Fiscal and Political DecentralizationFiscal and Political ...gent.uab.cat/oroca/sites/gent.uab.cat.oroca/files/ispwp...Fiscal and Political DecentralizationFiscal and Political Decentralization

International Studies ProgramInternational Studies ProgramWorking Paper 10-05February 2010

Fiscal and Political DecentralizationFiscal and Political Decentralization and Government Quality

Andreas P KyriacouAndreas P. KyriacouOriol Roca-Sagalés

International Studies Program Andrew Young School of Policy Studies Georgia State University Atlanta, Georgia 30303 United States of America Phone: (404) 413-0233 Fax: (404) 413-0244 Email: [email protected] Internet: http://isp-aysps.gsu.edu Copyright 2006, the Andrew Young School of Policy Studies, Georgia State University. No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means without prior written permission from the copyright owner.

International Studies Program Working Paper 10-05

Fiscal and Political Decentralization and Government Quality

Andreas P. Kyriacou Oriol Roca-Sagalés February 2010

International Studies Program Andrew Young School of Policy Studies The Andrew Young School of Policy Studies was established at Georgia State University with the objective of promoting excellence in the design, implementation, and evaluation of public policy. In addition to two academic departments (economics and public administration), the Andrew Young School houses seven leading research centers and policy programs, including the International Studies Program. The mission of the International Studies Program is to provide academic and professional training, applied research, and technical assistance in support of sound public policy and sustainable economic growth in developing and transitional economies. The International Studies Program at the Andrew Young School of Policy Studies is recognized worldwide for its efforts in support of economic and public policy reforms through technical assistance and training around the world. This reputation has been built serving a diverse client base, including the World Bank, the U.S. Agency for International Development (USAID), the United Nations Development Programme (UNDP), finance ministries, government organizations, legislative bodies and private sector institutions. The success of the International Studies Program reflects the breadth and depth of the in-house technical expertise that the International Studies Program can draw upon. The Andrew Young School's faculty are leading experts in economics and public policy and have authored books, published in major academic and technical journals, and have extensive experience in designing and implementing technical assistance and training programs. Andrew Young School faculty have been active in policy reform in over 40countries around the world. Our technical assistance strategy is not to merely provide technical prescriptions for policy reform, but to engage in a collaborative effort with the host government and donor agency to identify and analyze the issues at hand, arrive at policy solutions and implement reforms. The International Studies Program specializes in four broad policy areas: Fiscal policy, including tax reforms, public expenditure reviews, tax administration reform Fiscal decentralization, including fiscal decentralization reforms, design of intergovernmental

transfer systems, urban government finance Budgeting and fiscal management, including local government budgeting, performance-

based budgeting, capital budgeting, multi-year budgeting Economic analysis and revenue forecasting, including micro-simulation, time series

forecasting, For more information about our technical assistance activities and training programs, please visit our website at http://isp-aysps.gsu.edu or contact us by email at [email protected].

1

Fiscal and Political Decentralization and Government Quality

Andreas P. Kyriacou*1

Departament d’Economia, Universitat de Girona, Campus de Montilivi,

and

Oriol Roca-Sagalés Departament d’Economia Aplicada, Universitat Autònoma de Barcelona

Abstract In this paper we apply both cross-section and panel analysis to the relationship between fiscal and political decentralization and government quality. We find that fiscal decentralization has a positive impact. Moreover, political decentralization tends to reduce the positive impact of fiscal decentralization on the quality of government. This negative impact of political decentralization on government quality persists when controlling for the degree of democratic maturity of countries but disappears when controlling for the extent of experience with statehood or public administration. This suggests that it is more affordable, in terms of government quality, to combine fiscal and political decentralization in countries with a long history of statehood.

JEL Classification: D73, H11, H77, K42. Keywords: Quality of government, fiscal decentralization, political decentralization, democratic maturity, cross-section and panel data.

* Corresponding author. Tel.: + 34 972 418716; fax: + 34 972 418032. E-mail address: [email protected] (A. Kyriacou). 1 Paper prepared for the conference on “The Political and Economic Consequences of Decentralization” in Santiago de Compostela, Galicia, Spain on November 5-6, 2009.

2 International Studies Program Working Paper Series

1. Introduction

Theoretical arguments have been advanced to explain why fiscal and political decentralization may either reduce or increase government quality. Fiscal decentralization puts resources in the hands of better informed benevolent local governments and thus potentially allows them to better cater towards citizens preferences. If governments are leviathans, then fiscal decentralization can improve their performance insofar as it increases inter-jurisdictional competition. Alternatively, inter-jurisdictional competition may lead to a race to the bottom to the detriment of government quality or it may lead to fiscal or regulatory “over-grazing” in the context of multi-level government. Political decentralization in the form of sub-national elections empowers voters and so is likely to improve government quality but it may also lead to the capture of sub-national politicians by special interests thereby having the opposite effect. Political decentralization in the form of additional checks and balances may protect against bad government but it may also prevent improvements in government effectiveness. And political decentralization in the guise of the allocation policies to sub-national governments may either improve government quality (since sub-national governments are better informed) or reduce it, if it makes it more difficult for citizens to assign merit or blame to the different levels of government.

Most empirical studies have focused on the impact of fiscal decentralization and come out overwhelmingly in favor of a positive contribution from this type of decentralization on government quality. There has been less empirical work on the influence of political decentralization on government quality and it is less unanimous: either political decentralization has no effect on government quality, or if it does so, it may be positive or negative. In this paper, we apply a new data set to the relationship between decentralization and government quality. Based on both cross-section and fixed effects panel estimations, we find that fiscal decentralization increases government quality but this positive effect is mitigated when combined with political decentralization (measured in different ways). Of particular interest is the additional finding that the mediating impact of political decentralization on the relationship between fiscal decentralization and government quality does not depend on the degree of democratic maturity but may depend on the depth of experience with public or state administration.

The paper is structured as follows. In sections 2 and 3 we review the theoretical and empirical work which has focused on the relationship between fiscal and political decentralization on government quality. We then examine, in section 4, the possible role of maturity (democratic or other) on this relationship. We continue by discussing the data used in our estimations and introduce the empirical methodology employed in sections 5 and 6 respectively. In the last three sections of the paper we present our empirical results (section 7), relate them to previous work in this area (section 8) and conclude the paper (section 9).

2. Fiscal decentralization and government quality

Giving sub-central governments the power to both raise and spend fiscal resources has typically been seen in a favorable light independently of whether one ascribes to the notion of government as a benevolent social welfare maximizer or government as a leviathan. In the case of the former, fiscal decentralization should increase welfare since sub-central governments are better informed about local conditions and can find it easier

Fiscal and Political Decentralization and Government Quality 3

to adapt policies to better satisfy citizen preferences (Oates, 1972). On the other hand, if governments are power-seeking leviathans, then fiscal decentralization can constrain them basically because it opens up the possibility of inter-jurisdictional competition (Brennan and Buchanan, 1980; Weingast, 1995; Montinola et al., 1995). Fiscal decentralization must tie local revenues to expenditures for interjurisdictional competition to be effective. Vertical transfers may create incentives for local officials to ignore competitive pressures for better management (Oates, 1999; Zhuravksaya, 2000; Jin et al., 2005). More generally, sub-central governments will tend to be more inefficient if they face soft rather than hard budget constraints, or, in other words, if they can count on the central government for future bail-outs (Qian and Rolands, 1998).

But fiscal decentralization may also lead to a fall in government quality. The creation of additional levels of regulatory and fiscal authority may lead, in the absence of coordination, to “overgrazing” by the multi-tier government structure (Shleifer and Vishny, 1993). Moreover, competition may reduce tax pressure and thus the capacity of local governments to collect sufficient taxes to provide basic public goods (Keen and Marchand, 1997; Oates, 1999). Local governments may also compete for capital by helping agents cheat the central government in ways that reduce the latter’s capacity to enforce regulations and collect taxes (Cai and Treisman, 2004). Government’s in regions which are uncompetitive for some structural reason not easily responsive to policy in the short term, may give up on business-friendly policies and dedicate themselves to predation instead (Cai and Treisman, 2005). Also, while assigning local governments a greater share of revenue can improve their incentives, the inevitable reduction of revenues in the hands of central government, may reduce its stake in economic development and, ultimately, the quality of its policies. (Treisman, 2006).

The empirical evidence on the impact of fiscal decentralization on government quality is overwhelmingly supportive of the benign effect of the former on the latter. Huther and Shah (1998) find that fiscal decentralization is associated with greater citizen participation, more political and democratic accountability, social justice, improved economic management and reduced corruption. However, their cross-section study suffers from omitted variable bias since they only provide simple correlations among these variables. De Mello and Barenstein (2001) control for GDP per capita and population and find that more fiscally decentralized countries tend to be less corrupt. Fisman and Gatti (2002a) find a similar relationship but controlling for more factors (GDP per capita, civil liberties, population and government share of expenditure or share of GDP). In another paper they find that a reliance on federal budgetary transfers in the US context is related to more corruption (Fisman and Gatti, 2000b). Estache and Sarbajit (1995) find that expenditure decentralization is related to more infrastructure spending at sub-central levels in developing countries and, to a lesser extent, in developed ones. Khaleghian (2003) and Robalino et al. (2001) – the latter applying fixed effects estimation – show that expenditure decentralization is positively related to improved health outcomes in low and middle-income countries. Dreher (2006) and Kyriacou and Roca (2009) also control for country-fixed effects and find that fiscal decentralization has a positive effect on institutional quality (which includes measures of corruption, bureaucratic quality and rule of law) and that this effect diminishes as countries become wealthier.

Alternatively Treisman (2002) finds expenditure decentralization to be positively related with infrastructure but negatively correlated with health and education outcomes. Treisman (2007) finds that more fiscal decentralization reduces perceived corruption, but if one controls for the percentage of Protestants in the population, the negative


impact of fiscal decentralization on corruption disappears. This finding does not apply to those panel estimations cited which apply fixed effects which, by definition, control for such time invariant (or slowly changing) factors. In a more recent piece, Treisman returns to the impact of fiscal decentralization on corruption by way of both cross-country and industry-level evidence. He finds that fiscal decentralization reduces corruption (even when controlling for religious affiliation) while more tiers of government increase it, a result attributed to “overgrazing” (Fan et al, 2009).

