time. The
first type is illustrated by Rosen (1988) where Ramsey rules are proposed as the optimal
tax policy, by setting tax rates according to the inverse elasticity rule. Other theories have

explained tax changes as being responsive to business cycles (See Kee and Shannon

109

1982, Vogel and Trost, 1979). Political equilibrium theories have also been put forward
to explain tax policies that argue that political parties in power affect the tax legislation
that passes (See Hettich and Winer, 1988). Another position has been the view that tax
policies in one state affect policies in neighboring states. This has been termed ‘yardstick
competition’. See (Case 1993, Besley and Case, 1995 and Wilson 1986, 1999). Finally, a
public choice view argues that complex tax structures and tax bases may induce more
taxes in the future (Holcombe, 1998). It is apparent from this survey of the state tax
literature that the analysis so far has neglected the relevance of religion and religious
institutions in society, and their impact on the decisions of governments regarding
taxation and expenditure. This paper will attempt to study the function that religion plays

in the complex decision-making processes that eventually lead to tax changes over time.
3.3 Theoretical Model

In this model, we use a structure that is characteristic of contemporary politics and
religion. The political structure is one of political parties competing against each other in
an election. The ‘secular’ parties are at some stage declaring their platform or positions
on the various issues that interest the voters. The voters respond to this political
competition by voting for their preferred candidate. Thus, this structure may not be
applicable to many periods in history, where the forms of political activity differ
substantially. Also, we construct the religious world as one where denominations take the
religious preferences of all the individuals in the society as given and according to some
specified objective, locate optimally on that spectrum of religious preferences.

A unique feature of the model is the timeline of events it presents. In the pre-game

stage, the preferences of individuals are given. Denominations or parties cannot influence

llO

these. In the game’s stages, we model the denominations as the first players. Their
decisions precede those of the parties. We believe that this formulation is both realistic
and insightful. To a great extent, religion arrives at home earlier than politics. Young
individuals are exposed to religion at an earlier stage in their life than they are to politics.
Religion arrives via the parents, relatives, and friends and more importantly for us, the
church. Smith et a1 (1998) explain the influence of parents on their children in the area of
religiOn as an “investment in a child’s religious human capita ”. They argue that this
investment by parents serves as the driving force for continuing interest in religion in the
later stages of a child’s life (Smith et a1 1998). As such, our structure assumes that the
religious institution in general influences individuals before they become exposed to the
world of politics. Religious institutions fuel this perpetuating influence over generations
of new children through their impact on parents, who represent the main vehicles of
religious learning and investment.

With their own preferences given beforehand, the denominations compete and attract
adherents and then impose conformity to a standard that they choose. These standards
that represent the teachings of the denomination, will direct the political and economic
decisions of the adherents. In other words, the religious teachings will lead to political
and economic orientations. This assumption that religious doctrine directs political-
economic perspectives and actions is the crucial concept that underlines the model in
general. In more concrete terms, the model presents a structure where political and

economic decisions are endogenized and this endogeneity is due to religious doctrine.

lll

The model is thus presenting religion as a central force in the lives of individuals, both on
the private and public levels”. The utility function used is described as follows:
The Utility Function

The first part of this utility function represents the religious utility while the second
part represents the politico-economic utility.
i) Religious Utility

The religious utility is given by:

fltR- srx, — at’}

In the function above, R is the religious ‘club’ good that is exogenously provided. fl

represents the weight placed on religion in the population in general. It is the same for all
individuals“. We can regard it as an indicator of the ‘power’ of religious institutions or
religion in the society, as evidenced by its effect on the utility function of all individuals.
This power may represent the prominence of religious institutions or their history. This

parameter is not related to the individual-specific preference on ‘religious investment’.

The variable a,- represents the ideal position of individual i on ‘religious investment’55

and x, corresponds to the level of religious investment chosen by denomination j in stage

2. We denote the transportation cost incurred by individuals as a result of their obligation

 

53 We should note that the model does not adopt a strict monocausal relationship between religious and
economic preferences. Both religious and economic preferences affect each other as explained on pp. 54-
55.

54 We can have ,3 vary by individual, although this will not affect the results of the model. However, a
value of 0 may indicate no preference for religious utility which may indicate that the individual is an
atheist or an agnostic. This could be a way of structuring the preferences of these groups in the population.

55 In practice, a,- will usually have a range of small values such as [0, 0.2], since contributions to
denominations do not constitute high percentages of individuals’ incomes. Similarly is the case for the tax
rate. Private consumption is usually the largest component of income spending. The nature of the model
and its primary interest in the religious and political components makes them seem to command very big
portions of income.

112

to conform to the teachings or rules of the denomination, by S. It is the ‘price’ of

sacrifice. This value is the same for all individuals in the society.

From the form of this utility, we observe that the individual derives utility from his/her
input into the religious good production as given by {— s(xj — aJ2}56. The individual
loses utility as a result of confonnism to the denomination’s rules, but also gets utility
from giving his/her preferred investment level. As such, the total effect from this is given
in the quadratic form {— S(xj— 0J2}.

ii) Politico-economic Utility
(I-flh’ Ur,t(Z. CI) + vrt}
In the politico-economic utility above, Z is the public good produced by the party that

assumes office after the elections. (1- [3) this refers to the weight that society places on
political and economic issues in general. Also, it may refer to the ‘power’ of these
political and economic institutions in the society. 6,- denotes the private consumption of
individual i after investing in religion and paying taxes. Whatever is left after religious
investment and government taxes is used for private consumption. Also, v,,, is the fixed

position of individual i towards party k. This position is independent of the tax rate or the
religious preference. It refers to the bias of individual i towards party k regardless of the

positions of party k in the elections, thus the term ‘fixed’. This variable could be positive
as well as negative. A positive vik means that the individual has a positive bias towards

party k.

 

56 This is reasonable because in the religious domain, individuals feel spiritually satisfied by the act of
giving to the denomination as well as receiving from it.

113

Hence, U,-j,,( Z, c,- ) represents the politico-economic utility of individual i. The
individual derives utility from the public good provided by the government (winning
party) and from private consumption. This function is maximized in the transition stage.

The composite utility function is therefore:

U,-,- = flt’R- 509-022} + (1- fl){ Uryt(Z. Ci) + Vik}

We now examine the workings of the model given our assumptions and objective
function.
The Voting Stage

In this stage, voters will decide which parties to vote for, based on utility
considerations, after the parties have announced their platforms on the tax rate issue.
Each voter in this stage already belongs to a particular denomination.
Political Competition Stage

Politico-economic equilibrium: In this stage, both parties will converge to the same
position in equilibrium”. This should be expected since all the parties care about is
winning. The equilibrium tax rate will depend on the average of the optimal positions on
‘religious investment’ chosen by the denominations. Also, it will depend on the share of
the adherent population that each denomination holds. Given this equilibrium result, we
now turn to the preceding stages to derive the strategic decisions made there.
The Transition Stage

We have seen that in this stage, believers take their denomination’s policy as given and
derive their optimal tax rate. Their optimal or preferred tax rates are influenced primarily

by the denomination’s positions on the ‘religious investment’ issue. This is the tax rate

 

57 For detailed derivations, see Chapter 1, pp 23.

114

that they would prefer to see implemented and it is driven by their religious convictions.
This will form the main idea of the model and we will elaborate on it in more detail in
working out the details of the religious competition stage.

As we stated previously, individuals maximize their politico-economic utility, given
that they now belong to a particular denomination. In general, each individual solves the

following problem:-

Max (Iii/((21) Ci) S.t Y-Xj-T',‘ = C; , Z; = TIN , where 7:. tax rate.
{T,-} N .‘ population.
Y: income.
We are assuming income to be homogenous across the population.

Substituting in the constraints, we have :

Max Unfit-M Y-xy-Tr)
{TI}

F.0.C.'- (dUg-k/dZi) (dZ,/d T,) + (dU,-j,,/dc,) (dc/d T,)=0
We can then derive the optimal tax rate for i, (T,- *) from the above result.

