International Comparison of Government Size in Terms of Revenues and Expenditures

Transcription

1 International Comparison of Government Size 1 International Comparison of Government Size in Terms of Revenues and Expendures Sung Tai Kim * Myoung Kyu Kim ** Byung In Lim *** Taek-Dong Yeo **** Abstract Most governments are interested in finding the optimal government size of s own country, since the direction of fiscal policy depends on the current state of government size. In this paper, we suggest that we compare government size in terms of both revenues and expendures for international comparison. Until now we have used International Tax Comparison Index (ITC) for revenue side comparison of government size, and/or we have used International Expendure Comparison Index (IEC) for expendure side comparison of government size separately. The problem of sole use of eher ITC or IEC is such that we may learn the direction of only one side of fiscal policy, eher revenue side or expendure side. Also, we often encounter the problem that controlled variables are different for estimation of ITC and IEC. Based on (ITC, IEC) combination we classify sample countries into four categories (B, B), (B, S), (S, B), and (S, S). We may also keep track of government size of given country in terms of (ITC, IEC). Advantage of this method is that we can use (ITC, IEC) combination for planning the optimal direction of fiscal policy. In this paper we estimate ITC and IEC using the same control variables for 30 OECD countries during periods. Key Words: government size, International Tax Comparison Index, International Expendure Comparison Index, fiscal policy JEL Classification: H1, H2, H5 * Professor, Department of Economics, Cheongju Universy, stkim@cju.ac.kr ** Research Associate, Chungbuk Research Instute, econ@cri.re.kr *** Associate Professor, Department of Economics, Chungbuk National Universy, billforest@hanmail.net **** (Corresponding Author) Professor, School of International Economics and Business, Yeungnam Universy, tdyeo@yu.ac.kr, Tel:

2 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 2 I. INTRODUCTION Most governments are interested in finding the optimal government size of her own country, since the direction of fiscal policy depends on the current state of government size. In this paper, we intend to compare government sizes in terms of both revenues and expendures for international comparison and year by year comparison. Until now we have used International Tax Comparison Index (ITC hereafter) for revenue side comparison of government size, and/or we have used International Expendure Comparison Index (IEC hereafter) for expendure side comparison of government size separately. For each index 100 is crical level such that the current size is normal. The problem of sole use of eher ITC or IEC is such that we may learn the direction of only one side of fiscal policy, eher revenue side or expendure side. Also, we often encounter the problem that controlled variables are different for estimation of ITC and IEC. Based on (ITC, IEC) combination, we classify sample countries into four categories (BB), (BS), (SB), and (SS). We may also keep track of government size of given country in terms of (ITC, IEC) along the time path. Advantage of our method is such that we can use (ITC, IEC) combination for planning the optimal direction of fiscal policy. For instance, for the case of (BB) country the optimal direction of fiscal policy is to decrease tax revenues and expendures simultaneously. On the contrary, for the case of (SS) country the optimal direction of fiscal policy is oppose. In this paper we estimate ITC and IEC using the same control variables for 30 OECD countries during periods. Using the panel data we found the fixed effect model is optimal among three model specifications OLS, random effect model, and fixed effect model - for (government revenue/gdp) estimation equation and (government expendure/gdp) estimation equation. We can use (ITC, IEC) estimation results in various ways. Firstly, we can keep track of one country s ITC and IEC along the time path. Secondly, we can compare a group of countries given period. In section 4, for the country