3. Political decentralization and government quality

Political decentralization may take different forms or guises (Treisman, 2002; Fan et al, 2009). One is electoral decentralization or, the extent to which sub-central politicians are elected rather than appointed. Another is decision-making decentralization which implies the allocation of political decision-making power in at least one policy area exclusively to sub-central levels of government. It’s worth noting that decision-making decentralization tends to be a standard feature of federal systems. Yet another form of political decentralization is the existence of a bicameral system with a regionally chosen upper house that can block lower house legislation.

Most theoretical contributions which have examined the impact of political decentralization on government quality have focused on the role of subnational elections. The possibility that voters may be better informed about local government activity together with their ability to punish bad performance at election time should improve government quality (Seabright, 1996; Tabellini, 2000). In decentralized settings voters can, moreover, use the performance of each jurisdiction as a benchmark and this should increase government efficiency (Salmon, 1987; Breton, 1996). On the other hand, local officials subject to an electoral constraint may be more susceptible to capture by local interests for several reasons. First, national office being more prestigious and powerful may be more intensely monitored (Prud'homme, 1995; Tabellini, 2000; Bardhan, 2002). Second, corruption is more likely simply because relations tend to be more personal in nature due to the greater intimacy and frequency of interaction at the local level (Tanzi, 1995; Prud'homme, 1995). Further, locally elected governments may indulge in rent-seeking behavior if more immobile factors such as farmers, workers or consumers are more numerous than mobile ones such as investors (Rodden and Rose-Ackerman, 1997). More generally, whether local governments are more vulnerable to capture than central governments depends on several factors which are ultimately linked to the degree of socio-economic inequality at central versus sub-central levels (Bardhan and Mookherjee, 2000, 2006).

One way to approach the possible influence of decision-making decentralization is to recognize that, in a sense, this dimension of political decentralization is conceptually similar to fiscal decentralization since it points to a sub-central government’s ownership of a policy area. Like the case of fiscal decentralization, decision making decentralization may generate benefits from the possibility that decentralized governments may be better informed about citizen preferences and can adapt policies to local needs. Costs may emerge insofar as the existence of multiple levels of decision-making authority makes it more difficult to assign accountability to any particular level (Gerring and Thacker, 2004 and Fan et al, 2009)2.

2 Accountability is more difficult if each government level’s responsibility is not clearly defined (WDR, 1999/2000). In his discussion of measurement issues related to decentralization, Rodden (2004) highlights


The impact of a regionally elected upper house on government quality has been analyzed in the more general context of the relationship between government quality and checks and balances. Panizza (2001) and Treisman (2002) have argued that the existence of multiple veto points – a bicameral system being just one example, - can offer protection against arbitrary and corrupt government but it may just as well prevent improvements in government performance by locking in the status quo.

What does the empirical evidence say? Insofar as electoral decentralization is concerned, Blanchard and Shleifer (2001) explain the differential performance of federalism in Russia and China based on whether regional governors in each setting where elected or appointed. They argue that elected governments in Russia have been captured by local interests and suggest the need for a degree of political centralization (the right of central government to appoint or fire regional governors or the creation of national parties exercising influence on regionally elected governors). Bardhan (2002) criticize the Blanchard and Shleifer (2001) story as incomplete. Uniquely in the Chinese case, the lack of strong rural lobbies (owing largely to an egalitarian land distribution) may have reduced capture there independently of whether regional governors are appointed or elected. Moreover, before recommending political centralization they need to provide a plausible story of a benevolent non-rentier central authority. Notwithstanding this critique, Enikolopov and Zhuravskaya (2007) use both cross-section and panel data from developing and transition countries and find that fiscal decentralization combined with one form of political centralization, namely, strong national parties, significantly improves government quality measured both in terms of government efficiency, regulatory quality, control of corruption and rule of law, and in terms of public good provision (health and education outcomes).

In a cross section study Treisman (2002) finds no statistically significant effects between electoral or decision-making decentralization and corruption. Acknowledging that what matters is not so much political decentralization in isolation but its interaction with the fiscal resources available to government, he interacts his political decentralization measures with fiscal decentralization so as to examine if the former influence the latter’s relationship with government quality. Again the results are not statistically significant. He does find that bicameralism does reduce public good provision, especially in poorer countries. Alternatively, Panizza (2001) finds that the presence of more checks and balances increases government quality in a cross section of countries. A range of cross-section studies have tried to calibrate the statistical relationship between federalism and corruption. Goldsmith (1999), Treisman (2000), Gerring and Thacker (2004) and Kunicová and Rose-Ackerman (2005) all found that federal structure was associated with more corruption. However, in a more recent review of the empirical literature, Treisman (2007) suggested that the federal effect disappeared as the number of countries in the sample increased. The insignificant impact of federalism is also reported by Fisman and Gatti (2002a) and Bohara et al (2004).

The later application by Treisman of industry level data to the relationship between political decentralization and government quality reaffirms this author’s previous findings: neither federalism nor decision-making nor electoral decentralization have a statistically significant effect on corruption (Fan et al, 2009). On the other hand, the number of tiers of government – which, recall, they take to test the presence of the prevalence of shared authority in selected policy areas; only rarely is a single level of government solely responsible. Consequently, “[w]hen decentralization amounts to adding layers of government and expanding areas of shared responsibility, it might facilitate blame shifting or credit-claiming, thus reducing accountability” (494).


overgrazing but which could be useful in testing the possible blurring of accountability in multi-tier settings (see Rodden, 2004)3 – is robustly correlated with more corruption.

However, one important shortcoming of this analysis as well the others cited measuring the impact of federalism on corruption, is that at no time do they interact the political decentralization variables with fiscal decentralization. We agree with Treisman (2002) that it is important to consider how political decentralization interacts with the fiscal resources available to government. This would allow examination of whether the former are in any way mediating the relationship between the latter and government quality. This is done by Kyriacou and Roca (2009) who interact fiscal decentralization with decision-making and electoral decentralization and find that these forms of political decentralization reduce the positive effect of fiscal decentralization on government quality (measured through corruption, rule of law and bureaucratic quality) in poor and medium income countries.

4. A role for maturity?

Several empirical studies have detected the possible influence of democratic maturity on corruption. In his cross-section analysis of the determinants of corruption Treisman (2000) found that the current degree of democracy was not significant in explaining corruption but a long period of exposure to democracy was. He finds that “countries with at least 40 years of consecutive democracy behind them enjoyed a significant, though small, corruption dividend, and those with 20–30 years may also have benefited slightly.” (439). Both Gerring and Thacker (2004) and Bohara et al (2004) confirm the corruption reducing effect of long-term democracy.

Sandholtz and Koetzle (2000) find that both the current level of democracy and experience with democratic rule correlate negatively with perceived levels of corruption. Of particular interest is the justification offered by these authors as to why exposure to democracy may have this effect. The beneficial effects attributed to democracy depend on the extent to which democratic norms and practices have taken root among citizens. These norms are transmitted from one generation to the next through a process of socialization. The longer a country has experienced democratic rule, the more deeply rooted the associated norms and values should be since it implies that democratic norms have diffused more extensively and intensively.

A different interpretation of how maturity or experience may affect government quality has been advanced by Bockstette et al (2002). These authors argue that a longer history of statehood – in the period between 1 and 1950 CE – may improve organizational effectiveness for several reasons: learning by doing in public administration or the development of attitudes consistent with bureaucratic discipline and hierarchical control. While the aim of their paper is to fathom the effect of such experience on economic growth, they also present simple correlations between their indicator (called State Antiquity) and several government quality measures and find a positive and statistically significant correlation.

In light of this literature, an important question worth exploring is whether the effect of political decentralization on the relationship between fiscal decentralization and

3 Indeed, the definition of tiers includes both fiscal and decision-making decentralization since a “tier of government” exists if a state executive body at that level (1) was funded from the public budget, (2) had authority to administer a range of public services, and (3) had a territorial jurisdiction.


government quality may be potentially driven by the level of democratic maturity or state antiquity in a country. Kyriacou and Roca (2009) suggest, but don’t explore, the possibility that the differential impact of political decentralization in rich and poor countries may be driven by underdeveloped political institutions in the former which are unable to harness the positive effects expected from electoral and decision-making decentralization. This echoes Bardhan’s (2002) warning that electoral decentralization in democratically immature countries may not work as expected. Similarly, the World Development Report (WDR, 1999/2000) warns that while in theory fiscal and political decentralization may improve government performance, in practice in countries which have not been exposed to democracy over a long period it’s not realistic to assume that people, in a decentralized setting, will make their voices heard through the political process. Or, if state antiquity is relevant, then political decentralization may work better in countries with a long history of state institutions.

The review of both the theoretical and empirical work points to the utility of applying a new data set so as to examine the impact of fiscal and political decentralization taking into account the role of maturity. While the theoretical literature on the impact of fiscal decentralization on government quality is ambiguous, the empirical evidence is overwhelmingly suggestive of a positive relationship. The same ambiguity exists with regards to the theoretical priors on the impact of political decentralization on government quality and this ambiguity extends to the empirical evidence which reports either a positive or a negative relationship or no such relationship at all. Thus, further empirical evidence in this area seems useful. Finally, an important question which emerges from our literature review is the following: does democratic maturity or state antiquity affect the possible influence of political decentralization on the relationship between fiscal decentralization and government quality? In the next two sections, we review the data and empirical methodology which will be used to address these issues.

5. The Data

We measure government quality by way of the World Bank Government Indicators explained by Kaufmann et al (2006). In particular we focus on four aspects of government quality: control of corruption, rule of law, regulatory quality and government effectiveness (see appendix 1 for full definitions of these and all the variables mentioned in this paper). These indicators are based on several hundred individual variables from thirty-one data sources measuring perceptions of governance. They are weighted averages of the underlying data, with weights reflecting the precision of the individual data sources. These data are available from 1996 to 2006 and for more than two hundred countries and territories, something which makes them attractive for both cross-section and panel estimations. Since we are interested on the impact of decentralization on overall government quality, our dependent variable is the average of all the just-mentioned four dimensions. This reduces the influence of idiosyncratic ratings stemming from any one dimension but it does not allow us to explain the behavior of each separate dimension.

