Political Decentralization,Subnational Political Capital,and Intergovernmental Transfers in Korea

Abstract

Faced with ever increasing pressures for better performance and financial bottom line, governments around the world are devolving more responsibilities to subnational governments. Especially in developing countries, this trend has coupled with increased demands for greater democracy and disaffection with the services provided by the central government. This article examines what has happened in South Korea since its devolution reform in 1995. Specifically, it examines political capital factors in determining the amount of intergovernmental transfers to the regional governments. The analysis shows that political decentralization has resulted in financial weakening of the subnational governments of general purposes and that although the incidence of intergovernmental transfer is affected by the political considerations, Korea has substantially improved horizontal fiscal imbalance.

Keywords

devolution fiscal decentralization intergovernmental transfers fiscal imbalance political capital South Korea

Introduction

Faced with ever-increasing pressures for better performance and financial bottom line, governments around the world are devolving more responsibilities to subnational governments. Especially in developing countries, increased demands for greater democracy and disaffection with the services provided by the central government have prompted politicians to decentralize political authorities and fiscal resources to subnational governments (Rodden, Eskeland, & Litvack, 2003). More responsibilities assumed by the subnational government with a limited base of own source revenues spotlight the importance of the design of the fiscal dimension of intergovernmental system (Bird & Vaillancourt, 2006; Litvack, Ahmad, & Bird, 1998; Musgrave, 1959). Against this backdrop, intergovernmental transfers (IGTs) have drawn careful scholarly attention during last three decades in the United States as well as around the world. Especially, developing and transition countries have offered a valuable testing ground for theoretical and practical issues in designing intergovernmental fiscal relations (Bird & Vaillancourt, 1998; Boadway & Shah, 2007; Ichimura & Bahl, 2009; Kincaid & Shah, 2007).

The Republic of Korea (hereafter Korea) is not an exception: While different general and earmarked revenue sharing mechanisms have historically been tried, currently more than 60% of county governments’ general revenues are from various forms of IGT such as general revenue sharing, grants, and subsidies. This revenue dependency on the national government by the subnational governments has gained new dynamic since the introduction of local self-governance system in 1995, which neatly coincides with the worldwide trend for decentralization.¹ In Korea, the issue of decentralization has long been a political hotbed, dating back to even before the Korean War. In 1949 right after the nation’s establishment, the ambitious Korea’s 1st National Assembly (NA) passed the Local Autonomy Act only to be abolished by the military junta in 1961. During the 1970s and 1980s, political opposition parties one after another persistently called for a full-blown local governance system, during which subnational administrators were all appointed by the central government. The opposition argued that the military dictatorship had exploited the centralized system to rig the national elections over and over again. In 1989, the NA once again passed the Local Autonomy Act in the wake of successful democratic protests, and on the end of the military dictatorship Korea finally moved toward a full-fledged local self-governance system in 1995 when both subnational administrators and legislatures were elected for the first time. Since then, Korea has seen every 4 years a subnational general election for provincial and local legislatures and chief administrative officers for the 16 provincial and metropolitan city governments and more than 230 local governments under those regional governments.

The Korean case of decentralization is particularly interesting. First of all, Korea has moved fast from brutal developmental dictatorship to democracy during the latter half of the 20th century, while achieving remarkable economic growth over the same period. Under the ironfisted military regime, the Korean subnational sector was essentially nonexistent, acting only as an arm for implementing national policy priorities. However, the export-driven economic growth over the period has nurtured an increasingly thicker middle class who in turn demanded democratic governance and better services in the areas of social welfare, health, income protection, and local economic development. Second, Korea is one of the few countries that have embraced the New Public Management (NMP) movement initiated in the United States during the 1990s (Batley & Larbi, 2004). To make the government smaller and competitive, during the Kim Dae-Jung administration (1998-2002), NMP-oriented reformers heatedly urged the government to adopt measures to privatize key public corporations and to delegate responsibilities to the subnational governments (Ahn & Kim, 2000; Im, 2003; P. S. Kim, 2000). Although this package of policy initiatives has been proven to be ephemeral (Lee, 2004), it has a lasting imprint in Korea’s intergovernmental relations. Third, right after the introduction of the local autonomy system, Korea has been bashed by the Asian financial crisis in 1997, which sent the country on the verge of economic meltdown. It affected the national and subnational policy priorities as well as the amount and composition of revenues (including IGTs) and expenditures at all levels of the government (Han, 2002).

This article aims to survey the current state of Korea’s fiscal decentralization as well as how it has evolved and to examine how political factors including subnational election results affect the decisions of distributing IGT from the central government since 1995. Building on the literature on the determinants of IGT that focuses on political resource factors, this article investigates how the gubernatorial vote margin enjoyed by the successful candidate is related to changes in the transfer amount. In addition, it also examines how the political alignment between the central and provincial governments affects IGT including general revenue sharing and categorical grants, taking into account service needs and fiscal capacity factors.

This article is organized as follows. In the section “Fiscal Decentralization Today in Korea,” a brief overview will be provided on Korea’s subnational government structure, their revenues sources, and different types of intergovernmental grants. The section on “Determinants of Intergovernmental Transfers” reviews the empirical studies that examine determinants of IGT with a focus on political and equalization considerations, which is followed by a description of research design in the section on “Method.” The sections “Determinants of Intergovernmental Transfers: Analytical Findings” and “Conclusion” will present the analytical results with a discussion about their theoretical and practical implications.

Fiscal Decentralization Today in Korea

While being unitary, the Korean intergovernmental system has three tiers, each with substantial political and fiscal authority within the jurisdiction. There are seven metropolitan cities and nine provinces under which there are 232 local governments of general purpose: 77 cities, 70 autonomous districts, and 85 counties as shown in Table 1. The governors, mayors, county administrators, and council persons for each subnational unit of government have been elected every 4 years since 1995. As of 2010, there are 759 council persons at the provincial level and 2,888 at the municipal governments across the country in addition to governors, mayors, and county administrators.

Table 1.

Structure of Korean Subnational Governments

Central government
Special metropolitan city (districts)		Metropolitan cities (districts, counties)		Provinces (cities, counties)
Seoul	(25)	Busan	(15, 1)	Gyeonggi	(27, 4)
		Daegu	(7, 1)	Gangwon	(7, 11)
		Incheon	(8, 2)	North Chungcheong	(3,9)
		Gwangju	(5, 0)	South Chungceong	(7, 9)
		Daejeon	(5, 0)	North Jeonla	(6,8)
		Ulsan	(5, 1)	South Jeonla	(5, 17)
				North Gyeongsang	(10, 13)
				South Gyeongsang	(10,10)
				Jeju	(2, 2)
25 districts		45 districts and 5 counties		77 cities and 80 counties

Although it is not stipulated by law, the establishment of a metropolitan city is usually considered for cities with a population more than one million. The seven metropolitan cities including Seoul are organized mainly into autonomous districts. As of 2009, about 46.2% of 50 million people live in the metropolitan cities. There are nine provinces based on Korea’s long history and geography. Each province has its own unique historical and economic background with a distinct dialect. Particularly, the southeastern and southwestern regions have entangled vicissitudes of competition and antagonism, which is clearly imprinted in politics. On the other hand, there are 77 urban cities and 80 counties under the provinces. The Local Autonomy Act states that a city should be “of an urban type in most parts” with a population of not less than 50,000. Currently, about 82% of the Korean people live in the urbanized areas, which became a policy concern because of the disparities between cities and predominantly rural counties.

