fiscal decentralization and governance
TRANSCRIPT
http://pfr.sagepub.com/Public Finance Review
http://pfr.sagepub.com/content/40/1/66The online version of this article can be found at:
DOI: 10.1177/1091142111424276
2011 2012 40: 66 originally published online 13 NovemberPublic Finance Review
Yener Altunbas and John ThorntonFiscal Decentralization and Governance
Published by:
http://www.sagepublications.com
can be found at:Public Finance ReviewAdditional services and information for
http://pfr.sagepub.com/cgi/alertsEmail Alerts:
http://pfr.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://pfr.sagepub.com/content/40/1/66.refs.htmlCitations:
What is This?
- Nov 13, 2011OnlineFirst Version of Record
- Jan 29, 2012Version of Record >>
by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from by guest on October 11, 2013pfr.sagepub.comDownloaded from
FiscalDecentralizationand Governance
Yener Altunbas1 and John Thornton1
AbstractThe literature on the economics of fiscal decentralization stresses thepotential for both positive and negative effects on governance in a country.Using a data set comprising sixty-four developed and developing economiesand several different measures of fiscal decentralization, the authors find thatcountries in which a larger share of fiscal revenues and expenditures arelocated at the level of subnational governments appear to be less corrupt.The authors also find that the beneficial impact of fiscal decentralization oncorruption is mitigated in the presence of mechanisms enforcing verticaladministrative decentralization. The results indicate that fiscal decentraliza-tion appears to reduce corruption even in countries in which there is a highdegree of political representation. The results are robust to alternative esti-mation methodologies and to specifications that control for the influence ofvariables that have been identified as affecting governance.
Keywordsfiscal decentralization, governance, corruption, cross section, panelestimation
1The Business School, Bangor University, Hen Goleg, Gwynedd, United Kingdom
Corresponding Author:
John Thornton, The Business School, Bangor University, Hen Goleg, College Road, Bangor,
Gwynedd LL57 2DG, United Kingdom
Email: [email protected]
Public Finance Review40(1) 66-85
ª The Author(s) 2012Reprints and permission:
sagepub.com/journalsPermissions.navDOI: 10.1177/1091142111424276
http://pfr.sagepub.com
The literature on the economics of fiscal decentralization stresses the
potential for both positive and negative effects on the quality of govern-
ance in a country. Several economists have made the case for fiscal
decentralization as a means to promote better governance. For example,
Oates (1972) argues that decentralized governments will be better
informed about local conditions and better able to satisfy citizen prefer-
ences. Seabright (1996) and Tabellini (2000) make the same case based
on citizens being better informed about the activities of local govern-
ments and therefore better placed to reward or punish local politicians
according to their performance. Weingast (1995) and Montinola, Yin-
gyi, and Weingest (1995) argue that fiscal decentralization means that
economic agents have the ability to leave more corrupt regions, which
would tend to improve governance. Inman and Rubinfeld (1997) and de
Mello (2000) argue that fiscal decentralization strengthens social capital
and encourages political participation. Also, the World Bank (2004)
suggests that the resulting competition between centers of authority
reduces the risk that governments will expropriate wealth.
Other economists have argued that fiscal decentralization can under-
mine governance. Bardhan and Mookherjee (2000), Tanzi (1995), and
Prud’homme (1995) argue that the local officials are more susceptible
to capture by local economic interests. Prud’homme (1995) and Tabel-
lini (2000) further suggest that the harmful effects may result because
local government activities are less intensely monitored than central
government activities. Hommes (1995), the World Bank (1999), and
Fukasaku and de Mello (1999) argue that fiscal decentralization may
lead to poor accountability and governance if expenditures and revenue
mobilization functions are not clearly assigned across different levels of
government.
The few relevant empirical studies have had rather mixed results.
Cross-country studies in this vein include De Mello and Barenstein
(2001), who report that a range of governance indicators improve as the
share of subnational government spending in total spending increases;
Fisman and Gatti (2002), who find that revenue and expenditure decen-
tralization reduce corruption; Enikolopov and Zhuravskaya (2007), who
report that governance indicators improve when fiscal decentralization
is combined with strong national parties; and Kyria and Roca-Sagales
(2011) who find that fiscal decentralization has a positive impact on the
quality of government in Organisation for Economic Co-operation and
Development (OECD) countries, though these effects are mitigated in
the presence of regional elections and multilevel government. In
Altunbas and Thornton 67
contrast, Treisman (2002) finds that any relationship between corruption
and fiscal decentralization is highly sensitive to the control variables
included, and Dreher (2006) finds that fiscal decentralization is consis-
tent with improvements in a number of key governance indicators
mainly in low-income countries.
In this article, we focus on the relation between fiscal decentralization
and the corruption aspect of governance. We add substantially to the
empirical literature by presenting results using a revamped data set of sev-
eral different measures of fiscal decentralization in developed and develop-
ing economies, by taking account of vertical administrative arrangements
governing decentralization, by employing different estimation methodolo-
gies, and by controlling for variables representing the degree of political
accountability that Lederman, Loayza, and Soares (2005) have shown to
dominate alternative explanations of corruption. We report three findings.
First, countries in which a larger share of fiscal revenues and expenditures
are located at the level of subnational government appear to be less corrupt.
Second, the beneficial impact of fiscal decentralization on corruption is
mitigated in the presence of mechanisms enforcing vertical administrative
decentralization, specifically, a federal constitution that provides for lim-
ited autonomy at the subnational level or an electoral system in which the
bottom tier of government is directly elected. This suggests that higher
shares of general revenue and expenditure at the level of the subnational
government are most effective in reducing corruption when those
resources are nonetheless controlled by the central government. Third, fis-
cal decentralization appears to reduce corruption even in countries in
which there is a high degree of political representation, a factor that has
been shown in other studies to dominate explanations of corruption. These
results are robust to alternative estimation methodologies, including using
panel instrumental variable (IV) procedures to control for endogeneity and
other potential biases.
Our article makes several contributions. First, it adds to the large empiri-
cal literature that has studied the general economic effects of fiscal decen-
tralization. While these studies have often recognized the shortcomings of
subnational government revenue and expenditure shares as reliable indi-
cators of fiscal decentralization, our results testify to the importance of
also incorporating administrative arrangements for decentralization into
empirical work. Second, it acts as a counterweight to the literature on the
determinants of corruption that emphasizes the dominance of political
accountability as an explanation since our results are robust to the inclu-
sion of such variables. Finally, our article adds to the policy-oriented
68 Public Finance Review 40(1)
fiscal decentralization literature in that the results lend support to
economists who take a more positive view of fiscal decentralization as
a means of promoting better governance, though with the added qualifica-
tion that the extent to which it does so depends in part on the administra-
tive arrangements governing vertical decentralization.
Data and Summary Statistics
To conduct our analysis, we use measures of corruption, fiscal decentraliza-
tion, political decentralization, and political representation. In addition, we
employ a set of baseline control variables that other researchers have shown
to impact on corruption.