Fiscal decentralization refers to the distribution of tax revenues and expenditures among the different tiers of government. A system is more fiscally decentralized the greater the proportion of tax-revenues and expenditures “owned” by lower tiers of government. The most widely used measure of fiscal decentralization in the literature is provided by the IMF’s Government Finance Statistics (GFS) and refers to expenditure or


revenue share of sub-central governments in total public expenditures or revenues, or as a proportion of GDP. In this paper we define fiscal decentralization as revenue share of sub-central governments in total public revenues. We are well aware that this definition will tend to overestimate the degree of fiscal decentralization since it does not reflect on whether sub-central governments, own these resources (Rodden, 2004; World Bank, 2004; Boex and Simatupang, 2008 and Kyriacou and Roca, 2009)4. Despite its shortcomings, we adopt this indicator since it is available for the widest coverage of countries and time periods. It gives rise to a cross section sample of up to 101 countries and an unbalanced panel of 58 countries over the period 1998-2006 (see appendix 2 for the list of countries).

Since political decentralization takes different forms we measure it in several ways. We capture electoral decentralization via a dummy variable (Reg. Elections) generated for the year 1999 by the WDR (1999/2000) which takes the value of 1 if a county hold regional elections and zero otherwise. We also use Schneider’s (2003) measure of political decentralization which is based on the existence of elections at the municipal or state/provincial levels in 1996 (we call it Schneider’s Decentralization). He uses confirmatory factor analysis to generate an index between 0 and 1 with higher values representing more decentralization. We follow Fan et al (2009) in measuring decision-making decentralization based on a dummy variable (Autonomy) which takes the value of 1 if the country’s constitution assigned at least one policy area exclusively to subnational governments and/or gave such governments exclusive authority to legislate on matters not constitutionally assigned to one level. We recognize with these authors that this variable is less than ideal since informal behavior often diverges from what is written in the constitution and the number and relative importance of policy areas constitutionally assigned to sub national governments differs (and there is no obvious way to add these up). On the other hand, trying to capture these aspects through a numerical index would inevitably increase the subjectivity of the whole exercise. Since decision-making decentralization is a feature of federal systems, we also measure it through a dummy (Federal) which takes the value of 1 if the country can be described as federal and 0 otherwise following Elazar (1995). To calibrate the influence of bicameralism, we follow Treisman (2002) and use a dummy (Bicameral) to capture whether country has a regional chosen upper house that could block lower house financial legislation. Finally, we also employ Gerring and Thacker’s (2004) index (Unitarism) since it combines both decision-making decentralization and checks and balances. In particular the index ranges from 1 to 5 where 5 indicates a non-federal country with a unicameral legislature.

A look at the simple pairwise correlations between the different political decentralization variables we employ reveals a strong positive correlation between our Autonomy and Federalism variables which is as expected (see the correlation matrices for both the cross-section and panel samples in appendix 3). We can also identify a strong negative relationship between the Unitarism index and the Federalism, Autonomy and Bicameralism variables which is as expected since the latter variables are defined inversely to the former. But otherwise, the different variables are not strongly correlated thereby confirming our intuition that they capture different aspects of political 4 For example, revenues received from grants by the central government appear as sub-central revenues regardless of whether they are conditional or unconditional. Moreover, no distinction is made as to whether these revenues are collected through shared taxes, piggybacked taxes, or locally determined "own-source" taxes. Similarly, no consideration is given to whether the central government has the power to set the rate and the base, leaving sub-national governments just to collect revenues.


decentralization. Moreover, Schneider’s (2003) electoral decentralization variable differs markedly from that proposed by WDR (1999/2000) and as such can serve as a robustness check on the impact of this type of political decentralization on government quality.

We follow Treisman (2000) in measuring democratic maturity in the cross-section analysis via a dummy which takes the value of 1 if a country experienced uninterrupted democracy between 1950 and 1995 (we label it Democ. Maturity DV). Following Persson et al (2003) we also measure democratic maturity by way of a continuous variable called Age of Democracy which captures the number of years of uninterrupted democracy since 1800. In line with Bockstette et al (2002), the State Antiquity variable ranges between 0 and 1 and captures 3 dimensions over the period 1 to 1950 CE: the existence of government above the tribal level; whether this government is foreign or locally based and; the portion of territory of the modern country which is ruled by this government. These authors experiment with different discount rates to reduce the weight of periods in the more remote past and we follow them in choosing an intermediate rate for our empirical analysis. While Democratic Maturity and Age of Democracy are quite correlated, neither of these is strongly correlated with the State Antiquity indicator something which supports the idea that these variables represent different aspects of maturity.

6. Empirical Methodology

We analyze the effect of decentralization on government quality taking two distinct approaches. First, we explore the determinants of cross-sectional variation in the government quality across countries using the sample of 101 countries. Second, we study the relationship between government quality and decentralization by focusing on the within-country time variation as captured through panel fixed effects regressions based on an unbalanced panel that incorporates information from 58 countries for the period 1998-2006 (for information on the within-country time variation in our sample, see appendix 4).

For the purposes of cross-sectional analysis, we use the following general regression model:

Government Qualityi = α1 + 1 Decentralizationi + 2 Xi +εi (1)

where i refers to countries, α1 is the constant, Xi is the vector of control variables and εi is the error term.

Consistent with previous cross-section studies, our control variables include: logarithm of real GDP per capita, size of government, logarithm of population, religious or ethno-linguistic fractionalization, openness, legal origin, religious affiliation, party age and government fractionalization (both capture the degree of political centralization), democracy (also squared) and finally, number of tiers.

For the time variant variables we take the mean of the available annual data over the period 1998-2006, except in the case of fiscal decentralization where we average available data over the period 1973–2000 so as to maximize our cross section sample. Our democracy indicator is Voice and Accountability and comes from the World Governance Indicators. We agree with Treisman (2007) that a problem when you use


similar variables from the same source on both the RHS and LHS is that they are likely to be strongly correlated and this actually turns out to be the case. But the use of this indicator has two advantages. First, it allows us to maximize our sample size. Second, to the extent that we include Voice and Accountability and the other RHS variables of interest remain significant, we would have some evidence as to their robustness.

Our empirical strategy for the panel data is based on cross country fixed effects estimation. Bringing fixed effects estimation to bear on the relationship between decentralization and government quality has one important advantage over cross-section estimations exploring this relationship namely, the ability to control for cross-country heterogeneity due to time constant (or slowly changing) observable or unobserved factors (cross-section fixed effects). Time constant factors include those that we control for in the cross-section analysis and, more importantly, other non-observable ones which may be related to decentralization and government quality. The exclusion of these variables would induce omitted variable bias, affecting both the estimated impact of decentralization and its statistical significance. Indeed, Bardhan and Mookherjee (2006) identify this form of bias as one of the important methodological problems with existing cross-section studies estimating the effect of decentralization on governance. Our fixed effect estimation reduces the likelihood of omitted variable bias since it eliminates the possible influence coming from both observable and unobservable time constant variables5.

We therefore estimate the following equation:

Government Qualityit = αi + 1 Decentralizationit + 2 Xit +εit (2)

where i refers to countries and t to years, αi are country fixed effects, Xit is the vector of control variables and εit is the error term.

We include the following time variant control variables in the panel regressions: logarithm of real GDP per capita, size of government, logarithm of population, openness, democracy, party age and government fractionalization. The panel estimations employ the available data over the period 1998-2006.

The related literature alerts us to the problem of endogeneity in the sense that government quality may be affected by decentralization but decentralization may also be influenced by the quality of government – for example, corrupt officials at the central level may oppose fiscal decentralization because it may limit their capacity to extract rents (see, for example, Fisman and Gatti, 2002a and Treisman, 2002). Similar arguments have been made for other economic variables used to explain government quality (La Porta et al, 1999; Acemoglu et al, 2001).

The presence of endogeneity has generally been dealt with, in both cross-section and panel analyses, through the use of instrumental variables. We follow Enikolopov and Zhuravskaya (2007) and estimate equation (1) by two-stage least squares with the geographical area of countries (in logarithms) used as an instrument for fiscal decentralization. To be a good instrument, geographical area must satisfy three

5 Given that our key variables exhibit variability over time, the fixed effects method can be used. We favor fixed effects over random effects for several reasons: the latter does not allow us to deal with omitted variable bias in a similar way; the latter implies that our sample is a random one from a large population something which is not the case, since our cross-section units are not exchangeable (Hsiao, 2003). Moreover, the relevant statistical tests indicate that cross-section fixed effects are not redundant while period-fixed effects are.


conditions (Murray, 2006). First, it must be uncorrelated with the error term which in the context of our estimation means geographic area should not itself be affected by government quality (no reverse causality); our government quality indicators range over the period 1998-2006 and we assume that this is too short a period for government performance to have had an impact on the geographic area of the countries in our cross-section sample. Second, geographic area should be correlated with fiscal decentralization. In line with this, several multivariate analyses of the determinants of fiscal decentralization have reported a positive significant impact coming from geographic area (Panizza, 1999; Arzaghi and Henderson, 2005). Moreover, the F-statistics from our first stage regressions of fiscal decentralization on geographic area and all our explanatory variables indicate that geographic area is a sufficiently strong instrument (see appendix 5). Third, the only impact of geographic area on government quality must pass through decentralization (the exclusion restriction). This is suggested by the fact that including fiscal decentralization and geographic area in the same regressions yields no statistically significant effect of the latter on government quality (see also Treisman, 2002).

Like Enikolopov and Zhuravskaya (2007), we deal with endogeneity in equation (2), by instrumenting our endogenous variables with their values lagged one period and estimating via two-stage least squares. The use of lagged values as instruments is supported by the absence of serial correlation in the residuals of the estimated equations – the presence of serial correlation would violate the requirement of exogeneity for valid instruments (Arellano, 2002).

7. Empirical Results The variables included in the regressions reported in this section are those which are statistically significant or whose exclusion seriously changes the results. Insofar as the control variables are concerned in the tables that follow, GDP per capita and Voice and Accountability are positively related with the quality of government while Population and a Socialist legal origin are negatively related. These results are fully in line with previous regression results.

