Local government reform: is it effective on debt burdens?
Local governments adopt city-county consolidation to enhance infrastructure, facilities, and public service equity (Ostrom, Bish, Ostrom, & Institute for Contemporary, 1988). Miller (2002) defined city-county consolidation as the shift of local boundaries to reduce public service costs by increasing the population of who receives benefits and by creating improved economies of scale. Also, states consider city-county consolidation as an intergovernmental management strategy to resolve local financial problems by enlarging and increasing property tax revenues (Berman, 2003). Thus, city-county consolidation is an organizational change by local or state governments with the expectation of improved government efficiency.
Given this definition, researchers have examined the impact of city-county consolidation on public service costs, municipal disparity, and economic development. There are mixed views on the impact of consolidation. On one hand, it is argued that city-county consolidation, as a reform of urban management, helps reduce total expenditures (Blomquist & Parks, 1995) as well as general expenditures and personnel costs (Selden & Campbell, 2000). However, on the other hand it is argued that consolidation is associated with an increase in education expenditures (Cook, 1973), total expenditures and tax revenues (Durning & Nobbie, 2000), operating expenditures and public safety expenditures (Cain, 2009; Selden & Campbell, 2000), personnel costs (Vojnovic, 2000), public works expenditures, housing and family service costs, and parks and recreation service costs (Cain, 2009). Some researchers contend that city-county consolidation is not correlated with decreased public service costs (Benton & Gamble, 1984; Faulk & Grassmueck, 2012). Rosentraub (2000) concludes that city-county consolidation resolves the financial disparity of local governments by redistributing tax revenues. However, other researchers argue that consolidation is not associated with a decrease in fiscal disparity (Miller, Miranda, Roque, & Wilf, 1995), poverty disparity, and economic disparity (Savitch & Vogel, 2000). Furthermore, while it has been argued that city-county consolidation partly fosters the effectiveness of an economic development strategy (Blomquist & Parks, 1995), downtown revitalization (Rosentraub, 2000), citizens' income (Carr, Bae, & Lu, 2006), and employment (Faulk & Schansberg, 2009), Feiock and Carr (1997) conclude that the success or otherwise of local industries, such as manufacturing, retail, and service, are not correlated to city-county consolidation.
The literature allows this study to connect consolidation with debt burden. Although some consolidation effects, such as expenditures, tax revenues and local economic development, are also the crucial determinants of government debt burdens (Clingermayer, 1991; Day & Boeckelman, 2012; Hildreth & Miller, 2002; Krueger & Walker, 2010; Temple, 1994), little attention has been devoted to the correlation between city-county consolidation and the level of debt burden. This study attempts to fill this gap by examining the following research question:
How are city-county consolidations associated with the variation of debt burden maintained by local governments?
This research explores the factors affecting debt management and proposes a hypothesis to explain the correlation between city-county consolidation and debt burdens. Previous studies provide evidence that increased debt burdens are associated with large tax revenues (Clingermayer & Wood, 1995; Day & Boeckelman, 2012; Johnson, 1999), high income level (Ashworth, Geys, & Heyndels, 2005; Brecher, Richwerger, & Wagner, 2003; Clingermayer & Wood, 1995; Denison, Yan, & Zhao, 2007; Temple, 1994), low unemployment rate (Hildreth & Miller, 2002; Krueger & Walker, 2010), and elected officials (Baber & Sen, 1986; Clingermayer, 1991).
Some demographic factors influence the level of debt burden. For example, increasing debt burden of municipalities has been associated with an aging population (Clingermayer, 1991), younger voters (Ellis & Schansberg, 1999), and total population (Garcia-Sanchez, Mordan, & Prado-Lorenzo, 2012). In addition, Trautman (1995) concluded that a high degree of urbanization leads to increasing state debt burdens. By contrast, systemic investment decisions (1) (Temple, 1994), the mayor-council form of government (Clinger, Feiock, McCabe, & Park, 2008), and grants (Martell & Smith, 2004) are associated with a decrease in debt burdens. Cook (1973) argues that the borrowing costs of consolidated governments are higher than fragmented governments, but his conclusion is not directly connected to debt burdens because such costs may be different depending on individual municipal bond ratings (Robbins & Kim, 2003). Similar to city-county consolidation, weak governments, such as politically fragmented entities, have higher debt burdens compared with politically strong governments (2) (Ashworth et al., 2005). Nevertheless, their study did not approach city-county consolidation as a local government reform.
The research presented in this study examines the direct impact of consolidation on debt burdens. Specifically, 40 consolidated and 9 non-consolidated cities in South Korea were used as units of analysis. Because they were created as part of the reform of urban management in 1995 and 1998, this research can control for time and cross sectional effects when empirically examining the impact of city-county consolidation on debt burden, and comparing pre-consolidated and post-consolidated governments for samples of city-county consolidation.
The following section reviews the literature on the correlation between city-county consolidation and debt burden. Section 3 sketches the characteristics of city-county consolidation in South Korea. Section 4 and 5 provide data collection and method. Section 6 shows findings and discusses results. Section 7 offers a conclusion and future research directions.
2.1 DEBT BURDEN AND CONSOLIDATED LOCAL GOVERNMENTS: OBSERVATIONS
The theory of economies of scale is predicated on the belief that service providers can expect a decrease in average unit and service costs by increasing units of production and services (Hirsch, 1959). Following this theory, city-county consolidation may reduce debt burdens with decreased public service costs. However, many studies of city-county consolidation (discussed below) infer that merged governments may result in higher debt burdens (Blomquist & Parks, 1995; Cain, 2009; Cook, 1973; Durning & Nobbie, 2000; Rosentraub, 2000; Savitch & Vogel, 2004). Generally speaking, the existing literature prompts enough questions to warrant examining the correlation between city-county consolidation and the level of debt burden in more detail.