Example: Let us consider the following utility function for simplicity and in order to get

results that can be readily interpreted.
U,-j,,( Z, c) = ban + (I-b)lnc,-
U.-,r(Z. 6) = [Dina-N) +(1-b)ln(Y-39-Tt)
F.0.C:- b/T, + (1- b) /(Y-x,—T,).-I = 0
T." = th-xd (2)
It is evident that the preferred tax rate, T}, is inversely related to the individuals

religiosity level, xj .

115

The Religious Competition Stage
As previously stated, the denominations act as benevolent players who sincerely seek
to maximize the sum of the ‘religious welfare’ of their adherents. They also compete in a
Nash environment. Each denomination chooses the best response strategy to its
competitor”.
Religious Equilibrium: Both denominations will locate equidistant from the end
points”. The denomination locating on the left will be labeled the liberal denomination

while that on the right labeled the conservative denomination. This differentiation is

partly caused by the benevolent or democratic nature of the denomination’s objectives“).

 

‘ l l l ”
xa xts
(liberal) (conservative)

Figure 3.1: Religious Equilibrium
Promsition:

In equilibrium, the liberal denomination will induce its members to favor a high tax

rate while the conservative denomination will induce its members to favor a low tax

rate6 ’.
Corollary:

Liberal denominations favor more government spending than conservative
denominations“. Thus, we can reasonably say that liberal denominations favor ‘bigger’

. . . . , 63
governments and conservative denominations favor smaller governments .

 

58 This methodology is similar to the application of Hotelling’s Model by Barros and Garoupa (2002).
59 For detailed derivations, see Chapter 1, pp 25.

60 The denominations act as sincere social planners, thus we arrive at the optimal location equilibrium.
6' For proof, see Chapter 1, pp 27.

116

This corollary is insightful, since in real politics, we observe that liberal denominations
are more closely affiliated with the Democratic Party and are thus proponents of ‘bigger’
governments, while the conservative denominations are more closely affiliated with the

Republican Party and are thus proponents of ‘smaller’ governments.

Based on this theoretical framework, we will proceed to explain the empirical model

that will be used to test the model’s propositions.
3.4 Empirical Model and Results

The main result derived from our model is that individuals with high religiosity levels
will prefer to pay less taxes to the government, while those with low religiosity levels
will be in favor of (or at least not opposed to) higher taxes. The reasoning behind the
hypothesis is that individuals with stricter religiosity levels will tend to give a larger
percentage of their income to religious causes, and as a result and due to the fact that they
are faced with a budget constraint, are willing to give less to the government in the form
of taxes. This result assumes that all individuals in the population have equal income
levels. By controlling for income differences, we can then observe the direct link between
religious and economic preferences, which is our main question in this study. In other
words, do conservative adherents strictly prefer lower taxes, regardless of whether they
are poor or rich? The extension that we derived from this result is that conservative
believers will desire ‘smaller’ governments, while liberal believers will favor ‘bigger’

governments, given that income is homogenous across the population. We will test the

 

62 We reiterate that this does not mean that conservative believers are opposed to welfare programs.
They are specifically opposed to government provision of these services using tax revenue. This means
they prefer that such spending be done by their religious institutions using religious donations of adherents.

63 For proof, see Chapter 1, pp 29.

117

model’s hypotheses using US. state level panel data from 1971, 1980, 1990 and 2000.

Therefore, in the context of our data, our objective is to test whether changes in state
tax rates over time is influenced by a state’s religious composition. In other words, does
an increase in the proportion of a state’s population that is religiously conservative lead to

a decrease in the state’s tax rates over time?

In the context of our model, believers’ economic preferences (derived fiom their
religious preferences) will have an effect on a state’s tax policy if their preferences are
sincerely represented by the religious institutions or denominations that they subscribe to,
and by the political party that they vote for. This means that in order for the believers’
preferences to be influential, their religious preferences should first be optimally mapped
or translated into economic preferences by their denominations. If this mapping is
fulfilled, these economic preferences will then lead to their corresponding political
outcomes, provided that the political parties adhere to their promises highlighted in their

policy platforms during election campaigns.

For the measure of the religiosity of a state’s population or the composition of its
population that is conservative, we will use state-level data on church membership for the
years 1971, 1980, 1990 and 2000 for all fifty states“. In particular, we will use the
percentage of the state’s population that belongs to conservative denominations as our
measure of religiosity. In order to gain more insight into the question that our paper is
addressing, we‘will also use another measure of religiosity, given by state-level data on

contributions to religious charities and organizations. The two measures of religious

 

64 Data on church membership are taken from Johnson et al. (1974) for 1971, Quinn et al. (1982) for
1980, Bradley et a1. (1992) for 1990, and Jones et al. (2002) for 2000. These surveys were conducted by
the Glenmary Research Center. The data collection methods are uniform across all four surveys.

118

contributions that we will use are the percentage of state households who contributed to

religious causes, and the average per household religious contribution in a state.

Our measure of the conservative composition of a state’s population will have to be
based on some given classification of which denominations are conservative, and which
are not. This classification should itself be based on a careful and comprehensive study of
the beliefs structure of the various denominations and their positions on many social,
economic, moral and political issues. The resulting outcome would be a classification
mechanism by which we can categorize denominations on some given dimension. For
our analysis, a classification based on religiosity or conservatism would be the most
appropriate. Steensland et a1 (2000) state that, “the most widely used denominational

classification scheme in survey research [is] (T.W.Smith1990)”.

According to T.W. Smith (1990), researchers in the field of religion have found that the
most appropriate categorization method for denominations is one that classified
denominations along a spectrum from most conservative (fundamentalist, orthodox or

Evangelical) to most liberal (secular, modem or humanistic)“.

Due to the limitations of the data on church membership, we are initially using data on

eight denominations as the measure of the conservative composition of a state’s

 

‘5 However, Steensland et a1 (2000) argue that, “T.W. Smith (1990) does not adequately capture
essential historical differences between American religious traditions and offers no way of measuring some
recent trends in religious affiliation”. They therefore develop a different classification system that registers
mainline Protestants, evangelical Protestants, black Protestants, Roman Catholic, Jewish and others (e.g.
Mormon, Muslim, etc.) as separate denominational classifications. In our analysis, we have classified
denominations according to the schemes discussed in Smith (1990) and used by many other studies. As
evident from the Steensland et al (2000) classification above, their approach does not allow us to classify
denominations according to religiosity, and thus would make it practically difl'lcult to find a measure of the
religiosity or strictness of a state’s population, which is necessary for our study.

119

population“. The percentage of the total population that belong to these denominations
will serve as our measure of the conservative composition, and hence religiosity, of a
state’s population“. We will later include the Catholic Church in this measure and
observe the effect this has on the results. Although the small number of denominations
may seem to limit the scope of our analysis, it is important to note that, within the
survey’s data, the conservative denominations used in this analysis comprise a substantial
percentage of the total adherent population in the nation as a whole“. These eight
denominations represent the prominent conservative denominations in the US. See table
3.1 for the percentage of the adherent population that belongs to these eight

denominations”.

Table 3.2 comprises the eight conservative denominations, in addition to the Catholic

Church 70.

 

(’6 These are the only conservative denominations that had complete data for all four years and all fifty
states.

67 The eight denominations are as follows: Church of God (Anderson, Indiana), Church of God
(Cleveland, Tennessee), The Church of Jesus Christ of Latter-day Saints, Church of the Nazarene, Lutheran
Church-Missouri Synod, The Salvation Army, Seventh-day Adventist Church and the Southern Baptist
Convention.

68 Heath et a1 (1995) limit their analysis to denominations with 100,000 or more adherents, as the
inclusion of the smaller denominations will have little or no effect on the final results.

69 The adherent population refers to only the denominations included in the surveys. The study does not
include all denominations in the nation.