Fiscal Decentralization and Subnational Finance

Since the reform in 1995, the subnational government’s general account revenues have increased substantially with an annual growth rate of 8.2%. The overall size of the subnational general revenues changed from US$49.3 billion to US$147.8 billion in 2009 (in constant price of 2005). The growth rate is a bit higher for metropolitan districts (11.3%) than for counties and cities (7.6% and 7.7%, respectively). These numbers, however, must be put into perspective since the national government’s general account has also increased 8.8% per annum during the last decade. Both the national and subnational governments have expanded more rapidly than the national economy has. During the same period, the national GDP has expanded with an annual rate of 6.2% despite some unexpected interruptions.

Figure 1 below shows some key features of the growth in the Korean subnational government. Only 2 years after the shock of the Asian financial crisis, all types of the government started to expand. However, this trend got stalled during the leftist Roh Moo Hyun Administration (2003-2008). This is quite surprising an observation given that the late² president Roh was said to spearhead a series of policies to devolve the responsibilities to the subnational governments. This preliminary look at the trend reveals that the policy initiatives by the Roh administration encouraged subnational governments to assume more responsibilities, not necessarily with more fiscal resources.

Figure 1.

Total general revenues by subnational governments: 1996-2009

While each tier of the subnational governments has its own sources of revenues, as many as 17 different sources make it truly complicated as shown in Table 2. Ten different types of general local taxes and four earmarked taxes are levied by the provincial/metropolitan governments. These regional taxes are usually collected by the local governments. For example, property acquisition and registration taxes and inhabitant tax are collected by the local governments although the regional governments are main beneficiaries of the revenues from such sources. As discussed below, the local governments are allowed to retain certain portions of the tax revenues that they have collected on behalf of the regional governments. Counties and cities are levying seven different general taxes with two earmarked taxes. The bases of these taxes are separate from the central tax, which is governed by the National and Local Tax Adjustment Act. However, the separation gets fuzzy in practice as in the cases of Local Education Tax and Special Tax for Rural Development. Being national taxes, they are still imposed on local taxes (J. Kim, 2004).

Table 2.

Sources of Revenues by Types of Government

	General taxes	Earmarked taxes	Nontax
Metropolitan cities	- Property acquisition tax	- Facilities tax	- Charges
	- Property registration tax	- City planning tax	- Fees
	- Automobile tax	- Regional development tax	- Property rents
	- Inhabitant tax	- Local education tax	- Business operations
	- Motor fuel tax		- Collection grants
	- Agricultural income tax
	- Butchery tax
	- Tobacco tax
Provinces	- Property acquisition tax	- Facilities tax	- Charges
	- Property registration tax	- Regional development tax	- Fees
	- License tax	- Local education tax	- Property rents
	- Leisure tax		- Business operations
			- Collection grants
Metro autonomous districts	- License tax	- Business office tax	- Charges
	- Property tax		- Fees
			- Property rents
			- Business operations
			- Collection grants
Cities and counties	- Inhabitant tax	- City planning tax	- Charges
	- Property tax	- Business firm tax	- Fees
	- Automobile tax		- Property rents
	- Motor fuel tax		- Business operations
	- Agricultural income tax		- Collection grants
	- Butchery tax
	- Tobacco tax

Source: Fiscal Yearbook of Local Government, 2010. Revised by the author.

The complexity of the tax revenue sources and the fuzziness of the tax bases are affirmed by Table 3, which reports the subnational tax revenues for 2009. First of all, the total amount of subnational tax revenues is almost US$41 billion. It is about 39.3% of the central tax revenues, which is an increase from 35.4% in 1998. While real, it is less than impressive. Even with fiscal decentralization, Korea has maintained a strong central tax base in personal and corporate income taxes, inheritance and gift taxes, value-added tax, liquor tax, and composite property tax. In other words, rather than allowing the expansion of local tax base, the central government has resorted to IGT to embrace increasing demands for social and economic services (H. Kim, 2009).³

Table 3.

Local Tax Revenues by Sources 2009

	Total	Metro-cities/provinces	Districts	Cities	Counties
Local taxes	40,960.50	27,846.06	1,949.11	9,277.98	1,887.35
General taxes	32,856.23	21,489.62	1,543.15	8,078.45	1,745.01
Property acquisition tax	6,039.88	6,039.88	-	-	-
Property registration tax	6,483.07	6,483.07	-	-	-
License tax	66.78	31.11	35.67	-	-
Inhabitant tax	6,851.71	3,769.12	-	2,614.53	468.07
Property tax	3,480.04	44.54	1,481.09	1,663.53	290.87
Automobile tax	2,576.37	1,158.05	-	1,191.33	226.99
Butchery tax	50.64	9.10	-	28.89	12.65
Leisure tax	911.15	911.15	-	-	-
Tobacco tax	2,737.01	1,240.59	-	1,200.07	296.36
Motor fuel tax	2,988.25	1,199.63	-	1,338.59	450.03
Property tax metro-city allotment	629.96	603.39	26.57	-	-
Composite property tax	41.38	-	-0.19	41.52	0.05
Earmarked taxes	7,559.15	5,974.34	368.52	1,094.47	121.82
City planning tax	2,062.93	1,209.58	-	786.63	66.72
Facilities tax	537.47	537.47	-	-	-
Business firm tax	734.56	3.11	368.52	307.83	55.10
Regional development tax	82.54	82.54	-	-	-
Local education tax	4,141.64	4,141.64	-	-	-

Source: Fiscal Yearbook of Local Government, 2010. In million dollars.

Second, the metropolitan cities and provinces are heavily dependent on taxes imposed on property transactions. Of US$21.5 billion from the general taxes, about 58.3% (US$12.5 billion) is collected from property acquisition and registration taxes. This is in part because the real estate market in Korea has boomed during the 2000s. Third, for the cities and counties, the most important source of tax revenue is inhabitant tax. It has two components: one imposed uniformly on individuals/businesses at flat rate and the other differentially on individual and business incomes.⁴ Fourth, reflecting the fast urbanization and redevelopment of cities, the revenues from the city planning tax is most notable among special purpose taxes, except for the education tax.

