How Close is Your Government to its People? Worldwide Indicators on Localization and Decentralization

This paper is intended to provide an assessment of the impact of the silent revolution of the last three decades on moving governments closer to people to establish fair, accountable, incorruptible and responsive governance. To accomplish this, a unique data set is constructed for 182 countries by compiling data from a wide variety of sources to examine success toward decentralized decision making across the globe. An important feature of this data set is that, for comparative purposes, it measures government decision making at the local level rather than at the sub-national levels used in the existing literature. The data are used to rank countries on political, fiscal and administrative dimensions of decentralization and localization. These sub-indexes are aggregated and adjusted for heterogeneity to develop an overall ranking of countries on the closeness of their government to the people. The resulting rankings provide a useful explanation of the Arab Spring and other recent political movements and waves of dissatisfaction with governance around the world.


Policy Research Working Paper 6138
This paper is intended to provide an assessment of the impact of the silent revolution of the last three decades on moving governments closer to people to establish fair, accountable, incorruptible and responsive governance. To accomplish this, a unique data set is constructed for 182 countries by compiling data from a wide variety of sources to examine success toward decentralized decision making across the globe. An important feature of this data set is that, for comparative purposes, it measures government decision making at the local level rather than This paper is a product of the Poverty Reduction and Economic Management Unit, East Asia and the Pacific Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank. org. The author may be contacted at shah.anwar@gmail.com.
at the sub-national levels used in the existing literature. The data are used to rank countries on political, fiscal and administrative dimensions of decentralization and localization. These sub-indexes are aggregated and adjusted for heterogeneity to develop an overall ranking of countries on the closeness of their government to the people. The resulting rankings provide a useful explanation of the Arab Spring and other recent political movements and waves of dissatisfaction with governance around the world.

Introduction
A silent revolution has been sweeping the globe since the 1980s. Hugely complex factors such as political transition in Eastern Europe, the end of colonialism, the globalization and information revolution, assertion of basic rights of citizens by courts, divisive politics and citizens' dissatisfaction with governance and their quest for responsive and accountable governance have been some of the contributing factors in gathering this storm. The main thrust of this revolution has been to move decision making closer to people to establish fair, accountable, incorruptible and responsive (F.A.I.R.) governance. The revolution has achieved varying degrees of success in government transformation across the globe due to inhibiting factors such as path dependency accentuated by powerful political, military and bureaucratic elites. While there has been monumental literature dealing with various aspects of this revolution, there has not been any systemic study providing a time capsule of the changed world as a result of this revolution. Such an assessment is critical to providing a comparative world perspective on government responsiveness and accountability. This paper takes an important first step in this direction by providing a framework for measuring closeness of the government to its people and providing a worldwide ranking of countries using this framework.
The paper is organized as is four parts as follows. Part I is concerned with highlighting the conceptual underpinnings and developing a framework to measure closeness of the government to people. It presents a brief overview of conceptual underpinnings of moving governments closer to people. This is followed by a discussion of basic concepts in measuring government closeness to its people. It calls into question the methodologies followed by the existing literature and argues for a focus on the role and responsibilities of local governments as opposed to sub-national governments where intermediate order governments typically dominate. It is the first paper that advocates and treats various tiers of local governments (below the intermediate order of government) as the unit of comparative analysis for multi-order governance reforms.
Part II presents highlights of the unique dataset compiled for this study. It presents summary statistics on structure, size, tiers of local governments and security of their existence. It also presents summary statistics on the various subcomponents of political, fiscal and administrative decentralization.
Part III is concerned with empirical implementation of the framework presented in Part I. It begins by highlighting the relative importance and significance of local governments. This is followed by providing country rankings on various aspects of political, fiscal and administrative decentralization. By combining these measurements, an aggregate indicator of localization is developed for each country. This index is then adjusted for population size, area and heterogeneity. We also provides correlations of these indexes with the corruption perceptions 3 index, citizen-centered governance indicators, per capita GDP, size of the government and the ease or difficulty of doing business in the country.
Part IV provides concluding observations highlighting the strength and limitations of the constructed indexes.