Research has offered different perspectives on whether consolidated governments are better able to achieve administrative efficiency in terms of reduced per capita spending (Cain, 2009; Cook, 1973; Durning & Nobbie, 2000; Vojnovic, 2000). A consolidated government can find optimal costs to deliver public services by balancing expenditures of different jurisdictions. Nevertheless, public service costs may increase if a consolidated government adopts the high service standards of a particular jurisdiction. For example, standards of education services are prone not to aim at optimal costs, but to follow the costs of the high service standards (Cook, 1973). Local government expenditures may be generally conceived in this way: Public officials reflect the demand of residents who want to receive improved public services from a consolidated government; hence they cannot easily downgrade the level of public services to reduce per capita expenditures and to meet optimal service costs. Durning and Nobbie (2000) surveyed employees in the city of Athens and Clarke County, Georgia, and concluded city-county consolidation is not associated with the improvement of administrative efficiency. Similarly, Cain (2009) examined how much city-county consolidation leads to decreasing per capita expenditures in individual departments. The findings were consistent with other scholars: among four consolidated cities, three are correlated with an increase in their expenditures (see, for example, Cook, 1973; Durning & Nobbie, 2000). Cain (2009) argued that consolidated cities tend to spend more on public safety, public works, parks and recreation. The literature of debt management concludes that the large scale of expenditures and tax revenues leads to increasing government debt burdens (Clingermayer & Wood, 1995; Day & Boeckelman, 2012). Additionally, public infrastructure expenditures resulting from constructing municipal facilities have been viewed as important determinants correlated with an increase in debt burden (Trautman, 1995). Therefore, city-county consolidation may be associated with an increase in debt burden.
This research assumes that city-county consolidation may lead to increasing debt burden due to the correlation between economic development and debt burdens. City-county consolidation is associated with an increase in income level and employment (Blomquist & Parks, 1995; Carr et al., 2006; Lowery, 2000; Rosentraub, 2000). Researchers have concluded that high income results in large municipal debt burdens (Ashworth et al., 2005; Brecher et al., 2003; Clingermayer & Wood, 1995; Temple, 1994). Urban management studies argue that city-county consolidation is associated with high salaries in manufacture, retail, and service industries (Carr et al., 2006). Consolidated governments foster economic development by providing more benefits, such as tax exemptions, to enterprises (Rosentraub, 2000). The companies offer new job opportunities to residents and this may increase income level. Consequently, city-county consolidation results in high income levels (Carr et al., 2006). In this regard, Faulk and Schansberg (2009) conclude that consolidation is associated with a decrease in the unemployment rate of Lafayette City-Lafayette Parish in Louisiana, but the same effect was not found to be statistically significant in other merged governments. Increased unemployment rates could lead to low municipal bond ratings and thus, local governments may shirk from borrowing funds from external resources due to high borrowing costs (Krueger & Walker, 2010). The point can be put another way. A high bond rating allows a local government to issue more debt due to low borrowing costs. Thus, this research assumes city-county consolidation may be correlated with an increase in debt.
Another key factor correlating city-county consolidation and debt burden is assessed value. Blomquist and Parks (1995) argue that a consolidated city has the ability to invest enough funds for public infrastructure because the increased assessed value provides a larger property tax base, and this helps raise the debt limit. In this regard, city-county consolidation contributes to increasing local tax revenues (Durning & Nobbie, 2000). The literature argues that the large scale of per capita revenues, total revenues, and assessed value are associated with an increase in debt burden (Clingermayer & Wood, 1995; Johnson, 1999). Poor governments who do not possess a large tax base cannot borrow large sums of money from external funding sources. However, consolidated governments can borrow more funds with increased assessed value and tax revenues. This connection provides an insight into the correlation between city-county consolidation and debt burden. It is also aligned with Blomquist and Parks' argument (1995) on the financial impact of consolidation. Thus, this research assumes that city-county consolidation may lead to increasing debt burden.
Political strength is the last potential factor potentially linking city-county consolidation and debt burden. Blomquist and Parks (1995) argue that stable leadership and voter turnout determine the degree of political strength and that city-county consolidation results in stronger political power by stabilizing leadership and raising voter turnout. One-party government can simplify a decision-making process because there are not as many decision-makers (potentially veto-players) participating in the process and, therefore, the government has higher political power (Ashworth et al., 2005). Similarly, political leaders elected by a high turnout and majority vote can implement their policies with a small number of veto-players. Further, consolidated governments are more likely to have a stable political structure consistently supported by one political party. Blomquist and Parks (1995) pointed out that city residents are likely to have a variety of political tendencies, whereas suburban governments receive overwhelming support from residents advocating a particular political party. Thus, a merged government may follow the political tendency of the suburbs. For example, the Republican Party has controlled the Indianapolis-Marion consolidated government since consolidation, but that was not the case when it was fragmented (Blomquist & Parks, 1995). Additionally, consolidation allows a political leader to have more authority than an executive of a fragmented government (Savitch & Vogel, 2004). Clingermayer (1991) concluded that debt burdens tend to be higher in governments operated by elected officials than by city managers. The existing literature shows a positive relationship not only between city-county consolidation and political power but also between political strength and the level of debt burden. Thus, this research infers that city-county consolidation may increase debt burden.