70 Black churches have generally been hugely underrepresented in the survey data. The slight exception
is the 1980 survey where four denominations were included (African Methodist Episcopal Zion, Bible
Church of Christ, Christian Methodist Episcopal Church and Fire Baptized Holiness Church) and the 1990
survey that had the same denominations (as 1980) with the exception of the Christian Methodist Episcopal
Church. The majority of black churches were not included in the 1971 and 2000 surveys. However, the
churches included in 1980 and 1990 constitute a small percentage of total black membership in the US.
and a very small percentage of total adherents included in the surveys (about 1.6 and 0.8 percent in 1980
and 1990 respectively). Also, the two large denominations included (African Methodist Episcopal Zion and
Christian Methodist Episcopal Church) are classified as ‘moderate’ and ‘liberal’ respectively by Smith
(1990). Furthermore, in similar method to Hull and Bold (1998), we estimate the equation for the subset of
the data with proportion of African-Americans less than or equal to 10% of the population. We did not
detect any change in the results. It is reasonable therefore to assume that the missing data on black churches
will not have a profound impact on our final results, since the measurement error seems to be generally

120

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Year Percentage of total adherents
1971 21.9
1980 21.8
1990 20.4
2000 21.4
Table 3.1
Year Percentage of total adherents
1971 66.4
1980 64
1990 59.3
2000 65.3
Table 3.2

From the tables, we can observe that even with our data limitation to eight
denominations plus the Catholic Church, these denominations combined still represent a
significant percentage of total adherents in the population, especially if we take into
consideration that the population also comprises of liberal and moderate denominations

that also have significant adherent populations". Furthermore, in order to fully ensure

 

systematic. We do also include ‘black’ as a separate explanatory variable to capture the effect that this
group of the population has on government policies and decisions.

7' For instance, in 1971, the percentage of the adherent population that belongs to the main liberal and
moderate denominations is 20.6 percent Similar figures for 1980, 1990 and 2000 are 18.4, 14.4 and 12.8
percent respectively. The liberal denominations considered here are: Episcopal Church, Friends (Quakers),
Unitarian Universalist Association of Congregations, United Church of Christ and The United Methodist

121

that this limitation will not affect the reliability of our results, we created alternative
measures that include more conservative denominations72 for the years for which the data
is available, and found a very high correlation coefficient between our main measure and
these alternative measures”. This means that the conservative denominations we are
using sufficiently represent the complete set of conservative denominations in all states.

In other words, our measure of the conservative composition of a state’s population is a
very reliable and representative measure. In addition to this measure, we will later use a
broader measure that includes some moderate denominations“. These denominations,
although classified as generally moderate by Smith (1990), are similar to conservative
denominations in many respects. By examining the sources that Smith (1990) relied upon
to construct the religious classifications, it becomes clear that different studies classified
these denominations differently. For example, Johnson (1962) and Gay (1980) classify
Disciples of Christ (Christian Church) as ‘fundamentalist’, while McCutcheon (1984)
classifies it as ‘conservative’ and Wood (1970), Gill (1982) and Glock and Stark (1965)

classify it as ‘moderate’. Another example is Church of the Brethren that is classified as

‘fundamentalist’ by Chi (1982) and Gay (1980), while Gill (1982) classifies it as

 

Church. The moderate denominations are: American Baptist Churches in the USA, Christian Church
(Disciples of Christ), Christian Reformed Church in America, Church of the Brethren and the Reformed
Church in America.

72 The additional denominations are: Assemblies of God, Baptist General Conference, Baptist
Missionary Association of America, The Christian and Missionary Alliance, Churches of Christ, Free
Methodist Church of North America, International Church of the Foursquare Gospel, International
Pentecostal Holiness Church and the Wisconsin Evangelical Lutheran Synod. The combined eighteen
denominations comprise 68.1, 63.3 and 69.7 percent of total adherents in 1980, 1990 and 2000 respectively.

73 For example, the correlation coefficient between our main measure (8 conservative denominations)
and the alternative measure (17 conservative denominations) for the change in church membership from
1980 to 1990 was 0.9757. Thus, our religiosity measure is a very reliable measure of the conservative
‘nature’ of a state’s population.

74 These are: American Baptist Churches in the USA, Christian Church (Disciples of Christ), Christian
Reformed Church in America and Church of the Brethren. All the denominations included in this broad
measure constitute 69.6, 67.2, 61.8 and 67.6 percent of total adherents in 1971, 1980, 1990 and 2000
respectively.

122

‘moderate’. It is therefore useful to consider these denominations as conservative as an
extension to our analysis, in order to fillly assess the impact that conservative adherents

and denominations have on government policies.

It is likely that long-standing cross-state differences in both religiosity and tax policy
are affected by unobserved state characteristics. As a result, we employ both a first
differencing approach and a fixed effect approach to eliminate bias due to unobserved
effects in our data. The first-differenced equation will have the following form:

Estimating Equation:
ATAX RATE = [in + £1 ACONS.C + fl; AHIGH + #3 ACOLL + fl, AF EM +

p, ABLACK + p, AHISP + p, AUNEM + 5, D80 + p, 090 + Ae

where75 :

TAX RATE = ratio of total state tax collections to total state personal income76

CONS.C = percent of state’s population that belongs to the eight conservative
denominations and the Catholic Church77

HIGH = percent of state population with high school diploma

COLL = percent of state population with college diploma

FEM = percent of state households headed by females

BLACK = percent of state population that is black

HISP = percent of state population that is Hispanic

UNEM = state unemployment rate

D80 = a dummy variable taking the value one if change is from 1970 to 1980,
zero otherwise

D90 = a dummy variable taking the value one if change is from 1980 to 1990,

zero otherwise
e = random error term, independently and normally distributed

 

75 The data sources are given in Appendix B, pp. 138.

7° The change is a 10 year range, such as 1970—1980, etc.

77 We will also consider results when the Catholic denomination is not included in the conservative
measure. When separately considered, they are CONS and CATH.

123

 

 

 

Variable Mean SD Minimum Maximum
Year = 1971
TAX RATE .0635863 .0129562 .0351111 .1001471
CONS 13.398 13.73258 .6 75.8
CATH 19.67 13.67417 1.4 63.6
HIGH 53.108 8.076998 37.8 67.3
COLL 10.572 2.1517 6.6 14.9
FEM 10.09085 3.106082 6.158948 27.62543
BLACK 8.668 9.236008 .1 36.8
HISP 3.356 5.384147 . 30.3
UNEM 4.482 1.252864 2.7 9.2
Year = 1980
TAX RATE .0641083 .0276795 .0290217 .2396667
CONS 12.676 12.7575 .5 68.5
CATH 19.074 13.73348 1.6 63.7
HIGH 67.338 7.568326 51.9 82.8
COLL 16.072 3.004082 9.7 23
FEM 12.95715 2.510142 7.944283 18.35509
BLACK 9.14 9.221935 .2 35.2
HISP 4.29 6.594749 .4 36.6
UNEM 6.838 1.612311 4 12.4
Year = 1990
TAX RATE .0650941 .0143656 .0258696 .1226984
CONS 13.332 13.65942 .6 73
CATH 18.874 13.31946 2.2 63.1
HIGH 75.686 7.109609 45.6 86.6
COLL 19.756 3.746681 12.3 27.2
FEM 14.75217 2.65576 10.23253 20.92812
BLACK 9.534 9.249326 .3 35.6
HISP 5.372 7.519391 .5 38.2
UNEM 5.468 1.152006 2.2 8.4
Year = 2000
TAX RATE .0684919 .0125684 .0412652 .0986686
CONS 12.366 12.58314 .9 67.4
CATH 19.61 12.40547 3.2 51.7
HIGH 85.482 3.959782 77.1 91.8
COLL 24.932 4.312319 15.3 34.6
FEM 16.8481 2.966972 12.04819 24.23026
BLACK 9.902 9.580134 .3 36.3
HISP 7.786 8.914721 .7 42.1
UNEM 3.902 .9464197 2.2 6.7
Table 3.3
Summary Statistics

124

Based on the propositions that were derived fiom our theoretical model, we predict that
a higher concentration of conservative adherents (i.e. a higher religiosity level) in a state
will lead to a decrease in the state tax rate over time, all other things being equal. With
regards to the Catholic denominations, the signs seem unclear. This is because the
Catholic Church tends to be labeled a moderate denomination as in Smith (1990). As
such, we will estimate the model both when the Catholic denomination is excluded and
included in order to observe the effect it has on the results. The relationship between
education (both HIGH and COLL) and tax rates is not clear cut. This is because higher
state education levels would increase personal income78 and at the same time increase
state tax revenues. A definite relationship can be suggested if education has a direct or
indirect effect on tax policies or tax legislation changes.