Corruption
Following many other corruption-related studies (e.g., Fisman and Gatti
2002; Mauro 1995, 1998), our indicator of corruption is the index of corrup-
tion in government produced by the International Country Risk Guide
(2001). This variable is meant to capture the likelihood that government
officials will demand special payments. The extent to which illegal pay-
ments are expected throughout lower levels of government is subjectively
ranked by panels of international experts (Knack and Keefer 1995) on a
scale from 0 to 6, with higher values indicating less corruption; however,
to facilitate ease of interpretation of the coefficients, we have rescaled the
corruption index to take on values between 0 (least corrupt) and 1 (most
corrupt).
Fiscal Decentralization
To measure fiscal decentralization, we make use of recent revisions and
updates of the fiscal decentralization indicators in the International Mone-
tary Fund (IMF) Government Finance Statistics Yearbook described in
Dziobek, Mangas, and Kufa (2011). These authors present data on fiscal
decentralization for about eighty countries over periods of up to twenty
years (1990–2008). The data set allows four measures of fiscal decentrali-
zation: subnational (state and local) government revenues, tax effort
(defined as the sum of tax revenue and compulsory social security contribu-
tions), expenditure, and compensation of employees (the wage bill), each of
which can be expressed as a proportion of general government revenues and
expenditures. These data are available on an accruals and cash basis for a
Altunbas and Thornton 69
few countries, but countries typically report in either one or the other
format. As there can be substantial differences between the accrual and cash
series, we employ accrual accounting fiscal data because these match the
time of recording to actual resource flows and are consistent with other
macroeconomic data sets (e.g., national accounts data are prepared on an
accrual basis).1
While the IMF has long been the most popular data source for cross-
country studies of fiscal decentralization, it is well known that these indica-
tors are flawed in the sense that they do not fully distinguish between
‘‘administrative’’ and ‘‘substantive’’ decentralization because they do not
recognize that high subnational revenue and spending shares do not neces-
sarily indicate high local autonomy. For example, surveys of fiscal relations
by Joumard and Kongsrud (2003) and Darby, Muscatelli, and Roy (2003)
show that limits on the discretion of subnational governments to determine
tax rates and tax bases significantly reduce local fiscal autonomy, and
Thornton (2007a, 2007b) shows that empirical estimates that fail to take
account of the degree of local autonomy can give misleading results.
Administrative Decentralization
We try to control for the shortcomings of the IMF data as an indicator of
fiscal decentralization by interacting the fiscal decentralization indicators
with three indicators of administrative subnational autonomy. The first indi-
cator is a dummy variable to capture whether the country has a federal con-
stitution, on the assumption that such countries will have more devolved
fiscal responsibilities. Second, we include a dummy variable to indicate
whether the bottom tier of government is directly elected, on the assumption
that directly elected subnational governments are more likely to have been
devolved some fiscal responsibility. Finally, to account for the distribution
of political power among the central and subnational governments, we
include a dummy variable to signify whether the constitution allows for lim-
ited autonomy at the level of the subnational government. These variables
provide insight into whether institutional arrangements accentuate or mitigate
the impact of fiscal decentralization on corruption. For example, Dreher
(2006) and Kyria and Roca-Sagales (2011) have shown that vertical decen-
tralization, measured by the number of subnational government tiers, tends
to deteriorate governance. The dummy representing whether a country has
a federal constitution is constructed from the Central Intelligence Agency
(CIA) World Factbook 2010, and the dummies for an elected bottom tier
of government and limited local autonomy are from Treisman (2000).2
70 Public Finance Review 40(1)
Political Representation
Lederman, Loayz, and Soares (2005) show that political accountability
variables are not only robust determinants of corruption but they also dom-
inate alternative explanations of corruption. Accordingly, we include three
proxies for political accountability in our estimates: an index of press free-
dom, a dummy variable to indicate whether a country has maintained
democratic institutions for a continuous period since 1950, and a dummy
variable to indicate that a country is a presidential democracy. The first
two political representation variables are included because governance
is likely to be better in democratic countries and in countries with a free
press and more vigorous civic associations—that is, in countries where
politicians are more accountable (Adsera, Boix, and Payne 2003).
Lederman, Loayza, and Soares (2005) argue that a presidential democracy
would have a positive impact on corruption since the leaders of the exec-
utive branch are more difficult to remove than in parliamentary systems.
The index of press freedom is from Freedom House, the continuous
democracy dummy is from Treisman (2000), and the presidential democ-
racy series is from Beck et al. (2001).
Baseline Controls
Our baseline control variables are a country’s real gross domestic product
(GDP) per capita, its population, a dummy variable to indicate whether the
country is a major fuel exporter, and a country’s degree of ethnic fractiona-
lization. Real GDP per capita is included because richer countries tend to be
more decentralized (Panizza 1999) and have better government quality
(Islam and Montenegro 2002), and because these conditions tend to create
a demand for better government (La Porta et al. 1999). In addition, real
GDP per capita should capture increases in the volume and size of eco-
nomic transactions, which would increase the benefits of developing insti-
tutions such as commercial codes and their associated adjudication and
enforcement mechanisms (Knack 2001). Population is included as an indi-
cator of country size to capture economies of scale in establishing effective
institutions (Srinivasan 1986; Knack 2001) and also because there is some
evidence that more populous countries tend to be more decentralized and to
have lower government quality (Treisman 2002). Ethnic fractionalization is
included because political theories predict that, as ethnic heterogeneity
increases, governments become more interventionist and less efficient, the
quality of public goods falls, and political freedoms are restricted (La Porta
Altunbas and Thornton 71
et al. 1999). Finally, many economists have argued that governance could
be undermined if countries receive substantial nontax revenues (e.g., from
natural resource rents) because citizens are likely to be less motivated to
scrutinize how government revenues are collected and spent (Collier and
Hoeffler 2005; Leite and Weidmann 2002). Presumably, this argument
could apply even if these revenues were decentralized. We try to control for
this possibility by including a dummy to capture whether a country is a
major energy exporter. The data for real GDP per capita and population are
from Heston, Summers, and Aten (2009), and the ethnic fractionalization
data are from Alesina et al. (2003).
Summary Statistics and Correlations
The data set comprises annual observations for up to sixty-four developed
and developing economies for the period 1995–2008, though observations
are not available for all countries for all years. Summary statistics and cor-
relations are presented in table 1. The main points to note are the substantial
cross-country variation in the summary statistics (panel A), including in the
fiscal decentralization indicators, with the measures of decentralization
ranging from .14 to .82, and that the correlation coefficients are relatively
low (panel B), with the exception of those for the fiscal decentralization
indicators, which range from .75 to .97.
Econometric Methodology
We begin with cross-section regressions in which we use one observation
per country, so that a period is defined as the range of years for which
we have data for that country. The basic specification is as follows:
CORit ¼ aþ b1FDECit þ b2FDECit � INSi þ b3POLi þ b4Xi þ ei; ð1Þ
where CORit is the corruption index; FDECit is, alternatively, subnational
government revenues, tax effort, expenditures, or the wage bill; INSi is the
set of dummy variables capturing political and administrative arrangements
for decentralization; POLi is the set of political accountability variables;
and Xi is the set of baseline controls that comprises the natural log of real
per capita GDP, the natural log of a country’s population, the ethnic fractio-
nalization ratio, and a dummy variable indicating whether a country is a
major energy exporter.