Table 1 presents estimates of the regression of our government quality indicator on our indicators of fiscal and political decentralization. What is immediately obvious from this table is the positive and statistically significant impact of fiscal decentralization on government quality. In fact, this tends to be the case for all the regressions conducted and reported in this section. And this is consistent with the lion’s share of the empirical literature previously reviewed. Columns 2 and 3 of table 1 add Gerring and Thacker’s (2004) Unitarism indicator and Schneider’s (2003) electoral decentralization indicator to the cross-section base regression in column 1. The former is positively and statistically significantly related to government quality while the latter implies that a greater degree of electoral decentralization reduces government quality. Taken together, these results are suggestive of the negative impact of political decentralization on government quality. These results are clarified by the interaction between fiscal and political decentralization, the latter captured by way of the dummy variables, Federal, Autonomy, Regional Election and Bicameralism (regressions 4 to 7 respectively). Now, the negative effect of political decentralization passes through its effect on the relationship between fiscal decentralization and government quality: in all cases, political decentralization (whether in terms of electoral or decision-making


decentralization or checks and balances) mitigates the positive impact of fiscal decentralization on government quality.


Table 1. Fiscal and Political Decentralization and Government Quality. Cross Section and Panel estimates.

(1) CS (2) CS

(3) CS

(4) CS

Federal

(5) CS

Autonomy

(6) CS

Reg. Elec.

(7) CS

Bicameral

(8) PANEL

(9) PANEL

Federal

(10) PANEL

Autonomy

(11) PANEL

Reg. Elec.

(12) PANEL

Bicameral

Constant

-2,535

(0,63)***

-2.924

(0,73)***

-3,503

(0.87)***

-2.315

(1.15)**

-2.668

(1,02)**

-1.319 (1.71)

-2.548 (1.23)**

11,204

(3,34)***

11,410

(3,25)***

11,534

(3,24)***

15,745

(3,18)***

11,267

(3,33)***

Fiscal Decentralization

1,390

(0,75)*

2,083

(0.76)***

2,225

(0.71)***

5.959

(2,17)***

5.070

(1.64)***

8.114 (3.11)**

6.954 (2.49)***

0,567

(0,28)**

1,007

(0,37)***

1,054

(0,37)***

0,925

(0,38)**

0,698

(0,42)*

GDP per capita

0,334

(0,06)***

0.292

(0,07)***

0.411

(0,09)***

0.341

(0,11)***

0.359

(0,10)***

0.171

(0,18)

0.331

(0,12)***

0,482

(0,09)***

0,545

(0,09)***

0,551

(0,09)***

0,527

(0,09)***

0,489

(0,09)***

Population

-0,003

(0,03)

-0,011

(0,03)

-0,014

(0,03)

-0,096

(0,06)*

-0,072

(0,05)

-0,079

(0,08)

-0,078

(0,05)

-1,646

(0,38)***

-1,735

(0,37)***

-1,754

(0,37)***

-2,100

(0,35)***

-1,660

(0,38)***

Voice and accountability

0,567

(0,07)***

0,703

(0,08)***

0,748

(0,11)***

0,668

(0,13)***

0,647

(0,12)***

0,826

(0,21)***

0,649

(0,13)***

0,128

(0,05)**

0,129

(0,05)***

0,128

(0,16)***

0,154

(0,05)***

0,129

(0,05)**

Socialist

Legal Origin

-0,511

(0,11)***

-0,653

(0,11)***

-0,361

(0,11)***

-1,041

(0,27)***

-0,882

(0,20)***

-1,061

(0,31)***

-0,995

(0,25)***

German

Legal Origin

-0,101

(0,19)

-0,204

(0,22)

-0,174

(0,21)

-0,510

(0,48)

-0,395

(0,42)

0,088

(0,23)

-0,088

(0,18)

French

Legal Origin

-0,192

(0,10)*

-0,184

(0,10)*

-0,026

(0,12)

-0,378

(0,15)**

-0,270

(0,16)

-0,335

(0,18)*

-0,330

(0,14)**

Scandinavian

Legal Origin

0,055

(0,16)

-0,221

(0,18)

-0,098

(0,19)

-1,174

(0,52)**

-0,914

(0,40)**

-0,671

(0,53)

-1,069

(0,52)*

Unitarism

0.107

(0,04)***

Schneider’s Decentralization

-0.510

(0,19)***

Pol. Decentralization *

Fisc. Decentralization

-5.766

(2,91)**

-4.571

(1,92)**

-8.665

(3,33)**

-6.455

(2,59)**

-0.890

(0,53)*

-0.934

(0,53)*

-0.631

(0,48)

-0.182

(0,47)

Pol. Decentralization

0.697

(0,62)

0.480

(0,33)

1.106

(0,38)***

0.707

(0,31)**

Adjusted R2 0.86 0.85 0.90 0.65 0.70 0.40 0.60 0.99 0.99 0.99 0.99 0.99

Observations 99 90 64 98 85 85 99 365 365 365 365 365

All estimations are two-stage least squares. In the CS estimates Fiscal Decentralization is instrumented through geographical area (in logs). All regressions report White Heteroskedasticity-Robust Standard Errors in parentheses. *, **, *** measures statistical significance at the 10, 5 and 1% levels respectively. Legal Origin and Political Decentralization variables (Federal, Autonomy, Reg. Elections and Bicameral) are dummy variables.


Our cross-section results are confirmed by the fixed effects estimations also reported in table 1 (columns 8 to 12). The presence of political decentralization in the guise of Autonomy or Federalism tends to mitigate the positive incidence of fiscal decentralization on government quality. The panel regressions also moderate the economic impact of fiscal decentralization on government quality reported in our cross section estimates. Focusing on our base cross-section regression in column 1 of table 1 and based on the summary statistics (presented in appendix 6), our results indicate that a one standard deviation increase in the degree of fiscal decentralization increases our government quality indicator by 0.182 points. By way of comparison, consider the panel regression reported in column 8 of table 1. Now, a one standard deviation increase in fiscal decentralization increases our government quality indicator by only 0.072 or in other words only 40% of the increase reported by our cross section estimates. The possible existence of time invariant omitted variables in our cross section estimations is likely to be biasing upwards the estimated economic impact of fiscal decentralization on government quality.

Next, we turn to the possible impact of maturity on the relationship between fiscal and political decentralization and government quality. In table 2 we report cross-section estimates which include two indicators of democratic maturity in the form of the continuous variable Age of Democracy (columns 1 to 4) and the dummy Democratic Maturity (columns 5 to 8). In none of these regressions does democratic maturity emerge as statistically significant. More importantly, the inclusion of democratic maturity has no affect on the economic or the statistical significance of political decentralization. Political decentralization continues to exercise a negative influence on the relationship between fiscal decentralization and government quality. These results are confirmed when we measure democratic maturity by way of a third continuous variable proposed by Gerring and Thacker (2004) which captures the total number of years democratic since 1900 (results not shown in the table).

Things turn out very differently when we include State Antiquity in our cross-section regressions (columns 9 to 12). State Antiquity has a positive and statistically significant impact on government quality in two out of the four regressions (columns 11 and 12). Moreover, the inclusion of State Antiquity means that the interaction of fiscal decentralization with all our political decentralization variables ceases to be statistically significant. Taken together, the results in table 2 indicate that it may be state antiquity rather than democratic maturity that exerts a positive influence on government quality either directly or indirectly through its effect on the impact of political decentralization on government quality. In other words, what may matter for government quality is a very long run experience with public administration or statehood rather than exposure to democracy.


Table 2. Cross section estimates: Decentralization, Democratic Maturity and State Antiquity.

(1) Federal

(2) Autonomy

(3) Reg. Elec.

(4) Bicameral

(5) Federal

(6) Autonomy

(7) Reg. Elec.

(8) Bicameral

(9) Federal

(10) Autonomy

(11) Reg. Elec.

(12) Bicameral

Constant

-2,364 (1,06)**

-2,690 (0,94)***

-1,422 (1,58)

-2,564 (1,14)**

-2,274 (1,08)**

-2,789 (0,87)***

-1,504 (1,60)

-2,683 (1,11)**

-0,828 (1,86)

-1,031 (2,01)

-0,448 (1,90)

-0,282 (2,00)

Fiscal Decentralization

5,257 (2,21)**

4,474 (1,79)**

7,472 (3,04)**

6,298 (2,43)**

5,433 (2,13)**

3,962 (1,71)**

7,001 (3,09**

5,732 (2,39)**

6,987 (3,61)*

6,866 (3,34)**

6,430 (2,98)**

8,642 (3,85)**

GDP per capita

0,335 (0,11)***

0,350 (0,10)***

0,167 (0,17)

0,320 (0,12)***

0,327 (0,11)***

0,354 (0,09)***

0,172 (0,17)

0,332 (0,12)***

0,238 (0,17)

0,251 (0,18)

0,163 (0,18)

0,152 (0,18)

Population

-0,082 (0,05)

-0,058 (0,05)

-0,063 (0,08)

-0,065 (0,04)

-0,093 (0,05)*

-0,051 (0,04)

0,058 (0,08)

-0,061 (0,04)

-0,201 (0,10)*

-0,197 (0,10)*

-0,179 (0,11)

-0,218 (0,11)**


0,640 (0,14)***

0,618 (0,13)***

0,795 (0,21)***

0,611 (0,13)***

0,601 (0,15)***

0,567 (0,13)***

0,771 (0,21)***

0,579 (0,13)***

0,613 (0,18)***

0,610 (0,18)***

0,685 (0,21)***

0,640 (0,19)***

Socialist Legal Origin

-0,939 (0,31)***

-0,796 (0,25)***

-0,956 (0,36)***

-0,881 (0,27)***

-0,870 (0,29)***

-0,671 (0,23)**

-0,870 (0,35)**

-0,781 (0,27)***

-1,143 (0,98)

-1,079 (0,94)

-1,036 (0,94)

-1,586 (1,32)

German Legal Origin

-0,386 (0,47)

-0,272 (0,42)

0,191 (0,26)

0,035 (0,20)

-0,530 (0,39)

-0,340 (0,30)

0,159 (0,22)

-0,016 (0,16)

-0,845 (0,55)

-0,895 (0,54)

-0,533 (0,44)

-0,741 (0,44)*

French Legal Origin

-0,360 (0,15)**

-0,262 (0,16)*

-0,310 (0,17)*

-0,308 (0,13)**

-0,293 (0,16)*

-0,191 (0,16)