Table 4 glimpses at a break down of the subnational general revenues of 2009 into five different categories of local tax, nontax, local share tax, national subsidies, and other transfers. The regional governments collect get the lion’s share of the tax revenues (68%), while at the same time getting almost a half of national subsidies. These are two major sources of revenues for the metropolitan/province governments. On the other hand, local share tax from the central government is relatively evenly distributed among regional, city, and county governments, but autonomous city districts are getting more than 62% of other transfers, which include metropolitan city revenue sharing. Cities and counties are much more dependent on service fees and charges than the regional government, which makes sense because they are the frontline deliverers of health, welfare, and other services. Cities are more fiscally autonomous than counties and districts. More than half of their general revenues come from own sources, while districts and counties get less than 40% of general revenues from their own sources.

Table 4.

Structure of Korea’s Subnational Finance: General Revenues 2009

	Total	Local tax	Nontax	Local share tax	National subsidies	Other transfers
Metropolitan cities/provinces	69,470	27,846	9,025	8,479	24,119	0
	Column %	68.0	23.7	32.8	48.8	0.0
	Row %	40.1	13.0	12.2	34.7	0.0
Autonomous districts	19,386	1,949	5,476	572	7,220	4,169
	Column %	4.8	14.4	2.2	14.6	62.7
	Row %	10.1	28.2	3.0	37.2	21.5
Cities	43,690	9,278	14,299	8,073	10,014	2,027
	Column %	22.7	37.6	31.2	20.3	30.5
	Row %	21.2	32.7	18.5	22.9	4.6
Counties	28,383	1,887	9,265	8,721	8,058	452
	Column %	4.6	24.3	33.7	16.3	6.8
	Row %	6.6	32.6	30.7	28.4	1.6
Total	160,928	40,961	38,064	25,845	49,410	6,648

Unit: US$Million

Figure 2 traces the trends in the dependence on IGT by different types of governments. The vertical axis indicates the share of IGT out of total general revenues. Except for county governments, IGT shares have steadily increased for the other three types of government. The most notable increase is that for autonomous districts where the IGT share has soared from a mere 8% to 45%: Most of the increase came in the wake of the 1997 financial turmoil. Another key observation is that counties, mostly rural, are more dependent although annual changes are rather unpredictable. Lastly, it seems they are converging to a more narrowed range to the end of the decade.

Figure 2.

Dependence on IGT from the higher level of government

Key Intergovernmental Transfer Mechanisms

As shown in Table 4 and Figure 2, the Korean subnational governments are heavily dependent on transfers from the national government. The most notable mechanism is local share tax. Dating back to an interim legislation during the Korean War, this mechanism originally aimed to lessen fiscal disparity between urban and rural local jurisdictions. The annual transfer amount was originally subject to central budgetary considerations but is now regulated by law such that 19.24% of the central tax revenues⁵ are channeled to subnational government as local tax sharing. The portion of local tax sharing out of the central tax revenues has been increased from 13.29% in 1981 to 15% in 1999, and to 19.24% in 2006, which indicates that the central government has directed more resources toward lowering horizontal fiscal imbalance. The formula to distribute the fund is based on estimated demands of expenditures by subnational entities that consider population, needy families and elderly population, number of households, the size of public employees, numbers of people employed in manufacturing, commerce, and agriculture, the size of public parks, etc (Ministry of Public Administration and Security, 2010). As a general revenue sharing, it allows receiving governments to enjoy full authority in spending decisions of such revenues.

In 2009, a total US$25.8 billion has been distributed through this mechanism, mainly among regional, city, and county governments: The district governments are receiving control grants from the metropolitan city governments, which is essentially a local share tax at the metro-city level. Over the period, according to the author’s own analysis, the regional governments have increased their share of local share tax from 24% in 1996 to 35% in 2009, while cities have made concessions to the former. To the latter half of the last decade, the counties have also lost some share of the local share tax.

The central government provides national subsidies for specifically designated purposes such as capital investment, implementation of delegated projects or those with national priority such as health services, and supports for low-income families, covering more than 150 projects in diverse areas. The amount of subsidies for different projects is annually appropriated by the NA, which may make it institutionally vulnerable to political influences. In 2009, a total US$49.4 billion was channeled through this mechanism whose lion’s share has been distributed to the regional government. The regional governments are the main caretaker (implementer) of the national policy priorities across the country. The share for the other three types of local governments has never increased beyond 45% over the period.

In addition to these two key national transfer mechanisms, there are also comparable regional level transfer mechanisms in place. One is the metropolitan city revenue sharing, which is intended to resolve fiscal disparities across autonomous districts within a metropolitan city. The metropolitan government distributes certain percentages of revenues from the property acquisition tax to the autonomous districts within the jurisdiction (Local Finance & Taxation Bureau, 2010; Ministry of Public Administration & Security, 2008).⁶ The provincial governments also administer the same type of transfer mechanisms for the cities and counties within their jurisdictions. As with the local share tax, the funds come with no strings attached, giving the municipalities full discretion in spending decisions. Another transfer mechanism at the regional level is the so-called financial reservation (increased grants) where certain proportions of the tax revenues collected by the local governments at the third tier on behalf of the metropolitan and provincial governments are retained by the collecting governments. Cities with a population more than 500,000 can keep 47% of the provincial taxes they collect, while other cities/counties and autonomous districts can retain 27% of such collections.

Although some scholars observe that fiscal decentralization has been initiated even during the late 1980s strengthening the local tax base (Kwon, 2003; Lee, 2005),⁷ the significant fiscal decentralization has come to pass only with the 1995 subnational election (Chu & Norregaard, 1997; H. Kim, 2009; Lee, 2005). A key feature of the intergovernmental landscape after the reform is a sharp increase in the share of subnational expenditures (H. Kim, 2009), with increasing presence of IGTs as described above. In 2010, the subnational governments take as much as 49% of the public expenditures in Korea, even excluding educational spending. According to the Fiscal Yearbook of Local Government, published annually by the Ministry of Public Administration and Security (MPAS), subnational own source revenues⁸ have shrunk from 60% in 1998 to 56% in 2010, while the share of IGTs has increased from 31.5% to 41% over the same period.

One expected change with political and fiscal decentralization is lower vertical fiscal imbalance (Lee, 2005). Calculated as the difference between unity and the ratio of intergovernmental transfer to subnational expenditures, Lee (2005) reported that the coefficient of vertical fiscal imbalance has increased from .433 in 1994 to .580 in 2000. A further analysis by the author of the data for the longer period confirms this trend such that the vertical imbalance coefficient has been hovering around .60 throughout the last decade except for the year of 2007. This may mean that the fiscal imbalance between the central government and the subnational ones has been improved. However, the improvement may be misleading since it includes the special accounts for public enterprises with special missions. If we restrict the analysis to the general account revenues with which the general purpose subnational government produces local public goods, there has been a discernible trend toward increasing subnational dependence on the central government as shown in Figure 2 above. This is at odds with the expectation that with the advent of the subnational self-governance system come demands for greater subnational fiscal autonomy (Chu & Norregaard, 1997). Although Korea has seen a rapid expansion in welfare services during the first decade of this century, the central government, with its preference for IGTs over expanding local tax base, has resorted to IGTs to finance increased subnational expenditures on health, welfare, and infrastructure, rather than reinforcing the tax bases of the subnational governments.