3 International Comparison of Government Size 3 study example we analyze Korea for periods. Also we analyze OECD countries in terms of (ITC, IGC) given periods. 2. DETERMINANTS OF GOVERNMENT SIZE In this section we will briefly review the determinants of government size theoretically. Government size and her growth have been the focus of many public economists interests recently. Especially there have been many studies which try to explain the change in the size of public sector in the U.S.A. in the period after World War II. 1 Most of them can be viewed as part of eher a responsive government or an excessive government interpretation of public sector growth. The former views government as reacting to external demands for expansion, while the latter views government self as a major source of growth. Empirically in general the government sizes have grown implying that the roles of public sector become more crucial and diversified. Theoretically the government size is determined by the social preference over public and private good and capabily of the government to provide public goods. In two-good economy the government size is determined by the optimal allocation condion between private and public sector that the marginal rate of substution of public good for private good must equal the marginal rate of transformation of public good for private good. In this paper we will not set up the theoretical model for determinants of the government size. However, we assume that the optimal government size of a specific country is determined by her own socio-economic factors such as per-capa GDP, the size and the structure of population, development stage of the economy, openness of the economy, and industry structure etc. Thus, we call a country big government when the size of government of the country is bigger than that taking into account her socio-economic factors. Similarly we call small government when the size of government is smaller 1 See Tarchys(1975) and Lowery and Berry(1983) for instance.

4 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 4 than that taking into account her socio-economic factors. 3. EMPIRICAL ANALYSIS ON ESTIMATING ITC AND IEC 3.1 ITC Estimation Definion of ITC Definion of International Tax Comparison Index (ITC hereafter) is as follows. Actual( R / GDP) ITC (3-1) Normal ( R / GDP ) where R denotes national burden that is the sum of tax burden and all kinds of extra burden of cizen such as pension payment and med-care payment. 2 Normal (R/GDP) ratio of country i in year t represents the level of (R/GDP) ratio taking into account that country s socio-economic environments. Normal (R/GDP) ratio may be interpreted as optimal (R/GDP) ratio. However, we should be very careful about that. That kind of interpretation is valid only if the specification of ITC equation is correct. Estimation of ITC Equation In order to estimate the ITC for country i in year t we should estimate the normal (R/GDP) ratio. In the first stage of estimation of the ITC we postulate the assumption that certain group of countries follows the same pattern in determination of (R/GDP) ratio in the process of socio-economic development. In this paper we assume that OECD countries follow the same pattern of government s raising her revenues and in the process of their 2 This is a broadened concept of burden of people.

5 International Comparison of Government Size 5 socio-economic development. Based on the assumption we postulated, we may derive the estimation equation for (R/GDP) ratio. R c log( GDPPC) log( POP) DEP GDP RUPOP RUPOPGR AGR log SERR log( POP14) log( POP65) (3-2) FOREIGNOPEN 10 u Explanations of variables in equation (3-2) are as follows. R/GDP GDP. The dependent variable R/GDP denotes the ratio of national burden to Log(GDPPC) We can use the per capa GDP variable (Log(GDPPC)) in equation (3-2) for proxy variable to represent economic development stage for country i in year t. The channel of how the per capa GDP affects (R/GDP) ratio is as follows. As the per capa GDP increases tax revenues will increase accordingly. On the other hand, an increase in per capa GDP decrease (R/GDP) ratio. Thus the net effect of an increase in per capa GDP is theoretically ambiguous. Log(POP) The Log(POP) variable represents the logarhm of population. The population size of one country represents the size of that country. In general as the population size becomes larger, tax revenues will increase accordingly so that (R/GDP) ratio becomes larger. We may also expect that the size of public sector becomes larger as the size of the economy increases, since the

6 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 6 government provides more public goods such as highways, national defense expendures, and welfare expendures. Thus, the effect of the population size on the ratio of government expendures to GDP will be posive. DEP The DEP variable represents the dependency ratio which is the ratio of dependent population (the sum of under age 15 people and above 65 people) to total population. This variable represents the process of aging of one country. As aging proceeds in the economy, there will be both posive and negative effects on government revenues. On the one hand, aging implies more pension and medical related payments so that government revenues will increase. On the other hand aging of the economy implies weakening of potential economic growth resulting in decreased tax revenue. Theoretically the net effect of aging on (R/GDP) ratio is ambiguous. RUPOPP The process of social development of a country also affects (R/GDP) ratio posively. We use two proxy variables for social development stage. The variable RUPOP means the size of rural population. RUPOPGR The variable RUPOPGR means the growth rate of rural population. AGR Independent of economic development stage, the industrial structure of a country may affect the government size in terms of (R/GDP) ratio. We may consider the agricultural sector and service sector to take into account the industrial structure. The variable AGR means the ratio of agricultural output to total industry output. In agricultural sector earned incomes are not realized by monetary income and most of incomes are earned by so many small farmers. In addion is very difficult for the tax collector to collect taxes. Therefore (R/GDP) ratio will decrease as the portion of agricultural sector