2.2 THE IMPACT OF CITY-COUNTY CONSOLIDATION ON DEBT BURDENS: HYPOTHESIS
The exploration of the existing literature provides five observations in the correlation between city-county consolidation and debt burden. On the basis of economies of scale, city-county consolidation may be associated with a decrease in debt burden through the per capita reduction of local expenditures as well as revenues. By contrast, city-county consolidation may lead to increasing debt burdens because the level of expenditures and revenues in consolidated governments is relatively higher than in fragmented governments. Concerning economic development, consolidation may result in the increase of debt burdens by raising employment and personal income. Assessed value of consolidated governments may be correlated with an increase in municipal debt burdens. Lastly, high political power of merged governments may lead to increasing debt burdens. The impact of city-county consolidation on the level of debt burden is not framed by the foundation of a particular theory because the empirical results have been mixed. This research found five phenomena for developing the connection between city-county consolidation and debt burden, which are efficiency, inefficiency, economic development, assessed value, and political power. Although city-county consolidation is correlated with other factors, such as education expenditures, local taxes, citizen participation, and so forth, these are not shown to be related to the propensity to borrow, and therefore are excluded from this study. Thus, the hypothesis, based on the factors identified in the literature, is as follows:
Hypothesis: City-county consolidation is associated with an increase in debt burden of local governments
3. CITY-COUNTY CONSOLIDATION: THE CONTEXT OF SOUTH KOREA
The stated purpose of city-county consolidation in South Korea has been different from that advanced in other countries. Local governments of the United States have proceeded with consolidation in order to enhance efficiency (Miller, 2002; Rusk, 1993). However, consolidation of local government in South Korea has focused on how the central government efficiently controls sub-governments (municipalities) with integrated management (Hong, 2005). Most city-county consolidations in South Korea have been forced by the central government, and are centered on a particular period. From 1995 to 2012, city-county consolidation created 44 governments, which were 44 cities and 44 counties. Among them, 39 cities were consolidated in 1995 and one city (Yeosu city-Yeocheon county) was created in 1998 (See Table 1). During the same period, nine cities failed to merge with other jurisdictions due to the opposition of residents (See Table 2). Additionally, Jeju City-Bukjeju County and Seogiwipo City-Namjeju County were created by referendum in 2005. Recently, Changwon City-Masan City-Jihae City were consolidated in 2010 and Cheongju City-Cheongwon County reached a consensus on consolidation in 2012.
Scholars can empirically examine the impact of city-county consolidation in Korea because of the sample size that allows for comparison between pre-and post-consolidation, and also comparison between consolidated and non-consolidated governments. Most research on city-county consolidation in South Korea is aligned with the empirical results of other countries, concluding that city-county consolidation has not had an impact on government efficiency (Choi, 2005; Hong, 2005; J.-G. Park, 2001; Yoo, 2010). However, Park and Hong (2007) partly support the proposition that city-county consolidation may be associated with a high location quotient score in a long-term period. Thus, city-county consolidation needs to be examined in light of the urban management policy of the central government in South Korea.
4. DATA COLLECTION
To test the hypothesis, this research examined panel data from 1986 to 2009 by using government-published data derived from the "Financial Yearbook of Local Government," and "Statistical Yearbook of Local Government," which are widely employed to analyze financial capacity and the socioeconomic situation in local governments (National Assembly Library, 2013). The Ministry of Public Administration and Security in South Korea issues an annual report providing the financial information of local governments. Specifically, the report provides details of local financial capacity, budget, tax revenues, and local expenditures. This research used 'local tax revenues', 'local shared tax revenues' and 'local borrowing' in the Financial Yearbook of Local Government as a dependent variable and control variables. Individual provincial governments publish an annual report of local governments (see Table 1). These provide local socioeconomic information, such as population, education, transportation, government and so forth. This research used 'population' and 'assessed value' as control variables in the report.
The units of analysis are 39 consolidated governments created in 1995 (See Table 1), one consolidated government (3) created in 1998 (Yeosu City-Yeocheon-County), and nine governments (4) that tried, but failed to consolidate (See Table 2). A total of 49 groups of post-consolidation governments were identified. The number of units is different between pre-consolidation and post-consolidation because the pre-consolidation governments are divided into 80 city and county jurisdictions (See Table 1). This research used city and county data from 1986 to 1995 and consolidated city data from 1996 to 2009. Because pre-consolidated governments embrace 41 cities and 39 counties, pre-consolidation was aligned with post-consolidation level in an attempt to establish balanced panel data. Additionally, non-consolidated units were merged into potential four consolidated units. So, the data covers 24 years, and 40 consolidated and 4 non-consolidated governments (44*24 = 1056).
This research used fixed effects regression analysis with the Driscoll-Kraay estimator to examine the impact of city-county consolidation on debt burdens. The fixed effects model is widely employed to control for omitted variables in longitudinal data and to compare between pre-effect and post-effect with binary variables (Stock & Watson, 2007) and the Driscoll-Kraay estimator is used for controlling heteroscedasticity of a panel model (Hoechle, 2007). The model is as follows:
[LB.sub.it] = [[beta].sub.1][CON.sub.it] + [[beta].sub.2][POP.sub.it] + [[beta].sub.3][LTR.sub.it] + [[beta].sub.4][LSTR.sub.it] + [[beta].sub.5][TAV.sub.it] + [[alpha].sub.i] + [u.sub.it],t = 1, 2, ..., T.
where [LB.sub.it] is the per capita total amount of local borrowing in Korean Currency Won (KRW) 1,000 value in city i during year t; [CON.sub.it] is a dummy variable for consolidated governments in city i during year t; [POP.sub.it] is population in city i during year t; [LTR.sub.it] is the per capita total amount of local tax revenues in KRW1,000 value in city i during year t; [LSTR.sub.it] is the per capita total amount of local shared tax revenues in KRW1,000 value in city i during year t; [TAV.sub.it] is the per capita total assessed value in KRW1,000 value in city i during year t; [[alpha].sub.i] is the unobserved effect; and [u.sub.it] is the error term.