In the sections that follow, while examining relevant factors in the determination of IGTs, this article analyzes the reasons why even with the political and fiscal decentralization the vertical fiscal imbalance has been expanded over the study period.

Determinants of Intergovernmental Transfers

Typical textbook treatments of IGTs center on normative purposes that they are supposed to serve: correction for externalities, improvement in efficiency in fiscal decisions, explicit redistribution of resources across subnational jurisdictions, substitutes of one tax structure for another, and risk sharing among different jurisdictions (Fisher, 2007; Von Hagen, 2007). These normative considerations prescribe IGTs as a measure to achieve vertical and horizontal equalization of fiscal capacity as well as more efficient resource allocation. However, the empirical studies of IGTs have been driven mostly by public choice theorists with a key assumption that politicians at the central as well as subnational governments are at least in part motivated by their own self-interests regardless of whether it is reelection or influence pandering (Alperovich, 1984; Grossman, 1994; Wright, 1974). In this perspective, transfers are a strategic tool in the pursuit of such interests (Veiga & Pinho, 2007).

In this line of empirical research, scholars have used the concept of political capital or political resources. Political capital, an ill-defined concept with high currency in ordinary conversations, can be loosely defined as the extent of leverage that a politician can exert to win support from other politicians and/or voters in pursuing policy and political interests. Grossman (1994) operationalizes it as party alignment between the two tiers of government. Specifically, he measured party alignment in the U.S. context by the percentages of votes for Democratic governors and of Democratic seats at the state legislatures, given that the federal House of the Representatives was controlled by the Democrats during the study period. At the same time, Grossman complementarily uses the size of population of a jurisdiction as a proxy for its political capital vis-a-vis national politicians.

Political congruence between two levels of government is also a relevant factor of political capital for subnational politicians. Measured by either party affiliation of key central and subnational politicians (Veiga & Pinho, 2004, 2007), or by legislative seats held by the same party as the one controlling the central legislature (Grossman, 1994), the same political control of the key institutions by the subnational politicians as the controlling politicians at the upper level brings more transfers to the subnational jurisdictions. So does a higher percentage of subnational votes for the parties that form the central government (Alperovich, 1984). Another key element of political capital proven to be effective in securing intergovernmental grants is per capita representation by the subnational jurisdiction at the higher legislative body (Ansolabehere, Gerber, & Snyder, 2002; Knight, 2008; Porto & Sanguinetti, 2001).

The extent of political support that subnational politicians command is also an element of political capital about which there are two key competing hypotheses. According to the swing voter hypothesis (Lindbeck & Weibull, 1987, 1993), the ruling political party at the central level channels IGTs to jurisdictions with narrow vote margins in the last election to buy off the voters from the opposition. On the other hand, for reelection, the optimal strategy for the risk-averse politician is maybe to provide redistribution to her preexisting supporters (Cox & McCubbins, 1986). Against the American context where groups of various socioeconomic and ethnographic backgrounds identify themselves with two major parties, Cox and McCubbins (1986) argue that politicians continue to favor their traditional supporters. Empirical findings are mixed at best: Some studies report no evidence of the swing voter hypothesis (Coats, Karahan, & Tollison, 2006), while others do find support (Dahlberg & Johansson, 2002; Dixit & Londregan, 1998; Johansson, 2003). In testing Cox and McCubbins’ (1986) argument against the swing voter hypothesis, Veiga and Pinho (2007) support the latter over the former. Dixit and Londregan (1996) introduce a qualification such that the effectiveness of political parties in delivering benefits to different groups of voters mediates the causal link. However, evidence abounds that political parties reward supporters rather than buy off opponents (Alperovich, 1984; Case, 2001; Schady, 2000).

IGTs can be a tool to address fiscal disparities across subnational jurisdictions. Fiscal disparities can arise from a variety of factors regarding differences in the ability of raising revenues and expenditure needs, which is often referred to as “needs-capacity” or “resource-requirements” gap (Bahl, Martinez-Vazquez, & Sjoquist, 1992; Dafflon, 2007). This justifies IGTs to equalize fiscal capacity at a common tax rate or to provide equal fiscal treatment for the individuals regardless of the jurisdictions they live in. The equalization transfers can be explained by the electoral choice models, specifically, the median voter hypothesis. The hypothesis posits that democratically elected politicians will distribute IGTs to serve the preferences of the median voter to maximize reelection possibilities (Boex & Martinez-Vazquez, 2004). Depending on where the majority of the voters live, the transfers would be distributed differently. If it is predominantly rural, then rural jurisdictions will get more transfers than urban ones. As such, the IGTs of this nature will benefit the jurisdictions with lower fiscal capacity.

However, a review of the IGT distribution reveals that, contrary to the expectation, higher fiscal capacity as measured by average household income, human development index, or gross regional domestic product (GRPD) actually brings more transfers from the central government (Boex & Martinez-Vazquez, 2004). Still, Vaillancourt (2010) reports a mixed result regarding per capita personal and corporate income in Canada, where, depending on the statistical model (random vs. fixed effect models), the findings were different. Boex and Martinez-Vazquez (2004) and Boex (2003) regard this inconclusiveness as evidence that political considerations, that is, shortcomings in institutional mechanisms, may outweigh the efforts in the normative perspective.

Korea has a law, called Local Subsidy Act, which specifically stipulates the purposes and design of equalization transfers. The local share tax incorporates the gap between expenditure requirements and own source revenues prospect. In addition, the politicians at the NA also consider fiscal disparities in considering the amount and distribution of the national subsidies/grants.

To examine the political influences in the determination of the amount of IGTs at the regional level in Korea, this article will examine the following hypotheses that focus on the political considerations and fiscal capacity/needs of the regional governments, building mainly on the two streams of empirical studies by public choice theorists briefly touched on above: IGT will be greater for the jurisdictions (a) whose governors’ vote margin is greater, (b) under the same political control as the national government, (c) with lower per capita gross regional domestic product (GRDP), (d) that pay lower per capita central taxes, and (e) with more dependent population.

Method

In surveying research on IGT distribution, Boex and Martinez-Vazquez (2004) identify three major policy determinants such as normative, voter choice, and political considerations. Within this framework, according to them, the empirical studies have used proxies of such factors as expenditure needs, revenue capacities, political factors, and the jurisdiction’s relative population size. Following their review, this article identifies four types of potential determinants of IGT.

First of all, political capital factors are captured by the percentage vote margin at the gubernatorial election, congruence in political control as for the presidency–governorship and the NA–regional council pairs, and the number of NA persons per 500,000 people. Each of these refers to political resources that can be exerted in the process of IGT allocation. The first variable, gubernatorial vote margin, is intended to test the competing hypotheses by Cox and McCubbins (1986) and by Linbeck and Weibull (1987, 1993). The key issue to be examined is whether the national politicians favor their supporters or try to win converts from the nonsupporters at the last election. The match in political control between different levels of government are also found to be relevant in different countries as reviewed above, but up until now, it has not been put to test for the Korean case. Although the fiscal transfers including national subsidies are considered first by the Public Administration & Security Committee⁹ in the NA, the congruence in political control of the national and regional legislatures is still relevant since in Korea ruling party leaders have maintained tight control over the votes of individual members as well as policy agenda (Sin, 1999).