Moving Governments Closer to People: Conceptual Underpinning of the Rationale and an Empirical Framework for Comparative Analysis Why Closeness of Government to Its People Matters: Conceptual Underpinnings
Several accepted theories provide a strong rationale for moving decision making closer to people on the grounds of efficiency, accountability, manageability and autonomy. Stigler (1957) argued that that the closer a representative government to its people, the better it works. According to the decentralization theorem advanced by Wallace Oates (1972. P.55), "each public service should be provided by the jurisdiction having control over the minimum geographic area that would internalize benefits and costs of such provision", because:  local governments understand the concerns of local residents;  local decision making is responsive to the people for whom the services are intended, thus encouraging fiscal responsibility and efficiency, especially if financing of services is also decentralized;  unnecessary layers of jurisdictions are eliminated;  inter-jurisdictional competition and innovation are enhanced. An ideal decentralized system ensures a level and combination of public services consistent with voters' preferences while providing incentives for the efficient provision of such services. The subsidiarity principle originating from the social teaching of the Roman Catholic Church and later adopted by the European Union has argued for assignment of taxing, spending and regulatory functions to the government closest to the people unless a convincing case can be made for higher level assignment. Recent literature have further argued that such local jurisdictions exercising such responsibilities should be organized along functional lines while overlapping geographically do that individual are free to choose among competing service providers (see the concept of functional, overlapping and competing jurisdictions (FOCJ) by Frey and Eichenberger, 1999).
Moving government closer to people has also been advanced on the grounds of creating public value. This is because local governments have the stronger potential to tap some of the resources that come as free goods -namely, resources of consent, goodwill, good Samaritan values, community spirit (see Moore, 1996).
Moving government closer to people also matters in reducing transactions costs of individuals to hold the government to account for incompetence or malfeasance -a neo institutional economics perspective advanced by Shah and Shah (2006). Finally, a network form of governance is needed to forge partnership of various stakeholders such as interest based network, hope based network, private for profit or for non-profit provides and government providers to improve economic and social outcomes. Such network form of governance is facilitated by having an empowered government closer to people that plays a catalytic role in facilitating such partnerships (see Dollery and Wallis, 2001).
In summing, a strong non-controversial case has been made by the conceptual literature to move government decision making closer to people on efficiency, accountability and responsiveness grounds. The question that is relevant is to develop a methodology for a comparative global assessment of a government's closeness to its people. This is the focus of research in the next section.

