Академический Документы
Профессиональный Документы
Культура Документы
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH
ACROSS COUNTRIES
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 761
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
762 AMERICAN SOCIOLOGICAL REVIEW
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 763
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
764 AMERICAN SOCIOLOGICAL REVIEW
indexes
Some of the survey questions in 1980to
relate for at-
the countries used in the
tendance at religious services. We use
statistical these
analysis (as dictated by data avail-
responses to generate the fractions of with
ability). Countries thelow levels of plural-
ism (values
population that attended church of the index close to zero) in-
or analogous
houses of worship at least weekly
clude some thatand at
are predominantly Catholic
least monthly. Subsequently, we
(Spain, call
Italy, thisBelgium, Ireland, and
Portugal,
variable church attendance. Some of the sur- much of Latin America), Protestant Scan-
veys (the two ISSP waves) include questions dinavia, Orthodox Greece, and Muslim Pa-
about time spent at prayer. Other queries kistan and Turkey. Places that exhibit high
concern religious beliefs and attitudes; for levels of pluralism (values of the index of .5
example, do you believe in heaven, hell, life or higher) include the United States, Ger-
after death, and God (in various senses)?9 many, the Netherlands, Switzerland, Austra-
Other questions, which might be more ro- lia, Malaysia, Singapore, and South Africa.
bust across religions, are whether the re- We used tabulations by Barrett et al.
spondent considers himself or herself to be a (2001:834-35) to measure the presence or
religious person and whether religion plays absence of a state religion. These classifica-
an important role in one's life. tions are clearer for some countries than oth-
We used tables from the first edition of the
ers. In some of the straightforward situa-
World Christian Encyclopedia (Barrett tions, the constitution designates an official
1982) to assemble information on religious state church and restricts or prohibits other
adherence, as professed in surveys or cen- forms of religion. However, even without
suses in which people are asked to state the these designations or prohibitions, the gov-
religion, if any, to which they adhere. These ernment may systematically favor a selected
data, which apply to 1970 and 1980,10 allow religion through subsidies and tax collec-
us to construct measures of religious plural- tions or through the teaching of religion in
ism, based on Herfindahl indexes for the re-public schools. These considerations cause
ligion shares." Table 1 shows the pluralism Barrett et al. to classify some countries as
having a "state religion," despite the absence
of an official state church in the constitution.
adherence over the last 30 years. The 2001 wave
Controversial cases in this category include
of the WVS will provide more observations on
East Asian countries. Italy, Portugal, and Spain, which are consid-
9 The meaning of some of these questions isered by Barrett et al. to have a Catholic state
unclear for some religions. For example, Hindu-
ism and Buddhism view heaven and hell as inter- a country (among those adhering to some reli-
mediate stages between reincarnations, whereas gion) belong to the same religion. Hence, our
Christianity regards heaven and hell as ends, not pluralism index-defined to equal one minus the
means, to salvation (McCleary 2003). Herfindahl index-is the probability that they
10 For some of the eastern European countries, belong to different religions. This index can,
which did not exist or were not covered in therefore, be viewed as an indicator of religious
Barrett's (1982) first edition, the data come pluralism
from or diversity. (Implicitly, the differences
Barrett et al. (2001) and refer to 1970 and between
1990. the religious groupings are assumed to
An alternative procedure would be to use bethe
the same for all pairs. Otherwise, one could
countrywide averages of religion data in the vari-weight religions in accordance with the extent of
ous international surveys. We did not proceed differences between them.) If everyone belongs
this way because the categories of religion differto the same religion, the Herfindahl index equals
across the surveys and because of some coding
one, and hence, the pluralism indicator equals
problems for religions in the World Values zero.
Sur- If there are two religions of equal size, the
vey. Herfindahl and pluralism variables each equal
11 We grouped data on religious adherence
one-half. If there are a large number of religions,
from Barrett (1982) into nine major categories:
each of which covers a negligible fraction of the
Catholic, Muslim, Protestant, Hindu, Buddhist,
population, the Herfindahl index equals (almost)
other Eastern religions, Jewish, Orthodox, and and hence, the pluralism indicator equals
zero,
other religions. The Herfindahl index-the(almost)
sum one. With nine groupings, the lowest
possible value of the Herfindahl index is .11, so
of the squares of the population fractions belong-
ing to each religion-can be interpreted as thethe highest possible value of the pluralism
that
measure
probability that two randomly selected persons in is .89.