Second, variables such as per capita GRDP and per capita central tax payment are used to measure the jurisdiction’s fiscal capacity. While per capita GRDP and income of a jurisdiction are frequently used to operationalize fiscal capacity (e.g., see Grossman, 1994; and Johansson, 2003), per capita national tax payment is not so. A jurisdiction with higher GRDP may pay more central taxes as evidenced by the correlation (r = .80) in the data set for this article. However, this additional variable makes it possible to test if the jurisdictions that pay more central taxes get better payback from the central government.

Third, the percentage of the dependent population intends to capture some extent of service needs. Although dependent population is usually defined as those aged 65 or above and 14 or below (Alperovich, 1984), it is operationalized as the percentage of those aged 65 or higher and those aged 19 or below. In Korea, 95.5% of teenagers are attending secondary school, which essentially makes them dependent on parents and public services. The population size is included in the models. It can represent both service needs and economy of scale. The bigger the size of the jurisdiction, the average cost of providing services will decline. All of the empirical models with population included found a negative relationship with the IGT amount (Boex & Martinez-Vazquez, 2004), which may indicate lower average costs for bigger jurisdictions.

Lastly, since Korea has suffered a lot from the 1997 financial crisis whose effects on the subnational governments persisted over a substantial period of time, the models will include two dummy variables of years before the crisis (1996-1997) and after the crisis (2002-2009). The reference years will be 1998-2001. Accordingly, the basic model is specified as follows:

\ln {(PCIGT)}_{i t} = β 0 + β 1 {(Vote Margin)}_{i t} + β_{2} {(Gubernatorial Congruence)}_{i t} + β_{3} {(Legislative Congruence)}_{i t} + β_{4} {(No. of Representatives)}_{i t} + β_{5} \ln {(PC GRDP)}_{i t} + β_{6} \ln {(Central Tax Payment)}_{i t} + β_{7} \ln {(Population)}_{i t} + β_{8} {(% Dependent Population)}_{i t} + β_{9} (Before) + β_{10} (After) + β_{11} {(Lagged Dependent Variable)}_{i t} + ε_{i t}

To test the hypotheses, the models will use four different dependent variables: per capita total IGT, per capita local share tax, per capita national subsidies, and fiscal autonomy (FISAUTO). As discussed, the local share tax and national subsidies are designed and implemented differently: the former follows formula and the latter depends on annual appropriations. FISAUTO is calculated as FISAUTO = 1 – (Total Transfers ÷ Total Expenditures). The higher the value is, the more fiscally autonomous the jurisdiction is.

The fiscal data are from the Financial Year Book of Local Governments 1997-2010, published by the Local Finance and Taxation Bureau under MPAS. The Yearbook includes details of revenues and expenditures of each subnational government. The election results are compiled and published by the National Election Commission. The data for the socioeconomic variables are available at the Statistics Korea, which is the national commission on statistics. The key variables and their descriptions are provided in Table 5.

Table 5.

Key Variables

Variable Name	Description and coding	Expected relationship
Per capita total transfers	Per capita total transfers from the central government	DEV
Per capita local share tax	Per capita amount of local share tax allotted from the central tax revenues	DEV
Per capita national subsidies	Per capita subsidies from the central governments	DEV
Fiscal Autonomy	= 1-(total transfers÷ total expenditures)	DEV
Per capita GRDP	Per capita Gross Regional Domestic Product	(−)
Per capita national tax payment	Per capita central tax payment in W100,000 (Korean won)	(−)
Vote margin	Percentage of the vote margin between the winner and the runner-up out of total votes cast from the last election	(+)
Congruence in political control of governorship	= 1 if the same party controls both the Presidency and the Governorship; = 0 if otherwise	(+)
Congruence in political control of legislatures	= 1 if the same party controls both the NA and the subnational Legislature; = 0 if otherwise	(+)
Number of NA Persons	Number of representatives per 500,000 people (Nationally, there are 299 representatives.)	(+)
Population	Number of population measured in 100,000	(+)
Percentage of dependent population	Percentage of the population aged below 19 or above 65	(+)
BeforeAfter	Dummies to capture differences in the constant from the years of 1997 financial crisis: Years of 1998-2001 is the reference

Note. DEV = dependent variable.

*Per capita amounts in W1,000 (Korean won) except for the per capita national tax payment.

Covering 16 regional governments over a 14-year period, the data is a typical time series cross-sectional (TSCS) set. With relatively small number of years, OLS regression is inappropriate due to its temporal and spatial correlation of the error terms: They are serially correlated and heterogeneous across the panels. Since this issue can make the hypothesis testing misleading, Beck and Katz (1995) propose to use OLS with panel corrected standard errors (PCSE) while controlling temporal correlation of the error term. Following their suggestions, the analysis uses OLS with PCSEs while including the lagged dependent variable in the model to capture dynamic effects from the previous years and to adjust for serial correction of the error terms.¹⁰ Specifically, it will set the panel specific first-order autocorrelation structure. To allow for a consideration of the potential differences that different analytical methods might make, the hypotheses are also tested using random effect models.

Determinants of Intergovernmental Transfers: Analytical Findings

IGTs are an overdetermined phenomenon where political, economic, and sociodemographic factors are intermingled. At the same time, normative considerations by the top policy makers at the central level are also obviously prominent. When examining the determinants, scholars essentially assume that politicians are sensitive to changes in their reelection prospect. And, reelection prospect is a matter of how they get the support of subnational politicians as well as the voters. As is the case, political factors are all found relevant. Applying the median voter theorem that expects political parties to meet the demands of the median voter, studies have also reported that national politicians distribute IGTs to the jurisdictions with lower fiscal capacity to meet the demands by the median voter. Consistent with such findings from the literature, this article reports that subnational jurisdictions with more political capital and/or service needs will get more per capita transfers. The results from OLS with PCSE and random effect models essentially confirm that these previous findings are true to the Korean case as in Table 6.¹¹

Table 6.