The total amount of local borrowing is the dependent variable in the model. It has been employed to represent debt burden including borrowing costs in municipalities (Kriz & Wang, 2013). This research uses per capita local borrowing based on the existing literature (Clinger et al., 2008; Clingermayer, 1991; Clingermayer & Wood, 1995; Ellis & Schansberg, 1999; Hildreth & Miller, 2002; Temple, 1994; Trautman, 1995). U.S. local governments have authority to issue debt, whereas Korean municipalities do not borrow funds from external financial resources without the approval of the central government (Bae, Chang, & Joo, 2007). Given this circumstance, most local governments of South Korea are financially subordinate to the central government. The central government tends to approve bond issuance depending on financial capacity of municipalities, such as revenue sources. Thus, Korean municipal local borrowing is similar to General Obligation Bonds in the United States because they are guaranteed by local governments who are debt issuers (Lee & Johnson, 1998). This research defined local borrowing as the total amount of outstanding debt approved by the central government. The Financial Yearbook of Local Government annually provides the total amount of outstanding debt, which is accumulated local borrowing.
This research is intended to examine the impact of city-county consolidation on debt burden in South Korea. Given this research question, city-county consolidation is included as an independent variable in the model. The city-county consolidation is a dummy variable, meaning that pre-consolidated governments are coded as 0 and post-consolidated governments are coded as 1. Recently, Faulk and Grassmueck (2012) used a consolidation dummy variable to measure the impact of government consolidation on local expenditures. The city-county consolidation of 39 cities was completed in 1995 and one consolidated city was created in 1998 and therefore, this research can control for the effect of the treatment time, and has an advantage for measuring the difference between pre-consolidated and post-consolidated governments.
The exploration of the existing literature provides population (Ashworth et al., 2005; Clingermayer, 1991; Ellis & Schansberg, 1999; Garcia-Sanchez et al., 2012), per capita local tax revenues (Clingermayer & Wood, 1995; Johnson, 1999), per capita local shared tax revenues (Clingermayer & Wood, 1995; Martell & Smith, 2004), and per capita assessed value (Johnson, 1999) as control variables. Local tax revenue is mainly used for general expenditures. It includes the amount of inhabitant tax, property tax, automobile tax, agriculture income tax, butchery tax, tobacco consumption tax, motor fuel tax, city planning tax, facilities tax, and business firm tax. This research uses the per capita total amount of local tax revenues. The empirical research concludes that the level of debt burden fluctuates due to intergovernmental grants (Martell & Smith, 2004). Local shared tax revenue is an intergovernmental fund in terms of its purpose and usage. Its purpose is similar to lump-sum grants in the United States, which are distributed to sub-national governments depending on their spending and tax base (Fisher, 2007). Thus, this research also uses per capita total amount of local shared tax revenue as a control variable. The central government in South Korea financially supports local governments with local shared tax revenues and a subsidy for controlling deficits (Bae et al., 2007; Lim & Hong, 2012). Local shared tax revenues are not required to be spent for a particular government program, but the usage of the subsidy is determined by the central government. Further, subsidies are matching funds, meaning that local governments should partly share the funds at the same amount of subsidy as is supplied by the central government. As noted, the central government of South Korea does not tend to approve the plan of bond issuance if the fiscal health of a local government is not good, and therefore Korean municipal governments are likely to cover insufficient financial resources with a subsidy. Bae et al. (2007) conclude that the level of subsidy is correlated to high municipal debt burdens in South Korea. However, this research excluded a subsidy variable in the model due to an endogeneity problem. Lastly, per capita assessed value is employed as a control variable. City-county consolidation financially benefits local governments by increasing assessed value because they can borrow more external funds with the increased assessed value (Blomquist & Parks, 1995). Johnson (1999) agrees with this argument through the examination of a linkage between assessed value and bond issues. Although the existing literature provides other control variables, such as the level of private income, investment decisions, the number of public authorities, capital budget, public officials' turnover, unemployment rate and political strength, this research did not use them because the Korean census does not offer those variables at the level of city and county. This is one of the limitations of this research.
6. FINDINGS AND DISCUSSION
Table 3 shows the descriptive statistics of the model. The dependent variable is local borrowing, the independent variable is the consolidation effect, and the control variables are population, local tax revenues, local shared tax revenues, and total assessed value. All variables are per capita units excluding consolidation and population and are natural logarithms excluding consolidation. This research examined the impact of city-county consolidation on debt burden in 40 consolidated and four nonconsolidated cities from 1986 to 2009, so the number of sample units is 1,056. The results of the descriptive analysis appear in KRW1,000 value in per capita local borrowing, local tax revenues, local shared tax revenues, and total assessed value. This research tested the hypothesis with fixed effects regression analysis and the Driscoll-Kraay estimator.