-0,255 (0,18)

-0,248 (0,14)*

-0,283 (0,17)

-0,174 (0,20)

-0,271 (0,16)

-0,259 (0,18)

Scandinavian Legal Origin

-1,017 (0,54)*

-0,786 (0,44)*

-0,561 (0,54)

-0,942 (0,51)*

-1,490 (0,57)**

-1,103 (0,41)***

-0,621 (0,48)

-1,079 (0,47)**

-1,376 (0,82)*

-1,321 (0,75)*

-0,828 (0,44)*

-1,474 (0,79)*

Pol. Decentralization * Fisc. Decentralization

-5.529 (2,88)*

-4.491 (1,96)**

-8.399 (3,22)**

-6.294 (2,46)**

-7.329 (2,93)**

-5.648 (1,99)***

-8.189 (3,01)***

-6,447 (2,35)***

-4.724 (4,54)

-3.734 (3,99)

-4.383 (3,43)

-5.464 (3,39)

Pol. Decentralization

0.698 (0,60)

0.491 (0,32)

1.061 (0,37)***

0.692 (0,29)**

0.953 (0,55)*

0.622 (0,31)*

1.025 (0,35)***

0.705 (0,28)**

0.523 (0,81)

0.302 (0,55)

0.538 (0,37)

0.621 (0,32)*

Age of Democarcy

0,190 (0,32)

0,213 (0,31)

0,261 (0,35)

0,326 (0,28)

Democ. Maturity DV *

Fisc. Decentralization

2.006 (1,57)

2.194 (1,55)

0.609 (1,73)

1.328 (1,52)

Democ. Maturity DV

0.076 (0,22)

0.010 (0,25)

0.138 (0,27)

0.078 (0,23)

State Antiquity

0,623 (0,44)

0,645 (0,52)

0,788 (0,48)*

0,922 (0,47)*

Adjusted R2 0.70 0.74 0.47 0.65 0.68 0.77 0.52 0.69 0.51 0.49 0.58 0.35

Observations 98 85 85 99 98 85 85 99 74 63 66 75

All estimations are two-stage least squares. Fiscal Decentralization is instrumented through geographical area (in logs). All regressions report White Heteroskedasticity-Robust Standard Errors in parentheses. *, **, *** measures statistical significance at the 10, 5 and 1% levels respectively. “Democ. Maturity” and “Federal”, “Autonomy”, “Reg. Elections” and “Bicameral” are dummy variables.


8. Discussion

At the outset, it should be emphasized that our empirical analysis is necessarily limited by the quality of our data. As previously stated, our fiscal decentralization data may over-estimate the degree of fiscal decentralization since they do not really capture the crucial issue of ownership of fiscal resources by sub-central governments. We are aware of the existence of more precise fiscal data generated by the OECD but unfortunately they only cover OECD countries. In addition, we acknowledge that while our binary definitions of the different dimensions of political decentralization may be useful, they do not capture other aspects which may be equally important but which we are unable to explore due to the unavailability of data. These include: the extent to which the candidates in local (national) elections are chosen by central (sub-central) level politicians (Rodden, 2004); the nature of sub-central electoral rules or institutions (size of electoral districts, party lists, timing of sub-central and national elections, whether the system of representation is majoritarian or proportional); the presence of checks and balances at more decentralized levels (WDR, 1999/2000); the possibility that sub-central politicians in immature democracies may be elected on the basis of tradition rather than performance (Tendler, 1997) and; the relative socio-economic inequality between these levels (Bardhan and Mookherjee, 2006).

Notwithstanding these data limitations several points are in order. Our results contradict previous findings cited in section 4 above which report a positive and statistically significant impact of democratic maturity on government quality. One reason for this difference is that previous studies tend to control for democratic maturity but not for the current level of democracy in the same regression. The exception is Sandholtz and Koetzle (2000) and the difference here may be due to the use of different cross section samples. Our study employs a cross section of up to 99 countries while their cross section contains 50 countries. More to the point, almost all of the countries that we include and they exclude from their sample are ones which can be characterized as democratically immature and so their exclusion is very likely to have a strong bearing on the estimated relationship between government quality and democratic maturity. To check for the possibility that our democratic maturity indicators are not significant because of the use of Voice and Accountability as a measure of the current level of democracy (remember it is strongly correlated with government quality) we do two things. We drop Voice and Accountability and re-estimate our regressions with democratic maturity but again don’t find it to be significant. Second, we include democratic maturity with other current democracy variables available in the literature and again the former is not significant.

The policy implication which emerges here is that combining fiscal decentralization with political decentralization may be bad for government quality. But, of course, governments may decentralize for other reasons (Ebel and Yilmaz, 2002). Throughout post-communist Central and Eastern Europe, decentralization is intimately linked to the transition from a socialist system to a market economy and democracy. In Latin America, the origin of decentralization is political pressure from the people for democratization. All over the world, decentralization may be an important instrument for diffusing secessionist tendencies (WDR, 1999/2000). Our empirical evidence suggests that fiscal decentralization can improve government quality but not so if it is combined with political decentralization (as we measure it in this paper). If political decentralization is pursued to fulfill some other objective then, from the perspective of


government quality, it is likely to be more “affordable” in countries with a very long experience in public administration or statehood.

9. Conclusion

In this paper we applied both cross-section and panel analysis to the relationship between fiscal and political decentralization and government quality as measured by the World Bank’s World Governance Indicators. Our findings support the lion’s share of the empirical literature which finds that fiscal decentralization has a positive impact on government quality. We further find that political decentralization, either in the form of sub-national elections, checks and balances (bicameralism) and especially decision-making decentralization (federalism or autonomy), tends to mitigate the positive impact of fiscal decentralization on the quality of government. It could be that local officials endowed with fiscal resources and subject to an electoral constraint may be prone to capture by local interests. Moreover, the existence of a regionally elected upper house with the power to block the lower house’s financial legislation may be preventing improvements in government performance which would normally arise from fiscal decentralization. Finally, decision-making decentralization may make it more difficult to assign accountability between central and sub-central government levels in fiscally decentralized settings. This suggests that those aiming to improve government quality should avoid decentralizing both fiscally and politically.

Our estimates also indicate that a very long tradition in statehood or public administration may neutralize the negative impact exerted by political decentralization on fiscal decentralization’s positive contribution towards government quality. Alternatively, whether a country can be characterized as a mature democracy had no comparable effect. This would suggest that if political decentralization is undertaken to meet objectives distinct from government quality, it is likely to be less costly, in terms of government quality, in countries with a long history of statehood.

Acknowledgments

Kyriacou would like to acknowledge financial assistance from the project SEJ2007-62500 (Ministerio de Ciencia y Tecnología) as well as a research grant from the Institut d’Estudis Autonòmics (Generalitat de Catalunya). Roca-Sagalés benefited from projects SEJ2007-67911 (Ministerio de Ciencia y Tecnología) and 2005SGR-117 (Direcció General de Recerca, Generalitat de Catalunya).


Appendix 1. Data Description

Quality of government: Average of: 1) control of corruption (the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as capture of the state by elites and private interests); (2) rule of law (the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, the police, and the courts, as well as the likelihood of crime and violence); (3) regulatory quality (the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development) and; (4) government effectiveness (the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government’s commitment to such policies). Source: World Governance Indicators, World Bank at http://info.worldbank.org/governance/wgi/index.asp.

Fiscal decentralization: Revenue share of sub-central governments in total public revenues. Source: Government Finance Statistics, IMF.

Electoral decentralization: Regional Elections is a dummy variable generated for the year 1999 which takes the value of 1 if a country holds regional elections and zero otherwise. Source: WDR (1999/2000). Schneider Decentralization captures the existence of elections at the municipal or state/provincial levels measured by an index between 0 and 1 with higher values representing more decentralization. Source: Schneider (2003).

Decision-making decentralization: Autonomy is a dummy variable which takes the value of 1 if the country’s constitution assigned at least one policy area exclusively to subnational governments and/or gave such governments exclusive authority to legislate on matters not constitutionally assigned to one level. Source: Fan et al (2009). Federal is a dummy which takes the value of 1 if the country can be described as federal and 0 otherwise. Source: Elazar (1995).

Checks and balances: Bicameral is s dummy variable which is 0 if the state was not bicameral with a regionally-chosen upper house or if the upper house did not have an absolute veto over the lower house (i.e., a veto that could not be overridden even by a supermajority). Source: Treisman (2002).

Unitarism: An index ranging from 1 to 5 which adds up the following coding categories from federalism and bicameralism and then reverses the scale. Federalism has 3 coding categories: 1, non-federal; 2, semi-federal (where there are elective legislatures at the regional level but in which constitutional sovereignty is still reserved to the national government); and 3, federal (elective regional legislatures plus constitutional recognition of subnational authority). Bicameralism similarly has 3 categories, according to the predicted degree of asymmetry and incongruence: 0, unicameral (no or weak upper house); 1, weak bicameral (upper house has some effective veto power, though not necessarily a formal veto); and 2, strong bicameral (same as above but the two houses are also incongruent). Source: Gerring and Thacker (2004).

Democratic maturity: Democratic Maturity DV is a dummy variable which takes the value of 1 if a country experienced uninterrupted democracy from 1950 to 1995. Source: Treisman (2000). Age of Democracy is defined as (2006-DEM_AGE)/206 and varies between 0 and 1, with the US being the oldest democracy (value of 1). DEM_AGE is the


year of birth of the democracy and corresponds to the first year of an uninterrupted string of positive yearly POLITY values (from Polity IV) until the end of the sample, given that the country is also an independent nation (foreign occupation during WWII not considered an interruption of democracy). Source: Persson et al. (2003).

State Antiquity is an indicator ranging from 0 to 1 with higher values indicating higher state or institutional antiquity. It captures 3 dimensions over the period 1 to 1950 CE: the existence of government above the tribal level; whether this government is foreign or locally based and; the portion of territory of the modern country which is ruled by this government. Source: Bockstette et al (2002).

GDP per capita: Real Gross Domestic Product (RGDP) per capita. Source: World Penn Tables at http://pwt.econ.upenn.edu/.