Determinants of Intergovernmental Transfers in Korea

Dependent variable	Log per capita total IGT		Log per capita local share tax		Log per capita national subsidies		Fiscal Autonomy
Model	Panel, RE	OLS w/PCSE	Panel, RE	OLS w/PCSE	Panel, RE	OLS w/PCSE	Panel, RE	OLS w/PCSE
ercent; gubernatorial vote margin	.0016	.0023**	.0044***	.0047***	.0009	.0017	.0000	−.0006**
Political congruence of governorship	.1176**	.1159**	.0413	.0536	.1767***	.1635***	−.0118	−.0217**
Political congruence of legislature	.0867*	.0798	.0462	.0317	.1011**	.0927*	−.0099	−.0107
Representatives per 500,000 people	.2939***	.3030***	.4050***	.4658***	.3450***	.3587***	−.0215*	−.0524***
Log per capita GRDP	.6791***	.8584***	.6400***	.8289***	.8038***	1.0760***	−.1080**	−.1059***
Log per capita national tax payment	−.3880***	−.4206***	−.3340***	−.3701***	−.4608***	−.4707***	.0862***	.0951***
Log population	−.2426***	−.3143***	−.5386***	−.5989***	−.1411***	−.2456***	.1138***	.0576***
ercent; dependent population	.0306**	.0244	−.0291	−.0351*	.0437***	.0346**	.0004	−.0285***
Before crisis	−.4321***	−.3552***	−.4359***	−.3842*	−.5021***	−.4058***	.0241*	.0188
After crisis	.3931***	.3941***	.4242***	.4733**	.5007***	.4700***	−.00048	−.0432**
Lagged dependent variable	.4435***	.3874***	.5592***	.5051***	.3789***	.3157***	.4163***	.3310***
Model summary
Observations	220	220	220	220	220	220	219	219
R-squared (between)	.9793	.9754	.9790	.9289	.9681	.9705	.6927	.9173
Wald chi-square	1697.0	931.4	1269.6	1016.8	1406.0	868.3	192.5	1417.2
Prob> chi-square	.0000	.0000	.0000	.0000	.0000	.0000	.0000	.0000

p < .10. **p < .05. ***p < .01.

First of all, the vote margin enjoyed by the incumbent governor from the last election has a positive effect on the amount of local share tax but not on national subsidies. Specifically, 1% increase in the vote margin leads to 0.47% increase in the local share tax. Although the distribution of the local share tax is regulated by the specific formulae, it can still be affected by other considerations in the process of designing such formulae. Interestingly, it seems irrelevant in allocating national subsidies. The decisions about the types of subsidy programs and respective amounts are made annually by the NA. Obviously the NA does not take this factor seriously when working on the details of the national subsidy programs.

The literature suggests two conflicting possibilities: Politicians would either reward preexisting supporters (Cox & McCubbins, 1986) or try to gain converts from the opponents (Lindbeck & Weibull, 1987). Using presidential election results and data between 1988 and 1997, Kwon (2005) found that the Korean national government provided more national subsidies to provinces with lower vote margins, supporting the swing voter hypothesis. On the other hand, using gubernatorial election data, the extant research partially supports Cox and McCubbins (1986) in that they transfer more local share taxes to jurisdictions with wider vote margins: As described below, Korea indeed offers a very similar situation to the U.S. context since a majority of Korean voters have kept deep loyalty to two key political parties throughout its modern history, and politicians reward their preexisting supporters with geographical tension. However, the results from the current analysis add one key qualification: The effect of the greater vote margin may be different across different types of transfers. If the distribution of the transfer is controlled by the administrative body, the MPAS that elaborates on the formulae, thus by the President, the vote margin gets relevant. Alternatively, it may be related with the expected beneficiaries of two types of transfers. The local share tax is a way of general revenue sharing targeting the broader constituents, while national subsidies are directed specific and narrowly defined infrastructure and other programs of national priority. This may refer to the possibility of a mediator, the effectiveness of politicians in delivering benefits to different groups (Dixit & Londregan, 1996).

The effect of the congruence in political control depends also on what types of transfers we are talking about, but, unlike the effect of the vote margin, is significant only for the distribution of national subsidies. If the presidency and the governorship are controlled by the same party, the jurisdiction gets about 16.35% more subsidy than otherwise. The markup from the congruence in the legislature is about 9%. Jointly, the effect is as much as 25.6% increase in national subsidies. On the other hand, the allocation of the local share tax is found to be independent from this factor.

The effect of the alignment in political control may be better understood if the Korean regional politics is put into context. There has been a deep tension between two competing regions in Korea since from the beginning of Korean history, dating back to as early as the 3rd century. The tension has gained a refreshed twitch during the Park Jung Hee Administration (1961-1979), which has favored the southeastern region at the expense of the southwestern counterpart. During this period and beyond up until 1997, most of the key economic development initiatives shied away from the southwestern provinces. In the general election held in 1997 amid the unprecedented Asian economic crisis, a veteran politician and leader of democratic movement, Kim Dae Jung, from the southwestern region got elected. Both he and his successor, Roh Moo Hyun who shared his core political beliefs, come from the same party. The voters from the southwestern region have predominantly supported the Kim and Roh Administrations, while those from the southeastern part have broken for the opponent party that emphasizes economic development over redistribution and welfare.

Therefore, the fate of the politicians from the two grueling regions has been decided solely by which party’s ballot they are on. During the administrations of Kim and Roh, the governorships in the southwestern Geolla region were just controlled by their political disciples, while those in the southeastern Gyeongsang were all controlled by their political opponents. There was not much of a difference in the NA and the subnational councils. Given such considerations, the findings with the congruence in the political control just confirm what has happened throughout the modern Korean history. During the two leftist administrations (1998-2007), per capita national subsidies to the Geolla region has increased at an annual rate of 12.0%, compared to 10.2% of the Gyeongsang region, which explains in part the effects of the congruence in the political control of key governing institutions.

As for the number of politicians per 500,000 people at the national legislature, the analysis finds the same direction of the relationship as those in the previous studies (Ansolabehere et al., 2002; Knight, 2008; Porto & Sanguinetti, 2001). An additional lawmaker at the NA adds as much as 35.9% increase in national subsidies and 46.6% markup in local share tax in per person. It is noteworthy that the effect is actually stronger for the general revenue sharing than for the discretionary subsidies. This may indicate the importance and relevance of politicians in designing the distribution formulae.

The models have two proxies of fiscal capacity to test for some potential evidence of the equalization effect. If equalization takes place, the coefficient on GRPD should be negative. If the efforts fail and political interest dominates the decision, we expect to see a positive sign. The result is consistent with what others have reported (Boex & Martinez-Vazquez, 2004): One percent increase in GRPD is associated with as much as 1.07% increase in IGTs. Even controlling all other political and service needs factors, GRPD has an independent effect on IGTs. The effect is stronger for the national subsidies than for the local share tax, which makes sense since the former mechanism is subject to political consideration.

However, this finding should be balanced against the other observation: The jurisdictions that pay more central taxes tend to get smaller transfers, which is consistent across all models. It is noteworthy that although per capita GRDP and central tax payment are highly correlated (r = .80), their respective effects run to the opposite directions. The more taxes they pay to the central government, the less they tend to get back from it, which amounts to be an equalization effect. However, the attenuating effect of central tax outflows cannot offset the bringing-in effect of GRDP. If we restrict the concept of fiscal balances (Vaillancourt, 2010) with respect only to GRDP and central tax payment, the jurisdiction with higher GRDP enjoys net surplus: One percent increase in both per capita GRDP and central tax payment will lead to 0.44% increase in the per capita amount of total incoming transfers. On the other hand, the jurisdictions with more dependent people get more national subsidies that are subject to political considerations. Furthermore, population itself is negatively associated with the transfer amount, which is consistent with the literature.