Cross-sectional regressions are subject to the shortcoming that they do
not fully control for unobserved country-specific effects and do not exploit
72 Public Finance Review 40(1)
Tab
le1.
Sum
mar
ySt
atis
tics
and
Corr
elat
ions
for
Key
Var
iable
s,C
ross
-Countr
yD
ata
A.Su
mm
ary
Stat
istics
MSD
Min
imum
Max
imum
Corr
uption,In
tern
atio
nal
Coun-
try
Ris
kG
uid
e(I
CR
G)
index
0.5
49
0.2
14
0.1
67
1.0
00
Rev
enue
dec
entr
aliz
atio
n0.1
59
0.1
13
0.0
04
0.4
40
Tax
effo
rtdec
entr
aliz
atio
n0.1
44
0.1
16
0.0
08
0.4
74
Expen
diture
dec
entr
aliz
atio
n0.2
65
0.1
46
0.0
14
0.5
91
Wag
ebill
dec
entr
aliz
atio
n0.3
99
0.2
66
0.0
03
0.8
20
Pre
ssfr
eedom
,Fr
eedom
House
31.1
30
21.2
41
8.6
67
86.6
67
Log
ofre
algr
oss
dom
estic
pro
duct
(GD
P)
per
capita
9.2
50
0.9
12
7.3
57
10.8
15
Log
ofpopula
tion
9.3
77
1.6
54
5.9
32
14.0
11
Frac
tional
izat
ion
ratio
0.3
31
0.2
25
0.0
02
0.8
75
B.C
orr
elat
ion
coef
ficie
nts
Corr
uption
Rev
enue
dec
entr
aliz
atio
nT
axef
fort
dec
entr
aliz
atio
nExpen
diture
dec
entr
aliz
atio
nW
age
bill
dec
entr
aliz
atio
nPre
ssfr
eedom
Frac
tional
izat
ion
ratio
Log
of
popula
tion
Rev
enue
dec
entr
aliz
atio
n�
0.3
39
Tax
effo
rtdec
entr
aliz
atio
n�
0.2
64
0.9
71
Expen
diture
dec
entr
aliz
atio
n�
0.3
88
0.8
43
0.7
47
Wag
ebill
dec
entr
aliz
atio
n�
0.4
89
0.8
25
0.7
56
0.9
15
Pre
ssfr
eedom
,Fr
eedom
House
�0.6
57
0.1
88
0.1
42
0.2
41
0.3
02
Log
ofpopula
tion
0.2
75
�0.1
32
�0.0
09
�0.0
04
�0.0
27
�0.3
52
Frac
tional
izat
ion
ratio
0.1
89
0.4
35
0.3
91
0.4
18
0.4
43
�0.2
69
0.0
41
Log
ofre
alG
DP
per
capita
�0.7
54
0.3
38
0.3
18
0.3
15
0.4
30
0.7
09
�0.4
50
�0.1
24
Not
e:A
llva
lues
are
aver
ages
for
the
per
iod
1995–2008,t
hough
obse
rvat
ions
are
not
avai
lable
for
allc
ountr
ies
for
ally
ears
.The
corr
uption
index
isre
scal
edto
take
valu
esbet
wee
n0
and
1w
ith
0¼
leas
tco
rrup
tion.
Ahig
hsc
ore
on
the
pre
ssfr
eedom
index
indic
ates
less
pre
ssfr
eedom
,an
dth
ein
dex
isen
tere
din
the
corr
elat
ion
mat
rix
with
aneg
ativ
esi
gn.
73
the time-series dimension of the available data (though in many cases, there
is little or no variation in the series over time). Accordingly, we also provide
results from pooled time-series cross-sectional (panel) estimates for the four
indicators of fiscal decentralization. The annual data in these estimates
cover the period 1995–2008. As data for some countries are not available
for all years, the panel is unbalanced with the number of observations
depending on the choice of explanatory variable.
Endogeneity Considerations
The relationship between fiscal decentralization and changes in the quality
of governance might be driven by reverse causation. For example, Fisman
and Gatti (2002) note that corrupt officials of the central government might
be reluctant to allow fiscal decentralization, as this would attenuate their
ability to extract rents. The presence of endogeneity has generally been
dealt with through the use of instrumental variables (IVs). As instruments
for fiscal decentralization, we use dummies identifying countries’ legal ori-
gins (British, French, Socialist, and Germanic). Legal origins have been
used quite extensively and successfully as instruments in the fiscal decen-
tralization literature (Dreher 2006; Fisman and Gatti 2002; de Mello and
Barenstein 2001). The justification for their use is the affinity of a civil legal
code (as opposed to common code) for government decentralization and the
likelihood that legal origin is only indirectly related to governance. The IV
estimates would give consistent results under the assumptions that there is
no second-order serial correlation and that the instruments are uncorrelated
with the error terms. We test for the validity of these assumptions and pres-
ent the test results below.3
Empirical Results
We start with the cross-country cross-section results, which are reported in
tables 2 through 4. Basic results incorporating just the fiscal decentraliza-
tion indicators and baseline control variables are reported in table 2.
Columns 1 through 4 report results from ordinary least squares (OLS) esti-
mates, and columns 5 through 8 report the IV estimates. The two sets of
results are broadly similar. The coefficients on all the fiscal decentralization
indicators are negative and statistically significant. In terms of magnitudes,
the OLS estimates indicate that a 1 standard deviation increase in fiscal
decentralization will be associated with a reduction in corruption rating
of between 0.33 (tax effort decentralization) and 0.54 (wage bill
74 Public Finance Review 40(1)
Tab
le2.