Measuring a Government's Closeness to Its People: An Empirical Framework
A government is closer to its people if it encompasses a small geographical area and population, and it enjoys home rule and cannot be arbitrarily dismissed by higher level governments. This requires an understanding of the structure, size and significance of local governments including its legal and constitutional foundation of its existence. An empirical framework for a comparative assessment must incorporate assessment of these factors. The following paragraphs elaborate on the methodology adopted in this paper to capture these elements.
Unit of analysis. The literature to-date without exception takes sub-national governments as a unit of analysis for measuring closeness to people. This viewpoint is simply indefensible. 2 This is because states or provinces in large countries such as USA, Canada, India, Pakistan, Brazil, and Russia are larger in population size and area than a large number of small or medium size countries. Having empowered provinces and states in these countries means that decision making is still far removed from the people. Also intermediate orders of government in large federal countries may be farther removed from people than the central government in smaller unitary states. Therefore it would be inappropriate to compare provinces in Canada or states in Brazil, India, or the USA with municipalities, say, in Greece. This approach also vitiates against small countries such as Liechtenstein and Singapore as these countries would be mistakenly rated as having decision making far removed from people. In view of these considerations, local governments are the appropriate unit for measuring closeness to people as implemented here.
Local government tiers. Local government administrative structure varies across countries and the number of administrative tiers varies from 1 to 5. This has also a bearing on the closeness of the government and must be taken into consideration.
Local government size. Average size of local government in terms of population and area also varies across countries and it has a bearing on potential participation of citizens in decision making. An example of potentially misleading choice of units for comparative analysis is in Fan et al 2009, where the authors create a dummy variable, which is equal to 1 when the executive bodies at the lowest tier of government are elected. As a result, say Bangladesh gets 0, and Indonesia gets 1, which suggests that at the lowest tier Indonesia is more politically decentralized than Bangladesh. However, the average population of the local government unit in Indonesia is about 0.5 million, while in Bangladesh (according to the definitions in the paper) it is about 100 people. There are elected executive bodies in Bangladesh at a level of administrative units with population even less than 0.5 million, which implies that Bangladesh is more politically decentralized than Indonesia.
Significance of local government. Whether or not local governments command a significant share of national expenditures indicates their respective role in multi-order public governance. This is important in terms of their roles and responsibilities. For example, a local government may have autonomy but only a limited and highly constrained role as in India. This needs to be taken into consideration while making judgment on closeness of government decision making to people.
Security of existence of local governments. If local governments do not have any security of existence then their autonomy can be a hollow promise. Thus safeguards against arbitrary dismissal of local governments must be examined. This is to be assessed both by de-jure the legal and or constitutional foundations of local government creation and also de-facto working of such provisions. For example, local governments in India have constitutional backing, the same in Pakistan are creatures of the provinces and in China they simply are created by an executive order. While the legal and constitutional foundations of local government in India and Pakistan are much stronger, in practice and by tradition, local governments enjoy greater security of tenure in China.
Empowerment of local government. This is to be assessed on three dimensions -political, fiscal and administrative (see Boadway andShah, 2009 andThompson, 2004).
Political or democratic decentralization implies directly elected local governments thereby making elected officials accountable to local residents.. Political decentralization is to be assessed using the following criteria: direct popular elections of council members and the executive head; recall provisions for elected officials; popular participation in local elections and the contestability and competition in local elections.
Fiscal decentralization ensures that all elected officials weigh carefully the joys of spending some else's money as well as the pain associated with raising revenues from the electorate and facing the possibility of being voted out. Fiscal decentralization is to be evaluated using the criteria: range of local functions; local government autonomy in rate and base setting for local revenues; transparency and predictability and unconditionality of higher level transfers; finance follows function or revenue means more or less match local responsibility; degree of selffinancing of local expenditures; responsibility and control over municipal and social services; autonomy in local planning, autonomy in local procurement; ability to borrow domestically and from foreign sources; ability to issue domestic and foreign bonds; and higher level government assistance for capital finance.
Administrative decentralization empowers local governments to hire, fire and set terms of reference for local employment without making any reference to higher level governments, thereby making local officials accountable to elected officials. This is to be assessed using indicators for: freedom to hire, fire and set terms of reference for local government employment; freedom to contract out own responsibilities and forge public-private partnerships; and regulation of local activities by passing bye-laws.

Description of the Data
To implement the above framework, we have developed a unique and comprehensive dataset for 182 countries using data for the most recent year of availability (mostly 2005) on the relative importance of local governments, their security of existence and various dimensions of their empowerment. The following sections introduce and analyze various dimensions of these data.

Local Government -Basic Definitions
General government (GG) consists of 3 parts: Central Government (CG), State or Provincial Government (SG), and Local Government (LG). Each part consists of governmental units (in case of CG -only 1 unit), which are united into one or more tiers (in case of CG -1 tier). As far as data permits, Social Security Funds are consolidated with an appropriate part of GG. We use commonly accepted definitions of LG and SG as provided by the IMF Government Finance Statistics (GFS). These definitions are quite vague which results into countries deciding for themselves and reporting corresponding data. This sometimes leads to inconsistencies. For example, France with three sub-national tiers of government reports all of them as LG, whereas Spain -which in many ways has the same administrative structure as France -reports one tier of SG, and two tiers of LG. Giving more precise definitions for LG and SG, which could be applied to all countries, is a difficult task. In constructing a comparative data set, we have attempted to correct for these self-reporting biases by using country specific research studies where available to make a distinction between SG and LG tiers.