All political resources variables identified in the models obviously do not lower IGT amounts. Their coefficients are all positive if not significant. To examine how these political resources affect the fiscal health of the jurisdictions, the fiscal autonomy is regressed on the same independent variables. The results are largely consistent with what is just reported: The more political resources the jurisdiction commands vis-a-vis the central government, its fiscal autonomy gets lower. The increase in the political resources or capital directly leads to lower fiscal autonomy, which may imply that the jurisdictions with more IGT do not match with own source revenues. The subnational politicians are arguably leveraging their capital against their national counterparts, currying favors with the voters within the jurisdiction. If the national politicians behave as vote maximizers in distributing IGTs to buy political capital of subnational politicians (Grossman, 1994), so do subnational politicians in seeking as much transfer from the national government as possible and in keeping the subnational taxes as low as possible (Grossman, 1990).

Due to different political and socioeconomic factors involved in the incidence of IGTs, fiscal resources are redistributed across the regions. Some jurisdictions will be benefited from their higher political representations, while bearing out the situation that their political institutions are controlled by a different party than the one that controls the national government. Or, the voters within the jurisdiction may be tightly split between two parties as in the Chungcheong or Ganwon provinces. All of these considerations affect the jurisdictions differently and the net effect of redistribution of fiscal resources may be determined in an a posteriori manner. To assess how well different jurisdictions fare in this redistribution game, Bayoumi and Masson’s (1995) interregional redistribution index¹² has been estimated, which is 0.78. This particular number means that 78% of the relative difference in the initial income distribution remains after the IGTs. With the IGTs, the interregional fiscal disparities have reduced by 22%.

Even with the presence of political factors in the determination of the transfer amounts, the system introduced in 1995 has improved fiscal equality at the regional level. Provinces such as Gangwon, North and South Chuncheong, and North and South Geolla, where the majority of counties are subsumed, have benefited from the fiscal decentralization. On the other hand, the decentralization has built in a force against the fiscal autonomy of subnational governments. The extant analysis offers some explanations of this aggravating fiscal imbalance: While the Korean national government tends not to make concessions in tax base to subnational ones, the subnational politicians have played a better game in exploiting central sources of revenues.

Conclusion

Korea is not so ordinary a case in achieving both economic growth and political democratization of which subnational governance system has been an integral component. Although it is still not clear whether fiscal decentralization comes ahead of economic development or vice versa (Martinez-Vazquez & McNab, 2003; Oates, 1993; Rodriguez-Pose & Krøijer, 2009), Korea has been enjoying relatively high economic growth despite two financial crises in a decade while moving toward a political and thus fiscal decentralization. This achievement alone makes the issue of how the Korean fiscal decentralization has proceeded deserve a separate attention.

On taking stock of the development since the political decentralization in 1995, this article observes that, with a very complex set of own source revenues, Korean subnational governments have developed a habit of resorting to the central counterpart in financing their expenditures. While more and more responsibilities are devolved to the subnational entities, a comparable strengthening of their fiscal health has yet to follow. This extant analysis presents some evidence that jurisdictions with better political capital do a better job in drawing in IGTs without making comparable efforts to raise own source revenues. Almost all of the elements of the subnational political capital identified in the literature find support in Korean IGTs to the regional governments. Even those jurisdictions with higher income have successfully secure more IGT than not so well-to-do jurisdictions. However, this does not necessarily mean that institutions have failed on their normative grounds since, even with higher dependence of the central government, the subnational income disparities have been substantially mitigated.

In examining the political capital factors, the extant analysis has introduced a key qualification to the hypothesis proposed by Cox and McCubbins (1986). The vote margin at the gubernatorial election has been conducive in drawing more local share tax but not national subsidies that are essentially block grants. Given that Korean parties have resorted to their respective geographical strongholds, Cox and McCubbins’ model based on the American two-party system would find a comparable case in Korea. However, given that the political power of the Korean President is largely much stronger than his legislative counterparts, the vote margin factor may find its way through the design of the distribution formulae of the local share tax that are administered by the Ministry of Public Administration & Security under the President. The same understanding can be applied to the national subsidies since the Ministry of Strategy & Finance plays a key role comparable to that of the U.S. Office of Budget and Management. Presidential policy priorities that reflect the governors’ preferences tend to roll in the budgetary process where the NA rarely makes substantial changes to the President’s budget proposals. Accordingly, the types of IGTs and the relative influences from different political institutions (in this case, the Presidency and the NA) seem to jointly mediate the effect from the vote margin.

Although the empirical findings of this study are largely in concert with what has been reported by the previous studies in different countries with diverse institutional arrangements, it should be noted that they are regarding regional governments. The same hypotheses have yet to be tested against the entire Korean local governments, where the roles played by the central and regional level politicians need to be considered. Specifically, the roles played by individual politicians who are serving on the Public Administration & Security Committee at the NA and on its counterparts at the regional level should be focused, which requires a separate treatment.

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Notes

Bio

Jungbu Kim, obtained his PhD in Public Policy from the Georgia Institute of Technology and Georgia State University in 2007. His research interest centers on public budgeting and financial management issues with a focus on intergovernmental fiscal relations and public R&D expenditures.

References

Ahn

B. M.

Kim

I. C.

(2000). Reforming public enterprises in South Korea. International Review of Public Administration, 5(2), 67-80.

Alperovich

(1984). The economics of choice in the allocation of intergovernmental grants to local authorities. Public Choice, 44, 285-296.

Ansolabehere

Gerber

Snyder

(2002). Equal votes, equal money: Court-ordered redistricting and public expenditures in the American states. American Political Science Review, 96, 767-778.

Bahl

Martinez-Vazquez

Sjoquist

D. L.

(1992). Central city-suburban fiscal disparities. Public Finance Quarterly, 20, 420-432.

Batley

Larbi

(2004). The changing role of government: The reform of public services in developing countries. Basingstroke, UK: Palgrave Macmillian.

Bayoumi

Masson

P. R.

(1995). Fiscal flows in the United States and Canada: Lessons for monetary union in europe. European Economic Review, 39, 253-274.

Beck

Katz

J. N.

(1995). What to do (and not to do) with time-series cross-section data. American Political Science Review, 89, 634-647.

Bird

R. M.

Vaillancourt

(1998). Fiscal decentralization in developing countries. Cambridge, UK: Cambridge University Press.

Bird

R. M.

Vaillancourt

(2006). In Bird

R. M.

Vaillancourt

(Eds.), Perspectives on fiscal federalism. Washington, DC: World Bank.

10.

Boadway

R. W.

Shah

(2007). Intergovernmental fiscal transfers: Principles and practices. Washington, DC: World Bank.

11.

Boex

(2003). The incidence of local government allocations in Tanzania. Public Administration and Development, 23, 381-391.

12.

Boex

Martinez-Vazquez

(2004). The determinants of the incidence of intergovernmental grants: A survey of the international experience. Public Finance and Management, 4, 454-479.

13.