Cro
ss-S
ectional
Res
ults;
Corr
uption,
Fisc
alD
ecen
tral
izat
ion,
and
Bas
icC
ontr
ols
—O
rdin
ary
Leas
tSq
uar
es(O
LS)
Res
ults
Colu
mns
1–4
and
Inst
rum
enta
lV
aria
ble
sR
esults
Colu
mns
5–8
12
34
56
78
Const
ant
�0.2
306*
(0.1
200)
�0.2
668**
(0.1
201)
�0.2
385**
(0.1
132)
�0.2
387**
(0.1
037)
�0.1
857
(0.1
192)
�0.2
356**
(0.1
164)
�0.2
465**
(0.1
169)
�0.2
766**
*(0
.0995)
Subnat
ional
fisca
lva
riab
les
Rev
enue
�0.5
577**
*(0
.2028)
�0.8
173**
(0.3
988)
Tax
effo
rt�
0.4
909*
(0.1
834)
�0.8
409**
(0.3
676)
Expen
diture
�0.5
417**
*(0
.1530)
�1.0
765**
*(0
.3741)
Wag
ebill
�0.3
746**
*(0
.0847)
�0.5
379**
*(0
.1148)
Bas
elin
eco
ntr
ols
Popula
tion
0.0
395**
*(0
.0132)
0.0
460**
*(0
.0126)
0.0
452**
*(0
.0124)
0.0
470**
*(0
.0118)
0.0
634**
*(0
.0212)
0.0
656**
*(0
.0194)
0.0
810**
*(0
.0210)
0.0
664**
*(0
.0131)
Rea
lgr
oss
dom
estic
pro
d-
uct
(GD
P)
per
capita
�0.1
610**
*(0
.0360)
�01596**
*(0
.0358)
�0.1
446**
*(0
.0341)
�0.1
398**
*(0
.0331)
�0.1
684**
*(0
.0358)
�0.1
671**
*(0
.0343)
�0.1
181**
*(0
.0412)
�0.1
114**
*(0
.0362)
Fuel
export
er0.0
465
(0.0
819)
0.0
495
(0.0
835)
0.0
752
(0.0
764)
0.0
897
(0.0
737)
0.1
173**
(0.0
579)
0.1
411**
(0.0
606)
0.1
716**
*(0
.0613)
0.1
340**
*(0
.0494)
Frac
tional
izat
ion
ratio
0.1
384
(0.1
086)
0.0
956
(0.0
979)
0.1
236
(0.1
013)
0.1
015
(0.0
861)
�0.0
457
(0.1
285)
0.0
201
(0.1
249)
0.0
688
(0.1
337)
0.0
189
(0.0
977)
Adju
sted
R2
.532
.506
.574
.615
.555
.549
.573
.658
SE0.1
26
0.1
27
0.1
17
0.1
12
0.1
37
0.1
39
0.1
32
0.1
20
Han
sen
J-st
atis
tic
(pva
lue)
5.5
48
(.353)
12.6
66
(.124)
10.2
25
(.115)
7.8
25
(.166)
Durb
in–W
u–H
ausm
ante
st(p
valu
e)3.2
04
(.074)
6.2
52
(.012)
4.8
92
(.027)
3.9
31
(.047)
Not
e:W
hite
robust
stan
dar
der
rors
are
inpar
enth
eses
bel
ow
the
coef
ficie
nts
.***
,**,
and
*in
dic
ate
stat
istica
lsig
nifi
cance
atth
e1,5
,and
10
per
cent
leve
ls,
resp
ective
ly.T
he
corr
uption
index
has
bee
nre
scal
edto
take
on
valu
esbet
wee
n0
(lea
stco
rrup
t)an
d1
(mos
tco
rrup
t).
75
decentralization) of a standard deviation, which is substantially larger than
Fisman and Gatti’s (2002) finding of a 0.30 standard deviation reduction for
revenue decentralization and which underlines the need to make an assess-
ment based on a range of decentralization indicators.4 The coefficients on
the log of population and of real GDP per capita are statistically significant
and of the expected sign in all of the regressions: more highly populated
countries are more corrupt and wealthier countries are less corrupt. The
energy exporter dummy is statistically significant only in the IV results, where
the coefficient is consistent with natural resource rents promoting corruption.
The ethnic fractionalization ratio is not statistically significant. In the IV esti-
mates, the J-test statistic of overidentifying restrictions is not statistically sig-
nificant, suggesting that the legal origin variables are valid instruments. In
addition, the Durbin–Wu–Hausman (DWH) statistics are statistically signifi-
cant and indicate that endogeneity of fiscal decentralization with the instru-
ments has a highly significant effect on the estimate, which is consistent
with the exogeneity of fiscal decentralization with respect to corruption.5
Table 3 reports the IV estimates that include the interactions between the
fiscal decentralization measures and the different political and administra-
tive arrangements for decentralization. Four points are noteworthy. First,
the coefficients on the fiscal decentralization terms remain statistically sig-
nificant and negative except when the variable interacts with a country hav-
ing an elected bottom tier of government. Second, the political and
administrative arrangements interact to mitigate the beneficial impact of
fiscal decentralization on corruption, which is consistent with the findings
of Dreher (2006) and Kyria and Roca-Sagales (2011). Specifically, federal
structures and, to a lesser extent, countries with a constitution that grants
limited autonomy to subnational governments tend to reduce substantially
the beneficial impact that fiscal decentralization has in reducing corruption.
Third, in these estimates, population, real GDP per capita, and energy
exporter control variables generally remain statistically significant and of
the expected sign. Finally, though the J-statistics remain appropriate, the
DWH test statistics do not support the exogeneity of fiscal decentralization
with respect to corruption in three of the estimates (columns 5, 7, and 8).
The beneficial impact of fiscal decentralization on corruption remains
robust to the inclusion of the political representations control variables.
These results are reported in table 4. The size of the coefficient on the fiscal
decentralization indicators is reduced somewhat but remains negative and
statistically significant. There is some variation in the statistical signifi-
cance of the interaction terms, but the earlier conclusion holds: where the
interaction terms are statistically significant, their impact is to mitigate the
76 Public Finance Review 40(1)
Tab
le3.
Cro
ss-S
ectional
Inst
rum
enta
lV
aria
ble
sR
esults:
Fisc
alD
ecen
tral
izat
ion,A
dm
inis
trat
ive
Dec
entr
aliz
atio
n,an
dBas
elin
eC
ontr
ols
12
34
56
78
910
11
12
Const
ant
�0.0
786
(0.0
913)
�0.3
109**
(0.1
264)
�0.2
476**
(0.1
162)
�0.1
163
(0.1
117)
�0.1
958
(0.9
81)
�0.2
048**
(0.0
981)
�0.1
905*
(0.0
959)
�0.1
148
(0.1
256)
�0.2
251**
(0.1
037)
�0.1
800**
(0.0
877)
�0.2