The relationship between tax rates and the percentage of families that are female-
headed is slightly ambiguous. This is because one would expect that as the percentage of
female-headed households increases, the government would attempt to reduce the tax
burden on these families since they would be economically disadvantaged relative to the
rest of the population. However, according to a study by Kula (2004), changes in state
government expenditures on welfare programs (such as Medicaid), had no significant
effect on state tax laws over time. In other words, state governments did not alter tax laws
in response to changes in their expenditures on welfare programs. This may suggest that
an increase in the proportion of female-headed households, which depend to some extent
on welfare programs, will not necessarily lead to a change in tax rates over time.

Furthermore, this rejects any positive correlation, as one would have argued that an

 

78 Heath et a1 (1995) also argue that education should be positively related to income since it measures
human capital levels in each state.

125

increase in the percentage of female-headed households would actually raise tax rates in
order to finance the welfare programs. The signs on BLACK and HISP can also go either
way, due to similar reasoning to the FEM case. In the case of unemployment, Kee and
Shannon (1992) show that tax policy is responsive to business cycles, and in particular,
that states raise taxes when their unemployment rate increases. However, David
Merriman finds no significant correlation between state tax policy changes and
unemployment rates. The results of estimating our model by the first-differencing and

fixed effects methods are given in tables 3.4 and 3.5 respectively.

3.4a. Church Membership

Table 3.4: Estimation Results using “Church Membership”-First Differencing Approach

Estimation Results-First Differencing Approach- Dependent Variable: DTAX

 

 

Variable Name Excluding Catholic Including Catholic
DCONS (DCONSC) 0.000 -0.001
(0.34) (1.15)
DHIGH 0.000 0.000
(0.72) (.73)
DCOLL 0.002 0.002
(2.20)* (2.06)‘
DFEM 0.001 0.001
(1.11) (1.22)
DBLACK , -0.001 -0.001
(0.49) (0.48)
DHISP 0.000 0.000
(0.21) (0.21)
DUNEM -0.002 -0.001
(1.25) (1.13)
D80 0.001 -0.001
(0.11) (0.12)
D90 0.002 0.002
(0.36) (0.34)
Constant -0.016 -0.015
(1.96) (1.86)
Observations l 50 l 50
Adjusted R-squared 0.06 0.07

Absolute value of t statistics in parentheses
* Significant at 5%; " significant at 1%

126

The second column (in Tables 3.4 and 3.5) contains the results of the regression with
Catholics included in the religiosity measure. In both the first-differencing and fixed
effects approaches, the coefficients on the conservative measure are statistically
insignificant. The signs of the coefficients become negative when the Catholics
denomination is included in the religiosity measure. Although in this case the sign is
consistent with our hypothesis that tax rates decrease with religiosity, the coefficients are
insignificant. This means that, based on the evidence, we cannot reasonably conclude that
tax rates decrease as the conservative religious composition of a state increases. The
change in the sign of the religiosity coefficient when the Catholic Church was added hints
at the possibility that Catholics may be conservative on economic issues. The only
statistically significant coefficient in our analysis (fiom the first-differencing approach) is
the positive relationship between tax rates and the percentage of the population that has a
college diploma. Although insignificant, the sign on the FEM and BLACK variables are
positive and negative respectively. The positive relationship implies that as the
percentage of female-headed households increases, the government raises taxes to fund
more welfare programs. The negative relationship implies that tax rates decrease as the
black population in a state increases. The UNEM coefficient’s sign is consistent with
David Merriman’s results of no significant correlation between state tax policy changes
and unemployment rates. The sign on the coefficient however rejects the Kee and
Shannon (1992) result that states raise taxes when their unemployment rate increases.

The general observation fiom both regression methods is the absence of any significant
relationship between tax rates and the explanatory variables. This would imply that these

variables describe very little of the variation in tax rates over time within states.

127

Table 3.5: Estimation Results using “Church Membership”-Fixed Effects Approach

 

Estimation Results-Fixed Effects Approach— Dependent Variable: TAX

 

 

Variable Name Excluding Catholic Including Catholic
CONS (CONSC) 0.001 -0.000
(0.82) (0.33)
HIGH 0.000 0.000
(0.73) (0.64)
COLL 0.001 0.001
(1.29) (1.29)
FEM 0.001 0.001
(0.61) (0.62)
BLACK -0.001 -0.000
(0.42) (0.35)
HISP 0.000 0.000
(0.66) (0.38)
UNEM -0.002 -0.002
(1.49) (1.35)
Constant 0.018 0.034
(0.49) (0.65)
Observations 200 200
Adjusted R-squared 0.64 0.64

Absolute value of t statistics in parentheses
" Significant at 5%; ** significant at 1%

We also used a broader measure of the conservative nature of a state’s population by
including ‘moderate’ denominations in our list of conservative denominations. The
results are similar to those in Tables 3.4 and 3.5, with no significant relationship between
tax rates and the dependent variables.

In addition, we increased our date range from 10 to 20 years to observe whether any
notable relationship exists for specifications with longer time periods. The time intervals
are now (1970-1990) and (1980-2000)”. The results are similar to the above results with
no significant relationship present. This means that even over longer intervals of time,

there is no observed effect of religiosity on changes in tax rates.

 

79 This gives us 100 observations in a first-differencing approach.

128

Next we employ a different measure of religiosity, namely religious contributions.
This is to provide an alternative measure of religiosity, and also to observe differences

in the final results from both analyses.
3.4b. Religious Contributions80

Table 3.6: Estimation Results using “Religious Contributions”

 

Estimation Results-First Differencing Approach-Dependent Variable: DTAX

 

Variable Name Average household religious contribution % of households contributing (religious)

 

DAVERC 0.000
(1.29)
DPERRC 0.006
(0.30)
DHIGH 0.000 0.000
(0.92) (0.63)
DCOLL 0.002 0.001
(0.98) (0.75)
DFEM -0.001 -0.001
(0.14) (0.19)
DBLACK 0.001 0.001
(0.31) (0.51)
DHISP 0.001 0.001
(0.69) (0.67)
DUNEM -0.002 -0.002
(1.04) (1.05)
Constant -0.016 -0.013
(0.84) (0.65)
Observations 30 30
Adjusted R-squared 0.22 0.16

Absolute value of t statistics in parentheses
"' Significant at 5%; " significant at 1%

We will use two different measures for religious contributions, the average household
contribution by state8| per year to religious organizations and the percentage of state

households contributing to religious causes”. The results are given in table 3.6 for the

 

80 We are grateful to Tom DeLeire for making this data available to us.

8' Due to data limitations, we only have data for 30 states.

82 We are using averages over 1982-85 (to represent the 1990 religiosity measure) and 1995-98 (to
represent the 2000 religiosity measure). We are implicitly assuming that the impact of religiosity (measured

129

first-differencing approach. The first column uses average household religious
contribution by state in a year (AVERC) as the religiosity variable while the second
column uses the percentage of households in the state that have contributed at all to
religious organizations (PERRC) as the religiosity measure.

From the table, it is clear that no significant relationship can be detected between tax
rates and religious contributions. Furthermore, the sign on AVERC and PERRC are not
negative, which is different from our previous measure of religiosity. However, due to the
insigrrificance of any of the results, we can neither draw definite conclusions nor make
reasonable comparisons between the different coefiicients that we use as religiosity
measures.

To summarize, no significant relationship was observed between tax rates and any of
our explanatory variables. In particular, the results of our analysis have provided no
significant relationship between tax rates and religiosity levels, measured by percentage
membership in conservative denominations and household religious contributions. The
results therefore fail to support our main hypothesis that tax rates tend to decrease as the
religiosity of a population increases. In other words, from our results, there is no evidence
that as the conservative adherents in a population increase, the government would

respond by decreasing taxes or enacting new legislation to reduce taxes.
3.5 Conclusion

The theoretical model we developed in this paper aimed to model the relationship

between religious and economic preferences. In particular, we constructed a model to

 

by religious contributions) on tax rates will occur some years later. Thus, we have two sets of data and
variables at the state level.