Tiers of Local Government
Our dataset contains detailed information about administrative structure of every country. In particular, we report which tiers of GG are ascribed to a local government, and number of governmental units at each tier. Tiers are needed to calculate the average population of LG administrative unit as follows: where LG-pop is the average population of an LG unit, T is the number of tiers in the country, P is its population, and X is the number of LG units at the i'th tier.
Of the sample of 182 countries only 20 have state governments (SG), while the rest of the countries have only local and central governments. 26% of the countries have one tier of local government, 46% have two tiers, while 23% and 6% have three and four tiers respectively. 7

Figure 1: Number of Tiers of Local Government -World Map
Source: Authors' calculations based upon data sources reported in Annex Table A1.
Note: Color of a country corresponds to its percentile in the world's distribution: red -0-25 th , yellow -25-50 th , blue -50-75 th , green -75-100 th . Figure 1 shows the world map, where darker shades represent countries having more tiers of local government. Table 1 reports analysis of these tiers by geographic region and by country per capita income. World regions on average have two LG tiers with South Asia and the East Asia regions having above average number of tiers. High income countries however, tend to allow lower number of LG tiers as compared to lower income countries.

Average Population Size of Local Government Units
The average tiers-adjusted population of a local government unit ranges from about several thousand people (Equatorial Guinea, Switzerland, Czech Republic, Austria) to several hundred thousand people (Somalia, DR Congo, Indonesia, Korea), with the country-average population of 101,000 people. As shown in Figure 2 (see also Table 1) local governments in European and North American countries are significantly smaller in population size than the ones in the rest of the world, while the LG in Sub-Saharan Africa and East Asia are on average more than five times larger. Lower income countries have significantly larger population size governments.

Figure 2: Population of Local Governments -World Map
Source: Authors' calculations based upon data sources reported in Annex Table A1.

Average Area of Local Government Units
The average area of a local government unit ranges from 0.01 thousand square kilometers (TSK) in Czech Republic to 70 TSK in Libya, with the cross-country average of 2.1 TSK. European and South Asian countries have relatively much smaller area size local government units, while Africa and Middle East have average LG areas of up to 14 times larger.
LG in higher income countries are generally smaller in average area than the ones in lower income countries (see Table 1 and Figure 3).
The overall pattern observed here is that higher income countries on average tend to have smaller size (both in terms of population and area) local governments with fewer tiers than lower income countries.

Figure 3: Area of Local Government -World Map
Source: Authors' calculations based upon data sources reported in Annex Table A1.

The Significance of Local Government: Relative Importance and Security of Their Existence
Measurement of relative importance of local government and constitutional safeguards regarding arbitrary disbandment are critical to reaching a judgment about closeness of the government to its people. The following paragraphs highlight the variables used in this measurement.

(a) Relative Importance of Local Governments
The relative importance of local governments is measured by share of LG expenditures(lgexpdec) in consolidated general government expenditures for all orders of government (GG). This is obviously an imperfect measure of relative importance of local governments as a significant part of local government expenditures may simply be in response to higher level government mandates with little local discretion. However, data on autonomous local government expenditures are simply not available.

Figure 4: Relative Importance of Local Governments and Their Independence -World Maps
Source: Authors' calculations based upon data sources reported in Annex Table A1.
LG share of GG expenditures varies greatly over our sample -from virtually zero percent in a number of countries (Guyana, Mozambique, Haiti, etc.) to 59 percent in Denmark, and have near chi-square distribution with one degree of freedom. A large majority of countries (63 percent) have local government expenditure shares less than the sample average of 13 percent, and only 11 percent of the countries have LG expenditures shares higher than 30 percent. Only in Europe, East Asia and North America, local governments are important players in the public sector.
An alternate variable that could serve as a proxy for the relative importance of LG is LG employment (lgempl): share of LG employment in GG employment. The available data on this variable are however much less reliable and shows a great deal of year to year volatility for most developing nations. In view of this, we are left with no alternative but the use of expenditure shares as the only variable to measure the relative importance of local governments.
LG employment is used in calculation of administrative decentralization index.

(b) Security of Existence of Local Governments
Local government security of existence is measured by LG independence(lgindep). This measure attempts to capture the constitutional and legal restraints on arbitrary dismissal of local governments.
Only in 6 out of 182 countries, local governments have significant safeguards against arbitrary dismissal. . LG in 48 percent of the countries have limited independence and for the remaining 49 percent of countries in our sample, local governments can be arbitrarily dismissed by higher order governments. Europe, North America and Brazil receive relatively higher scores on this indicator whereas local governments in Africa and the Middle East have almost no security of existence.
LG Relative Importance LG Independence

Local Government Empowerment
Local government empowerment is measured on fiscal, political, and administrative dimensions as discussed below.