437**
*
(0.0
930)
�0.2
082**
(0.0
932)
Subnat
ional
fisca
lva
riab
les
Rev
enue
�1.1
544**
*
(0.2
679)
�1.2
811
(1.5
396)
�1.1
247**
*
(0.2
924)
Tax
effo
rt�
1.0
656**
*
(0.3
562)
�0.5
595
(2.6
432)
�0.6
243**
*
(0.2
22)
Expen
diture
�1.1
395**
*
(0.3
511)
�0.6
312
(1.1
065)
�0.9
862**
*
(0.2
550)
Wag
ebill
�0.4
546**
*
(0.0
735)
�0.1
799
(0.3
573)
�0.4
501**
*
(0.1
001)
Adm
inis
trat
ive
dec
entr
aliz
atio
nin
tera
ctio
ns
Fisc
alva
riab
le�
feder
al
const
itution
0.9
100**
*
(0.2
470)
0.7
618**
*
(0.2
847)
0.3
973**
*
(0.1
456)
0.2
420**
*
(0.0
735)
Fisc
alva
riab
le�
elec
ted
bott
om
tier
0.5
488
(1.4
221)
0.2
916
(2.5
141)
�1.0
022
(0.9
325)
�0.1
742
(0.2
778)
Fisc
alva
riab
le�
limited
auto
nom
y
0.3
501
(0.2
785)
0.4
978**
*
(0.1
660)
0.2
420
(0.1
683)
0.1
726*
(0.1
002)
Bas
elin
eco
ntr
ols
Popula
tion
0.0
307**
(0.0
151)
0.0
449**
(0.0
152)
0.0
582**
(0.0
148)
0.0
499**
(0.0
157)
0.0
563**
(0.0
118)
0.0
565**
(0.0
162)
0.0
499**
*
(0.0
105)
0.0
499**
*
(0.0
105)
0.0
579**
*
(0.0
109)
Rea
lG
DP
per
capita
�0.1
733**
*
(0.0
283)
�0.1
194**
*
(0.0
417)
�0.1
226**
*
(0.0
358)
�0.1
781**
*
(0.0
304)
�0.1
889**
*
(0.0
381)
�0.0
153**
*
(0.0
298)
�0.1
133**
*
(0.0
475)
�0.1
954**
*
(0.0
475)
�0.1
131**
*
(0.0
335)
�0.1
419**
*
(0.0
289)
�0.1
352**
*
(0.0
378)
�0.1
408**
*
(0.0
311)
Fuel
export
er0.0
683
(0.0
504)
Frac
tional
izat
ion
ratio
�0.0
904
(0.0
994)
Adju
sted
R2
.665
.458
.432
.591
.613
.617
.488
.686
.509
.732
.617
.691
SE0.1
17
0.1
28
0.1
31
0.1
30
0.1
27
0.1
24
0.1
29
0.1
12
0.1
26
0.1
03
0.1
16
0.1
11
Han
sen
Jst
atis
tic
(pva
lue)
9.8
30
(.198)
12.9
36
(.165)
7.7
01
(.261)
9.1
53
(.329)
8.3
41
(.138)
11.8
30
(.159)
7.7
62
(.170)
12.2
77
(.139)
10.0
49
(.186)
18.3
67
(.019)
13.2
42
(.104)
10.9
81
(.139)
Durb
in–W
u–H
ausm
an
test
(pva
lue)
3.4
12
(.065)
2.4
64
(.104)
7.3
82
(.007)
8.1
27
(.004)
1.3
50
(.245)
7.2
63
(.007)
6.1
14
(.013)
0.1
91
(.662)
7.0
32
(.008)
1.0
49
(.306)
0.3
23
(.570)
3.8
07
(.051)
Not
e:W
hite
robust
stan
dar
der
rors
are
inpar
enth
eses
bel
ow
the
coef
ficie
nts
.**
*,**
,an
d*
indic
ate
stat
istica
lsi
gnifi
cance
atth
e1,5,an
d10
per
cent
leve
ls,
resp
ective
ly.T
he
corr
uption
index
has
bee
nre
scal
edto
take
on
valu
esbet
wee
n0
(lea
stco
rrup
t)an
d1
(mos
tco
rrup
t).
77
Tab
le4.C
ross
-Sec
tional
Inst
rum
enta
lV
aria
ble
sR
esults:
Fisc
alD
ecen
tral
izat
ion,A
dm
inis
trat
ive
Dec
entr
aliz
atio
n,Polit
ical
Rep
rese
nta
tion,
and
Bas
elin
eC
ontr
ols
12
34
56
78
910
11
12
Const
ant
�0.6
526**
*
(0.1
323)
�0.6
192**
*
(0.1
870)
�0.5
686**
*
(0.1
683)
�0.6
788**
*
(0.1
637)
�0.7
368**
*
(0.1
837)
�0.6
559**
*
(0.1
591)
�0.6
437**
*
(0.1
201)
�0.1
148
(0.1
256)
�0.6
711**
*
(0.1
037)
�0.1
800**
(0.1
173)
�0.5
970**
*
(0.1
477)
�0.6
350**
*
(0.0
942)
Subnat
ional
fisca
lva
riab
les
Rev
enue
�0.8
246**
*
(0.1
812)
�0.9
733
(1.6
572)
�0.6
923**
*
(0.2
119)
Tax
effo
rt�
0.8
309**
*
(0.2
743)
�1.7
986*
(2.6
432)
�0.8
854**
*
(0.2
999)
Expen
diture
�0.4
666**
*
(1.1
221)
�0.5
067
(0.5
322)
�0.5
174**
*
(0.1
802)
Wag
ebill
�0.4
402**
*
(0.1
188)
�0.3
972
(0.4
045)
�0.3
534**
*
(0.0
859)
Adm
inis
trat
ive
dec
entr
aliz
atio
nin
tera
ctio
ns
Fisc
alva
riab
le�
fed-
eral
const
itution
0.3
386*
(0.1
696)
0.3
314**
*
(0.2
645)
0.0
722
(0.1
039)
0.1
358
(0.0
845)
Fisc
alva
riab
le�
elec
ted
bott
om
tier
0.4
048
(1.5
485)
1.1
818
(0.8
598)
0.0
290
(0.4
676)
0.0
266
(0.3
308)
Fisc
alva
riab
le�
lim-
ited
auto
nom
y
0.3
668**
(0.1
865)
0.4
274*
(0.2
544)
0.2
093*
(0.1
236)
0.1
713*
(0.0
603)
Polit
ical
repre
senta
tion
vari
able
s
Pre
siden
tial
syst
em0.1
644**
*
(0.0
498)
0.1
531**
*
(0.0
537)
0.1
547**
*
(0.0
545)
0.1
738**
*
(0.0
514)
0.1
369**
(0.0
562)
0.1
494**
*
(0.0
516)
0.1
447**
*
(0.0
458)
0.1
578**
*
(0.0
415)
0.1
665**
*
(0.0
397)
0.1
087**
(0.0
469)
0.1
377**
*
(0.0
454)
0.1
615**
*
(0.0
407)
Continuous
dem
ocr
acy
�0.0
709*
(0.0
391)
�0.1
053*
(0.0
531)
�0.1
037**
(0.0
459)
�0.0
730*
(0.0
436)
�0.1
450**
(0.0
630)
�1.1
038**
(0.0
480)
�0.0
745**
(0.0
371)
�0.0
794*
(0.0
442)
�0.0
852**
(0.0
381)
�0.0
984**
(0.0
425)
�0.1
153**
(0.0
569)
�0.0
985**
*
(0.0
365)
Pre
ssfr
eedom
0.0
040**
*
(0.0
013)
0.0
027*
(0.0
017)
0.0
025
(0.0
017)
0.0
041**
*
(0.0
015)
0.0
033*
(0.0
018)
0.0
036*
(0.0
016)
0.0
037**
*
(0.0
012)
0.0
039**
*
(0.0
013)
0.0
040**
*
(0.0
012)
0.0
014
(0.0
014)
0.0
021
(0.0
013)
0.0
031**
*
(0.0
010)
Bas
elin
eco
ntr
ol
Popula
tion
0.0
280**
(0.0
114)
0.0
258*
(0.0
138)
0.0
269**
*
(0.0
099)
0.0
269**
(0.0
135)
0.0
463**
*
(0.0
169)
0.0
454**
*
(0.0
015)
0.0
305**
*
(0.0
109)
0.0
338**
*
(0.0
117)
0.0
312**
(0.0
121)
0.0
335**
(0.0
134)
0.0
374**
*
(0.0
081)
0.0
325**
*
(0.0
071)
Adju
sted
R2
.779
.590
.614
.757
.586
.667
.796
.792
.807
.765
.756
.732
SE0.0
89
0.1
14
0.1
10
0.0
94
0.1
25
0.1
12
0.0
86
0.0
87
0.0
83
0.0
94
0.0
96
0.0
89
Han
sen
Jst
atis
tic
(pva
lue)
9.5
44
(.216)
12.0
42
(.340)
8.6
21
(.375)
10.1
96
(.251)
3.8
75
(.790)
5.7
09
(.574)
8.9
70
(.345)
8.5
61
(.286)
9.4
34
(.223)
11.8
05
(.107)
8.0
72
(.233)
10.6
37
(.223)
Durb
in–W
u–H
ausm
an
test
(pva
lue)
7.1
97
(.007)
3.7
45
(.053)
4.0
12
(.045)
6.6
62
(.010)
8.2
02
(.004)
5.9
95
(.014)
3.8
61
(.050)
4.2
62
(.039)
3.2
31
(.072)
2.9
36
(.087)
4.1
06
(.043)
2.9
62
(.085)
Not
e:W
hite
robust
stan
dar
der
rors
are
inpar
enth
eses
bel
ow
the
coef
ficie
nts
.**
*,**
,an
d*
indic
ate
stat
istica
lsi
gnifi
cance
atth
e1,5,an
d10
per
cent
leve
ls,
resp
ective
ly.T
he
corr
uption
index
has
bee
nre
scal
edto
take
on
valu
esbet
wee
n0
(lea
stco
rrup
t)an
d1
(mos
tco
rrup
t).