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 765
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
766 AMERICAN SOCIOLOGICAL REVIEW
Table 2. Meansfive
Each system in Table 3 comprises and Standard
or Deviations for
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 767
Per capita GDP (logged) .04 (.16) -.28 (.21) -.28 (.18)
Years of education .192*** (.036) .253*** (.040) .234*** (.038)
Urbanization rate -1.49*** (.34) -1.65*** (.43) -2.11*** (.41)
1/(life expectancy at age 1) 2.21 (1.20) -3.52 (2.00) -2.49 (1.72)
Population share over age 65 .32 (3.19) -9.84* (4.74) -12.51** (4.18)
Population share under age 15 5.51** (2.03) 4.88 (2.92) 1.49 (2.59)
Religious pluralism 1.35*** (.32) 1.24** (.38) 1.13** (.37)
Number of observations 22, 36, 22, 21, 33, 16, 21, 33, 16,
for each equation 32, 28, 41 30, 30 30, 30
R-squared for each equation .87, .77, .80, .69, .79, .90, .77, .75, .71,
.89, .84, .74 .85, .84 86, .71
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
768 AMERICAN SOCIOLOGICAL REVIEW
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 769
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
770 AMERICAN SOCIOLOGICAL REVIEW
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 771
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
772 AMERICAN SOCIOLOGICAL REVIEW
average
formance mostly indirectly, that ratios
is, of investment and government
through
influences on the religious consumption
beliefs. Hence,to GDP; an international open-
we envision a chain whereby church
ness measure,atten-
based on the ratio of exports
plus
dance affects religious beliefs, imports
which to GDP; changes in the terms
affect
individual traits, which affectof individual and
trade; subjective measures of maintenance
aggregate economic outcomes. of the rule of law and democracy; and the
To put it another way, we inflation
view religious
rate. In the present analysis, we add
measuressector,27
beliefs as an output of the religion of religiosity to this list of ex-
and we view church attendance as an
planatory input
variables.
to this sector. Thus, if we holdThe
fixed the be-
regressions include all countries that
liefs, as we do in the regressions
have data that
on all ofwe the variables. We began
highlight here, an increase inwith
church
a sampleatten-
of 87 countries, but the inclu-
sion sector
dance signifies that the religion of data on is
church
lessattendance and reli-
productive. That is, more resources in lowered
gious beliefs termsthe sample to 41 coun-
of time and goods are beingtries.consumed
We end up withfor 2 countries in sub-Sa-
given outputs (beliefs). Hence,haran
ourAfrica, 8 in Latin America, 9 in Asia,
anticipa-
2 in
tion is that for given religious Eastern higher
beliefs, Europe, and 20 that have been
church attendance would show members
up asof a the
nega-
Organization for Economic
Cooperation and Development (OECD)
tive influence on economic performance.
since the 1960s (not
A different perspective, represented byincluding Japan and
Sacerdote and Glaeser (2001)Turkey,
and which
Putnamare included with Asia).
(2000), argues that houses ofOur worship are appear in Table 4.
empirical results
Each system
important as civic organizations, per se.includes
That as explanatory vari-
ables afostered
is, the networks and interactions measure of monthly
by church atten-
churches are important elements dance and of
a measure
social of religious belief-in
the present
capital. Thus, if this social capital case, belief in heaven or hell.
is produc-
tive, we might find that, for(These givenvariables enter, as before, in the form
religious
beliefs, church attendance would have a log[x/(1-x)], where x is the fraction of per-
separate, positive effect on economic sons attending or believing.) To minimize
growth. the loss of observations, we entered into the
Church attendance might also proxy for regressions a single measure of monthly
the influence of organized religion on laws church attendance or religious belief for
and regulations that affect economic incen- each period for countries that had data on
tives. Adverse examples would be restric- these variables at any point in time. Specifi-
tions on credit and insurance markets and cally, we defined a measure of attendance or
belief for a country to be the value from
general discouragement of the profit motive.