130

describe the relationship between the religiosity of an individual and the tax rate he/she
prefers. The model proposed that individuals that prefer high levels of religiosity and as
such belong to conservative denominations will prefer to give less to the government in
the form of taxes. This is because these individuals will instead prefer to give more to
religious causes as an expression of their strict commitment to the demanding rules of the
doctrine they adopt, and the institution they belong to. This rationale assumes that
individuals face a budget constraint that forces them to allocate their income into
preferred levels of religious investment (money and time to religious causes) and
‘secular’ investment (government taxes and private consumption). These religious
preferences of individuals will therefore translate into economic and political preferences.
Conservative denominations that 'sincerely represent the interests of their adherents will
induce them to support political parties and government policies that advocate lower
taxes and government spending.

We then tested these propositions using US. state-level data for a period spanning
from 1970 to 2000. We used two sets of data as measures of the religiosity of state
populations. The first measure was church membership in conservative denominations
and the second was religious contributions. The measure of church membership was
broadened to include more denominations to ensure that we capture the impact of
religiosity levels as best as possible. The empirical results did not include significant
relationships to support our hypothesis that tax rates decrease as religiosity levels
increase, although we did sometimes obtain coefficient signs that are consistent with our
predictions. We also obtained similar results when we increased the date range to monitor

changes over longer periods of time. The absence of results to support our hypothesis

131

suggests the possibility that certain assumptions of our model do not hold in practice.

One such assumption is that religious authorities sincerely represent the preferences of
their adherents in the economic and political spheres. It could be that the assumption of
the ‘benevolence’ of religious authorities does not fully reflect the reality of religious
institutions in the US. and their true objectives. Another assumption is that these
religious preferences will eventually map themselves into corresponding political choices
on elections day and eventually into government policies that fulfill campaign promises.
This is also an overly ambitious assumption, since we do know that politics is a very
complex and unpredictable process, and post-election policies do not genuinely reflect
pre-election promises. Furthermore, elections and subsequently, government policies
depend on many other factors other than religious preferences. Sometimes, other factors
may outweigh the impact of religion on the political and economic decisions of the
government. Finally, changes in tax legislation and tax policies over time are sometimes
very slow and minimal due to the complex nature of tax structures and the many
obstacles that stand in the way of such changes.

The paper’s findings encourage a deeper examination of the question at hand, and not a
conviction that the results are final and conclusive. Further research should attempt to
consider different measures of tax policy changes and also an attempt to correct for
possible deficiencies in the data. An alternative would be to measure the incidence of tax
legislation as a proxy for tax preferences. It should also incorporate political factors and
outcomes into the analysis, such as preferences of parties and politicians. Finally, one can
also consider the preferences of religious leaders and authorities in order to examine the

correlation between their preferences and those of their respective congregations.

132

CONCLUSION

The three essays in this dissertation have attempted to study different aspects of the
relationship between religion, economics and politics, using neoclassical economic
theory as our analytical apparatus. In addition, we incorporated into our models concepts
and theories from political science, religion and sociology. The use of these concepts and
modeling techniques has proved to be both useful and informative.

In formal modeling, a significant portion of the overall picture is lost when a one-
dimensional or reductionistic approach is employed. This is naturally due to the overly
restrictive perspective the assumptions impose on the model, and this extends to a
corresponding restriction on the potential descriptiveness of the results and propositions.
Despite these reservations, economic theory has much to offer in terms of specificity,
exactness and insights.

Economics models allow us to zoom in on the specific object of inquiry that is of
interest. This is evident in our essays, since we were capable of focusing on questions of
interest, such as the impact of religiosity on preferences for taxes and government
spending. Similarly, we studied the impact of government funding of religious charities
on religiosity levels in the population and how this affects the competition between
religious denominations. In addition, this specificity allows us to derive exact
propositions regarding the relationship between different institutions in society. This
helps us in overcoming the real complexities that so often prevent us from arriving at
practical and comprehensible insights. By addressing such a complex system as the
interaction of religion, economics and politics, the models presented us with useful

insights and important policy implications for the present and the future. In light of the

133

above, we strongly believe that the models and results of the three essays have much to
offer.

The main contribution of the first essay is its effort to explain the behavior of political
parties and religious denominations explicitly. As evident in the review of literature on
the economics of religion, the behaviors of parties and denominations have not been
jointly studied in a framework that explains both their independence and their
interdependence. The model explains the behavior of religious denominations in the
religious domain and the behavior of political parties in the political-economic domain.
The model offers an economic perspective for the ‘Culture Wars’ thesis that we observe
in society. As such, this is the first effort to formally model this real world phenomenon.

The second essay also presents the first formal examination of the relationship between
the government and religious charities. The model studies the impact of government
funding on religious charities and religion in general. We develop an original model to
describe the provision of a social service by religious clergy. Using the concept of
compensating wage differentials, since the clergy derive additional utility from
proselytizing their particular doctrine, they are thus willing to demand a lower supply
price for their labor. The model derives important policy implications for present and
future government faith-initiatives.

In the third essay, we empirically test the hypothesis that higher levels of religiosity in
states will lead to lower tax rates over time. The insignificance of the results gives a
possible indication that some assumptions do not hold in practice. As such, this
encourages an extended study of the hypothesis, by investigating the empirical reliability

of the model’s assumptions and any alternative formulation that may prove more useful.

I34

APPENDICES

135

APPENDIX A: Religiosity

In this appendix, we discuss the issue of religiosity in more detail. The question as to
whether ‘religious investment’ can be used to represent religiosity is one of potential
controversy. This is so because it introduces the problems associated with ‘fiee-riding’.
There is a tendency for consumers of religious goods to free ride. This is due to the fact
that religious goods produce an extemality effect that is unaccounted for by the religious
market. Iannaccone (1994) argues that the collective nature of religious consumption
induces some members or adherents of a denomination to contribute less than others. The
‘club good’ effect yields contribution levels that are disproportionate to the individual
consumption levels. Ianaccone (1994) states that “people with low levels of religious
commitment tend to free ride off those with higher levels; they tend to take more than
they give.” If this is the case, the problem arises as to whether one can use ‘religious
investment’ as a proxy for ‘religiosity”. Iannaccone (1994) states that “the aspects of
religious participation that confer the greatest external benefits (effort, enthusiasm,
solidarity, etc.) are intrinsically difficult to monitor and reward. The willingness to pay
membership dues is a poor proxy for these qualities because income correlates weakly
with most dimensions of religious commitment. . .”

Zaleski and Zech (1995) acknowledge that, “economists are aware that people’s
willingness to pay for a product, even through voluntary contributions, measures their
satisfaction with that product. Analyzing contributions can offer séme valuable insights

into the effect of competition in the religious market place.”

136

However, they argue that, “to the extent that churches are populated by free riders,
the use of contributions as a measure of religiosity might be misleading.” (Zaleski and
Zech (1995)

These perspectives thus question the modeling of religiosity as a function of
voluntary contributions and attendance. The arguments we present in response to these
concerns are as follows:

The problem that the correlation between income and “religious commitment” can
pose is not present in the framework of our model, because we are assuming that all
individuals have the same income level regardless of their actual beliefs and preferences.

Also, the model is assuming honesty and consistency on the part of religious
consumers. The willingness to contribute to charity or attend sermons is a sincere
representation of an individual’s demand for religion. This means that free riding may be
characterized as a ‘sinful’ act. One may also argue that it is an act of selfishness and
therefore, counter-religious. The issue of free riding is essentially distinguishing theory
from practice, and preferences correspond to theory and not practice. If in practice,
individuals do free ride, then that can be modeled as an extension to the model in the
form of uncertainty, where religious investment follows some specified probability
distribution. Then, we can distinguish between ideal preferences and actual deeds.