78
benefits of fiscal decentralization on corruption. The political representation
controls are all statistically significant and of the expected sign: countries
with long periods of continuous democracy and a free press are associated
with less corruption, and countries with presidential democracies are associ-
ated with more corruption (compared to parliamentary democracies). Inclu-
sion of the political representation variables appears to affect the baseline
control variables, with only population remaining statistically significant.
In general, the J-test statistics and DWH test statistics remain satisfactory.
The panel estimation results are reported in table 5 and confirm the find-
ings of the cross-Sectional estimates. The coefficients on each of the fiscal
decentralization variables are negative and statistically significant, and the
beneficial impact of fiscal decentralization on corruption is again substan-
tially mitigated by the interaction variable. The impact of the political rep-
resentation variables on corruption is statistically significant and of the
expected sign, and the coefficients on the baseline control variables are gen-
erally statistically significant and of the expected sign. The J-test statistics
for overidentifying restrictions and DWH statistics are again generally
satisfactory in regard to the validity of the instruments and the exogeneity
of fiscal decentralization.
Conclusion
We have focused in this article on the impact of fiscal decentralization on
the corruption aspect of governance employing a wider range of decentra-
lization measures and a broader range of control variables than used in
much of the empirical literature. Our results provide strong support to the
view that fiscal decentralization has a beneficial impact on improving gov-
ernance in a country by reducing corruption. This result is robust to control-
ling for interactions between fiscal decentralization and administrative
arrangements for decentralization, for political representation variables that
other studies have shown to dominate explanations of corruption, and for
the level of development, country size, ethnic fractionalization, and a coun-
try being a major energy exporter. Our results testify to the importance of
also incorporating formal institutional arrangements for decentralization
into empirical work. Subnational government autonomy appears to reduce
the beneficial impact of fiscal decentralization on corruption, which sug-
gests that fiscal decentralization is most effective in reducing corruption
when these resources are nonetheless largely directed by the central govern-
ment. Our results also act as a counterweight to the literature on the deter-
minants of corruption that emphasizes the dominance of political
Altunbas and Thornton 79
Tab
le5.
Pan
elIn
stru
men
tal
Var
iable
sR
esults:
Corr
uption,
Fisc
alD
ecen
tral
izat
ion,
Adm
inis
trat
ive
Dec
entr
aliz
atio
n,
and
Polit
ical
Rep
rese
nta
tion
12
34
56
78
910
11
12
Const
ant
�0.4
511**
*
(0.0
843)
�0.6
516**
*
(0.0
842)
�0.6
792**
*
(0.0
893)
�0.4
456**
*
(0.0
824)
�0.6
420**
*
(0.0
853)
�0.4
591**
*
(0.0
778)
�0.4
200**
*
(0.0
949)
�0.5
728**
*
(0.0
827)
�0.4
022**
*
(0.0
839)
�0.6
146**
*
(0.1
113)
�0.6
152**
*
(0.1
036)
�0.6
581**
*
(0.1
027)
Subnat
ional
fisca
lva
riab
les
Rev
enue
�0.9
036**
*
(0.1
314)
�1.9
841**
*
(0.5
691)
�0.7
601**
*
(0.1
172)
Tax
effo
rt�
0.8
252**
*
(0.1
164)
�2.5
718**
*
(0.8
726)
�0.7
450**
*
(0.1
114)
Expen
diture
�0.6
140**
*
(0.0
965)
�0.6
544**
*
(0.1
798)
�0.5
281**
*
(0.0
915)
Wag
ebill
�0.3
660**
*
(0.0
983)
�0.5
359**
*
(0.1
508)
�0.3
229**
*
(0.0
721)
Adm
inis
trat
ive
dec
entr
aliz
atio
nin
tera
ctio
ns
Fisc
alva
riab
le�
fed-
eral
const
itution
0.2
711**
*
(0.1
039)
0.1
955*
(0.1
087)
0.0
054
(0.0
684)
�0.0
542
(0.0
573)
Fisc
alva
riab
le�
elec
ted
bott
om
tier
1.3
509*
(1.5
485)
1.9
247**
(0.8
179)
0.3
197*
(0.1
633)
0.2
441*
(0.1
325)
Fisc
alva
riab
le�
limited
auto
nom
y
0.4
876**
*
(0.1
168)
0.2
098**
(0.1
039)
0.3
249**
*
(0.0
641)
0.1
872*
(0.0
544)
Polit
ical
repre
senta
tion
vari
able
s
Pre
siden
tial
syst
em0.1
097**
*
(0.0
261)
0.1
221**
*
(0.0
255)
0.1
401**
*
(0.0
272)
0.0
634**
*
(0.0
234)
0.0
790**
*
(0.0
231)
0.0
755**
*
(0.0
226)
0.0
932**
*
(0.0
266)
0.0
924**
*
(0.0
247)
0.1
669**
*
(0.0
247)
0.1
087**
(0.0
317)
0.0
986**
*
(0.0
312)
0.1
595**
*
(0.0
314)
Continuous
dem
ocr
acy
�0.0
487**
(0.0
207)
�0.0
912**
*
(0.0
218)
�0.0
769**
*
(0.0
226)
�0.0
593**
*
(0.0
202)
�0.0
965**
*
(0.0
228)
�0.0
623**
*
(0.0
200)
�0.0
329
(0.0
229)
�0.0
744**
*
(0.0
215)
�0.0
432**
(0.0
209)
�0.0
961**
*
(0.0
294)
�0.1
205**
*
(0.0
320)
�0.0
705**
*
(0.0
268)
Pre
ssfr
eedom
0.0
015
(0.0
010)
0.0
020**
(0.0
010)
0.0
009
(0.0
011)
0.0
021**
(0.0
010)
0.0
026**
*
(0.0
010)
0.0
021**
(0.0
010)
0.0
003
(0.0
011)
0.0
019*
(0.0
011)
0.0
010
(0.0
011)
0.0
001
(0.0
014)
0.0
003
(0.0
013)
0.0
016
(0.0
013)
Bas
elin
eco
ntr
ols
Popula
tion
0.0
237**
*
(0.0
066)
0.0
228*
(0.0
063)
0.0
255**
*
(0.0
071)
0.0
238**
*
(0.0
062)
0.0
228**
*
(0.0
039)
0.0
257**
*
(0.0
060)
0.0
322**
*
(0.0
072)
0.0
214**
*
(0.0
063)
0.0
222**
*
(0.0
067)
0.0
515**
*
(0.0
129)
0.0
406**
*
(0.0
122)
0.0
224*
(0.0
119)
(con
tinue
d)
80
Tab
le5.