Case

(2001). Election goals and income redistribution: Recent evidence from Albania. European Economic Review, 45, 405-423.

14.

Chu

Norregaard

(1997). Korea. In Ter-Minassian

(Ed.), Fiscal federalism in theory and practice (pp. 540-569). Washington, DC: International Monetary Fund.

15.

Coats

R. M.

Karahan

Tollison

R. D.

(2006). Terrorism and pork-barrel spending. Public Choice, 128, 275-287.

16.

Cox

G. W.

McCubbins

M. D.

(1986). Electoral politics as a redistributive game. Journal of Politics, 48, 370-389.

17.

Dafflon

(2007). Fiscal capacity equalization in horizontal fiscal equalization programs. In Boadway

Shah

(Eds.), Intergovernmental fiscal transfers: Principles and practice (pp. 361-396). Washington, DC: World Bank.

18.

Dahlberg

Johansson

(2002). On the vote-purchasing behavior of incumbent governments. American Political Science Review, 96(1), 27-40.

19.

Dixit

Londregan

(1996). The determinants of success of special interests in redistributive politics. Journal of Politics, 58, 1132-1155.

20.

Dixit

Londregan

(1998). Ideology, tactics, and efficiency in redistributive politics. Quarterly Journal of Economics, 113, 497-529.

21.

Fisher

R. C.

(2007). State and local public finance (3rd ed.). Chicago, IL: Thompson South-Western.

22.

Grossman

P. J.

(1990). The impact of federal and state grants on local government spending: A test of the fiscal illusion hypothesis. Public Finance Review, 18, 313-327.

23.

Grossman

P. J.

(1994). A political theory of intergovernmental grants. Public Choice, 78, 295-303.

24.

Han

I. S.

(2002). An analysis of the impacts of Asian financial crisis on the revenues of subnational governments. Korean Public Administration Review, 36, 229-247.

25.

Ichimura

Bahl

(2009). Decentralization policies in Asian development. Singapore: World Scientific.

26.

(2003). Bureaucratic power and the NPM reforms in Korea. International Review of Public Administration, 8(1), 89-102.

27.

Johansson

(2003). Intergovernmental grants as a tactical instrument: Empirical evidence from Swedish municipalities. Journal of Public Economics, 87, 883-915.

28.

Keele

Kelly

N. J.

(2006). Dynamic models for dynamic theories: The ins and outs of lagged dependent variables. Political Analysis, 14, 186-205.

29.

Kim

(2009, September). The contemporary needs of general and earmarked grants in Korea: An assessment. Copenhagen Seminar on Designing Grants to Subcentral Governments, Copenhagen, Denmark.

30.

Kim

(2004). Fiscal decentralization in Korea. Seoul: Korea Institute of Public Finance.

31.

Kim

P. S.

(2000). Administrative reform in the Korean central government: A case study of the Dae Jung Kim administration. Public Performance & Management Review, 24, 145-160.

32.

Kincaid

Shah

(2007). The practice of fiscal federalism: Comparative perspectives. Montreal, Quebec, Canada: McGill-Queen’s University Press.

33.

Knight

(2008). Legislative representation, bargaining power and the distribution of federal funds: Evidence from the US congress. Economic Journal, 118, 1785-1803.

34.

Kwon

H. Y.

(2005). Targeting public spending in a new democracy: Evidence from South Korea. British Journal of Political Science, 35, 321-341.

35.

Kwon

(2003). The effects of fiscal decentralization on public spending: The Korean case. Public Budgeting & Finance, 23(4), 1-20.

36.

Lee

(2004). Searching for an alternative paradigm for Korean public administration after the NPM movement. Seoul: Korean Association for Public Administration

37.

Lee

(2005). Adequacy of Korea’s tax assignment: A fiscal federalism approach. International Review of Public Administration, 9, 139-148.

38.

Lindbeck

Weibull

J. W.

(1987). Balanced-budget redistribution as the outcome of political competition. Public Choice, 52, 273-297.

39.

Lindbeck

Weibull

J. W.

(1993). A model of political equilibrium in a representative democracy. Journal of Public Economics, 51, 195-209.

40.

Litvack

J. I.

Ahmad

Bird

R. M.

(1998). Rethinking decentralization in developing countries. Washington, DC: World Bank.

41.

Local Finance and Taxation Bureau. (2010). Financial yearbook of local government 2010. Seoul, The Republic of Korea: Ministry of Public Administration and Security.

42.

Martinez-Vazquez

McNab

R. M.

(2003). Fiscal decentralization and economic growth. World Development, 31, 1597-1616.

43.

Ministry of Public Administration and Security. (2008). Enforcement Decree of the Local Autonomy Act. Seoul, The Republic of Korea: Author.

44.

Musgrave

(1959). The theory of public finance: A study in public economy. New York, NY: McGraw-Hill.

45.

Oates

(1993). Decentralization and economic development. National Tax Journal, 46, 237-243.

46.

Porto

Sanguinetti

(2001). Political determinants of intergovernmental grants: Evidence from Argentina. Economics & Politics, 13, 237-256.

47.

Rodden

Eskeland

G. S.

Litvack

(2003). 1. Introduction and overview. In Rodden

Eskeland

G. S.

Litvack

(Eds.), Fiscal decentralization and the challenge of hard budget contraints (pp. 3-31). Cambridge, MA: MIT Press.

48.

Rodriguez-Pose

Krøijer

(2009). Fiscal decentralization and economic growth in central and eastern Europe. Growth and Change, 40, 387-417.

49.

Schady

N. R.

(2000). The political economy of expenditures by the Peruvian social fund (FONCODES), 1991-95. American Political Science Review, 94, 289-304.

50.

Sin

(1999). Mass politics and culture in democratizing Korea. Cambridge, UK: Cambridge University Press.

51.

Vaillancourt

(2010). Inter-regional fiscal flows: Interpretation issues. In Bosch

Espasa

Olle

A. S.

(Eds.), The political economy of inter-regional fiscal flows: Measurement, determinants and effects on country stability (pp. 39-58). Northampton, MA: Edward Elgar.

52.

Veiga

L. G.

Pinho

M. M.

(2004). The determinants of intergovernmental grants in Portugal: A public choice approach (ERSA Conference Papers No. ersa04p192). Vienna, Austria: European Regional Science Association.

53.

Veiga

L. G.

Pinho

M. M.

(2007). The political economy of intergovernmental grants: Evidence from a maturing democracy. Public Choice, 133, 457-477.

54.

Von Hagen

(2007). Achieving economic stabilization by sharing risk within countries. In Boadway

Shah

(Eds.), Intergovernmental fiscal transfers (pp. 107-132). Washington, DC: World Bank.

55.

Wright

(1974). The political economy of new deal spending: An econometric analysis. Review of Economics and Statistics, 56(1), 30-38.

56.

Zhao

Z. J.

(2009). Fiscal decentralization and provincial-level fiscal disparities in China: A Sino–US comparative perspective. Public Administration Review, 69, S67-S74.