In most religious traditions, voluntary contributions and attendance at sermons is an
integral part of the doctrine. To give charity is thus a religious obligation. And the
amount one prefers to give represents his/her commitment to these religious obligations,
assuming income is the same for all. Given the above, we argue that our use of ‘religious

investment’ as a proxy for religiosity is both reasonable and informative.

137

APPENDIX B: Data Sources

Church Membership]:
1971: Churches and Church Membership in the United States 1971, Table 2, pp. 3-14.
1980: Churches and Church Membership in the United States 1980, Table 3, pp. 10-27.
1990: Churches and Church Membership in the United States 1990, Table 3, pp. 12-36.
2000: Churches and Church Membership in the United States 2000, Table 3, pp. 14-42.

State Tax Collections:
1970: Statistical Abstract of the United States, 1971, Table 629, p.412.
1980: Statistical Abstract of the United States, 1981, Table 490, p.293.
1990: Statistical Abstract of the United States, 1992, Table 463, p.290.
2000: Statistical Abstract of the United States, 2002, Table 429, p.282.

Personal Income:
1970: Statistical Abstract of the United States, 1972, Table 519, p.319.
1980: Statistical Abstract of the United States, 2002, Table 642, p.425.
1990: Statistical Abstract of the United States, 2002, Table 642, p.425.
2000: Statistical Abstract of the United States, 2002, Table 642, p.425.

Percent of state population with high school diploma:
1970: Statistical Abstract of the United States, 1977, Table 223, p.139.
1980: Statistical Abstract of the United States, 1981, Table 227, p.144.
1990: Statistical Abstract of the United States, 2001, Table 219, p.141.
2000: Statistical Abstract of the United States, 2001, Table 219, p.141.

Percent of state population with college diploma:
1970: Statistical Abstract of the United States, 1977, Table 223, p.139.
1980: Statistical Abstract of the United States, 1981, Table 227, p.144.
1990: Statistical Abstract of the United States, 2001, Table 219, p.141.
2000: Statistical Abstract of the United States, 2001 , Table 219, p.141.

Percent of state households headed by females:
1970, 1980 and 1990: State Statistical Abstracts’. _
2000: Statistical Abstract of the United States, 2001, Table 56, p.50.

Percent of state population that is black:
1970: US Census Bureau, 2002, Table A-7.
1980: US Census Bureau, 2002, Table A-3.
1990: US Census Bureau, 2002, Table A-1.
2000: Statistical Abstract of the United States, 2001, Table 24, P27.

 

' Glenmary Research Center.
2 These percentages were manually computed from each state’s individual Census Population Survey.

138

Percent of state population that is Hispanic:
1970: US Census Bureau, 2002, Table A-7.
1980: US Census Bureau, 2002, Table A-3.
1990: US Census Bureau, 2002, Table A-1.
2000: Statistical Abstract of the United States, 2001, Table 23, P25.

State unemployment rate:
1970: Census Population Survey, 1970.
1980, 1990 and 2000: Bureau of Labor Statistics: Local Area Unemployment
Statistics, 2004.

Religious Contributions:
Tom DeLeire, personal communication.

139

BIBLIOGRAPHY

140

BIBLIOGRAPHY

Aaranson, William E.; Zinn Jacqueline S. and Rosko, Micheal D. (1994). “Do For-Profit
and Not-for- Profit Nursing Homes Behave Differently?”, Gerontologist, vol. 34. pp.
775-86.

Anderson, G. M. (1988). “Mr. Smith and the Preachers: The Economics of Religion in
the Wealth of Nations.” The Journal of Political Economy, vol. 96, pp. 1066-1088.

Anderson, G. M., Ekelund, R. B., Hebert, R. F. and Tollison, R. D. (1992). “An
Economic Interpretation of the Medieval Crusades”, Journal of European Economic
History, vol. 21, pp. 339-63.

Andreoni, James. (1989). “Giving with Irnpure Altruism: Applications to Charity and
Ricardian Equivalence”, Journal of Political Economy, vol. 97, pp.1447-58.

. (1990). “Impure Altruism and Donations to Public Goods: A Theory of
Warrn-Glow Giving”, Economic Journal, vol. 100, pp.464-77.

 

Azzi, C. and Ehrenberg, R. G. (1975). ‘Household Allocation of Time and Church
Attendance’, Journal of Political Economy, vol. 84, pp.27-56.

Barro, Robert J. and Rachel McCleary. (2003). “Religion and Economic Growth.” NBER
working paper, May 2003.

Barros, P. P. and Garoupa, N. (2002). “An Economic Theory of Church Strictness”, The
Economic Journal, vol. 1 12, pp. 559-576.

Berger, PL. (1967). The Sacred Canopy: Elements of a Sociological Theory of Religion,
Garden City, Doubleday.

Besley, Timothy and Anne Case. (1995). “Incumbent Behavior: Vote-seeking, tax-setting
and yardstick competition.” American Economic Review, vol. 85, pp. 25-45.

Becker, Gary S. (1965). “A Theory of the Allocation of Time”, Economic Journal, vol.
75(299), pp. 493-517.

. (1974). “A Theory of Social Interactions”, The Journal of Political
Economy, vol. 82, pp. 1063- 93.

 

Becker, Gary S. and Kevin M. Murphy. (1988). “A Theory of Rational Addiction”, The
Journal of Political Economy, vol. 96, pp. 675-700.

Berrien, J., 0. Mc Roberts, and C. Winship. (2000). “Religion and the Boston Miracle:
The Effect of Black Ministry on Youth Violence. In Who Will Provide? The Changing

141

Role of Religion in American Social Welfare, edited by M. J. Bane, B. Coffin, and R.
Thiernann, pp. 266-85. Boulder, CO: Westview Press.

Case, Anne. (1993). “Interstate Tax Competition after TRA86.” Journal of Policy
Analysis and Management, vol. 12, pp. 136-148.

Chaves, M. (1994). “Secularization as Declining Religious Authority.” Social Forces,
March, pp. 749-774.

Chi, Kenneth. (1982). “Religious Affiliation and Marital Success: Questioning the
Established View Point.” Unpublished Master’s Thesis, Ohio State University.

Clarkson, Kenneth W. and Martin, Donald L., eds. (1980) The Economics of
Nonproprietary Organizations. Greenwich, CT: J AI Press.

Comes, Richard, and Sandler, Todd. (1984). “Easy Riders, Joint Production, and Public
Goods”. Economic Journal, vol. 94, pp. 580-98.

. (1994). “The Comparative Static Properties of the Irnpure Public Good
Model”. Journal of Public Economics, vol. 54, pp. 403-21.

 

Davidson, A. B. and Ekelund, R. B. (1997). “The Medieval Church and Rents fi'om
Marriage Market Regulations”, Journal of Economic Behavior and Organization,
vol. 32, pp. 215-45.

Davis, Nancy J ., and Robert V. Robinson. (1996). “Are the Rumors of War Exaggerated?
Religious Orthodoxy and Moral Progressivism in America", American Journal of
Sociology, vol. 102, pp. 756-787.

Desmond, D. and J. Maddux. (1981). “Religious Programs and Careers of Chronic
Heroin Users”, American Journal of Drug and Alcohol Abuse, vol. 8, pp. 71 -83.

Durkin, J. T. and Greeley, A. (1991). "’A Model of Religious Choice Under
Uncertainty”, Rationality and Society, vol. 3, pp. 178-96.

Ebaugh, H. R., Pipes, P. F ., Chafetz, J. S. and Daniels, M. (2003). “Where’s the
Religion? Distinguishing Faith-Based from Secular Social Service Programs”, Journal
for the Scientific Study of Religion, vol. 42, pp. 411-26.

Ekelund, R. B., Hebert, R. F. and Tollison, R. D. (1989). “An Economic Model of the
Medieval Church: Usury as a Form of Rent Seeking”, Journal of Law, Economics and
Organization, vol. 5, pp.307-31.

. (1992). “The Economics of Sin and Redemption: Purgatory as a Market-
pull Innovation?” Journal of Economic Behavior and Organization, vol. 19, pp. 1-15.