(co
nti
nu
ed
) 12
34
56
78
910
11
12
Rea
lgr
oss
dom
estic
pro
duct
(GD
P)
per
capita
�0.0
607**
*
(0.0
215)
�0.0
008
(0.0
247)
�0.0
511**
(0.0
215)
�0.0
738**
*
(0.0
211)
�0.0
198
(0.0
255)
�0.0
731**
*
(0.0
204)
�0.0
643**
*
(0.0
234)
�0.0
366*
(0.0
222)
�0.0
716**
*
(0.0
205)
�0.0
144
(0.0
298)
�0.0
138
(0.0
300)
�0.0
018
(0.0
284)
Fuel
export
er0.0
518
(0.0
445)
0.0
439
(0.0
444)
0.0
835*
(0.0
477)
0.0
896*
(0.0
450)
0.0
692
(0.0
450)
0.0
935**
(0.0
455)
0.0
932*
(0.0
474)
0.0
531
(0.0
448)
0.0
908**
(0.0
452)
0.0
465
(0.0
582)
0.0
440
(0.0
589)
0.0
228
(0.0
542)
Frac
tional
izat
ion
ratio
0.1
456**
*
(0.0
451)
0.2
776**
*
(0.0
639)
0.1
101**
(0.0
506)
0.1
219**
*
(0.0
427)
0.2
537**
*
(0.0
641)
0.1
031*
(0.0
439)
0.2
107**
*
(0.0
479)
0.2
433**
*
(0.0
558)
0.0
830
*
(0.0
492)
0.2
704**
*
(0.0
696)
0.2
828**
*
(0.0
738)
0.2
310**
*
(0.0
719)
Adju
sted
R2
.509
.507
.531
.488
.486
.498
.474
.500
.522
.460
.425
.533
Han
sen
Jst
atis
tic
(pva
lue)
2.3
41
(.505)
6.7
56
(.239)
9.8
07
(.200)
4.7
15
(.452)
4.1
56
(.385)
4.3
98
(.355)
6.3
18
(.097)
4.7
16
(.194)
7.2
93
(.121)
7.2
77
(.064)
3.5
85
(.310)
6.9
00
(.075)
Durb
in–W
u–H
ausm
an
test
(pva
lue)
20.1
77
(.000)
3.0
66
(.0799)
14.3
33
(.000)
41.5
84
(.000)
8.8
03
(.003)
11.6
44
(.001)
25.8
49
(.000)
17.4
50
(.000)
9.0
00
(.003)
3.2
65
(.071)
11.0
11
(.001)
0.0
25
(.876)
Not
e:W
hite
robust
stan
dar
der
rors
are
inpar
enth
eses
bel
ow
the
coef
ficie
nts
.**
*,**
,an
d*
indic
ate
stat
istica
lsi
gnifi
cance
atth
e1,5,an
d10
per
cent
leve
ls,re
spec
tive
ly.T
he
corr
uption
index
has
bee
nre
scal
edto
take
on
valu
esbet
wee
n0
(lea
stco
rrup
t)an
d1
(mos
tco
rrup
t).
81
accountability variables. Finally, our results lend support to economists
who take a more positive view of fiscal decentralization as a means to pro-
mote better governance.
Acknowledgments
The authors are grateful to the editor, three anonymous referees of the journal, and
to Mariana Mc Loughlin for comments that improved the article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or
publication of this article.
Notes1. For example, in the case of Colombia, the cash data indicate that about 26 percent
of general government revenue is at the subnational level of government,
whereas the cash data indicate about 16 percent.
2. The CIA data can be downloaded from https://www.cia.gov/library/publications/
the-world-factbook/.
3. The first stage equation is as follows
FDECit ¼ ai þ b1LEG1;
where LEG is the set of dummy variables indicating British, French, Socialist, or
Scandinavian legal origin.
4. To determine the magnitude of the impact of fiscal decentralization on corrup-
tion, we estimate beta coefficients in the regression model by transforming pre-
dictor variables standard scores (or Z scores). As the beta coefficients are all in
the same standardized units, we can compare these coefficients to assess the rela-
tive strength of each of the predictors. For example, in our OLS cross-sectional
estimates (table 2), subnational government revenue estimation has a beta
coefficient of �.3522. Thus, a 1 standard deviation increase in revenue leads
to a 0.35 standard deviation decrease in predicted corruption.
5. F tests for the joint significance of the legal origin dummies show that they are
good predictors of the fiscal decentralization indicators, though the J-test statis-
tics for overidentifying restrictions are not always accepted.
82 Public Finance Review 40(1)
References
Adsera, Alicia, Carlos Boix, and Mark Payne. 2003. ‘‘Are You Being Served? Polit-
ical Accountability and Quality of Government.’’ The Journal of Law Economics
and Organization 19:445-90.
Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, and
Romain Wacziarg. 2003. ‘‘Fractionalization.’’ Journal of Economic Growth
8:155-94.
Bardhan, Pranab, and Dilip Mookherjee. 2000. ‘‘Capture and Governance at Local
and National Levels.’’ The American Economic Review 90:135-39.
Beck, Thorsten, George Clark, Alberto Groff, Philip Keefer, and Patrick Walsh.
2001. ‘‘New Tools in Comparative Political Economy.’’ World Bank Economic
Review 15:165-76.
Central Intelligence Agency. 2009. The World Factbook. Washington, DC: Central
Intelligence Agency.
Collier, Paul A., and Anke Hoeffler. 2005. ‘‘Democracy and Natural Resource
Rents.’’ Oxford University Department of Economics Working paper GPRG-
WPS-016.
Darby, Julia, R., Anton Muscatelli, and Graeme Roy. 2003. ‘‘Fiscal Decentralization
in Europe: A Review of Recent Experience.’’ In Regional Public Finance: Eur-
opean Research in Regional Science, vol. 13, edited by Jan Monnesland, 1-32.
London: Pion Limited.
de Mello, Luiz. 2000. ‘‘Can Fiscal Decentralization Strengthen Social Capital?’’
International Monetary Fund Working Paper 00/129. Washington, DC.
de Mello, Luiz, and Matias Barenstein. 2001. ‘‘Fiscal Decentralization and Govern-
ance: A Cross-Country Analysis.’’ International Monetary Fund Working Paper
01/71. Washington, DC.