Size of government: Government Share of Real GDP. Source: World Penn Tables at http://pwt.econ.upenn.edu/.

Population: Total Population is based on the de facto definition of population, which counts all residents regardless of legal status or citizenship — except for refugees not permanently settled in the country of asylum, who are generally considered part of the population of their country of origin. Source: World Penn Tables at http://pwt.econ.upenn.edu/.

Openness: Exports plus Imports divided by Real GDP. Source: World Penn Tables at http://pwt.econ.upenn.edu/.

Fractionalization: the probability that two randomly selected individuals belong to different ethno-linguistic or religious groups and so increases with the number of groups. Source: Alesina et al (2003).

Legal origin: Dummy variable which identifies the legal origin of the company law or commercial code of each country. There are five possible origins; (1) English Common Law; (2) French Commercial Code; (3) German Commercial Code; (4) Scandinavian Commercial Code; (5) Socialist Communist Laws. Source: La Porta el al (1999).

Religious affiliation: Identifies the percentage of the population of each country that belongs to the three most widely spread religions in the world in 1980. For countries of recent formation, the data is available from 1990-1995. The numbers are in percent (scaled from 0 to 100). The three religious identified are: (1) Roman Catholic; (2) Protestant and; (3) Muslim. The residual are “Other Religions”. Source: La Porta el al (1999).

Party age: This is the average of the age of the 1st and 2nd government parties and the 1st opposition party, or the subset of these for which age of the party is known. Source: Beck et al. (2001).

Government fractionalization: The probability that two deputies picked at random from among the government parties will be of different parties. Source: Beck et al. (2001).

Voice and accountability: the extent to which a country’s citizens are able to participate in selecting their government, as well as freedom of expression, freedom of association,


and free media. Source: World Governance Indicators, World Bank at http://info.worldbank.org/governance/wgi/index.asp.

Nº of tiers: Number of tiers of government. A tier is coded as a “tier of government” if state executive body at that level (1) was funded from the public budget, (2) had authority to administer a range of public services, and (3) had a territorial jurisdiction. Source: Fan et al (2009).


Appendix 2. List of Countries

Cross Section: 101 countries

Panel: 58 countries

Albania Gambia, The Papua New Guin.

Argentina Malaysia

Argentina Germany Paraguay Armenia Malta Australia Greece Peru Australia Mauritius Austria Guatemala Philippines Austria Moldova Azerbaijan Honduras Poland Belarus Morocco Bahrain Hungary Portugal Belgium Netherlands Bangladesh Iceland Romania Bolivia New Zealand Belarus India Russia Bulgaria Norway Belgium Indonesia Senegal Canada Peru Benin Iran Slovak Republic Chile Poland Bolivia Ireland Slovenia Costa Rica Portugal Botswana Israel Somalia Croatia Romania Brazil Italy South Africa Cyprus Russia Bulgaria Japan Spain Czech Rep Slovakia Burkina Faso Jordan Sri Lanka Denmark Slovenia Cameroon Kazakhstan Sudan El Salvador South Africa Canada Kenya Swaziland Estonia Spain

Chile Korea, Dem. Rep. Sweden

Finland Sweden

China Latvia Switzerland France Switzerland Colombia Lithuania Tajikistan Georgia Thailand Congo, Dem. Rep. Luxembourg Thailand

Germany Ukraine

Congo, Rep. of Madagascar Trinidad &Tobago

Greece UK

Costa Rica Malawi Tunisia Hungary USA Croatia Malaysia Turkey Iceland Cyprus Mauritius Uganda India Czech Republic Mexico United Kingdom Iran Denmark Moldova United States Ireland Dominican Rep. Mongolia Uruguay Israel Ecuador Morocco Venezuela Italy El Salvador Netherlands Zambia Jamaica Estonia New Zealand Zimbabwe Kazakhstan Ethiopia Nicaragua Latvia Fiji Norway Lesotho Finland Pakistan Lithuania France Panama Luxembourg


Appendix 3.1. Correlation Matrix. Cross Section Sample

GOV

QUAL FISC DESC LGDP LPOP

VOICE ACC

SCHN DEC UNITAR

STATE ANT

DEM MATU

AGE DEM

LEGAL SO

LEGAL GE

LEGAL FR

LEGAL SC FED AUTON

REG ELEC BICAM

GOV QUAL 1.0000 0.2045 0.8450 -0.1495 0.8839 0.3621 -0.0782 0.3457 0.7211 0.7239 -0.1975 0.2993 -0.1324 0.3917 0.2245 0.1895 0.4367 0.2347

FISC DESC 1.0000 0.2129 0.3350 0.1217 0.0703 -0.3162 0.2821 0.2618 0.2431 0.1861 0.3390 -0.2695 0.2334 0.4078 0.2600 0.2343 0.1469

LGDP 1.0000 -0.1517 0.7654 0.4175 -0.1745 0.3369 0.5821 0.6195 -0.0281 0.2605 -0.1264 0.2918 0.2754 0.2442 0.4227 0.3032

LPOP 1.0000 -0.1900 0.2102 -0.5601 0.3367 0.0102 -0.0091 -0.0770 0.1055 0.1130 -0.1977 0.4031 0.3155 0.3043 0.2817

VOICE ACC 1.0000 0.4754 -0.0817 0.1448 0.6111 0.6421 -0.1039 0.2382 -0.0585 0.3227 0.1942 0.1752 0.4353 0.2473

SCHN DEC 1.0000 -0.4109 0.1072 0.2888 0.3207 -0.2066 0.1785 0.1309 0.0819 0.3419 0.3213 0.5429 0.4034

UNITAR 1.0000 -0.0376 -0.1601 -0.1962 0.2037 -0.2315 -0.0410 0.1417 -0.7559 -0.7189 -0.5365 -0.7223

STATE ANT 1.0000 0.3289 0.1950 0.2822 0.3393 -0.0890 0.1991 0.0698 0.0315 0.1346 0.1025

DEM MATU 1.0000 0.8603 -0.2782 0.2529 -0.1707 0.4202 0.2800 0.2640 0.4136 0.1782

AGE DEM 1.0000 -0.2925 0.2017 -0.0855 0.2809 0.3148 0.2762 0.4383 0.2287

LEGAL SO 1.0000 -0.1040 -0.4322 -0.1169 -0.1580 -0.2139 -0.2254 -0.1647

LEGAL GE 1.0000 -0.1713 -0.0463 0.3285 0.2492 0.2675 0.2265

LEGAL FR 1.0000 -0.1925 -0.0310 0.0544 0.0336 0.0520

LEGAL SC 1.0000 -0.1001 -0.1477 0.0437 -0.1127

FED 1.0000 0.8401 0.5057 0.4923

AUTON 1.0000 0.5483 0.4509

REG ELEC 1.0000 0.5130

BICAM 1.0000

GOV QUAL is Government Quality, FISC DESC is Fiscal Decentralization, LGDP and LPOP are logs of GDP and Population, VOICE ACC is Voice and Accountability, SCHN DEC is Schneider Decentralization, UNITAR is Unitarism, STATE ANT is State Antiquity, DEM MATU is Democratic Maturity DV, AGE DEM is Age of Democracy, LEGAL SO-GE-FR-SC are Soviet-German-French-Scandinavian Legal Origin respectively, FED is Federal, AUTON is Autonomy, REG ELEC is Regional Elections and BICAM is Bicameralism.


Appendix 3.2. Correlation Matrix. Panel Sample

GOV QUAL FISC DESC LGDP LPOP VOICE ACC FED AUTON REG ELEC BICAM

GOV QUAL 1.0000 0.2387 0.8397 -0.1808 0.9144 0.2037 0.2417 0.5031 0.1224

FISC DESC 1.0000 0.2952 0.3694 0.1698 0.5811 0.4251 0.4109 0.3025

LGDP 1.0000 -0.1288 0.7363 0.2416 0.3267 0.5163 0.0921

LPOP 1.0000 -0.2104 0.4692 0.3621 0.3928 0.5129

VOICE ACC 1.0000 0.1247 0.1832 0.5065 0.1050

FED 1.0000 0.8275 0.4957 0.5107

AUTON 1.0000 0.5644 0.4629

REG ELEC 1.0000 0.4953

BICAM 1.0000

GOV QUAL is Government Quality, FISC DESC is Fiscal Decentralization, LGDP and LPOP are logs of GDP and Population, VOICE ACC is Voice and Accountability, FED is Federal, AUTON is Autonomy, REG ELEC is Regional Elections and BICAM is Bicameralism.


Appendix 4. Panel data description

code

Countries

Obs*

Quality of Government Fiscal Decentralization mean max min mean max min

1 Argentina 3 -0.6155 -0.7945 -0.5313 0.3885 0.4073 0.3768 2 Armenia 4 -0.2843 -0.3373 -0.2490 0.0310 0.0381 0.0260 3 Australia 8 1.8007 1.968 1.7693 0.2671 0.3056 0.2525 4 Austria 6 1.8192 1.8640 1.7396 0.2009 0.2285 0.1659 5 Belarus 4 -1.2335 -1.2741 -1.1809 0.3146 0.3583 0.2529 6 Belgium 7 1.4949 1.6320 1.3919 0.1426 0.1564 0.1233 7 Bolivia 8 -0.4476 -0.8134 -0.2125 0.2053 0.2455 0.1874 8 Bulgaria 9 0.0697 0.1834 -0.1184 0.0872 0.1169 0.0415 9 Canada 7 1.8310 1.8685 1.7867 0.5332 0.5290 0.5149