7 International Comparison of Government Size 7 increases. SERR The variable SERR means the ratio of service sector output to total industry output. Log(POP14) The variable Log(POP14) means the ratio of population size under age 14. Log(POP65) The variable Log(POP65) means the ratio of population size above age 65. FOREIGNOPEN In general as the trade volume of one country increases meaning that the degree of openness becomes larger, the GDP will increase proportionally. However, taxes and tariff revenues do not increase significantly as the trade volume increases, since the tariff rate and commody tax rates on expert goods are not that high in developing and developed countries. Thus, the degree of openness of one country does not significantly affect the (R/GDP) ratio We may notice that in equation (3-2) the subscript is attached to all variables where i denotes the country, t denotes the year and t=1970,, In the case of panel data the disturbance term u may be decomposed into two parts. u i (3-3) where i represents the individual effect and represents the residual term. Two alternative specification of the model are possible in their treatment of the individual effect. The so-called fixed effect model treats i as a fixed but

8 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 8 unknown component differing across individual cross-section uns. In this paper the cross-section uns are thirty OECD countries. Thus i may include the characteristics of tax scheme and pension scheme as well as normative value judgments on government role. The alternative is known as the random effect or variance component model in which i is assumed to be drawn from i.i.d. distribution. Table 3-1 presents the estimation results for equation (3-2). In Table 3-1 the OLS (Ordinary Least Square) estimation result in column (1), the fixed effect model estimation result in column (2), and the random effect model estimation result in column (3). Hausman(1978) suggested to test the null hypothesis E[ i x i ]=0 against the alternative hypothesis E[ i x i ] 0, where x i represents a vector of explanatory variables in equation (3-2). The chi-square test statistic value is which well exceeds that is 1% crical value wh degree of freedom 10, we reject the null hypothesis. Therefore we may conclude that the fixed effect model is optimal model. We explain the estimation results based on the fixed effect model. Theoretically as per capa GDP increases the national burden relative to GDP may eher increase or decrease. Empirically turns out that (R/GDP) ratio decreases as per capa GDP increases. The size of country in terms of population affects (R/GDP) ratio posively. The effect of aging on (R/GDP) ratio is posive since aging of society will increase pension payments. RUPOPGR representing the development stage of country have posive coefficients following theoretical arguments. The variables AGR and SERR representing the industry structure of country have expected signs of coefficients.

9 International Comparison of Government Size 9 Table 3-1 Estimation Results for ITC Equation Explanatory Variables Constant (1) OLS *** (12.332) Dependent Variable: Revenue/GDP (2) Fixed Effect Model *** (28.725) (3) Random Effect Model *** (13.868) Log(GDPPC): Per Capa GDP 0.262*** (0.098) *** (0.778) (0.417) Log(POP): Population *** (0.177) *** (1.913) (0.602) DEP: Dependency Ratio 0.320** (0.124) 0.722*** (0.099) 0.243*** (0.088) RUPOP: Population of Rural Area RUPOPGR: Population Growth Rate of Rural Area AGR: Portion of Agriculture Sector SERR: Portion of Service Sector *** (1.893) (0.151) *** (0.074) (0.036) (4.535) 0.637*** (0.081) *** (0.055) *** (0.028) * (4.101) 0.492*** (0.083) *** (0.055) (0.026) Log(POP14): (2.972) *** (2.312) *** (2.089) Log(POP65) 7.299*** (1.424) *** (1.609) ** (1.449) FOREIGNOPEN *** (0.006) *** (0.006) *** (0.006) Adjusted R-squares (0.3954) (0.3379) F-Statistic(Wald χ 2 ) *** 63.77*** (504.88***) Number of Observations 1,017 1,017 1,017 Note: Absolute t-values are in parentheses. denotes the optimal model specification. ***: 1%, **: 5%, *: 10% significance levels respectively.