Table 4 depicts the fixed effects model results of log per capita local borrowing. It shows that consolidation, log per capita local tax revenues, and log per capita total assessed value have a statistically significant effect on per capita local borrowing and these variables are positively related to the dependent variable. Specifically, the effect of local tax revenues supports the existing literature on debt burdens (Clingermayer & Wood, 1995; Johnson, 1999). Further, the finding of per capita total assessed value supports the argument that the size of assessed value leads to increasing debt burdens (Johnson, 1999). The results of the model support the hypothesis that city-county consolidation is associated with an increase in debt burdens of local governments. However, log population and log per capita local shared tax revenues are not statistically related to debt burdens. There is elasticity between the dependent and independent variable. Post-consolidated governments have more than 32.3 % higher per capita local borrowing compared with pre-consolidated governments. Thus, a consolidation effect is important for the level of debt burden and this result indirectly shows that city-county consolidation does not lead to decreasing but to increasing public service costs (Cain, 2009; Cook, 1973; Durning & Nobbie, 2000). This research additionally used time fixed effects and random-effects analysis to examine whether consolidation is consistently effective on debt burden.
Table 5 depicts the time fixed effects model results of log per capita local borrowing. The results are similar to the fixed effects model with Driscoll-Kraay estimator. The findings show that consolidation and log per capita total assessed value are correlated with an increase in per capita local borrowing and have a statistically significant effect on the dependent variable. The model reflects the history of economic recession in South Korea. 1997 and 2009 have a statistically significant effect on debt burden of local governments (p-value < 0.1). Both periods had economic recession in South Korea. Also, the recession resulted in low debt burden because it seems that the central government was unlikely to approve issuance of municipal debt in economic crisis.
GLS random effects regression analysis was used to compare differences between consolidated and non-consolidated cities. Table 6 shows that consolidation, consolidated cities, log per capita local tax revenues, log per capita local shared tax revenues, and log per capita total assessed value have a statistically significant effect on local borrowing. Consolidated governments have more than 18.9% higher per capita local borrowing compared with non-consolidated governments.
As a result, all models proposed by this research argue that city-county consolidation is associated with an increase in the debt burden of local governments. Although the results of population and local shared tax revenues are not aligned with previous conclusions, the results of local tax revenues and total assessed value verified the existing theories on debt burden. The model infers that the concept of economies of scale no longer provides a theoretical foundation for city-county consolidation as local government reform. Thus, this research suggests that urban managers need to approach the policy of city-county consolidation not in an attempt to reduce public service costs and debt burdens but for other purposes, such as the quality of public services, responsiveness, accountability, economic development, transparency, and so forth.
Consolidation might be one of the solutions for accomplishing financial sustainability in terms of economies of scale. Debt management is one of the important public tasks to make sustainable financial conditions. Municipalities have issued bonds to construct public infrastructure. In this regard, consolidated governments could save financial resources by sharing public infrastructure and achieve financial sustainability. However, this research suggests that municipalities should not adopt consolidation to accomplish financial sustainability in terms of debt management.
In summary, this research defined the concept of consolidation and suggested a research question to fill the gap between city-county consolidation and debt burden. To verify the research question, this research found the linkage between city-county consolidation and debt burden through the existing literature. The exploration of the studies provided impact factors for connecting city-county consolidation to the level of debt burden, which are public expenditures and revenues, economic development, assessed value, and political strength. Fixed effects regression analysis was employed to examine the hypothesis that city-county consolidation may be associated with an increase in debt burden. The results of the model support the hypothesis.
Some limitations of the model should be noted. First, this research did not include other control variables affecting debt burdens because the data is not available in South Korea. The existing literature provides investment decision, the number of local authorities, capital budget, public official's turnover, unemployment rate, personal income, and political strength as potentially significant variables. Thus, this research cannot fully exclude a random error effect, systematic bias, and less confidence in the model. If the relationship between city-county consolidation and debt burdens is examined in the context of the United States, those factors need to be applied to a research model. Second, although this research proposed and tested the hypothesis by offering theoretical underpinnings based on the existing literature, the results in this research are confined to the context of South Korea. Hence, this research does not assure findings that can be applied to the cases of city-county consolidation in other countries. Third, the model does not have timely autocorrelation in the Wooldridge test, but this research does not fully control for serial correlation in terms of the Durbin-Watson test. Finally, this research did not use other dependent variables, such as borrowing costs, representing debt burden, because the Korean census does not provide them. Future research needs to consider the limitations in examining the relationship between city-county consolidation and debt burdens in local governments.
The author is indebted to Dr. Carol Ebdon, Dr. Kenneth Kriz, Dr. Dale Krane, and Dr. Jooho Lee for their helpful comments on the manuscript. I also wish to thank my colleagues--James Harrold, Michael Pippin, Anthony Campbell, and James Haavisto--for reviewing and commenting on early drafts. I would like to thank PFM's editors and the anonymous reviewers for their constructive comments and criticisms.
Ashworth, J., Geys, B., & Heyndels, B. (2005) "Government weakness and local public debt development in Flemish municipalities," International Tax and Public Finance, 12(4): 395-422.
Baber, W. R., & Sen, P. K. (1986) "The political process and the use of debt financing by state governments," Public Choice (1986-1998), 48(3): 201.
Bae, S., Chang, D., & Joo, K. (2007) "The impact of goverment subsidiary on debt issurance," Paper presented at the Korean Association of Public Administration.
Benton, J. E., & Gamble, D. (1984) "City/County consolidation and economies of scale: Evidence from a time-series analysis in Jacksonville, Florida," Social Science Quarterly, 65(1): 190-198.
Berman, D. R. (2003) Local government and the states : autonomy, politics, and policy. Armonk, N.Y.: M.E. Sharpe.
