Types of decentralisation and governance: evidence from across the world

ACADEMIC PAPER Types of decentralisation and governance: Evidence from across the world Pedro J. Camões CICP/Universidade do Minho, Portugal Correspondence Pedro J. Camões, CICP/Universidade do Minho, Portugal. Email: [email protected].pt Funding information Fundaç~ ao para a Ciência e a Tecnologia The relationship between decentralisation and quality of government is theoretically ambivalent and remains empirically mixed. The classical theory argues that decentralisation improves the ability of governments to fully respond to collective preferences. A different perspective points to some disadvantages related to problems of coordination and inertia. This paper provides a worldwide cross-sectional analysis that tests these two contrasting views and the effect of democracy. The results robustly support that (a) decentralisation has an independent and positive effect on the quality of government, but that (b) political, fiscal and administrative types of decentralisation have quite different effects and that (c) the effect is attenuated when the effect of democracy is considered. Together, these results shed light with regard to the optimal design of reforms towards decentralisation across the world. 1|INTRODUCTION Decentralisation, the territorial organisation of the state and public sectors, and its relation with government performance is an important element in the recent debate on institutional design. But it is not a new debate. Small governments in the form of small territories have been praised since Plato, Aristotle, Rosseau, Montesquieu and J. S. Mill, among others. These arrangements are regarded as appropriate to protect the individual rights against tyranny of the majority and to encourage political participation and the accommodation. Particularly, liberal authors were specially keen on praising the civic virtues of local democracy. For example, de Tocqueville claimed that “municipal institutions constitute the strength of free nations. Town meetings are to liberty what primary schools are to science. A nation may establish a free government, but without municipal institutions, it cannot have the spirit of liberty.”To J. S. Mill, ‘free and popular municipal institutions’are part of the “peculiar training of a citizen”, and the practical part of the political education of a free people. Accordingly, ‘modern governmental systems, even those of unitary states, are all multilevel systems’(Breton, 1998, p. 181). These arguments about democracy and liberty are mostly normative. On positive grounds, the question is whether decentralisation is important and how it makes a difference. Do more decentralised countries have better quality of government? Moreover, do all institutional designs of decentralisation (types) have the same effect? According to the classical theory, decentralisation improves the ability of governments to fully respond to the collective preferences. A different perspective points to some disadvantages related to problems of coordination and inertia. In sum, the relationship between decentralisation and quality of government is theoretically ambivalent and remains empirically mixed. This paper provides a worldwide empirical analysis that tests these two contrasting views. In particular, it addresses the specific impact of political, fiscal and administrative decentralisation and the role played by democratic institutions. The results show that (a) decentralisation has indeed an independent and positive effect on the quality of government, but that (b) the three types of decentralisation have substantively different effects and that (iii) the effect of decentralisation is attenuated when the effect of democracy is considered. Together, these results shed light with regard to the optimal design of reforms towards decentralisation across the world. The next section presents a brief picture of the general literature on the costs and benefits of decentralisation, including the huge amount of empirical results. The third section focuses solely on the This study was conducted at Research Center in Political Science (UID/CPO/0758/2019), University of Minho and supported by the Portuguese Foundation for Science and Technology and the Portuguese Ministry of Education and Science through national funds. Received: 5 June 2020 Accepted: 20 August 2020 DOI: 10.1002/pa.2435 J Public Affairs. 