10 Chile 7 1.3109 1.3479 1.2794 0.0761 0.0898 0.0682 11 Costa Rica 5 0.5221 0.6162 0.4145 0.0326 0.0350 0.0305 12 Croatia 5 0.2410 0.2998 0.2035 0.0990 0.1060 0.0945 13 Cyprus 9 1.0135 1.0733 0.9798 0.0285 0.0652 0.0179 14 Czech Rep 7 0.7300 0.8094 0.5979 0.1544 0.1696 0.1348 15 Denmark 9 1.9897 2.1287 1.8787 0.3451 0.3636 0.3087 16 El Salvador 5 -0.2603 -0.3691 -0.2007 0.0359 0.0485 0.0158 17 Estonia 3 1.0095 1.0747 0.9307 0.1394 0.1412 0.1374 18 Finland 8 2.0638 2.1134 1.9870 0.2519 0.2547 0.2500 19 France 8 1.3291 1.3971 1.2519 0.1382 0.1457 0.1331 20 Georgia 4 -0.7443 -0.8818 -0.3368 0.2259 0.3061 0.1943 21 Germany 8 1.6976 1.7814 1.6154 0.3311 0.3393 0.3194 22 Greece 8 0.7637 0.8029 0.6425 0.0073 0.0157 0.0020 23 Hungary 6 0.8615 0.9202 0.7943 0.1396 0.1479 0.1274 24 Iceland 9 1.9467 2.1081 1.7146 0.2531 0.2611 0.2321 25 India 4 -0.1533 -0.2235 -0.1036 0.3007 0.3113 0.2857 26 Iran 6 -0.7481 -0.9260 -0.6263 0.0611 0.0694 0.0504 27 Ireland 8 1.6468 1.6847 1.5688 0.0903 0.1027 0.0824 28 Israel 7 0.9288 1.0432 0.8224 0.0690 0.0730 0.0655 29 Italy 6 0.7989 0.8869 0.6038 0.1844 0.1904 0.1766 30 Jamaica 4 -0.1963 -0.2322 -0.1471 0.0109 0.0130 0.0087 31 Kazakhstan 7 -0.8118 -0.9305 -0.6728 0.2777 0.3785 0.1692 32 Latvia 5 0.6112 0.6710 0.5105 0.1810 0.1860 0.1738 33 Lesotho 8 -0.2314 -0.3455 -0.1394 0.0010 0.0025 0.0000 34 Lithuania 7 0.6067 0.7395 0.3959 0.1263 0.1957 0.0935 35 Luxembourg 8 1.9726 2.0940 1.8411 0.0693 0.0763 0.0595 36 Malaysia 6 0.5203 0.5761 0.4882 0.1067 0.1303 0.0855 37 Malta 8 1.1566 1.2759 1.0363 0.0024 0.0043 0.0000 38 Mauritius 5 0.5879 0.6522 0.5499 0.0178 0.0206 0.0162 39 Moldova 5 -0.6840 -0.7657 -0.6189 0.1815 0.2323 0.1393 40 Morocco 3 -0.0739 -0.0969 -0.0575 0.0472 0.0493 0.0460 41 Netherlands 9 1.9389 2.0027 1.8300 0.1154 0.1188 0.1052 42 New Zealand 4 1.9119 2.0704 1.9382 0.0766 0.0775 0.0751 43 Norway 7 1.8417 1.9366 1.7259 0.1553 0.1774 0.1381 44 Peru 8 -0.2367 -0.4390 -0.0570 0.0550 0.0581 0.0491 45 Poland 6 0.4813 0.5527 0.3794 0.1751 0.2168 0.1581 46 Portugal 8 1.1975 1.2951 1.1146 0.0849 0.0924 0.0808 47 Romania 4 -0.1551 -0.2074 -0.0940 0.2031 0.2085 0.1918 48 Russia 5 -0.6980 -0.8447 -0.5236 0.3333 0.3681 0.3141 49 Slovakia 7 0.5390 0.7526 0.2609 0.0808 0.1192 0.0497 50 Slovenia 7 0.8584 0.9281 0.7208 0.1004 0.1067 0.0959 51 South Africa 9 0.4625 0.5630 0.4129 0.1749 0.1300 0.1934 52 Spain 9 1.3716 1.4663 1.0965 0.2543 0.3218 0.1823 53 Sweden 8 1.8912 1.9680 1.8106 0.3514 0.3904 0.3155 54 Switzerland 8 1.9632 1.9970 1.9049 0.4378 0.4947 0.3474 55 Thailand 9 0.1279 0.2219 0.0533 0.0759 0.0795 0.0726 56 Ukraine 6 -0.6869 -0.4728 -0.8121 0.1719 0.2055 0.1459 57 UK 9 1.8469 1.9069 1.7081 0.0870 0.1034 0.0803 58 USA 7 1.6316 1.7326 1.4938 0.4331 0.4664 0.3963

* The number of observations considered for each country is limited by the availability of the Fiscal Decentralization data.


Appendix5. First-stage regressions. Revenue decentralization regressed on geographic area.

(1) (2)

(3)

(4) Federal

(5) Autonomy

(6) Reg. Elec.

(7) Bicameral

(8) Federal

(9) Autonomy

(10) Reg. Elec.

(11) Bicameral

Constant -0,320

(0,16)**

-0.224

(0,23)

-0,119

(0.31)

-0.165

(0.19)

-0.205

(0.21)

-0.295

(0.23)

-0.273

(0.20)

-0.077

(0.23)

-0.154

(0.27)

-0.249

(0.28)

-0.277

(0.22)

Geographic Area 0,036

(0,01)***

0,040

(0.01)***

0,044

(0.01)***

0.024

(0,01)***

0.032

(0,01)***

0.030

(0,01)***

0.028

(0,01)***

0.024

(0,01)***

0.028

(0,01)**

0.031

(0,01)**

0.027

(0,01)**

GDP per capita 0,029

(0,02)*

0.025

(0,02)

0.058

(0,03)**

0.015

(0,02)

0.020

(0,02)

0.023

(0,02)

0.029

(0,02)*

0.002

(0,02)

0.009

(0,02)

0.015

(0,02)

0.022

(0,02)

Population 0,001

(0,01)

-0,007

(0,01)

-0,004

(0,01)

0,001

(0,01)

-0,002

(0,01)

0,006

(0,01)

0,001

(0,01)

0,003

(0,01)

0,004

(0,01)

0,007

(0,02)

0,008

(0,01)


-0,016

(0,02)

-0,018

(0,02)

-0,001

(0,03)

-0,016

(0,02)

-0,019

(0,02)

-0,008

(0,02)

-0,017

(0,02)

0,011

(0,02)

0,012

(0,02)

0,016

(0,03)

0,007

(0,02)

Socialist

Legal Origin

0,078

(0,02)***

0,058

(0,02)**

0,058

(0,03)**

0,092

(0,02)***

0,081

(0,02)***

0,088

(0,03)***

0,075

(0,02)***

0,191

(0,11)*

0,193

(0,10)*

0,195

(0,10)**

0,179

(0,10)*

German

Legal Origin

0,230

(0,05)***

0,235

(0,05)***

0,245

(0,07)***

0,266

(0,05)***

0,239

(0,05)***

0,254

(0,05)***

0,253

(0,05)***

0,244

(0,04)***

0,238

(0,05)***

0,240

(0,05)***

0,234

(0,05)***

French

Legal Origin

-0,020

(0,02)

-0,019

(0,02)

-0,026

(0,03)

-0,004

(0,02)

-0,016

(0,03)

-0,019

(0,03)

-0,019

(0,02)

-0,014

(0,02)

-0,022

(0,03)

-0,020

(0,03)

-0,029

(0,02)

Scandinavian

Legal Origin

0,142

(0,04)***

0,138

(0,04)***

0,129

(0,05)**

0,185

(0,03)**

0,165

(0,04)***

0,176

(0,04)***

0,144

(0,04)***

0,163

(0,04)***

0,149

(0,04)***

0,157

(0,05)***

0,133

(0,04)***

Unitarism -0.003

(0,01)

Schneider’s Decentralization

-0.051

(0,05)

Pol. Decentralization *

Geographic Area

0.029

(0,01)**

0.010

(0,01)

0.019

(0,01)

0.019

(0,01)*

0.032

(0,02)**

0.022

(0,01)

0.017

(0,02)

0.014

(0,01)

Pol. Decentralization -0.132

(0,09)

-0.023

(0,07)

-0.115

(0,07)

-0.122

(0,05)

-0.152

(0,10)

-0.111

(0,08)

-0.103

(0,09)

-0.096

(0,06)*

State Antiquity 0.026

(0,06)

0.007

(0,06)

-0.006

(0,07)

-0.023

(0,05)

Adjusted R2 0.47 0.46 0.49 0.50 0.45 0.44 0.48 0.51 0.45 0.44 0.48

Observations 99 90 64 98 85 85 99 74 63 66 75

F-statistics 11.80 9.33 7.60 10.65 7.78 7.67 10.01 7.93 5.63 5.53 7.29

All regressions report White Heteroskedasticity-Robust Standard Errors in parentheses. *, **, *** measures statistical significance at the 10, 5 and 1% levels respectively. Legal Origin and Political Decentralization variables (Federal, Autonomy, Reg. Elections and Bicameral) are dummy variables.


Appendix 6. Summary Statistics

Cross Section Sample

GOV

QUAL FISC DESC LGDP LPOP

VOICE ACC

SCHN DEC UNITAR

STATE ANT

AGE DEM

DEM MATU

LEGAL SO

LEGAL GE

LEGAL FR

LEGAL SC FED AUTON

REG ELEC BICAM

Mean 0.1995 0.1515 8.7522 9.4220 0.1694 0.5660 3.9891 0.4555 0.1461 0.2277 0.2079 0.0396 0.4158 0.0495 0.1600 0.2588 0.4022 0.4455

Median -0.0467 0.1001 8.8834 9.2380 0.1524 0.5100 5.0000 0.4050 0.0306 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

Maximum 2.0285 0.5217 10.716 14.054 1.5493 1.0000 5.0000 1.0000 1.000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000

Minimum -1.9871 0.0109 6.0204 5.6487 -2.0942 0.0100 1.0000 0.0700 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

Std. Dev. 0.9965 0.1306 1.1115 1.5790 0.9535 0.2698 1.3787 0.2635 0.2264 0.4214 0.4078 0.1960 0.4953 0.2180 0.3684 0.4405 0.4932 0.4995

Observations 101 99 101 101 101 65 92 76 101 101 101 101 101 101 100 85 87 101

GOV QUAL is Government Quality, FISC DESC is Fiscal Decentralization, LGDP and LPOP are logs of GDP and Population, VOICE ACC is Voice and Accountability, SCHN DEC is Schneider Decentralization, UNITAR is Unitarism, STATE ANT is State Antiquity, DEM MATU is Democratic Maturity DV, AGE DEM is Age of Democracy, LEGAL SO-GE-FR-SC are Soviet-German-French-Scandinavian Legal Origin respectively, FED is Federal, AUTON is Autonomy, REG ELEC is Regional Elections and BICAM is Bicameralism.