WVS 1990 if available, then we filled in
Positive forces might include legislation and
regulation related to fraud and corruption.with the value from WVS 1981 if the 1990
Our empirical framework for studyingvalue was unavailable. (We adjusted for the
economic growth is the one described average
in discrepancy between the two values
Barro and Sala-i-Martin (2004, chap. 12).
among countries that had information for
The dependent variable is the growth rateboth
of years.) If neither of these values were
available, we used in an analogous way the
per capita GDP over three 10-year periods:
1965-1975, 1975-1985, and 1985-1995. value for ISSP 1991, then WVS 1995, and
The explanatory variables, suggested by pre-then ISSP 1998. This procedure means that
vious studies, are the values at the start of
the systems for economic growth incorpo-
each period of the log of per capita GDP, rate
life on the right-hand side measures of reli-
expectancy (measured inversely), yearsgious of attendance and beliefs that sometimes
postdate the growth rates.
school attainment, and the total fertility rate;
One reason that the use of later observa-
27 The religion sector has outputs other thantions on the right-hand side of the growth
equations might be satisfactory is that
religious beliefs. For example, organized religion
can be important in social welfare activities andchurch attendance and religious beliefs ex-
education. hibit a lot of persistence over time. Hence,
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 773
Table 4. Regressions for Economic Growth, 1965 -1975, 1975 -1985, 198
Religion Shares
Eastern religion share -.010 - .003 - -.013
(.011) (.009) (.012)
Hindu share -.034* - -.030 - -.040*
(.016) (.017) (.018)
Jewish share -.004 - .006 - -.006
(.014) (.013) (.015)
Muslim share -.032* - -.012 - -.034*
(.015) (.010) (.015)
Orthodox share -.050* - -.029 - -.051"
(.021) (.017) (.021)
Protestant share -.015* - -.018* - -.012
(.007) (.008) (.008)
Other religion share -.001 - -.001 - -.001
(.015) (.016) (.015)
p-value for religion .001 - .001 - .001
shares jointly
Number of countries and 41, 118 41, 118 41, 118 41, 118 41, 118 41, 118
total observations
Number of observations 38, 41, 39 38, 41, 39 38, 41, 39 38, 41, 39 38, 41, 39 38, 41, 39
for each time period
R-squared values for .39, .77, .67, .63, .36, .68, .66, .60, .39, .78, .65, .63,
each time period .36 .25 .32 .36 .35 .19
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
774 AMERICAN SOCIOLOGICAL REVIEW
growth for
later values may proxy satisfactorily equations.
the The main result, which
missing earlier ones, which can be viewed
typifies our findings, is that the estimated
as the true causes of differences in growth
coefficient for monthly church attendance is
rates. However, the results innegative,
Table whereas
3 sug-that for religious belief, in
gest substantial responses of religiosity
this case belief to
in hell, is positive. These es-
economic variables. Thus, the timated
religion coefficients
mea- are individually and
jointly statistically
sures might pick up reverse causation from significant (p-value for
economic development to religiosity,
the two together
rather is .000). The estimated co-
than the reverse. efficient for monthly attendance, -.0095
To deal with the reverse causation issue, (S.E. = .0018), means that an increase in this
we use the instrumental variables suggested variable by one standard deviation (1.20, as
by the regressions in Table 3, in which the shown in Table 2) would reduce the eco-
religion measures were the dependent vari- nomic growth rate on impact by 1.1 percent
ables. Specifically, we use as instruments the per year. The estimated coefficient for belief
dummy variable for the presence of a state in hell, .0094 (S.E. = .0025), means that an
religion (in 1970), the dummy variable for increase in this variable by one standard de-
the existence of state regulation of religion viation (1.06) would increase the economic
(in the 1970s), the shares in 1970 or 1980 of growth rate on impact by 1.0 percent per
religious adherence among the major reli- year.29 Similar numerical calculations can be
gions, and the religious pluralism index for applied to the other results in Table 4, using
1970 or 1980. These variables are highly sta- the standard deviations of variables shown
tistically significant for explaining monthly in Table 2.