2022;22:e2435. wileyonlinelibrary.com/journal/pa © 2020 John Wiley & Sons Ltd 1of12 https://doi.org/10.1002/pa.2435 14791854, 2022, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pa.2435 by Cochrane Portugal, Wiley Online Library on [22/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License link of decentralisation and quality of government performance and reviews the most important empirical arguments as well as derives three main hypothesis. Section four presents the data and methods, and section five discusses the findings. 2|THEBENEFITSANDCOSTSOF DECENTRALISATION The advantage of small governments of small territories is seen as the appropriate arrangements to protect against tyranny of the majority, promote political participation, accommodate different and often conflicting views and to respond to citizens’demands (Inman & Rubinfeld, 1997). In very simple terms, the most important theoretical argument concerning decentralisation is that it can make government more accountable and responsive to the governed (Faguet, 2014). The arguments concerning proximity are later developed in explicit economic terms by Tiebout (1956). He formulated the advantages of decentralisation and to the virtues of smaller units of government because they were more capable of promoting a closer match between tastes and preferences of the citizens of a locality and a local government that is able to adjust the policies to those preferences. The differences of policies among local governments promote horizontal competition among local governments. But for the match to occur, a necessary condition has to be met: geographic mobility of citizens to search for the most preferred policy mix. In Tiebout's words, citizens “vote with their feet”. According to the model, efficiency of public provision is achieved. A new “second-generation theory of fiscal federalism”emerged to provide new insights into the structure and working of decentralised unitary and federal systems (Oates, 2005), particularly the implications of incentives created by political and fiscal institutions (Weingast, 2014). This different perspective points out some important caveats that ultimately make the net effect of decentralisation on output as indeterminate (Treisman, 2006). The main disadvantages of decentralisation are related to problems of coordination and inertia. For example, Treisman (1999) formalises a game theory model in which decentralised countries, particularly in the group of less developed, may have more difficulty in abandoning vicious cycles of inefficiency. Treisman's argument is that ‘decentralization may complicate attempts at macroeconomic stabilisation and economic reforms’.In the end, the net effect of these benefits and cost on output (economic performance and governance) is indeterminate. In consequence, the empirical analysis of the effect of decentralisation also produce very mixed results. 2.1 |Empirical results The first strand of empirical studies focused on the effects of fiscal decentralisation. This line of research was produced by economists stressing the role of self-interested and overspending politicians and became known as the “searching for Leviathan”literature. The idea of a Leviathan that overspends permanently was inspired in Brennan and Buchanan (1980). Marlow (1988) analysed time series data (1946–1985) for the USA and found that fiscal decentralisation is associated with slowing or falling public sector size. In a panel analysis of 18 OECD (1980–2000), Baskaran (2011) found that decentralisation leads to an increase in the size of government. This results were irrespective of the ideology of the central government, with no support for his initial hypothesis that the effect of fiscal decentralisation on public sector size depends on government's ideology; however, the magnitude of the marginal effect is larger under a leftthan under a right-wing government. Some analysis present an unequivocally positive findings. Lessmann (2009), both with cross section and panel data of OECD countries (1982–2000), finds that higher a degree of fiscal decentralisation is associated with lower regional disparities. Hence, poor regions experience no disadvantages from decentralisation and, instead, appear to benefit. Ligthart and Van Oudheusden (2015) look at whether fiscal decentralisation is associated with citizens’trust in government and related institutions, with the argument that more decentralised fiscal systems improve perceived governments’responsiveness to citizens’preferences. Their survey data from up to 42 countries over the period 1994–2007 confirm this positive relationship, although this positive association with fiscal decentralisation does not extend to other, non-government-related institutions. Baskaran, Feld, and Schnellenbach (2016) suggested that decentralisation may also increase growth by raising the ability of the political system to innovate and carry out reforms. However, when political decentralisation is considered, the results appear more negative or, at least, mixed. The results in Rodden and Wibbels (2002), comparing 46 federal and unitary developing countries in the 197,995 period, suggest that poor performance is more likely in federal countries. His argument is that “devolved political and fiscal institutions create incentives for sub-national governments to avoid or delay the political costs of fiscal adjustment.”This argument of avoidance or delay was confirmed by Treisman (2000) that found no direct effect on inflation, but that political decentralisation makes it more difficult to change the patterns of evolution of inflation or fiscal policy because it creates more veto players leading to “lock-in”effect. Those decentralised countries that start with low inflation rates, inflation increases slowly; the decentralised countries that start with higher levels of inflation tend to remain with higher levels and to experience more difficulty in reducing it. A different view is presented in Blanchard and Shleifer (2001), which attribute success of China, as opposed to Russia to the presence of political decentralisation. 3|DECENTRALISATION AND QUALITY OF GOVERNMENT The relationship between decentralisation of government activities and their performance or quality is an important element in the recent debate on institutional design. The first empirical paper to address 2of12 CAMÕES 14791854, 2022, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pa.2435 by Cochrane Portugal, Wiley Online Library on [22/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License empirically the broad concept of quality of government was La Porta, Lopez-de Silanes, Shleifer, and Vishny (1999) considering a large number of measures of government intervention, public sector efficiency, public good provision, size of government and political freedom and studied a large cross section of countries. Their econometric analysis suggests that countries that are poor, close to the equator, ethnolinguistically heterogeneous, use French or socialist laws or have high proportions of Catholics or Muslims exhibit inferior government performance. Additionally, larger governments tend to performing better. A well-defined body of empirical literature introduces specifically decentralisation as key independent variable. As review in the previous section, the results also globally mixed. The first study tested the relationship between five measures of (political and fiscal) decentralisation and the level of corruption and the provision of public services as their measures of the quality of government (Treisman, 2002). The findings, using a large sample of 67–88 countries, showed a weak relationship between decentralisation and corruption and low social service provision. In this decade, Andreas Kyriacou, Oriol Roca-Sagalés, and their colleagues published a series of three papers analysing this relation. The first joins a cross-sectional and panel analysis to find that fiscal decentralisation improves government quality but vanes when it is accompanied by political decentralisation Kyriacou and RocaSagalés (2011a). Later, they found that this positive effect is mitigated in the presence of regional elections and multilevel government (Kyriacou & Roca-Sagalés, 2011b). In their last piece, they suggested that a process of fiscal decentralisation, accompanied by measures to improve the quality of government, would be an effective strategy for reducing regional inequalities (Kyriacou, Muinelo-Gallo, & RocaSagalés, 2017). Under this broad and very often imprecise theme of quality of government (Rothstein & Teorell, 2008), the concept and measures of corruption are probably the most empirically studied dimension. Fisman and Gatti (2002), with a relatively small sample of 28–55 countries, found that fiscal decentralisation of government expenditure is strongly and significantly associated with lower corruption. These results persist when decentralisation is instrumented by the origin of a country's legal system. Fan, Lin, and Treisman (2009) studied how political decentralisation can affect the frequency and costliness of bribe extraction on the part of corrupt officials and found that countries with a larger number of government or administrative tiers and a larger number of local public employees, reported bribery was more frequent. When local or central governments received a larger share of GDP in revenue, bribery was less frequent. Overall, they interpreted these results as suggesting the danger of uncoordinated rent-seeking as government structures become more complex. 