 

I42

. (1996). Sacred Trust: The Medieval Church as an Economic Firm.
Oxford: Oxford University Press.

 

Ewing, Bradley T. (2000). “The wage effects of being raised in the Catholic Religion:
does religion mater.” The American Journal of Economics and Sociology, vol. 59, pp.
419-32.

Finke, Roger. (1997). “The Consequences of Religious Competition: Supply-Side
Explanations for Religious Change.” in Rational Choice Theory and Religion:
Summary and Assessment. Ed. Lawrence A. Young. New York, NY: Routledge.

Finke, R. and Stark, R. (2000). Acts of Faith, Los Angeles, CA: University of California
Press.

. (1992). The Churching of America, 1 776-1990, New Brunswick, NJ:
Rutgers University Press.

 

Gay, David A. (1980). “Do Protestant denominational subcultures exist? The issue and its
dimensions.” Master Thesis, Texas Tech University.

Gill, Anthony. “The Political Origins of Religious Liberty: Initial Sketch of a General
Theory,” Forthcoming.

. (1998). Rendering Unto Caesar: The Catholic Church and the State in
Latin America, Chicago, IL: The University of Chicago Press.

 

Gill, Sandra K. (1982). “Supporting equality for women: An analysis of attitudes towards
the equal rights Amendment.” Unpublished Ph. D. dissertation, University of Oregon.

Glock, Charles Y. and Rodney Stark (1965). Religion and Society in Tension, Chicago:
Random House.

Glazer, Amihai, and Konrad, Kai A. (1996). “A Signaling Explanation for Charity”,
American Economic Review, vol. 86, pp. 1019-28.

Green, John C., James L. Guth and Cleveland R. Fraser, (1991). “Apostles and
Apostates? Religion And Politics among Party Activists.” In The Bible and the Ballot
Box: Religion and Politics in the 1988 Election, ed. James L. Guth and John C. Green.
Boulder, CO: Westview.

Green, John C., James L. Guth, Lyman A. Kellstedt, and Corwin E. Smidt. (1996).
Religion and the Culture Wars: Dispatches from the Front. Lanharn, MD: Rowman &
Littlefield.

Grier, Robin. (1997). “The effect of religion on economic development: a cross national
study of 63 former colonies.” Kyklos, vol. 50, pp. 47-62.

143

Grossman, G. M. and Helpman, E. (2001). Special Interest Politics, Cambridge, MA: The
MIT Press.

Guidance to Faith-Based and Community Organizations on Partnering with the Federal
Government. Washington, DC: White House Office of Faith-Based and Community
Initiatives. Retrieved September 5, 2004, httn://www.whitehouse.gov/govemment/fbci/

Guiso, Luigi, Paola Sapienza and Luigi Zingales. (2003). “People’s opium? Religion and
economic attitudes.” Journal of Monetary Economics, vol. 50, pp. 225-282.

Guth, James L., and John C. Green. (1986). “Faith and Politics: Religion and Ideology
among Political Contributors.” American Politics Quarterly, vol. 14, pp. 186-99.

. (1987). “The Moralizing Minority: Christian Right Support among
Political Contributors.” Social Science Quarterly, vol. 68, pp. 598-610.

 

Hawes, Catherine and Phillips, Charles D. (1986). “‘The Changing Structure of the
Nursing Home Industry and the Impact of Ownership on Quality, Cost, and Access”, in
F or-Profit Enterprise in Health Care. Ed.: Bradford H. Gray. Washington, DC:
National Academy Press, pp. 492-541 .

Heath, W. C., M. S. Waters and J. K. Watson. (1995). “Religion and Economic Welfare:
An empirical analysis of state per capita income.” Journal of Economic Behavior and
Organization, vol. 27, pp. 129-142.

Heittich, Walker and Stanley L. Winter. (1988). Democratic Choice and Taxation: A
Theoretical and Empirical Analysis, Cambridge: Cambridge University Press.

Hoge, Dean R. and Fenggang Yang. (1994). “Determinants of Religious Giving in
American Denominations: Data from Two Nationwide Surveys,” Review of Religious
Research, vol. 36, pp. 123-48.

Holcombe, Randall G. (1998). “Tax Policy from a Public Choice Perspective.” National
Tax Journal, vol. 51, pp. 359-371.

Hotelling, H. (1929). ‘Stability in Competition’, Economic Journal, vol. 39, pp.41-57.

Hull, Brooks B. and Frederick Bold. (1998). “Product Variety in Religious Markets”,
Review of Social Economy, vol. LVI, pp.1-l9.

Hunter, James Davidson. (1991). Culture Wars: The Struggle to Define America. New
York: Basic.

144

Iannaccone, L. R. (1988). “A Formal Model of Church and Sect”, American Journal of
Sociology, vol. 94 (Supplement), pp. 5241-68.

. (1990). “Religious Participation: a Human Capital Approach”, Journal
for the Scientific Study of Religion, vol.29, pp.297-314.

 

. (1991). “The Consequences of Religious Market Structures: Adam
Smith and The Economics of Religion.” Rationality and Society, April, pp. 156-177.

 

. (1992). “Sacrifice and Stigma: Reducing Free-riding in Cults,
Communes and Other Collectives”, Journal of Political Economy, vol. 100, pp.271-92.

 

. (1994). “Why Strict Churches are Strong”, American Journal of
Sociology, vol. 99, pp. 1180-1211.

 

. (1995). “Risk, Rationality and Religious Portfolios”, Economic
Inquiry, vol. 38, pp.285-95.

 

. (1997). “Rational Choice: Framework for the Scientific Study of
Religion.” in Rational Choice Theory and Religion: Summary and Assessment. Ed.
Lawrence A. Young. New York, NY: Routledge.

 

. (1998). “An Introduction to the Economics of Religion”, Journal of
Economic Literature, Vol. 36, pp.1465-95.

 

James, Estelle and Rose-Ackerman, Susan. (1986). The Nonprofit Enterprise in Market
Economies, Chur: Harwood Academic Press.

Jeavons, T. H. (1998). “Identifying Characteristics of ‘Religious’ Organizations: An
Exploratory Proposal”. In Sacred Companies: Organizational Aspects of Religion and
Religious Aspects of Organizations, Edited by NJ. Demerath III et. al. New York:
Oxford U. Press.

Johnson, Benton. (1962). “Ascetic Protestantism and Political Preference.” Public
Opinion Quarterly, vol. 26, pp. 35-46.

Kagan, Sharon L. (1991). “Examining Profit and Nonprofit Child Care: An Odyssey of
Quality and Auspices”, Journal of Social Issues, vol. 47, pp. 87-104.

Kee, James Edwin and John Shannon. (1992). “Crisis and Anti-Crisis: Why recessions
weaken Washington And strengthen state and local governments.” p. 3-10 in
Proceedings of the 84“ Annual Conference on Taxation of the National Tax
Association, Columbus, Ohio: National Tax Association.

Kellstedt, Lyman A., Corwin E. Smith, and Paul M. Kellstedt. (1991). “Religious
Tradition, Denomination, And Commitment: White Protestants and the 1988 Election.”

145

In The Bible and the Ballot Box: Religion and Politics in the 1988 Election, ed. James
L. Guth and John C. Green. Boulder, CO: Westview.

Kisker, Ellen Eliason et a1. (1991). A Profile of Child Care Settings: Early Education
and Care in 1990. Princeton, NJ: Mathematica Policy Research, Inc.

Kula, Maria Comachione. (2004). “US. States, the Medicaid Program, and Tax
Smoothing.” Southern Economic Journal, vol. 70, pp. 490-511.

Layman, Geoffrey C., (1997). “Religion and Political Behavior in the United States: The
Impact of Beliefs, Affiliations, and Commitment From 1980 to 1994”, Public Opinion
Quarterly, vol.6l, pp.288-316.

Leege, David C. ( 1993). “The Decomposition of the Religious Vote: A Comparison of
White, Non-Hispanic Catholics with Other Ethnoreligious Groups, 1960-1992.” Paper
Presented at the Annual Meeting of the American Political Science Association,
Washington, D.C., September 2-5.