church attendance, belief in heaven, and be- One test of over-identification is whether
lief in hell-in each case, the p-value for a group of three instruments-the dummy
joint significance is .000. Thus, we are not variables for the presence of a state religion
concerned that the instruments are weak. and for state regulation of the religion mar-
It is possible that the instruments-al- ket and the religious pluralism index-mat-
though prior in time to the growth observa- ter directly for growth. To carry out this
tions-are themselves endogenous with re- test, we maintained the hypothesis that the
spect to economic development. For ex- religious adherence shares do not enter di-
ample, over long periods, development rectly into the economic growth equations.
could induce countries to drop their official (This assumption allows us to identify the
state religions or to alter their regulatory sys- systems even when the group of three in-
tems for religion. For the composition of re- struments is added to the growth equations.)
ligions, the major concern is that the religion The p-value for joint significance of the
shares would matter directly for economic three variables in the growth equations is
growth, not just through their predictive con- .25. Similar results apply to the other cases
tent for the religiosity measures.28 To check shown in Table 4-we always accept with a
out these possibilities, the analysis includes high p-value the exclusion of the three vari-
tests of over-identifying restrictions of the ables from the growth equations. Therefore,
model. for the rest of the analysis, we include these
Columns 1 and 2 of Table 4 use the "be- three variables-the dummy variables for
lief in hell" variable. The first column ex-
state religion and state regulation of reli-
gion and the pluralism index-as instru-
cludes the religious adherence shares in the
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 775
ments but exclude them from the economic implied by the point estimates are larger than
growth equations.30 in column 1.
If we include the religious adherence Columns 3 and 4 of Table 4 replace belief
shares directly in the growth equations, we in hell with belief in heaven. The results are
get the estimated coefficients shown in col- similar to those found in columns 1 and 2,
umn 2 of Table 4. The coefficients for each except that the quantitative effect and statis-
religion share should be interpreted as thetical significance for belief in heaven are
effect relative to that for Catholic. The esti- somewhat less than those for belief in hell.
mated coefficients that are individually sta- For example, when the religion shares are
tistically significant are the negative values excluded from the growth system in column
shown for Hindu, Muslim, Orthodox, and 3, the estimated coefficient for monthly
Protestant. Some of these signs are surpris- church attendance is -.0104 (S.E. = .0023),
ing, but a full interpretation has to consider whereas that for belief in heaven is .0069
that the systems include monthly church at- (S.E. = .0029). The p-value for joint signifi-
tendance and belief in hell, and these two cance of these variables is .000. When the
variables are strongly related to religious religion shares are included in column 4, the
composition (as shown in Table 3).31 The re- estimated coefficients are -.0123 (S.E. =
ligion shares are jointly statistically signifi- .0043) for monthly attendance and .0076
cant with a p-value of .001. (S.E. = .0048) for belief in heaven. In this
Since some of the religion shares matter case, the p-value for joint significance is
directly for economic growth, we also exam- .009. Thus, the results are again robust to the
ine the estimated growth effects of church inclusion of religion shares in the growth
attendance and belief in hell when the reli-
system.
gion shares are included in the growth equa- We also looked at other measures of reli-
tions. The resulting coefficients, shown in gious belief. A pattern of results similar to
column 2 of Table 4, are -.0156 (S.E. = those shown in Table 4 emerges if we use
.0044) for monthly church attendance and belief in an afterlife, rather than the more
.0140 (S.E. = .0058) for belief in hell. The specific beliefs in hell or heaven.32 For ex-
p-value for joint significance of these vari- ample, in a specification comparable to
ables is .001. Therefore, the results for those in columns 1 and 3 of the table, the
church attendance and belief in hell remain estimated coefficients are -.0094 (S.E. =
statistically significant when the religious.0021) for monthly church attendance and
adherence shares are included in the growth .0079 (S.E. = .0035) for belief in an after-
equations. Moreover, the quantitative effectslife.33
The results are much weaker if we use be-
30 The religious pluralism index plays a minor lief in God rather than the more concrete be-
role as an instrument. If we exclude this variable liefs in hell, heaven, or an afterlife. In a
from the instrument list, the estimated coeffi-
cients in the system for economic growth become
-.0104 (S.E. = .0020) for monthly church atten- 32 We again used the functional form log[x/(1-
dance and .0085 (S.E. = .0028) for belief in hell. x)], where x is now the fraction of persons who
These results are close to those shown in column expressed a belief in an afterlife. For the WVS
1 of Table 4. data, we coded an individual's dummy variable
31 As an example, if we do not include religi- for belief in an afterlife as 1 if the respondent an-
osity variables in the growth regressions, the es- swered yes to the question "Do you believe in life
timated coefficient for the Muslim adherence after death?" For the ISSP data, we coded an
fraction is close to zero. Muslim adherence is as- individual's dummy variable as 1 if the respon-
sociated with high belief in hell relative to church dent answered yes, definitely, or yes, probably,
attendance (Table 3, columns 1 and 3), and thisto the question "Do you believe in life after
mechanism tends to encourage growth, accord-death?" The afterlife question was not included
ing to the estimates in columns 1 and 2 of Table in the Gallup Millennium Survey. As before, in
4. The results in column 2 imply that, once thethe economic growth regressions, we used only
connections with belief in hell and church atten- one observation per country for belief in an af-
dance are held constant, the remaining effect of terlife.