Panel Sample

GOV QUAL FISC DESC LGDP LPOP VOICE ACC FED AUTON REG ELEC BICAM

Mean 0.7084 0.1631 9.4320 9.2398 0.7287 0.2068 0.2758 0.4705 0.5000

Median 0.7371 0.1374 9.5186 9.1692 1.0084 0.0000 0.0000 0.0000 0.5000

Maximum 2.1287 0.5290 10.911 13.909 1.8264 1.0000 1.0000 1.0000 1.0000

Minimum -1.2741 -0.0652 7.4803 5.6193 -1.7112 0.0000 0.0000 0.0000 0.0000

Std. Dev. 0.9346 0.1270 0.7634 1.5737 0.7981 0.4054 0.4473 0.4996 0.5004

Observations 522 383 522 522 522 522 522 459 522

GOV QUAL is Government Quality, FISC DESC is Fiscal Decentralization, LGDP and LPOP are logs of GDP and Population, VOICE ACC is Voice and Accountability, FED is Federal, AUTON is Autonomy, REG ELEC is Regional Elections and BICAM is Bicameralism.


Bibliography

Acemoglu, D., Johnson, S., Robinson, J., 2001. The colonial origins of comparative development: An empirical investigation. American Economic Review 91(5), 1369-1401.

Alesina, A., Arnaud D., William, E., Sergio, K., Romain, W., 2003. Fractionalization. Journal of Economic Growth 8, 155-194.

Arellano, M., 2002. Sargan's instrumental variables estimation and the generalized method of moments. Journal of Business & Economic Statistics 20, 450-459.

Arzaghi, M., Henderson, V., 2005. Why countries are fiscally decentralizing. Journal of Public Economics 89, 1157–1189.

Bardhan, P., 2002. Decentralization of governance and development. Journal of Economic Perspectives 16 (4), 185–205.

Bardhan, P., Mookherjee, D., 2000. Capture and governance at local and national levels. American Economic Review 90(2), 135-139.

Bardhan, P., Mookherjee, D., 2006. Decentralization, corruption and government accountability: An overview, in: S. Rose-Ackerman (Ed.), International Handbook on the Economics of Corruption, Edward Elgar, pp. 161-188.

Beck, T., Clarke, G., Groff, A., Keefer, P., Walsh, P., 2001. New tools in comparative political economy: The database of political institutions. World Bank Economic Review 15(1), 165-176.

Blanchard, O., Shleifer, A., 2001. Federalism with and without political decentralization: China versus Russia. IMF Staff Papers 48(4), 171-179.

Bockstette, V., Chanda, A., Putterman, L., 2002. States and markets: The advantage of an early start. Journal of Economic Growth 7, 347-369.

Boex, J., Simatupang, R., 2008. Fiscal decentralization and empowerment: Evolving concepts and alternative measures. Fiscal Studies 29(4), 435-465.

Bohara, A., Mitchell, N., Mittendorff, C., 2004. Compound democracy and the control of corruption: A cross-country investigation. The Policy Studies Journal 32(4), 481-499.

Brennan, G., Buchanan, J., 1980. The Power to Tax. Analytical Foundations of a Fiscal Constitution. Cambridge: Cambridge University Press.

Breton, A., 1996. Competitive Governments. A Economic Theory of Politics and Public Finance. New York: Cambridge University Press.

Cai, H., Treisman, D., 2004. State corroding federalism. Journal of Public Economics 88, 819-843.

Cai, H., Treisman, D., 2005. Does competition for capital discipline government? Decentralization, globalization and public policy. American Economic Review 95(3), 817-830.

Dreher, A., 2006. Power to the people? The impact of decentralization on governance. Mimeo, Swiss Federal Institute of Technology: Department of Management, Technology and Economics.

Ebel, R., Yilmaz, S., 2002. Concept of fiscal decentralization and worldwide overview. Mimeo: World Bank Institute.

Enikolopov, R., Zhuravskaya, E., 2007. Decentralization and political institutions. Journal of Public Economics 91(11-12), 2261-2290.

Elazar, D., 1995. From statism to federalism: A paradigm shift. Publius 25(2), 5-18. Estache, A., Sarbajit, S., 1995. Does decentralization increase spending on public

infrastructure. World Bank Policy Research Working Paper 1457. Fan, S., Lin, C., Treisman, D., 2009. Political decentralization and corruption: evidence

from around the world. Journal of Public Economics 93, 14-34.


Fisman, R., Gatti, R. 2002a. Decentralization and corruption: evidence across countries. Journal of Public Economics 83, 325-345.

Fisman, R., Gatti, R., 2002b. Decentralization and corruption: Evidence from U.S. federal transfer programs. Public Choice 113 (1-2), 25-35.

Gerring, J., Thacker, T., 2004. Political institutions and corruption: The role of unitarism and parliamentarism. British Journal of Political Science 34 (2), 295-330.

Goldsmith, A., 1999. Slapping the grasping hand: correlates of political corruption in emerging markets. American Journal of Economics and Sociology 58(4), 865-883.

Huther, J., Shah, A., 1998. Applying a simple measure of good governance and its application to the debate on the appropriate level of fiscal decentralization. World Bank Working Paper Series 1894, Washington D.C.

Jin, H.; Yingyi, Q., Weingast, B., 2005. Regional decentralization and fiscal incentives: Federalism Chinese style. Journal of Public Economics 89(9-10), 1719-1742.

Kaufmann, D., Kraay, A., Mastrutzzi, M., 2006. Governance matters V: Aggregate and individual governance indicators for 1996-2005. World Bank Policy Research Paper 4012.

Keen, M., Marchand, M., 1997. Fiscal competition and the pattern of public spending. Journal of Public Economics 66, 33–53.

Khaleghian, P., 2003. Decentralization and public services: The case of immunization. World Bank Policy Research Working Paper 2989.

Kuniková, J., Rose-Ackerman, S., 2005. Electoral rules and constitutional structures as constraints on corruption. British Journal of Political Science35(4), 573-606.

Kyriacou, A., Roca-Sagalés, O., 2009. Fiscal decentralization and the quality of government: Evidence from panel data. Hacienda Pública Española 189(2), 131-156.

La Porta, R.; Lopez-de-Silanes, F.; Shleifer, A., Vishny, R., 1999. The quality of government. Journal of Law, Economics and Organization 15(1), 222-279.

de Mello, L., Barenstein, M., 2001. Fiscal Decentralization and governance: A cross-country analysis. IMF Working Paper 01/71.

Montinola, G.; Yingyi, Q., Weingast, B., 1995. Federalism Chinese style: The political basis for economic success. World Politics 48(1), 50-81.

Murray, M., 2006. Avoiding invalid instruments and coping with weak instruments 20(4), 111-132.

Oates, W., 1972. Fiscal Federalism. New York: Harcourt Brace Jovanovich. Oates, W., 1999. An essay on fiscal federalism. Journal Economic Literature 37, 1120-

1149. Qian, Y., Roland, G., 1998. Federalism and the soft budget constraint. American

Economic Review 88(5), 1143-1162. Panizza, U., 1999. On the determinants of fiscal centralization: Theory and evidence.

Journal of Public Economics 74, 97-139. Panizza, U., 2001. Electoral rules, political systems and institutional quality. Economics

and Politics 13(3), 311-342. Persson, T.; Tabellini, G., Trebbi, F., 2003. Electoral rules and corruption. Journal of

European Economic Association 1 (4), 958-989. Prud’homme, R., 1995. On the dangers of decentralization. World Bank Research

Observer 10(2), 201-220.


Robalino, D.; Picazo, O., Voetberg, A., 2001. Does fiscal decentralization improve health outcomes? Evidence from a cross country analysis. World Bank Policy Research Working Paper 2565.

Rodden, J., 2004. Comparative federalism and decentralization. On meaning and measurement. Comparative Politics 36(4), 481-500.

Rodden, J., Rose-Ackerman, S., 1997. Does federalism preserve markets? University of Virginia Law Review 83, 1521-1572.

Salmon, P., 1987. Decentralization as an incentive scheme. Oxford Review of Economic Policy 3(2), 24-43.

Sandholtz, W., Koetzle, W., 2000. Accounting for corruption: Economic structure, democracy, and trade. International Studies Quarterly 44, 31-50.

Schneider, A., 2003. Decentralization: conceptualization and measurement. Studies in Comparative International Development 38(3), 32-56.

Seabright, P., 1996. Accountability and decentralization in government: An incomplete contracts model. European Economic Review 40(1), 61-89.

Shleifer, A., Vishny, R., 1993. Corruption. Quarterly Journal of Economics 108, 599-616.

Tabellini, G., 2000. Constitutional determinants of government spending. Mimeo: Bocconi University.

Tanzi, V., 1995. Fiscal federalism and decentralization: A review of some efficiency and macroeconomic aspects. Annual World Bank Conference on Development Economics, 295-316.

Tendler, J., 1997. Good Government in the Tropics. Baltimore, London: John Hopkins University Press.

Treisman, D., 2000. The causes of corruption: A cross-national study. Journal of Public Economics 76, 399-457.

Treisman, D., 2002. Decentralization and the quality of government. Mimeo: University of California.

Treisman, D., 2006. Fiscal decentralization, governance, and economic performance: A reconsideration. Economics and Politics 18 (2), 219-235.

Treisman, D., 2007. What we have learned about the causes of corruption from ten years of cross-national empirical research? Annual Review of Political Science 10: 211-244.

Weingast, B., 1995. The economic role of political institutions: Market-preserving federalism and economic development. Journal of Law, Economics, and Organization 11, 1-31.

World Bank, 2004. Measuring Fiscal Decentralization. Data Note. World Development Report, 1999/2000. Entering the 21st Century. Washington DC:

World Bank. Zhuravskaya, E., 2000. Incentives to provide local public goods: Fiscal federalism,

Russian style. Journal of Public Economics 76(3), 337-368.

Documents

Fiscal and Political DecentralizationFiscal and Political ...gent.uab.cat/oroca/sites/gent.uab.cat.oroca/files/ispwp...Fiscal and Political DecentralizationFiscal and Political Decentralization