(a) Fiscal Decentralization
The following variables are used to assess local government fiscal autonomy. • LG vertical fiscal gap(lgvergap). Vertical fiscal gap refers to the fiscal deficiency arising from differences in expenditure needs and revenue means of local government. These deficiencies are partially or fully overcome by higher level financing. Therefore, vertical fiscal gap is a measure of fiscal dependence of local government on higher level financing. The design and nature of higher level financing has implications for fiscal autonomy of local governments. It must therefore be recognized that vertical fiscal gap while being a useful concept cannot be looked in isolation of a number of related indicators to have a better judgment on local fiscal autonomy as done here. The average vertical gap in the world is 52 percent. It is somewhat higher in African and Latin American countries. However, in all regions there are local governments with high share of expenditures and high reliance on financing from above (e.g. Brazil), as well as almost non-existent LG governments that rely solely on their own financing (Togo, Niger).

•
LG taxation autonomy (lgtaxaut). This measure reflects upon a local government's empowerment and access to tools to finance own expenditures without recourse to higher level governments. It measures its ability to determine policy on local taxation (determining bases and setting rates) and as well as autonomy in tax collection and administration. Only 16 percent of the countries in our sample grant significant taxation autonomy to their LGs, while the rest grant limited or no tax autonomy to their local governments.

13
• LG unconditional transfers (lgtransf). Unconditional, formula based grants preserve local autonomy. Such grants are now commonplace yet conditional grants still dominate. Europe and North America, Latin America and Southern Asia regions have high percentage of countries with high scores on this indicator.


LG Expenditure Autonomy. Measured by share of LG expenditures in total GG expenditures this variable does not fully reflect the actual expenditure discretion that local governments have. First, LG may be simple distributors of the funding transferred to them from an upper-tier government, and have little choice over how the money in their budget should be spent. If the LG vertical gap (difference between LG expenditures and LG non-transfer revenues) is wide, and if the transfers from upper-tier governments are earmarked and discretionary, the actual spending power of LG may be much lower than it would be indicated by lgexpdec. Second, even the own revenues of LG (tax revenues or borrowed funds) may strongly depend on CG policy. If LG are not allowed to regulate taxes without CG interference (usually in such cases they receive a revenue-share of a tax, which is regulated by CG), then they cannot fully rely on the revenues from these taxes, and their policy would still be partly dependent on CG.
We adjust for the first argument -that the real LG expenditure autonomy depends on the vertical gap and the structure of intergovernmental grants -by defining LG expenditure autonomy variable (lgexpdiscr): Note from (2), that even if a country has widest possible vertical gap (1), and smallest possible share of unconditional formula-based transfers (0) it still keeps 0.25 share of its original expenditure decentralization. This is to reflect the fact that discretionary conditional grant from 14 CG still gives more autonomy to the LG than the direct spending of CG. At the same time, country with a positive vertical gap and best possible set of transfers gets less than lgexpdecshare of it. This is to reflect the fact that even the best set of transfers does not give LG as much fiscal independence as its own revenues. • LG borrowing freedom (lgborrow). Can LG borrow money to satisfy their capital finance needs? Can the borrowing be done without consent or regulation of CG? 89 of 160 countries in our sample forbid any kind of borrowing by LGs, while only in 22 countries LGs are allowed to borrow without any restrictions. Local borrowing rules are more accommodating in Europe and Latin America.
The descriptions, definitions and sample distributions of fiscal decentralization variables that we use are reported in Tables 7 and 8, and Figure 6 displays corresponding world maps.  Annex Table A1. Note: Color of a country corresponds to its percentile in the world's distribution: red -0-25 th , yellow -25-50 th , blue -50-75 th , green -75-100 th .