Muslim adherence on economic growth is nega- 33 The sample of observations was selected to
tive. be the same as that used in Table 4.
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
776 AMERICAN SOCIOLOGICAL REVIEW
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 777
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
778 AMERICAN SOCIOLOGICAL REVIEW
0 1 0 .042 - ? 0
C 0
o o
S04.- 0 o0 0
x C
.02 0 o
-coa 80o V 02 ',-, c, . 0
Drr
O ........ ..... 0 i0 _ 0 ..
- o ... - 0
0- 00O0 O.
-3 2 -1 0 1 2
I
3 4
E -.04 - o
W 0
-.06 -.06- I
-3 -2 -1 0 1 2 3 4 -3 -2 -1 0 1 2 3 4
a. a. 0 o
.02
o --
8. .
o o o
x 0
So 00 o
X 02- O 00
0 0
S-.02 - o
0 00
0..02 0 .o
O o8
0
- 60
-0 0 0 0
0 0 o
E -.04
C
- o o0
-04 0
0
LU06 , L .06 , I.
-3 -2 -1 0 1 2 3 4 -3 -2 -1 0 1 2 3 4
Note: Figure 3a holds constant belief in hell; Figure 3c holds constant belief in heaven.
vian countries and Japan-values from the Figures Ic and id come from the estimates
1990 WVS of monthly church attendance, shown in column 3. The graphs are con-
belief in heaven, and belief in hell are: .11, structed to hold constant the influence on
.55, and .27 in Finland; .13, .44, and .19 in economic growth of all the explanatory vari-
Norway; .09, .57, and .12 in Iceland; and ables other than the one shown on the x-
.14, .43, and .32 in Japan. axis.39 These kinds of diagrams are useful
Figure 1 depicts partial relations between
the rate of economic growth and the vari-
39 To calculate the value plotted on the y-axis,
ables that measure monthly church atten- we took the growth rate of per capita GDP (the
dance, belief in hell, and belief in heaven. dependent variable) over each of the three time
Figures la and lb are computed from the es- periods used in Table 4 and subtracted the esti-
timates shown in column 1 of Table 4, and mated effect of all the explanatory variables aside
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 779
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
780 AMERICAN SOCIOLOGICAL REVIEW
Other useful research would assess the ef- "Religion and Political Economy in an Inter-
fects of religiosity on political and social national Panel" (Working Paper No. 8931,
variables, including democracy, the rule of May). Cambridge, MA: National Bureau of
Economic Research.
law, fertility, and health. Finally, it would be
Barro, Robert J. and Xavier Sala-i-Martin. 2004.
valuable to extend our analysis of religiosity
Economic Growth. 2d ed. Cambridge, MA:
at a countrywide level to the behavior of in- MIT Press.
dividuals.
Berger, Peter L. 1967. The Sacred Canopy: Ele-
ments of a Sociological Theory of Religion.