3.1 |Which type of decentralisation? It is clear that in this significant and rich empirical literature, the consensus is still to be achieved, largely because a significant number of questions remain open. Nevertheless, some relevant points can be brought to the surface. Above all, there is a difference between fiscal and political decentralisation which is expressed in the different effects in quality of government shown in empirical literature. This is no surprise because these are not alternative measures for the same concept as some wrongly literature still assumes. Rather, these are two different institutional mechanisms that have different theoretical and empirical implications. For example, O'Dwyer and Ziblatt (2006) found that the different types of decentralisation have differing and sometimes opposing impacts on the quality of governance. Specifically, political decentralisation is associated with higher government efficiency among high GDP per capita countries, while it is associated with lower government efficiency among low GDP per capita countries. Moreover, the literature, by focusing almost exclusively on the political and fiscal types, ignores the concept of administrative decentralisation, that is, the distribution of administrative structures and employers and/or functions (Lessmann & Markwardt, 2016). As for long has been acknowledged, this is a different institutional mechanism that is expected to have still different implications for the quality of government (Kaufman, 1969). Figure 1shows the scatters of the relation between the three different types of decentralisation—fiscal, administrative and political—and quality of government. The positive bi-variate relationships are common, in accordance with the general expectation of the literature, but the the patterns are not entirely coincident. Therefore the first two hypotheses are: H1 :Decentralization has an independent and positive effect on the quality of government; H2 :Political, fiscal and administrative types of decentralization have different effects on the quality of government. 3.2 |The role of democracy and accountability? The specific role played of specific political and social institutions also cannot be neglected either. Karlström (2015) notes that the potential for decentralisation to curb corruption is dependent on the presence of institutions that give citizens information on government behaviour and the capacity to act upon the given information. In the same fashion (Lessmann and Markwardt (2010)) presents empirical results suggesting that the benefits of decentralisation only occur if there is a supervisory body such as a free press, which is often lacking in a substantial number of countries. His policy implication is that a free press is a necessary pre-condition for successful decentralisation programmes. Democracy plays a determinant role in this link between decentralisation and governance because it is the most important institutional setting to have citizens informed about government behaviour. Again, the empirical results on role of democracy remain uncertain. For one hand, O'Dwyer and Ziblatt (2006) found that the effects of decentralisation are minimal when controlling for basic structural variables such as the degree of democracy and per capita GDP. For the CAMÕES 3of12 14791854, 2022, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pa.2435 by Cochrane Portugal, Wiley Online Library on [22/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License other, the effect of democracy is however not found in Kyriacou and Roca-Sagalés (2011a) study. Their main conclusion, that fiscal decentralisation has no effect on government quality when accompanied by political decentralisation, does not change even when controlling for the degree of democratic maturity across countries. Figure 2depicts two different patterns in the relationship observed in Figure 1depending on which group we look at: democratic and nondemocratic countries. Therefore, in line with Karlström (2015), the second hypothesis considers how the level of democracy moderates the relationship between decentralisation and the quality of government. Lastly, the following third hypothesis concerningtheroleofdemocracyis: H3 :The effect of decentralization is attenuated when the effect of democracy is considered. 