(b) Political Decentralization
Political decentralization refers to home rule for local self-governance. This is examined using the following criteria. • LG legislative election(lglegel). Are legislative bodies at the local level elected or appointed? Is the truth somewhere in between? (For example, part of council members is appointed, part is elected, or members of councils are elected from preapproved by CG list.) Elected local councils are now commonplace around the world with only 34 percent of the countries in the sample having any restraints on popular elections of legislative councils at the local level, and only 14 countries have appointed local councils . Middle East and Sub-Saharan Africa are lagging behind the rest of the world in permitting directly elected local councils.

•
LG executive election (lgexel). Are executive heads (mayors) at the local level elected - LG Vertical Gap LG Unconditional Transfers LG Taxation Autonomy LG Expenditure Autonomy LG Borrowing Freedom 16 directly or indirectly -or appointed? Direct elections of mayors are not yet commonplace with some restrictions on direct elections in 79 percent of the countries. Thirty-six countries have no restrictions, while in 36 countries mayors are appointed at all LG tiers. While Africa and Middle East are traditionally lagging behind, European countries also receive relatively low scores on this indicator as most of the countries have some tiers of local government with appointed or indirectly elected mayors.
• Direct democracy provisions (lg_dirdem). Are there legislation provisions for obligatory local referenda for major spending, taxing and regulatory decisions, recall of public officials, and requirement for direct citizen participation in local decision making processes?
Only three countries in our sample (Switzerland, Japan and USA) have direct democracy provisions (as defined in Table 5 Table A1. Note: Color of a country corresponds to its percentile in the world's distribution: red -0-25 th , yellow -25-50 th , blue -50-75 th , green -75-100 th .

(c) Administrative Decentralization
Our concern here is to measure the ability of local governments to hire and fire and set terms of employment of local employees as well as regulatory control over own functions. As the latter data are not available, we are constrained to measure administrative decentralization simply by the first set of variables as follows.
LG Legislative Election LG Executive Election LG Direct Democracy

•
LG HR policies (lghrpol). Are LG able to conduct their own policies regarding hiring, firing and setting terms of local employment? Only 43 of 158 countries allow their LGs full discretion regarding whom and at what terms to hire or fire. Europe, North America, Australia, and Latin America are leaders on this indicator. Many more countries (77) make this kind of decisions only at the central level even for local employees.
LG employment (lgempl): share of LG employment in GG employment. Country average for LG employment is estimated to be 26 percent. However, about 34 percent of the countries in our sample report more than 30 percent of public workforce to be employed at the local level.
The descriptions, definitions and sample distributions of administrative decentralization variables are reported in Tables 9 and 10, Figure 6 displays corresponding world maps.

Figure 7: Administrative Decentralization Variables -World Maps
Source: Authors' calculations based upon data sources reported in Annex Table A1. Note: Color of a country corresponds to its percentile in the world's distribution: red -0-25 th , yellow -25-50 th , blue -50-75 th , green -75-100 th .

Worldwide Ranking of Countries on Various Dimensions of Closeness of Their Governments to the People
Our main assumption is that decentralization to local governments matters only when local governments are important players in the public sector as measured by their share of general government expenditures, and have security of existence. Indeed, it is hard to believe that local governments -however politically or administratively independent they are from the centerhave little ability to serve their residents if they do not command significant budgetary resources and if they can be dissolved at will by a higher order government. These two variables adjusted by the degree of political, fiscal and administrative decentralization form the basis of our aggregate country rankings on "closeness" or "decentralization" nexus.
In the following, political, fiscal and administrative decentralization sub-indexes are first constructed for sample countries. These indexes are then aggregated to develop a composite index of government's closeness to its people -the so-called "decentralization index". Finally this index is adjusted for heterogeneity and size of LGs.

Fiscal Decentralization Index
The formula for our fiscal decentralization index (fdi) is the following: Where lg_expaut is local expenditure autonomy, lg_taxaut is tax autonomy and lg_borrow represents legal empowerment for local borrowing. This index penalizes those countries, where LG do not have taxation autonomy nor borrowing freedom, however, it may still be positive for these countries (equal to 0.25 share of lg_expaut) reflecting the fact that own revenues do grant some degree of discretion to LG. At the same time, countries with full taxation autonomy and borrowing freedom get an index, which is equal to lg_expaut.
If there is no data on lg_taxaut or lg_borrow then the worst possible values are assumed: lg_taxaut=lg_borrow=0.
LG HR Policies LG Employment 20 Figure 8: Fiscal, Political, Administrative Decentralization Indexes -World Maps Source: Authors' calculations based upon data sources reported in Annex Table A1.