10 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 10 Estimation of ITC We can estimate the values of ITC for country i in year t as follows. ITC y 100 (3-5) yˆ where y is the actual value of (R/GDP) ratio for country i in year t and y ˆ is the fted value of (R/GDP) ratio for country i in year t. For instance in the case of Korea we can use the estimation results from the fixed effect model in Table 3-1 to obtain the fted value of (R/GDP) ratio during periods. 3.2 IEC Estimation Definion of IEC Definion of International Government Expendure Comparison Index (IEC hereafter) is as follows. Actual( G / GDP) IEC (3-6) Normal ( G / GDP ) where G denotes government expendure. Normal (G/GDP) ratio of country i in year t represents the level of (G/GDP) ratio taking into account that country s socio-economic environments. Normal (G/GDP) ratio may be interpreted as optimal (G/GDP) ratio. However, we should be very careful about that. That kind of interpretation is valid only if the specification of IEC equation is correct. Estimation of IEC Equation In order to estimate the IEC for country i in year t we should estimate the normal (G/GDP) ratio. In the first stage of estimation of the IEC we postulate the assumption that certain group of countries follows the same pattern in

11 International Comparison of Government Size 11 determination of (G/GDP) ratio in the process of socio-economic development. In this paper we assume that OECD countries follow the same pattern of government s expendures in the process of their socio-economic development. Based on the assumption we postulated, we may derive the estimation equation for (G/GDP) ratio. G c log( GDPPC) log( POP) DEP GDP RUPOP RUPOPGR AGR log SERR log( POP14) log( POP65) FOREIGNOPEN 10 u (3-7) Explanations of variables in equation (3-7) are the same as above. Estimation Results Table 3-2 presents the estimation results for equation (3-7). In Table 3-2 the OLS estimation result in column (1), the fixed effect model estimation result in column (2), and the random effect model estimation result in column (3). We executed a Hausman specification test for equation (3-7). Based on the Hausman test, we conclude that the fixed effect model is optimal for IGC estimation equation (3-7). We can interpret the estimation results for the fixed effect model. All explanatory variables except RUPOP variable and DEP variable are statistically significant. The signs of coefficients cope wh the expected signs based on theory Estimation of IEC We can estimate the values of IEC for country i in year t using the similar method to ITC.

12 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 12 Table 3-2 Estimation Results for IEC Equation Explanatory Variables Constant Dependent Variable: Expendure/GDP (1) OLS ** (19.677) (2) Fixed Effect Model *** (58.069) (3) Random Effect Model *** (24.693) Log(GDPPC): Per Capa GDP (0.153) *** (1.834) *** (0.544) Log(POP): Population *** (0.255) *** (4.055) *** (0.749) DEP: Dependency Ratio (0.225) (0.241) (0.212) RUPOP: Population of Rural Area RUPOPGR: Population Growth Rate of Rural Area AGR: Portion of Agriculture Sector ** (2.606) *** (0.206) (0.117) (9.240) 0.283* (0.161) *** (0.121) ** (6.941) (0.163) ** (0.117) SERR: Portion of Service Sector (0.051) 0.392*** (0.061) 0.231*** (0.056) Log(POP14): ** (4.939) (5.224) * (4.776) Log(POP65) *** (2.655) *** (3.709) 9.483*** (3.115) FOREIGNOPEN (0.008) *** (0.014) *** (0.012) Adjusted R-squares F-Statistic(Wald χ 2 ) *** 44.96*** (364.04***) Number of Observations Note: Absolute t-values are in parentheses. denotes the optimal model specification. ***: 1%, **: 5%, *: 10% significance levels respectively..