Blomquist, W. & Parks, R. B. (1995) "Fiscal, service, and political impacts of Indianapolis-Marion County's Unigov", Publius, 25(4): 37-54.
Brecher, C., Richwerger, K., & Wagner, M. V. (2003) "An approach to measuring the affordability of state debt," Public Budgeting & Finance, 23(4): 65-85.
Cain, B. (2009) "The impact of city-county consolidation on local Government finance," Capstone Paper.
Carr, J. B., Bae, S., & Lu, W. (2006) "City-county government and promises of economic development: A tale of two cities," State and Local Government Review, 38(3): 131-141.
Choi, Y. (2005) "The financial impacts of city-country consolidation, with special reference to major public facilities in Jeju Province," Journal of the Korean Urban Management Association, 18(3): 147-169.
Clinger, J. C., Feiock, R. C., McCabe, B. C., & Park, H. (2008) "Turnover, transaction costs, and time horizons: An examination of municipal debt financing," The American Review of Public Administration, 38(2): 167-179.
Clingermayer, J. C. (1991) "An intergenerational transfer model of state debt financing," Public Choice (1986-1998), 72(1): 13-13.
Clingermayer, J. C., & Wood, B. D. (1995). "Disentangling patterns of state debt financing," The American Political Science Review, 89(1): 108-120.
Cook, G. C. A. (1973) "Effect of metropolitan government on resource allocation: The case of education in Toronto," National Tax Journal (pre-1986), 26(4): 585.
Day, J. & Boeckelman, K. (2012) "The impact of legislative term limits on state debt: Increased spending, flat fevenue," Politics & Policy, 40(2): 320-338.
Denison, D. V., Yan, W., & Zhao, Z. (2007) "Is management performance a factor in municipal bond credit ratings? The case of Texas school districts," Public Budgeting & Finance, 27(4): 86-98.
Drukker, D. M. (2003) "Testing for serial correlation in linear panel-data models," The Stata Journal, 3(2): 168-177.
Durning, D. & Nobbie, P. (2000) "Post-transition employee perspectives of city-county unification: The case of Athens-Clarke County," Public Administration Quarterly, 24(2): 140-168.
Ellis, M. A. & Schansberg, D. E. (1999) "The determinants of state government debt financing," Public Finance Review, 27(6): 571-587.
Faulk, D. & Grassmueck, G. (2012) "City-county consolidation and local government expenditures," State and Local Government Review, 44(3): 196-205.
Faulk, D. & Schansberg, E. (2009). "An Examination of Selected Economic Development Outcomes from Consolidation," State and Local Government Review, 41(3): 193-200.
Feiock, R. C., & Carr, J. B. (1997) "A reassessment of city/county consolidation: Economic development impacts," State and Local Government Review, 29(3): 166-171.
Fisher, R. C. (2007). State & local public finance (3e [ed.]. ed.). Mason, OH, USA: Thomson South-Western.
Garcia-Sanchez, I. M., Mordan, N., & Prado-Lorenzo, J. M. (2012) "Effect of the political system on local financial condition: Empirical evidence for Spain's largest municipalities," Public Budgeting & Finance, 32(2): 40-68.
Hildreth, W. B. & Miller, G. J. (2002) "Debt and the local economy: Problems in benchmarking local government debt affordability," Public Budgeting & Finance, 22(4): 99-113.
Hirsch, W. Z. (1959). "Expenditure implications of metropolitan growth and consolidation," The Review of Economics and Statistics, 41(3): 232-241.
Hoechle, D. (2007) "Robust standard errors for panel regressions with cross-sectional dependence," Stata Journal, 7(3): 281-312.
Hong, J. (2005) "Intra-and inter-regional variations in the balanced growth effects of city-county consolidation," Korean Ssociety and Public administration, 16(1): 299-324.
Johnson, C. L. (1999) "Tax increment debt finance: An analysis of the mainstreaming of a fringe sector," Public Budgeting & Finance, 19(1): 47-67.
Kriz, K. A., & Wang, Q. (2013). Measuring and Monitoring Debt Capacity and Affordability: Market-and Nonmarket-based Models. In H. Levine, J. B. Justice & E. A. Scorsone (Eds.), Handbook of Local Government Fiscal Health. Burlington, MA: Jones and Bartlett Learning.
Krueger, S. & Walker, R. W. (2010) "Management practices and state bond ratings," Public Budgeting & Finance, 30(4): 47-70.
Lee, Robert D., & Johnson, Ronald Wayne. (1998). Public budgeting systems (6th ed.). Gaithersburg, Md.: Aspen Publishers.
Lim, S. & Hong, S. (2012) "Analysing the determinants of local borrowings in Korea," The Korean Journal of Local Finance, 17(1): 83-106.
Lowery, D. (2000) "A transactions costs model of metropolitan governance: allocation versus redistribution in urban America," Journal of Public Administration Research and Theory, 10(1): 49-78.
Martell, C. R., & Smith, B. M. (2004) "Grant levels and debt issuance: Is there a relationship? Is there symmetry?" Public Budgeting & Finance, 24(3): 65-81.
Miller, D., Miranda, R., Roque, R., & Wilf, C. (1995) "The fiscal organization of metropolitan areas: The Allegheny County case reconsidered," Publius: The Journal of Federalism, 25(4): 19-35.
Miller, D. Y. (2002). The regional governing of metropolitan America. Boulder, Colo. ;Oxford: Westview Press.
National Assembly Library (2013) Financial yearbook of local government. Accessed January 15, 2013 (www.nanet.go.kr).