Lipford, Jody W. and Robert D. Tollison. (2003). “Religious Participation and Income.”
Journal of Economic Behavior and Organization, vol. 51, pp. 249-260.

Martin, D. (1978). A General Theory of Secularization, Oxford, Basil Blackwell.

McCutcheon, Allen L. (1984). “Denominations and religious intermarriage: trends among
white Americans in the twentieth century.” Paper presented to the Society for the
Scientific Study of Religion.

Merriman, David. (Working Paper). “When Do States Legislate Tax Changes?”
<//wc.wustl.edu/>

Miller, Arthur H., and Martin P. Wattenberg. (1984). “Politics from the Pulpit:
Religiosity and the 1980 Elections.” Public Opinion Quarterly, vol. 48, pp. 301-17.

Montalvo, Jose G. and Marta Reynal-Querol. (2003). “Religious polarization and
economic development.” Economics Letters, vol. 80, pp. 201-210.

Monsma, Stephen V. (1996). When Sacred and Secular Mix: Religious Nonprofit
Organizations and Public Money. Lanharn, MD: Rowman and Littlefield Publishers.

Montgomery, J. D. (1996). “The Dynamics of the Religious Economy: Exit, Voice and
Denominational Secularization”, Rationality and Society, vol. 8, pp.83-l 10.

Paldam, Martin. (2001). “Corruption and religion adding to the economic model.”
Kyklos, vol. 54, pp. 383-413.

146

Putnam, Robert D. (2000). Bowling Alone: The Collapse and Revival of American
Community. New York: Simon & Schuster.

Reinikka, Ritva and Svensson, Jakob. (2003). “Working for God?: Evaluating Service
Delivery of Religious Not-for-Profit Health Care Providers in Uganda”, World Bank
Working Paper.

Roberts, Russell D. (1984). “A Positive Model of Private Charity and Public Transfers”,
Journal of Political Economy, vol. 92, pp. 136-48.

Roemer, John E. ( 1998). “Why the Poor Do Not Expropriate the Rich in Democracies,”
Journal of Public Economics, vol. 70, pp.399-442.

. (2001). Political Competition: Theory and Applications, Cambridge,
MA: Harvard University Press.

 

Rose-Ackerman, Susan. (1996). “Altruism, Nonprofits, and Economic Theory”, Journal
of Economic Literature, vol. 34, pp.701-28.

Rosen, Harvey S. (1988). Public Finance 2“d ed. Homewood, Illinois: Irwin.

Rosen, S. (1986). “The Theory of Equalizing Differences,” in Ashenfelter, CC. and R.
Layard, editors, Handbook of Labor Economics, North-Holland, Vol. I, chap. 12.

Rozell, Mark J ., and Clyde Wilcox, eds. (1995). God at the Grass Roots: The Christian
Right in the 1994 Elections. Lanham, MD: Rowman & Littlefield.

Sen, Amartya K. (1977) “Rational Fools: A Critique of the Behavioral Foundations of
Economic Theory”. Philosophy and Public Ajfairs, vol. 6, pp.317-44.

Smith, Adam. (2000). The Wealth of Nations, New York, NY: The Modern Library.

Smith, 1., Sawkins, J. W., and Seaman, P. T. (1998). “The Economics of Religious
Participation: A Cross-Country Study,” Kyklos, vol. 51, pp.25-43.

Smith, Ian. and Sawkins, John W. (2003). “The Economics of Regional Variation in
Religious Attendance” Applied Economics, vol. 35, pp. 1577-88.

Smith, Steven R. and Sosin, Michael R. (2001). “The Varieties of Faith-Based Agencies”,
Public Administration Review, vol. 61, pp. 651-670.

Smith, Tom W. (1990). “Classifying Protestant Denominations.” Review of Religious
Research, vol. 31, pp. 225-45.

Stark, R. and W.S. Brainbridge (1987). A Theory of Religion, New York: Lang.

 

147

Steensland, Brian, Jerry Z. Park, Mark D. Regnerus, Lynn D. Robinson, W. Bradford
Wilcox and Robert D. Woodberry. (2000). “The Measure of American Religion:
Toward Improving the State of the Art.” Social Forces, vol. 79, pp. 291-318.

Steinberg, Richard S. (1987). Voluntary Donations and Public Expenditures in a
Federalist System”, American Economic Review, vol. 77, pp. 24-36.

Steinberg, Richard and Gray, Bradford H. (1993). “‘The Role of Nonprofit Enterprise’ in
1993: Hansmann Revisited”, Nonprofit and Voluntary Sector Quarterly, vol. 22,
pp.297-316.

Stigler, George J. and Becker, Gary S. (1977). “De Gustibus Non Est Disputandum”,
The American Economic Review, vol. 67, pp. 76-90.

Sugden, Robert. (1982). “On the Economics of Philanthropy”, Economic Journal, vol.
92, pp. 341-50.

. (1984). “Reciprocity: The Supply of Public Goods Through Voluntary
Contributions”, Economic Journal, vol. 94, pp. 772-87.

 

Tawney, R. H. (1926). Religion and the Rise of Capitalism, New York: Harcourt, Brace
& World, Inc.

Unruh, H. R. (2001). Religious Elements of Faith-Based Social Service Programs: Types
and Integrative Strategies. Paper presented at the Society for the Scientific Study of
Religion Meetings. Columbus, OH.

. (2004). Religious Elements of Church-Based Social Service Programs:
Types, Variables, and Integrative Structures. Review of Religious Research, vol. 45, pp.
317-35.

 

Vogel, Robert C. and Robert P. Trost. (1979). “The response of state government receipts
to economic fluctuations and the allocation of counter-cyclical revenue sharing grants.”
The Review of Economics and Statistics, vol. 61, pp. 389-400.

Warr, Peter G. (1982). “Pareto-Optimal Redistribution and Private Charity”, Journal of
Public Economics, vol. 19, pp. 131-38.

Weisbrod, Burton. (1988). The Nonprofit Economy. Cambridge, MA: Harvard U. Press.

. (1989). “Rewarding Performance That is Hard to Measure: The
Private Nonprofit Sector”, Science, vol. 244. pp.541-46.

 

. (1994). “Does Institutional Form Matter: Comparing the Behavior of
Private Firms, Church-Related Nonprofits, and Other Nonprofits” Draft. Northwestern
U.

 

148

 

 

Wilcox, Clyde. (1992). God ’s Warriors: The Christian Right in Twentieth Century
America. Baltimore: Johns Hopkins University Press.

Wilson, B. (1966). Religion in Secular Society: A Sociological Comment, London, Watts.

Wilson, J. D. (1986). “A Theory of Interregional Tax Competition.” Journal of Public
Economics, vol. 19 pp. 296-315.

. (I999). “Theories of Tax Competition.” National Tax Journal, vol. 52, pp.
269-304. pp. 296-315.

 

Wood, James R. (1970). “Authority and controversial policy: The churches and civil
rights.” American Sociological Review, vol. 35, pp. 1057-1069.

Wood, Richard L. and Mark R. Warren. (2002). “A Different Face of Faith-Based
Politics: Social Capital and Community Organizing in the Public Arena.” International
Journal ofSociology and Social Policy, vol. 22, pp. 6-54.

Wuthnow, Robert. (1988). The Restructuring of American Religion: Society and Faith
since World War 11. Princeton, NJ: Princeton University Press.

Wuthnow, R., Hackett, C., and Yang Hsu, B. (2004). “The Effectiveness and
Trustwortl‘iness of Faith-Based and Other Service Organizations: A Study of
Recipients’ Perceptions”. Journal for the Scientific Study of Religion, vol. 43, pp. 1-17.

Zaleski, R. P. and Zech, Ch. E. (1992). “Determinants of Contributions to Religious
Organizations: Free-riding and Other Factors”, American Journal of Economics and
Sociology, vol. 51, pp. 459-72.

. (1995). “The Effect of Religious Market Competition on Church
Giving”, Review of Social Economy, vol. 53, pp. 350.

 

149

 

 

ullll'nliiililljljluwith

12