4|EMPIRICAL STRATEGY In accordance with the hypotheses presented in the previous section, the main objective of this paper is to empirically test (a) the independent effect of decentralisation on the quality of government; (b) to distinguish the differences among the types of decentralisation and (c) address the role played by democracy in the relationship. A worldwide analysis is required, and the following stylised model is tested. QoGi=αDeci+βDemi+γXi+ϵ, where QoG is the quality of government, Dec is the decentralisation (political, fiscal and administrative), Dem is the index of democracy, and Xis a set of control variables. Following the point made in Adam, Delis, and Kammas (2014) that the effect of decentralisation on governance is not linear but determined by an inverted U-shaped relationship, decentralisation is included in the regressions as a second-degree polynomial Deci=Di+D2 i: Most of the data used in the analysis comes from the QoG Standard data set of the Quality of Government Institute (University of Gothemburg), which assembles an impressive number of variables from many different sources (Teorell et al., 2019). The exception was variables taken directly from Ivanyna and Shah (2012) who produced 0 .2 .4 .6 .8 1 Quality of Government 0 .2 .4 .6 .8 1 Political Decentralization Index 0 .2 .4 .6 .8 1 Quality of Government 0 .2 .4 .6 .8 1 Fiscal Decentralization Index 0 .2 .4 .6 .8 1 Quality of Government 0 .2 .4 .6 .8 1 Administrative Decentralization Index (a) (c) (b) FIGURE 1 A–C, Political, fiscal, and administrative decentralisation and quality of government 4of12 CAMÕES 14791854, 2022, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pa.2435 by Cochrane Portugal, Wiley Online Library on [22/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Albania Argentina Australia Austria Bahamas Bangladesh Belgium Bolivia Botswana BrazilBulgaria Canada Chile Colombia Costa Rica Croatia Cyprus Czech Republic Denmark Dominican Republic Ecuador El Salvador Estonia Finland France Germany Ghana Greece Guatemala Guyana Honduras Hungary Iceland India Indonesia Ireland Israel Italy Jamaica Japan Kenya Korea, Rep. Latvia Liberia Lithuania Luxembourg Malawi Mali Malta Mexico Mongolia Netherlands New Zealand Nicaragua Norway Pakistan Panama Papua New Guinea Paraguay Peru Philippines Poland Portugal RomaniaSenegal Serbia Sierra Leone Slovenia South Africa Spain Suriname Sweden Switzerland Taiwan Trinidad and Tobago Turkey Ukraine United Kingdom United States Uruguay ZambiaAlgeria Angola Armenia Azerbaijan Bahrain Belarus Burkina Faso Cameroon China Cuba Egypt Ethiopia Gabon Gambia Guinea Guinea−Bissau Haiti Iraq Jordan Kazakhstan Kuwait Lebanon Madagascar Malaysia Morocco Mozambique Namibia Niger Nigeria Oman Qatar Saudi Arabia Singapore Somalia Sri Lanka Thailand Togo Tunisia Uganda United Arab Emirates Venezuela Yemen Zimbabwe 0 .2 .4 .6 .8 1 Quality of Government 0 .2 .4 .6 .8 1 Political Decentralization Index Albania Argentina Australia Austria Bahamas Bangladesh Belgium Bolivia Botswana BrazilBulgaria Canada Chile Colombia Costa Rica Croatia Cyprus Czech Republic Denmark Dominican Republic Ecuador El Salvador Estonia Finland France Germany Ghana Greece Guatemala Guyana Honduras Hungary Iceland India Indonesia Ireland Israel Italy Jamaica Japan Kenya Korea, Rep. Latvia Liberia Lithuania Luxembourg Malawi Mali Malta Mexico Mongolia Netherlands New Zealand Nicaragua Norway Pakistan Panama Papua New Guinea Paraguay Peru Philippines Poland Portugal RomaniaSenegal Serbia Sierra Leone Slovenia South Africa Spain Suriname Sweden Switzerland Taiwan Trinidad and Tobago Turkey Ukraine United Kingdom United States Uruguay ZambiaAlgeria Angola Armenia Azerbaijan Bahrain BelarusBurkina Faso Cameroon ChinaCuba Egypt Ethiopia Gabon Gambia Guinea Guinea−Bissau Haiti Iraq Jordan Kazakhstan Kuwait Lebanon Madagascar Malaysia Morocco Mozambique Namibia Niger Nigeria OmanQatar Saudi Arabia Singapore Somalia Sri Lanka Thailand Togo Tunisia Uganda United Arab Emirates Venezuela Yemen Zimbabwe 0 .2 .4 .6 .8 1 Quality of Government 0 .2 .4 .6 .8 1 Fiscal Decentralization Index Albania Argentina Australia Austria Bahamas Bangladesh Belgium Bolivia Botswana BrazilBulgaria Canada Chile Colombia Costa Rica Croatia Cyprus Czech Republic Denmark Dominican Republic Ecuador El Salvador Estonia Finland France Germany Ghana Greece Guatemala Guyana Honduras Hungary Iceland India Indonesia Ireland Israel Italy Jamaica Japan Kenya Korea, Rep. Latvia Liberia Lithuania Luxembourg Malawi Mali Malta Mexico Mongolia Netherlands New Zealand Nicaragua Norway Pakistan Panama Papua New Guinea Paraguay Peru Philippines Poland Portugal RomaniaSenegal Serbia Sierra Leone Slovenia South