National Assembly Library (2013) Statistical yearbook of local government. Accessed January 25, 2013 (www.nanet.go.kr).
Ostrom, V., Bish, R. L., Ostrom, E., & Institute for Contemporary, Studies. (1988). Local government in the United States. San Francisco, Calif.: International Center for Economic Growth affiliated with the Institute for Contemporary Studies.
Park, J. & Hong, J. (2007) "Regional economic growth effects of city-county consolidation in Korea," Journal of the Korean Association for Policy Studies, 16(1): 167-197.
Park, J. (2001) "Can the industrial merger theory explain the structure/process of local government consolidations?" Korean Public Administration Review, 34(4): 303-322.
Robbins, M. D., & Kim, D. (2003) "Do state bond banks have cost advantages for municipal bond issuance?" Public Budgeting & Finance, 23(3): 92-108.
Rosentraub, M. S. (2000) "City-county consolidation and the rebuilding of image: The fiscal lessons from Indianapolis's UniGov program," State and Local Government Review, 32(3): 180-191.
Rusk, D. (1993). Cities without suburbs. Washington, D.C. :Baltimore, Md.: Woodrow Wilson Center Press ;Distributed by the Johns Hopkins University Press.
Selden C. S. & Campbell, R. W. (2000) "The expenditure impacts of unification in a small Georgia county: A contingency perspective of city-county consolidation," Public Administration Quarterly, 24(2): 169-201.
Savitch, H. V., & Vogel, R. K. (2004) "Suburbs without a city: power and city-county consolidation," Urban Affairs Review, 39(6): 758-790.
Savitch, H. V., & Vogel, R. K. (2000) "Metropolitan consolidation versus metropolitan governance in Louisville," State and Local Government Review, 32(3): 198-212.
Stock, J. H., & Watson, M. W. (2007). Introduction to econometrics (2nd ed.). Boston: Pearson/Addison Wesley.
Temple, J. (1994) "The debt/tax choice in the financing of state and local capital expenditures," Journal of Regional Science, 34(4): 529-547.
Trautman, R. R.. (1995). "The impact of state debt management on debt activity," Public Budgeting & Finance, 15(2): 33-51.
Vojnovic, I. (2000) "The transitional impacts of municipal consolidations," Journal of Urban Affairs, 22(4): 385-417.
Yoo, J. (2010) "Explaining city-county consolidation initiatives in Korea and proposing a policy alternative," Korean public administration review, 44(1): 179-202.
Ji-Hyung Park, University of Nebraska Omaha
(1.) Temple (1994) illustrates two types of public capital investment decisions. First, in systematic investment decisions, public officials try to find an optimal level of planned capital spending that reflects the financial capacity of local governments, such as tax revenues. Second, voters determine the level of capital expenditures without any consideration of how municipalities can spend funds for constructing public infrastructure. She concludes the former investment pattern tends to decrease government debt burdens (1994).
(2.) Ashworth et al. (2005) contend that the number of political parties determines the degree of government fragmentation. Fragmented governments may not easily initiate public programs because they need to convince political parties who have different perspectives on policies. Following this view, they define weak governments as politically fragmented governments (2005). The findings show that weak governments have more debt than strong governments supported by one party (Ashworth et al., 2005).
(3.) The Yeosu City-Yeocheon-County consolidation occurred later than the other 39 consolidated governments, but is included here to increase sample size.
(4.) Non-consolidated governments were employed to examine the difference between consolidated and non-consolidated groups in this research, but the findings do not show any effects because fixed effect regression analysis automatically omitted the dummy variable of non-consolidated cities. This research tried to use GLS random effects regression analysis to compare differences between consolidated and non-consolidated cities. Consolidated governments have more than 18.9% higher per capita local borrowing compared with non-consolidated governments. However, this research does not consider the finding because of the Hausman test (See Table 4).
(5.) There is serial correlation based on the Durbin-Watson Test in the model because the budgetary decision-making process in South Korea is based on incrementalism. However, the Wooldridge test is applicable because the panel is a long time series. The result of the Wooldridge test shows that the model does not have serial correlation (Drukker, 2003).
(6.) All governments and periods do not issue local borrowing in South Korea. Thus, the fixed effects regression analysis automatically excludes missing values.