13 International Comparison of Government Size GOVERNMENT SIZE IN TERMS OF ITC AND IEC 4.1 Classification of Government Size in Terms of ITC and IEC In this section we analyze the government size in terms of both ITC and IEC in general. We can classify the government size of a specific country in a specific year into four categories in terms of ITC and IEC. In Figure 4-1 X- axis represents the ITC and Y-axis represents the IEC. Figure 4-1 Four Quadrants of ITC and IEC Space IEC 115 Ⅱ (SB) type 110 Ⅰ (BB) type ITC 95 Ⅲ (SS) type 90 Ⅳ (BS) type 85 As shown in Figure 4-1, there are four types of government size, namely (BB), (BS), (SB), and (SS) type. The first B means that the ITC is bigger than 100, while the first S means that the ITC is smaller than 100. Similarly

14 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 14 the second B means that the IEC is bigger than 100, while the second S means that the IEC is smaller than 100. Thus (BB) means that both ITC and IEC are bigger than 100. The type (BB) government size is belonging to the quadrant I in Figure 1. Firstly, In the case of (BB), we may say that the government is too big. It implies that the government collects more taxes than normal level and she spends more expendures than normal level. By normal level sometimes we mean the optimal level of government size. For this type of government size of a country in a year the government could have achieved higher level of social welfare by decreasing both tax revenues and government expendures in that year. Secondly, the type (SS) means that both ITC and IEC are smaller than 100 representing, so called, the small government. The type (SS) government is belonging to the quadrant III in Figure 4-1. The type (SS) government collects less tax than the normal level and spends less expendure than the normal level. For this type of government size of a country in a year the government could have achieved higher level of social welfare by increasing both tax revenues and government expendures in that year. Thirdly, the type (BS) means that the ITC is bigger than 100 while the IEC is smaller than 100. The type (BS) government is belonging to the quadrant IV in Figure 4-1. The type (BS) government collects more tax than the normal level and spends less expendure than the normal level. For this type of government we cannot say that the government is eher too big or too small. Usually the type (BS) government may go through fiscal surplus, because of more tax and less expendure relative to normal levels. Lastly, the type (SB) means that the ITC is smaller than 100 while the IEC is bigger than 100. The type (SB) government is belonging to the quadrant II in Figure 1. The type SB government collects less tax than the normal level while she spends more expendure than the normal level. For this type of government we may not say that the government is eher big or small for similar reason to (BS) type. Often the type (SB) government should go through fiscal defic because of less tax and more expendure relative to

15 International Comparison of Government Size 15 normal levels. 4.2 Year by year Comparison of Government Size: the Case of Korea In Table 4-1 the ITC and IEC are shown for Korea during periods. Table 4-1 ITC and IEC for Korea: Revenue / GDP Govt Expendure / GDP year ITC year ITC year IEC year IEC Note: ITC: International Tax Comparison Index. IEC: International Government Expendure Comparison Index In year 1975 the Korean government size started from BB type. During periods the Korean government size followed the BB type.

16 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 16 During periods the government size showed the SS type which implies the small government. For periods the government size showed the (BS) type. That period was in some sense transional period from small government to big government. Since 1991 the government size showed the BB type until 2005 except For that period Korea went through financial crisis followed by real economic crisis in Fiscally the government activies were weakened during the period. For periods the government size showed the (BB) type again. Table 4-2 Type of Government for Korea: year Type of Gov t year Type of Gov t year Type of Gov t year Type of Gov t 1973 BB 1982 BB 1991 SS 2000 SS 1974 BB 1983 BB 1992 SS 2001 SS 1975 BB 1984 BB 1993 SS 2002 SS 1976 BB 1985 SS 1994 SS 2003 BB 1977 BB 1986 SS 1995 SS 2004 SS 1978 BB 1987 SS 1996 SS 2005 SS 1979 BB 1988 SS 1997 SS 2006 BB 1980 BB 1989 BS 1998 SS 2007 BB 1981 BB 1990 BS 1999 SS 2008 BB Note: ITC: International Tax Comparison Index. IEC: International Government Expendure Comparison Index From the starting point 1975 (ITC=109.2, IEC=102.0) year by year the Korean government size is shown as one point in the ITC-IEC space. During periods Korea remained in the first quadrant showing (B, B). During the period the posion moved toward the origin wh decreasing ITC and IEC. In 1985 the government size posion moved into the quadrant III. Since 1985 the points remained in the (S, S) quadrant until 2002 except for 1989 and 1990 in which Korea belonged to the (B, S) quadrant. Since 2006 the Korean government posion came back in the (B, B) quadrant. It seemed that recently the Korean government has been heading for big government