Dreher, Axel. 2006. ‘‘Power to the People? The Impact of Decentralization on
Governance.’’ KOF Swiss Economic Institute Working Paper No. 06-121, Zurich.
Dziobek, Claudia, Carlos Gutierrez Mangas, and Phebby Kufa. 2011. Measuring
Fiscal Decentralization – Exploring the IMF’s Databases. International Mone-
tary Fund Working Paper 11/126, Washington, DC.
Enikolopov, Ruben, and Ekaterina Zhuravskaya. 2007. ‘‘Decentralization and
Political Institutions.’’ Journal of Public Economics 91:261-90.
Fisman, Raymond, and Roberta Gatti. 2002. ‘‘Decentralization and Corruption:
Evidence across Countries.’’ Journal of Public Economics 83:325-45.
Fukasaku, Kiichiro, and Luiz R. de Mello. 1999. Fiscal Decentralization in
Emerging Economies: Governance Issues. Paris: Inter-American Development
Bank/Development Centre of the Organisation for Economic Co-operation and
Development.
Altunbas and Thornton 83
Heston, Alan, Robert Summers, and Bettina Aten. 2009. Penn World Table Version
6.3, Center for International Comparisons of Production, Income and Prices at the
University of Pennsylvania. http://pwt.econ.upenn.edu/php_site/pwt_index.php.
Hommes, Rudolf. 1995. ‘‘Conflicts and Dilemmas of Decentralization.’’ Paper pre-
pared for the World Bank’s Annual Bank Conference on Development Economics,
Washington, DC. http://website1.wider.unu.edu/lib/pdfs/WB-ACDE-1995.pdf.
Inman, Robert P., and Daniel L. Rubinfeld, 1997. ‘‘Rethinking Federalism.’’ The
Journal of Economic Perspectives 11:43-64.
International Country Risk Guide. 2001. ‘‘Financial, Political and Economic Risk
Ratings for 140 Countries.’’ PRS Group. http://www.prsgroup.com/icrg/icrg.html.
Islam, Roumeen, and Claudio E. Montenegro. 2002. ‘‘What Determines the Qual-
ity of Institutions.’’ World Bank Policy Research Working Paper 2764,
Washington, DC.
Joumard, Isabelle, and Per Mathis Kongsrud. 2003. ‘‘Fiscal Relations across
Government Levels.’’ OECD Economic Studies 36:155-228.
Knack, Stephen. 2001. ‘‘Aid Dependence and the Quality of Governance:
Cross-Country Empirical Tests.’’ Southern Economic Journal 66:310-29.
Knack, Stephen, and Philip Keefer. 1995. ‘‘Institutions and Economic Performance:
Cross-Country Tests Using Alternative Institutional Measures.’’ Economics and
Politics 7:207-27.
Kyria, Andreas P., and Oriol Roca-Sagales. 2011. ‘‘Fiscal Decentralization and
Government Quality in the OECD.’’ Economics Letters 111:191-3.
La Porta, Rafael, Florencio Lopez de Silanes, Andrei Shleifer, and Robert Vishny.
1999. ‘‘The Quality of Government.’’ Journal of Law Economics and Organiza-
tion 15:222-79.
Lederman, Daniel, Norman Loayza, and Ries R. Soares. 2005. ‘‘Accountability and
Corruption: Political Institutions Matter.’’ Economics and Politics 17:1-35.
Leite, Carlos A., and Jens Weidmann. 2002. ‘‘Does Mother Nature Corrupt? Natural
Resources, Corruption, and Economic Growth.’’ In Governance, Corruption,
and Economic Performance, edited by George T. Abed and Sanjeev Gupta,
159-96. Washington, DC: International Monetary Fund.
Mauro, Paolo. 1995. ‘‘Corruption and Growth.’’ Quarterly Journal of Economics
110:681-712.
Mauro, Paolo. 1998. ‘‘Corruption and the Composition of Public Expenditure.’’
Journal of Public Economics 69:263-79.
Montinola, Gabriella, Qian Yingyi, and Barry R. Weingast. 1995. ‘‘Federalism Chi-
nese Style: The Political Basis for Economic Success.’’ World Politics 48:50-81.
Oates, Wallace E. 1972. Fiscal Federalism. New York: Harcourt Brace Jovanovich.
Panizza, Ugo. 1999. ‘‘On the Determinants of Fiscal Centralization: Theory and
Evidence.’’ Journal of Public Economics 74:97-139.
84 Public Finance Review 40(1)
Prud’homme, Remy. 1995. ‘‘On the Dangers of Decentralization.’’ World Bank
Research Observer 10:201-20.
Seabright, Paul. 1996. ‘‘Accountability and Decentralization in Government: An
Incomplete Contracts Model.’’ European Economic Review 40:61-89.
Srinivasan, T. N. 1986. ‘‘The Costs and Benefits of Being a Small, Remote, Island,
Landlocked, or Mini-State Economy.’’ World Bank Research Observer 1:205-18.
Tabellini, Guido. 2000. ‘‘Constitutional Determinants of Government Spending.’’
CESifo Working Paper No. 265, Munich, Germany.
Tanzi, Vito. 1995. ‘‘Fiscal Federalism and Decentralization: A Review of some
Efficiency and Macroeconomic Aspects.’’ In Annual World Bank Conference
on Development Economics, edited by Michael Bruno and Boris Pleskovic,
295-316. Washington, DC: The World Bank.
Thornton, John. 2007a. ‘‘Fiscal Decentralization and Economic Growth Reconsid-
ered.’’ Journal of Urban Economics 61:64-70.
Thornton, John. 2007b. ‘‘Further Evidence on Revenue Decentralization and
Inflation.’’ Economics Letters 95:140-5.
Treisman, Daniel. 2000. ‘‘The Causes of Corruption: A Cross-National Study.’’
Journal of Public Economics 76:399-457.
Treisman, Daniel. 2002. ‘‘Decentralization and the Quality of Government.’’
Unpublished paper, University of California. http://imf.org/external/pubs/ft/
seminar/2000/fiscal/treisman.pdf.
Weingast, Barry R. 1995. ‘‘The Economic Role of Political Institutions: Market-
Preserving Federalism and Economic Development.’’ Journal of Law Economics
and Organization 11:1-31.
World Bank. 1999. World Development Report. Washington, DC: The World Bank.
World Bank. 2004. World Development Report. Washington, DC: The World Bank.
Bios
Yener Altunbas is a professor of Banking and Finance at Bangor University Busi-
ness School. His main research interests are the study of European banks, bank effi-
ciency, stock market analysis, corporate governance, electoral studies, and regional
and urban economics. He holds a PhD from the University of Wales, Bangor.
John Thornton is a professor of Global Finance and Head of Bangor University
Business School. His research interests are macroeconomic policy in emerging mar-
ket economies, finance and economic development, financial market regulation, and
macroeconomic management and fiscal decentralization. He has been a staff mem-
ber of the Economics Department of the OECD and an assistant director at the Inter-
national Monetary Fund. He holds a PhD from CASS Business School, London.
Altunbas and Thornton 85