Table 1. Consolidated Regions Former Jurisdiction Consolidated Province City County City Kyounggi Pyoungtaek Pyoungtaek Pyoungtaek /Songtan Migum Namyangju Namyangju Gangwon Chuncheon Chuncheon Chuncheon Wonju Wonju Wonju Gangneung Myoungju Gangneung Samcheok Samcheok Samcheok Chungbuk Chungju Chungwon Chungju Jecheon Jecheon Jecheon Chungnam Cheonan Cheonan Cheonan Gongju Gongju Gongju Daecheon Boryoung Boryoung Onyang Asan Asan Seosan Seosan Seosan Jeonbuk Gunsan Okgu Gunsan Iri Iiksan Iiksan Jeongju Jeongeup Jeongeup Namwon Namwon Namwon Gimje Gimje Gimje Jeonnam Yeosu Yeocheon Yeosu Suncheon Seungju Suncheon Naju Naju Naju Dong- Gwangyang Gwangyang gwangyang Kyoung-buk Pohang Youngil Pohang Gyoungju Gyoungju Gyoungju Gimcheon Geumneung Gimcheon Andong Andong Andong Gumi Seonsan Gumi Youngju Youngpung Youngju Youngchun Youngchun Youngchun Sangju Sangju Sangju Jeomchon Munkyoung Munkyoung Gyoungsan Gyoungsan Gyoungsan Kyoung-nam Changwon part of Changwon Changwon Masan part of Changwon Masan Jinju Jinyang Jinju Chungmu Tongyoung Tongyoung Samcheonpo Sacheon Sacheon Gimhae Gimhae Gimhae Milyang Milyang Milyang Jangseungpo Geojae Geojae Sources: Hong (2005). "Intra- and Inter-Regional Variations in the Balanced Growth Effects of City-County Consolidation," Korean Society and Public Administration, 16(1): 299-324. Table 2. Non-Consolidated Regions Target Areas Cheongju city Cheongwon county Sokcho city Yangyang county Mokpo city Muan county Dongducheon city Uijeongbu city Yangju county Sources: Hong (2005). "Intra- and Inter-Regional Variations in the Balanced Growth Effects of City-County Consolidation," Korean Society and Public Administration, 16(1): 299-324. Table 3. Descriptive Statistics of the Model Standard Variable N Mean Deviation Local Borrowing 1056 21.94 43.27 Consolidation (consolidated city = 1056 0.63 0.48 1, nonconsolidated city=0) Population 1056 236798.10 137920.60 Local Tax Revenues 1056 189.51 169.07 Local Shared Tax Revenues 1056 409.86 515.63 Total Assessed Value 1056 80.19 864.86 Variable Minimum Maximum Local Borrowing 0.00 499.70 Consolidation (consolidated city = 0.00 1 00 1, nonconsolidated city=0) Population 70791 792944.00 Local Tax Revenues 6.92 2032.83 Local Shared Tax Revenues 8.71 4497.70 Total Assessed Value 0.00 23032.17 Table 4. Model Results (5) Driscoll--Kraay log per capita local borrowing coefficient Standard Errors Consolidation 0.323 0.106 log population 0.320 0.380 log per capita local tax revenues 0.502 0.216 log per capita local shared tax 0.399 0.255 revenues log per capita total assessed value 0.104 0.033 constant -2.745 2.243 R-square 0.546 F 147.20 Hausman test 0.003 Wooldridge test 0.489 N [683.sup.6] log per capita local borrowing t p value Consolidation 3.04 0.004 log population 0.84 0.405 log per capita local tax revenues 2.32 0.025 log per capita local shared tax 1.56 0.125 revenues log per capita total assessed value 3.17 0.003 constant -1.22 0.228 R-square F Hausman test Wooldridge test N Table 5. Time Fixed Effects Model Results Standard log per capita local borrowing coefficient Errors Consolidation 1.537 0.549 log population 0.083 0.386 log per capita local tax revenues -0.286 0.275 log per capita local shared tax revenues 0.055 0.224 log per capita total assessed value 0.084 0.028 year 1987 -0.345 0.213 1988 -0.217 0.193 1989 0.107 0.224 1990 0.042 0.249 1991 0.511 0.275 1992 0.659 0.303 1993 0.704 0.327 1994 0.658 0.345 1995 -0.679 0.249 1996 -0.453 0.236 1997 -0.383 0.214 1998 -0.240 0.208 1999 -0.181 0.203 2000 -0.211 0.214 2001 -0.591 0.194 2002 -0.014 0.182 2003 -0.174 0.171 2004 -0.200 0.154 2005 -0.264 0.142 2006 -0.017 0.139 2007 -0.126 0.135 2008 0.000 (omitted) 2009 0.302 0.124 Constant 0.254 2.358 R-square 0.566 F 33.95 N 683 log per capita local borrowing t p value Consolidation 2.8 0.005 log population 0.21 0.831 log per capita local tax revenues -1.04 0.299 log per capita local shared tax revenues 0.24 0.807 log per capita total assessed value 2.97 0.003 year 1987 -1.62 0.107 1988 -1.12 0.262 1989 0.48 0.633 1990 0.17 0.865 1991 1.86 0.064 1992 2.18 0.030 1993 2.15 0.032 1994 1.91 0.057 1995 -2.73 0.007 1996 -1.92 0.055 1997 -1.79 0.074 1998 -1.16 0.248 1999 -0.89 0.374 2000 -0.98 0.326 2001 -3.05 0.002 2002 -0.08 0.938 2003 -1.02 0.309 2004 -1.3 0.195 2005 -1.87 0.063 2006 -0.12 0.903 2007 -0.94 0.350 2008 2009 2.43 0.016 Constant 0.11 0.914 R-square F N Table 6: Random-Effects GLS Regression Model Results Standard log per capita local borrowing coefficient Errors Consolidation 0.330 0.064 Consolidated cities 0.189 0.106 log population 0.053 0.172 log per capita local tax 0.438 0.118 revenues log per capita local shared 0.495 0.100 tax revenues log per capita total assessed 0.070 0.023 value constant -1.604 0.998 R-square 0.537 Wald 794.83 N 683 log per capita local borrowing z p value Consolidation 5.12 0.000 Consolidated cities 1.78 0.075 log population 0.31 0.760 log per capita local tax 3.73 0.000 revenues log per capita local shared 4.94 0.000 tax revenues log per capita total assessed 3.09 0.002 value constant -1.61 0.108 R-square Wald N
|Printer friendly Cite/link Email Feedback|
|Publication:||Public Finance and Management|
|Date:||Jun 22, 2013|
|Previous Article:||The impact of amalgamation on the financial sustainability of municipalities in Estonia.|
|Next Article:||An evaluation of amalgamation and financial viability in Australian local government.|