Africa Spain Suriname Sweden Switzerland Taiwan Trinidad and Tobago Turkey Ukraine United Kingdom United States Uruguay ZambiaAlgeria Angola Armenia Azerbaijan Bahrain BelarusBurkina Faso Cameroon ChinaCuba Egypt Ethiopia Gabon Gambia Guinea Guinea−Bissau Haiti Iraq Jordan Kazakhstan Kuwait Lebanon Madagascar Malaysia Morocco Mozambique Namibia Niger Nigeria OmanQatar Saudi Arabia Singapore Somalia Sri Lanka Thailand Togo Tunisia Uganda United Arab Emirates Venezuela Yemen Zimbabwe 0 .2 .4 .6 .8 1 Quality of Government 0 .2 .4 .6 .8 1 Fiscal Decentralization Index (a) (b) (c) FIGURE 2 Decentralisation and democracy TABLE 1 Summary statistics Variable Mean Std. Dev. Min. Max. N Quality of government 0.532 0.21 0.083 0.972 130 Government effectiveness −0.043 0.993 −2.27 2.168 177 Control of corruption −0.047 0.998 −1.609 2.41 177 Regulatory quality −0.05 0.982 −2.518 1.965 177 Democracy index 34.78 27.029 0 92.423 173 Political decentralisation 0.488 0.235 0 1 154 Fiscal decentralisation 0.35 0.248 0.06 1 154 Administrative decentralisation 0.32 0.28 0 0.9 154 GDP per capita, log 8.871 1.298 5.806 11.594 177 Government expenditure (% GDP) 33.553 11.349 13.966 73.252 164 Schooling years 8.227 3.131 1.241 13.424 132 Population density, log 4.29 1.432 0.61 9.84 177 Area, log 11.299 2.696 0.693 16.611 177 CAMÕES 5of12 14791854, 2022, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pa.2435 by Cochrane Portugal, Wiley Online Library on [22/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License TABLE 2 Decentralisation and quality of government: cross-sectional results Variables Political Decentralisation Fiscal Decentralisation Administrative Decentralisation Decentralisation 0.053*(0.031) 0.003 (0.025) 0.181*** (0.031) 0.067*** (0.024) 0.083** (0.038) 0.038 (0.024) Effective democracy index 0.006*** (0.001) 0.005*** (0.001) 0.005*** (0.001) GDP per capita, log 0.079*** (0.024) 0.053*** (0.014) 0.060*** (0.018) 0.048*** (0.013) 0.084*** (0.025) 0.054*** (0.015) Government expenditure (percent of GDP) 0.005** (0.002) 0.002 (0.001) 0.007*** (0.002) 0.003** (0.001) 0.006*** (0.002) 0.003*(0.001) Average schooling years 0.008 (0.009) −0.005 (0.006) 0.004 (0.008) −0.006 (0.006) 0.005 (0.009) −0.007 (0.007) Population density, log 0.008 (0.015) 0.001 (0.010) −0.003 (0.012) −0.004 (0.008) 0.011 (0.015) 0.000 (0.009) Area, log −0.006 (0.009) 0.000 (0.006) −0.018** (0.009) −0.006 (0.006) −0.010 (0.011) −0.003 (0.006) Colonial origin = Dutch 0.078*(0.041) 0.011 (0.031) 0.085** (0.037) 0.022 (0.032) 0.105** (0.043) 0.023 (0.032) Colonial origin = Spanish −0.022 (0.074) 0.022 (0.051) 0.014 (0.057) 0.028 (0.047) −0.016 (0.073) 0.022 (0.049) Colonial origin = US 0.060 (0.049) 0.052 (0.032) 0.083** (0.041) 0.056*(0.031) 0.066 (0.056) 0.043 (0.035) Colonial origin = British −0.002 (0.045) 0.045 (0.036) 0.021 (0.044) 0.048 (0.036) −0.004 (0.049) 0.045 (0.037) Colonial origin = French −0.044 (0.054) 0.019 (0.049) −0.030 (0.049) 0.016 (0.048) −0.045 (0.054) 0.018 (0.048) Colonial origin = Portuguese −0.044 (0.097) −0.050 (0.058) −0.122 (0.083) −0.078 (0.058) −0.062 (0.100) −0.058 (0.063) Colonial origin = Australian −0.161*** (0.058) 0.038 (0.050) −0.038 (0.053) 0.055 (0.047) −0.122** (0.058) 0.047 (0.046) REGION = Oceania 0.197** (0.075) 0.026 (0.051) 0.109*(0.064) 0.016 (0.049) 0.182** (0.079) 0.023 (0.050) REGION = America −0.072 (0.059) −0.065 (0.042) −0.079 (0.054) −0.069*(0.042) −0.074 (0.064) −0.068 (0.042) REGION = Asia −0.035 (0.063) 0.059 (0.043) −0.022 (0.054) 0.055 (0.042) −0.061 (0.063) 0.050 (0.042) REGION = Europa −0.039 (0.075) −0.046 (0.044) −0.074 (0.067) −0.056 (0.044) −0.103 (0.087) −0.071 (0.050) Constant −0.367*(0.207) −0.186 (0.140) −0.094 (0.209) −0.089 (0.139) −0.364 (0.249) −0.160 (0.146) Observations 102 100 100 100 100 100 R 2 0.716 0.885 0.797 0.895 0.728 0.890 Note: Robust standard errors in parentheses. *p< .1. **p< .05. ***p< .01. 6of12 CAMÕES 14791854, 2022, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pa.2435 by Cochrane Portugal, Wiley Online Library on [22/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License TABLE 3 Decentralisation and quality of government: alternative measures Variables Government Effectiveness Control of Corruption Regulatory Quality Political decentralisation −0.030 (0.083) −0.117 (0.096) −0.061 (0.099) Fiscal decentralisation 0.271*** (0.102) 0.417*** (0.141) 0.191 (0.115) Administrative decentralisation 0.204*** (0.076) 0.272** (0.105) 0.179** (0.087) Effective democracy index 0.024*** (0.002) 0.022*** (0.002) 0.023*** (0.002) 0.032*** (0.002) 0.028*** (0.003) 0.031*** (0.002) 0.023*** (0.003) 