17 International Comparison of Government Size 17 probably due to increasing government expendures to overcome global financial crisis. Figure 4-2 Loci of (ITC, IEC) for Korea: Korea, Rep._ITC_IEC iec_fe c_fe 4.3 Cross-country Comparison of Government Size We calculated the ITC and IEC values for 30 OECD countries in each year during periods based on the fixed effect model estimation results shown in Table 3-1 and 3-2. Using the estimates of the ITC and IEC we can compare the government size for a group of countries given period. For example, recently (BB) type countries are Korea, Denmark, Norway, Chile, and Denmark, On the other hand, (SS) type countries are Japan, Australia, Austria, Belgium, Canada, Swzerland, Czech Republic, Germany, Iceland, Sweden, Turkey, Netherlands, New Zealand, and Poland.

18 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 18 The (SB) type countries are Spain, Finland, France, UK, Greece, Ireland, Luxemburg, Slovenia, and USA. We pointed out that the (SB) type countries are likely to experience fiscal defics and the hypothesis is supported by empirical evidence that two thirds of the (SB) type countries are currently going through fiscal defic problem. Finally, (BS) type countries are Hungary and Italy. 4.4 Example: the case for the PIGS The methodology of this paper is original and we expect that our methodology measuring the government size in terms of both revenues and expendures be useful for evaluation fiscal status of specific countries. In Figures 4-3 we show the loci of ITC and IEC for the PIGS Portugal, Italy, Greece, and Spain which have gone through fiscal defic problem recently. The four PIGS countries show different individual characteristics in the loci of (ITC, IEC). However, there are a couple of common movements of the loci for the PIGS. Firstly, all four show a clockwise spinning movements in (ITC, IEC) space. For example Portugal started from the (SB) quadrant II and went through the quadrant I and IV, then stayed in the quadrant III, showing (SB) (BB) (BS) (SS) movement. Italy showed the clockwise movement (SS) (BS) (BB) (BS), then stayed in the (BS) quadrant since Greece started from the (SS) quadrant and moved up to the (SB) quadrant and the (BB) quadrant back and forth, then moved down to the (SS) quadrant again. Thus, Greece showed (SS) (SB) (BB) (SS) ended at the (SB) quadrant in Spain also showed the clockwise movement (SS) (BS) (BB) (BS) (SB). Wh a ltle twist the PIGS show a clockwise spinning movements in (ITC, IEC) space. It will be very crucial and interesting research theme why the countries like PIGS show a clockwise spinning movements. We conjecture that such a clockwise movement is related to polical business cycle. Secondly, the PIGS eher ended up wh or moved toward the (SB)

19 80 90 iec_fe 95 iec_fe International Comparison of Government Size 19 quadrant. Figure 4-3 Loci of (ITC, IEC) for Portugal, Italy, Greece, Spain: Portugal_ITC_IEC c_fe 1992 Italy_ITC_IEC c_fe

20 iec_fe iec_fe Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 20 Greece_ITC_IEC c_fe Spain_ITC_IEC c_fe