Political Decentralization Index
This index is constructed by simply taking the average variables described in the earlier section: Every variable discussed above is an essential and independent part of political decentralization. Therefore, taking the average of all variables seems to be a reasonable measure.
The index is calculated for 182 countries.

Administrative Decentralization Index
Administrative decentralization index (adi) is constructed as follows: The index is built for 182 countries. The Aggregate Decentralization Indexes The aggregate index (di) incorporates the relative importance of LG (measured by lg_expdec), the security of existence of LG (measured by lg_indep), and fiscal, political and administrative indexes. It is constructed as follows: The index penalizes countries with low political and administrative decentralization, but even if pdi=adi=0 the index is still positive if LG have some fiscal autonomy and security of existence. It reflects the fact that even fully subordinated LG without any considerable administrative responsibilities still makes fiscal decisions in more decentralized way than the CG. It also smoothes measurement errors that can be contained in our measures of political, administrative decentralization, and security of existence.
This index is constructed for 158 countries worldwide. Together they comprise 98% of the world's GDP, and 99% of the world's population. The Figure 8 depicts distribution of the decentralization index on the World map. The darker the color of a country, the more decentralized it is. European countries, North America, Brazil, and China receive high scores on this index. Countries from Latin America, former Soviet Union, and East Asia receive average decentralization index, while Middle East and African countries are the least decentralized.  Table A1.

Developing the Government "Closeness" Index by Adjusting the Decentralization Index for Heterogeneity of Size and Preferences
Our main premise is that the decentralization brings government decision making closer to the people. The decentralization indexes reported earlier indicate the significant local governments are in policymaking and public service delivery responsibilities in any country. These indexes do not fully capture the actual closeness of local governments to people. This is because local governments vary widely in population, area and diversity of preferences of residents. For example, Indonesia has average LG unit population size of 0.5 mln people, while in Switzerland, for instance, the average local government population size is only 3 thousands. Population of such countries as Malta, Iceland, Belize, Maldives, etc. is lower than 0.5 mln people. It is obvious that in most aspects, e.g. accounting for heterogeneous preferences, being accountable and known to people, etc., even central governments in these countries are closer to people that the LG in Indonesia. Therefore, the decentralization indexes need to be adjusted for LG population and area and other measures of a country's heterogeneity.
Our procedure of the adjustment is the following. Suppose we have a country with decentralization index β, average population of LG unit N, and heterogeneity index α. Heterogeneity index is based on average area of LG unit, ethno-linguistic, age, income, urbanization composition of the country's population, as well as its geographical features (relief, versatility of climatic zones, etc.). Each resident of the country has different preferences regarding the level of governmental services provided. If an average LG provides x units of the service then the disutility of a resident i is , where f is some function of two arguments. Disutility increases with the distance between the decision of the government and the preference of the resident, and all things equal, disutility increases with heterogeneity of the country, i.e. residents are more distant in their preferences in more heterogeneous countries. Governments are assumed to be benevolent, and minimize the aggregate disutility of all residents in a region they are in charge of. Since we assume symmetric distribution of preferences in the region, benevolent government would provide N/2 units of the service -a level preferred by the median resident.
Given the assumptions above, the question we ask is what decentralization index should (β,N,α)country have in order to produce a disutility of an average resident equal to the one in (β, )country, a country with the same decentralization index β, but some benchmark levels of average LG unit population and heterogeneity index? The answer to this question is follows from the identity below: where AD(N,α) is the disutility of an average resident in LG with population N and heterogeneity index α, given that the government sets its service to satisfy the median resident. AD can be found from the following expression: where in the above equation we use approximation of a sum with the integral (to simplify calculations), and our assumption about symmetric around median preferences.
For our calculation of decentralization index adjustment we take the following f : where parameter A allows us to control the sensitivity of our results to large differences in average LG unit population. Given f, the AD from (8) becomes: First, we assume there is no heterogeneity, i.e. α=0. By choosing different A's we consider three scenarios: sensitive (A=0.01), moderate (A=0.1), and conservative (A=1). Then we introduce heterogeneity in the moderate scenario. First, our α is only based on the average LG unit area.