0.020*** (0.003) 0.022*** (0.003) GDP per capita, log 0.293*** (0.078) 0.278*** (0.075) 0.298*** (0.081) 0.276*** (0.082) 0.249*** (0.077) 0.279*** (0.086) 0.206** (0.081) 0.194** (0.080) 0.209** (0.083) Government expenditure (% of GDP) −0.002 (0.004) −0.000 (0.004) −0.001 (0.004) 0.001 (0.006) 0.004 (0.005) 0.003 (0.006) −0.016*** (0.006) −0.014*** (0.005) −0.014** (0.005) Average schooling years 0.011 (0.026) 0.008 (0.024) −0.000 (0.026) −0.043 (0.027) −0.050** (0.025) −0.060** (0.028) 0.012 (0.028) 0.008 (0.027) 0.001 (0.028) Population density, log 0.019 (0.038) −0.002 (0.035) 0.010 (0.035) −0.067 (0.040) −0.105*** (0.039) −0.085** (0.037) 0.034 (0.038) 0.016 (0.038) 0.023 (0.037) Area, log −0.011 (0.023) −0.039 (0.024) −0.032 (0.020) −0.059** (0.026) −0.108*** (0.027) −0.094*** (0.025) −0.019 (0.027) −0.042 (0.028) −0.041 (0.027) Colonial origin = Dutch −0.213* (0.110) −0.180* (0.106) −0.163 (0.101) −0.377*** (0.129) −0.319** (0.129) −0.303** (0.124) −0.360** (0.141) −0.333** (0.142) −0.313** (0.144) Colonial origin = Spanish 0.125 (0.196) 0.119 (0.185) 0.113 (0.194) 0.176 (0.214) 0.160 (0.195) 0.152 (0.211) 0.149 (0.251) 0.141 (0.241) 0.135 (0.247) Colonial origin = US 0.164 (0.117) 0.144 (0.114) 0.073 (0.111) 0.018 (0.139) −0.044 (0.131) −0.138 (0.144) −0.200 (0.153) −0.232 (0.144) −0.295** (0.138) Colonial origin = British 0.094 (0.145) 0.112 (0.139) 0.090 (0.139) 0.110 (0.169) 0.146 (0.164) 0.113 (0.169) −0.110 (0.153) −0.092 (0.153) −0.110 (0.149) Colonial origin = French 0.090 (0.172) 0.092 (0.164) 0.085 (0.163) 0.156 (0.191) 0.162 (0.187) 0.153 (0.192) −0.046 (0.196) −0.043 (0.192) −0.048 (0.187) Colonial origin = Portuguese 0.172 (0.231) 0.062 (0.243) 0.127 (0.267) 0.153 (0.206) −0.015 (0.187) 0.093 (0.214) 0.267 (0.193) 0.191 (0.184) 0.228 (0.198) Colonial origin = Belgian 0.697*(0.363) 0.677*(0.376) 0.658*(0.364) 0.911 (0.754) 0.875 (0.775) 0.852 (0.755) 0.385 (0.288) 0.368 (0.290) 0.348 (0.281) Colonial origin = Australian 0.353** (0.175) 0.409** (0.165) 0.368** (0.161) −0.428* (0.223) −0.350* (0.206) −0.418* (0.210) −0.078 (0.201) −0.043 (0.205) −0.069 (0.198) REGION = Oceania −0.191 (0.167) −0.219 (0.155) −0.191 (0.151) 0.206 (0.198) 0.158 (0.180) 0.201 (0.187) 0.153 (0.207) 0.130 (0.205) 0.150 (0.200) REGION = America −0.195 (0.150) −0.187 (0.143) −0.211 (0.145) −0.221 (0.152) −0.215 (0.134) −0.249* (0.144) −0.220 (0.183) −0.218 (0.170) −0.236 (0.176) REGION = Asia 0.216 (0.168) 0.210 (0.160) 0.185 (0.165) 0.346*(0.185) 0.344** (0.172) 0.312*(0.181) 0.092 (0.206) 0.092 (0.204) 0.068 (0.202) REGION = Europa −0.150 (0.163) −0.176 (0.157) −0.279* (0.161) −0.230 (0.198) −0.271 (0.181) −0.404* (0.206) 0.124 (0.200) 0.104 (0.198) 0.009 (0.202) (Continues) CAMÕES 7of12 14791854, 2022, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pa.2435 by Cochrane Portugal, Wiley Online Library on [22/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License data on the three different types of decentralisation, but only as cross-sectional measures. 4.1 |Dependent variable: QOG The quality of government is complex and multidimensional concept. La Porta et al. (1999) use the term “good”in their article to stand for good-for-economic development, that is, good government performance. They use measures as diverse as government intervention, public sector efficiency, public good provision, size of government and political freedom in their analysis. The Quality of Government Institute developed a specific definition of QoG along the concept of the impartiality of institutions that exercise government authority Rothstein and Teorell (2008). Accordingly, in Charron and Lapuente (2013), quality of government is understood as: (a) control of corruption; (b) impartial treatment of citizens and (c) government effectiveness. This way, the primary dependent variable for the quality of government is a composite measure by the International Country Risk Guide which includes three components: ‘Corruption’,‘Law and Order’, and ‘Bureaucracy Quality’scaled 0–1. For robustness purposes, I also test three variables included in the dataset of the World Governance Indicators released by the World Bank: (a) government effectiveness; (b) control of corruption and (c) regulatory quality. 