21 International Comparison of Government Size CONCLUSION AND POLICY IMPLICATIONS In this paper we introduced a new methodology by which we can compare the government size in terms of both revenues and expendures. Up to now we compare the government size in terms of eher revenues or expendures. In this paper we analyzed the government size in terms of both ITC and IEC in general. We classify the government size of a specific country in a specific year into four categories in terms of ITC and IEC. There are four types of government size, namely (BB), (BS), (SB), and (SS) type. In the case of (BB), we may say that the government is too big. Secondly, the type (SS) means that both ITC and IEC are smaller than 100 representing, so called, the small government. Thirdly, the type (BS) means that the ITC is bigger than 100 while the IEC is smaller than 100. The type (BS) government collects more tax than the normal level and spends less expendure than the normal level. Usually the type (BS) government may go through fiscal surplus, because of more tax and less expendure relative to normal levels. Fourthly, the type (SB) means that the ITC is smaller than 100 while the IEC is bigger than 100. The type SB government collects less tax than the normal level while she spends more expendure than the normal level. The type (SB) government should go through fiscal defic because of less tax and more expendure relative to normal levels. One of the advantages of using our method of comparing the government size is such that we may derive the optimal direction of fiscal policy for a specific country and specific year. For the (BB) type government, for instance, we suggest that the government decrease both revenue and expendure, vice versa. For (SB) type government we may suggest that the government try to increase tax revenue and to decrease expendure. We illustrated three examples for our new methodology for comparison of government size. Firstly, we traced the loci of (ITC, IEC) for periods for Korea. We found that for periods turned out that the government type is (BB) meaning the too big government. The policy implication of this empirical finding is such that we are now heading for the

22 Sung Tai Kim Myoung Kyu Kim Byung In Lim Taek-Dong Yeo 22 big government. Secondly, using the estimates of the ITC and IEC we can compare the government size for a group of countries given period. We classify 30 OECD countries in 2008 into four groups (BB), (SB), (BS), (SS). We noted that the (SB) type countries are likely to experience fiscal defics. The hypothesis is supported by empirical evidence that two thirds of the (SB) type countries are currently going through fiscal defic problem. Thirdly, we examined the loci of ITC and IEC for the PIGS Portugal, Italy, Greece, and Spain which have gone through fiscal defic problem recently. There are a couple of common movements of the loci for the PIGS. First, all four show a clockwise spinning movements in (ITC, IEC) quadrants. Second, the PIGS eher ended up wh or moved toward the (SB) quadrant.

23 International Comparison of Government Size 23 REFERENCES Aschauer, David Alan, (1985), Fiscal Policy and Aggregate Demand, American Economic Review 75, Berry, William D., Lowery, David. (1987), Explaining the Size of the Public Sector: Responsive and Excessive Government Interpretations (Book Review), Journal of Polics, Vol. 49, Issue 2, p Berry, William D., Lowery, David. (1984), The Measurement of Government Size: Implications for the Study of Government Growth, Journal of Polics, Vol. 46, Issue 4, p Garrett, Thomas A., Rhine (2006), On the Size and Growth of Government, Russell M. Review, Vol. 88, Issue 1, pp , Ghali, Khalifa H. (1999), Government Size and Economic Growth: Evidence from a Multivariate Cointegration Analysis, Applied Economics, Vol. 31, Issue 8, pp Hausman, J. A. (1978), Specification tests in econometrics, Econometrica, Vol. 46, pp Heller, Peter S. and J. Diamond, 1990, International Comparison of Government Expendure Revised-The Developing Countries-, , October paper No. 69, Washington: International Monetary Fund. Karras, Georgios. (1997), On the Optimal Government Size in Europe: Theory and Empirical Evidence, Manchester School of Economic and Social Studies, Vol. 65, Issue 3, pp Karras, Georgios. (1994), Government Spending and Private Consumption: Some International Evidence, Journal of Money, Cred, and Banking, Vol. 26, Issue 1, pp Lowery, David and William D. Berry (1983), The Growth of Government in the Uned States, American Journal of Polical Science 97, World Bank (2004), World Development Indicator.