Figure 10: Government Closeness Index -World Map
Source: Authors' calculations based upon data sources reported in Annex Table A1. Note: Color of a country corresponds to its percentile in the world's distribution: red -0-25 th , yellow -25-50 th , blue -50-75 th , green -75-100 th .
Then the heterogeneity index is extended to account for additional variables. These are age, residency, income, ethnic, religious, linguistic structure of population, country's area, relief heterogeneity (difference between highest and lowest points), and climate heterogeneity (difference between highest and lowest latitude). Table 13 presents top ten leaders in each of the five new indexes (columns 2-6), each corresponding to adjustments presented above. The decentralization index without adjustments is presented in column 1. As is suggested by the name, the conservative scenario adjustment (A=1) results in the smallest changes. Yet, Finland, Switzerland, USA, Iceland move up the ladder as the countries with traditionally small governments. On the other hand, countries with large average LG population e.g. China, Japan, and Republic of Korea have their rankings lowered. Moving from conservative to sensitive scenario, countries with small LG continue to get relatively higher indexes. Switzerland is the most decentralized country with this kind of adjustment, Iceland is the second. More European countries (Hungary, Georgia, Czech Republic) enter the list of leaders instead of Asian countries. Adjustment for area and heterogeneity do not change the ranking much, which may suggest that the adjustment procedure is too conservative. The only notable difference is that Switzerland gets lower index (moves down from 1st to 2nd place) because of its linguistic and ethnic heterogeneity. Figure 10 shows the distribution of our final Government Closeness Index in the world.

Relationship of the Decentralization Indexes with Government Size, Incidence of Corruption, Ease of Doing Business and Incomes and Good Governance
In the Table 14 we present simple OLS regressions of our decentralization indexes (and lg_expdec -a standard measure of decentralization in the literature) on disaggregate decentralization indicators, corruption measures (TI corruption perception index), development measures (GDP per capita), size of the government (GG consumption, % of GDP), number of procedures in a country needed to start a new business (Start of business, # proc.), number of civil conflicts in a country (# civil conflicts), strength of country's democratic institutions (Democraty score), durability of political regime in a country (Durability of regime), and citizencentric governance indicators (CGI) as reported in . We report both regressions with no other controls apart from corresponding economic indicator and regressions, where we also control for level of development of a country (measured by GDP per capita). These regressions indicate that decentralized governance is associated with higher per capita GDP , lower incidence of corruption (higher corruption perception index), better environment for doing business, and higher durability of political regime -even controlling for the level of development. We also find that decentralization is associated with lower government consumption, higher citizen-centric governance performance, and stronger democracy institutions, although the relationship with these variables looses significance (but keeps sign) when controlling for the level of development.
When decentralization is measured only by lg_expdec the statistical associations between decentralization and our selected economic indicators have generally lower significance (i.e. have lower t-statistics). At the same time, decentralization index adjusted for heterogeneity and LG population generally produces higher regression coefficients than unadjusted decentralization index.

Concluding Remarks
The silent revolution of the past two decades has attracted strong policy and research attention worldwide. The assessment of the impact of this revolution in moving decision making closer to the people, however, remains an unanswered question. This paper takes an important first step in this direction by providing a framework of comparative measurement and developing worldwide ranking of countries on people empowerment on various aspects of government decision making. While there is a crying need for systematic collection of quality data needed for the application of the comparative framework presented here, the integration of available diverse dataset as done here has yielded promising results. For example, the closeness indexes presented here show that one could have predicted well in advance with a fair degree of accuracy countries that were ripe for popular people revolt such as the one experienced through the Arab Spring or similar movements across the globe. The indexes also provide useful barometers of the enabling environment for doing business or promoting growth and economic development and good governance. Overall they provide useful aggregate measures of government closeness to their people. We hope this paper will stimulate further research to improve upon the data and the methodology presented here as well as facilitate building common consensus in countries poorly ranked here for fundamental governance reforms.