4.2 |Key independent variables: decentralisation and democracy With regard to decentralisation, we already argued for the three different types of decentralisation that expected to produce somewhat different effects (Manor, 1999). I use the three variables developed by Ivanyna and Shah (2012), the only authors that collected these types of data. 4.2.1 |Political decentralisation Political decentralisation is the transfer of power to lower levels of authority that are largely democratic and independent of the higher levels. It is the case when central government is giving up effective authority (Manor, 1999). For example, the creation of a new level of regional or local government that is democratically elected constitutes a case of political decentralisation. In the same way, when an already existing level of sub-national government has governance bodies that were previously appointed by central government but are later democratically elected, there is an increase in political decentralisation. The key aspect of political decentralisation is the increasing influence or power of organised interests at lower levels. For this reason, some authors claim this type of decentralisation to be the only genuine type of decentralisation. Although commonly used in analyses, the federal status is a very poor measure of political decentralisation because federalism is a category of systems rather than a single system of TABLE 3 (Continued) Variables Government Effectiveness Control of Corruption Regulatory Quality Constant −3.363*** (0.610) −2.911*** (0.610) −3.120*** (0.574) −2.263*** (0.577) −1.485** (0.609) −1.849*** (0.599) −1.933*** (0.681) −1.568** (0.692) −1.680** (0.665) Observations 113 113 113 113 113 113 113 113 113 R 2 0.901 0.908 0.906 0.876 0.889 0.883 0.847 0.850 0.851 Note: Robust standard errors in parentheses. *p< .1. **p< .05. ***p< .01. 8of12 CAMÕES 14791854, 2022, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pa.2435 by Cochrane Portugal, Wiley Online Library on [22/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License TABLE 4 Decentralisation and quality of government: spatial analysis Variables Political Decentralisation Fiscal Decentralisation Administrative Decentralisation Decentralisation 0.062** (0.031) 0.009 (0.021) 0.187*** (0.026) 0.077*** (0.022) 0.082*** (0.029) 0.038** (0.019) Effective democracy index 0.006*** (0.000) 0.005*** (0.001) 0.005*** (0.000) GDP per capita, log 0.076*** (0.016) 0.052*** (0.011) 0.057*** (0.014) 0.047*** (0.010) 0.083*** (0.016) 0.054*** (0.010) Government expenditure (% of GDP) 0.006*** (0.002) 0.003** (0.001) 0.008*** (0.001) 0.004*** (0.001) 0.007*** (0.002) 0.003*** (0.001) Average schooling years 0.007 (0.008) −0.005 (0.005) 0.002 (0.007) −0.006 (0.005) 0.004 (0.008) −0.007 (0.005) Population density, log 0.013 (0.012) 0.003 (0.008) 0.003 (0.010) −0.000 (0.008) 0.016 (0.012) 0.003 (0.008) Area, log 0.000 (0.008) 0.002 (0.005) −0.011 (0.007) −0.003 (0.005) −0.005 (0.008) −0.001 (0.006) Colonial origin = Dutch 0.100 (0.117) 0.022 (0.075) 0.111 (0.098) 0.039 (0.071) 0.124 (0.115) 0.034 (0.074) Colonial origin = Spanish 0.001 (0.066) 0.031 (0.042) 0.041 (0.055) 0.043 (0.040) 0.004 (0.065) 0.031 (0.042) Colonial origin = Italian 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) Colonial origin = US 0.040 (0.119) 0.043 (0.076) 0.067 (0.097) 0.049 (0.071) 0.054 (0.115) 0.038 (0.073) Colonial origin = British 0.003 (0.044) 0.046 (0.029) 0.027 (0.037) 0.050*(0.027) −0.001 (0.043) 0.046 (0.028) Colonial origin = French −0.041 (0.063) 0.019 (0.040) −0.026 (0.052) 0.015 (0.038) −0.042 (0.061) 0.018 (0.040) Colonial origin = Portuguese −0.034 (0.095) −0.045 (0.060) −0.114 (0.079) −0.076 (0.058) −0.055 (0.093) −0.054 (0.059) Colonial origin = Belgian 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) Colonial origin = British-French 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) Colonial origin = Australian −0.144 (0.143) 0.041 (0.092) −0.013 (0.120) 0.063 (0.088) −0.107 (0.141) 0.051 (0.091) REGION = Oceania 0.179*(0.096) 0.022 (0.063) 0.086 (0.081) 0.008 (0.059) 0.167*(0.095) 0.019 (0.061) REGION = America −0.081 (0.057) −0.070*(0.036) −0.089*(0.047) −0.075** (0.034) −0.081 (0.056) −0.072** (0.036) REGION = Asia −0.040 (0.049) 0.054*(0.032) −0.029 (0.040) 0.047 (0.031) −0.067 (0.048) 0.045 (0.032) REGION = Europa −0.005 (0.066) −0.030 (0.042) −0.038 (0.055) −0.038 (0.040) −0.075 (0.068) −0.058 (0.043) Constant −0.440** (0.188) −0.223*(0.123) −0.178 (0.160) −0.133 (0.118) −0.436** (0.183) −0.200*(0.120) CW (quality of government) −0.150** (0.066) −0.066 (0.042) −0.170*** (0.055) −0.090** (0.040) −0.125*(0.064) −0.061 (0.041) Observations 102 100 102 100 102 100 Observations 100 100 100 100 100 100 Note: Standard errors in parentheses. *p< .1. **p< .05. ***p< .01. CAMÕES 9of12 14791854, 2022, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/pa.2435 by Cochrane Portugal, Wiley Online Library on [22/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License