Robert J. Barro is Robert C. Waggoner Profes- Garden City, NJ: Doubleday.
sor of Economics at Harvard University, a senior Chaves, Mark. 1994. "Secularization as Declin-
fellow of Stanford University's Hoover Institu- ing Religious Authority." Social Forces 72:
tion, a research associate of the National Bureau 749-74.
of Economic Research, and a viewpoint colum- Chaves, Mark and David E. Cann. 1992. "Regu-
nist for Business Week. He is President-Elect of lation, Pluralism, and Religious Market Struc-
the Western Economic Association and was re- ture." Rationality and Society 4:272-90.
cently Vice President of the American Economic Chaves, Mark and Philip S. Gorski. 2001. "Reli-
Association. His recent research involves the gious Pluralism and Religious Participation."
determinants of economic growth across coun- Annual Review of Sociology 27:261-8 1.
tries, and he is currently writing a new interme- Davie, Grace. 1994. Religion in Britain since
diate textbook on macroeconomics. He is pleased 1945: Believing without Belonging. Oxford,
to be publishing an article in a sociology jour- England: Blackwell.
nal. Finke, Roger and Laurence R. Iannaccone. 1993.
"Supply-Side Explanations for Religious
Rachel M. McCleary is Director of the Religion,
Change." Annals of the American Academy of
Political Economy, and Society Project atPolitical and Social Sciences 527:27-39.
Harvard University's Weatherhead Center Finke, for Roger and Rodney Stark. 1992. The
Churching of America, 1776-1990. New
International Affairs and is also a lecturer in the
Department of Government. She holds a Ph.D. Brunswick,
in NJ: Rutgers University Press.
moral philosophy from the University of ChicagoFox, Jonathan and Shmuel Sandler. 2003. "Sepa-
and a Master of Theological Studies from Emory ration of Religion and State in the 21st Cen-
tury: Comparing the Middle East and Western
University. Her current research deals with reli-
giously based private voluntary organizations Democracies." Paper presented at the Interna-
and their interactions with the U.S. government. tional Studies Association conference, Febru-
She is also studying the economic incentives con-ary 25, Portland, OR.
tained in the doctrines of the world's major Heston,
reli- Alan, Robert Summers, and Bettina
gions. Aten. 2002. Penn World Tables Version 6.1
(October). Philadelphia, PA: University of
Pennsylvania, Center for International Com-
REFERENCES parisons (CICUP).
Huntington, Samuel P. 1996. The Clash of Civi-
Anderson, John. 2002. "Social, Political, and and
lizations In- the Remaking of World Order.
stitutional Constraints on ReligiousNew York: Simon and Schuster.
Pluralism
in Central Asia." Journal of Contemporary Iannaccone, Re-
Laurence R. 1991. "The Conse-
ligion 17:181-96. quences of Religious Market Structures: Adam
Barrett, David B. 1982. World Christian Smith andEncy-
the Economics of Religion." Ration-
clopedia. 1st ed. Oxford, England:ality Oxford
and Society 3:156-77.
University Press. Inglehart, Ronald and Wayne E. Baker. 2000.
Barrett, David B., George T. Kurian, and Todd Cultural Change, and the Per-
"Modernization,
M. Johnson. 2001. World Christian sistence of Traditional Values." American So-
Encyclo-
pedia. 2d ed. Oxford, England: Oxford ciological
Univer- Review 65:19-51.
sity Press. Koran, The. 1994. Translated by John M.
Barro, Robert J. 1991. "Economic Growth in a Rodwell. London, England: Everyman.
Cross Section of Countries." Quarterly Jour- Landes, David S. 1999. The Wealth and Poverty
nal of Economics 106:407-43. of Nations: Why Some Are So Rich and Some
Barro, Robert J. and Jong-Wha Lee. 2001. "In- So Poor. New York: Norton.
ternational Data on Educational Attainment: Martin, David. 1978. A General Theory of Secu-
Updates and Implications." Oxford Economiclarization. Oxford, England: Basil Blackwell.
Papers 53:541-63. McCleary, Rachel M. 2003. "Salvation, Damna-
Barro, Robert J. and Rachel M. McCleary. 2002.tion, and Economic Incentives" (Working Pa-
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms
RELIGION AND ECONOMIC GROWTH 78 1
This content downloaded from 106.201.13.205 on Thu, 11 Jun 2020 16:20:19 UTC
All use subject to https://about.jstor.org/terms