An investigation into the relationship between government budget deficits, defense expenditure and transfer payments in Iran

Transcription

1 An investigation into the relationship between government budget deficits, defense expenditure and transfer payments in Iran Dr. Ali Hussein Samadi Assistant Professor of Economics, Shiraz University, Shiraz, Iran. Mohammad Behboodi M.S in Economics, Shiraz University, Shiraz, Iran Abstract The main aim of this paper is to investigate the relationship between government budget deficits, defense expenditure and transfer payments in Iran during the period of To this end a 5-variable- VECM and IRF have been used. Results of VECM reveal the positive effect of defense expenditure, educational expenditure, and transfer payments on budget deficits and also the negative effect of health expenditures on budget deficits in the long-run. Also evidence indicates the rejection of Ricardian Equivalence Theorem (RET) in Iran. Therefore, government can t use budget deficits and defense expenditures to transfer incomes between different social-income and inter-generation groups. Keywords: Budget Deficit, Defense Expenditure, Income Redistribution, Cointegration Tests, Vector Error Correction Model (VECM), Impulse Response Function (IRF), IRAN JEL Codes: H51, H52, H56, H62, C13, C30, I38 21

2 Introduction In recent years, the growth of financial deficit, which accounts for the increase of government s debt, has begun to interest financial policy-makers. While a few economists consider government s budget deficits as the reason for economic problems such as high real interest rate, imbalances in the balance of payments, among others, a few other economists consider that creating deficits in government budgets through expansionary fiscal policies as the reason for economic development and is useful to surmount recession and unemployment. From the Keynesian point of view, an increase in government s debt justifies expansionary fiscal policies (including transfer payments) resulting in an increase in the private sector income. An increase in private sector income stimulates consumption and causes a decrease in national saving, which in longterm causes an increase in price level, production, and interest rate and also causes a decrease of private investment. Ricardian Equivalence Theorem (RET) in contrast to Keynesian point of view points out that increases in government s debt cause no increase in private sector wealth, but lead to increases in private consumption and decreases of saving in current period. So as a result, price level, production, and interest rate won t change, even if it is possible that private sector saving increases for the purpose of supplying future tax liability in the long-run. From this perspective, we can examine whether budget deficit can be used as a policy to transfer income among different income groups and generations. This paper is organized as follows. Section 1 explores the rationale and objectives of the study. In the second section we review some related studies briefly. In section third the methodology and data are explained. Major findings and discussions of empirical results are presented in the fourth section and conclusions are presented in fifth section. 1. Rationale and Objectives of the Study If budget deficit influences real economic variables, it can be used as a mean for redistributing income among different generations by government (Seiglie, 1997). In this case, tax policies, government subsidies, and government s transfer payments have a critical role in successful transfer of income among different income and inter-generation groups in the current period (Meltzer and Richard, 1981). So the budget deficit can be used as a mean for redistributing income (Cukierman and Meltzer, 1989). It is believed that defense expenditures impose a great pressure on government s budget; and especially in third-world countries, it causes the budget deficit to increase. In Iran, due to the income structure and dependence on oil revenues, budget deficits have been considered as an inseparable part of governmental budgets for many years. On the other hand, in third- 22

3 world countries such as Iran, a great part of economic activities is handled by government. Therefore the ongoing budget deficit has significant effects on the formation of gross national product, influencing investment, trade balance, and balance of payments. This paper aims to analyze the relationship between the government s budget deficit, defense expenditures and transfer payments in Iran during the period of To accomplish that process, the main objective of this paper is to investigating that whether the budget deficit due to various factors such as increase in defense expenditure can help to improve income distribution or not. 2. Related Literature Hess (1989) estimated the relationship between defense expenditures and economic growth within simultaneous equations (equations for military burden, economic growth, and Human Suffering Index) framework. He estimated this model using 2SLS method for 66 third-world countries for the period Results of this study reveal the direct relationship between defense expenditures and economic growth. Looney (1990) reviewed the military/nonmilitary regime and economic efficiency from the comparative budgetary process point of view. He especially dealt with whether budgeting models among third-world military/nonmilitary countries are different, and if so, whether this difference is to the extent that military expenditures in different regimes influence social-economic expenditures and consequently, quality of life. He concluded that third-world military regimes increase their own military power severely and a great amount of government s budget is allocated to military defense. Later, Frederickson and Looney (1994) analyzed Pakistan s defense expenditures during the period They concluded that defense expenditures and budget deficit are correlated and that predicting the effect of any special plan is difficult. However, they concluded that once military burden increases, government decreases fundamental plans and social issues and adjust budget deficit by decreasing social expenditures. In 1994, Using the USA s time-series data during the period of , Abel dealt with defense expenditures and income distribution by controlling macroeconomic variables such as tax, economic growth, interest rate, inflation, and nonmilitary expenditures. He concluded that defense expenditures cause more income imbalances among different social groups. Melman (1974) also showed that economic cost of continued war is the increase of income imbalances. In another survey Seiglie (1997) investigated the relationship between transfer programs and government s defense expenditures on the American Federal s budget deficit in the period of

4 He concluded that defense expenditures and budget deficit are correlated and also both are sources of income transfer among different social groups. He claimed that during the middle of the 1960s America s budget deficit began, and revealed a structural transfer, which caused a mechanism for redistribution of wealth among different social groups and future generations. Özsoy (2008) reviewed the relationship between the budget deficit, defense expenditures, and redistribution of income among different social-income groups in Turkey by means of a VAR model covering the period Results of the VAR estimation and Impulse Response Function show that governments can use the budget deficit and defense expenditures as instruments for income transfer among different social groups and inter-generations. More recently, Elveren (2012) examines the relationship between defense spending and income inequality in Turkey during He used the Theil Index as a proxy for overall income distribution. The findings show a causality running from defense spending to income inequality, as expected. Which means curtailing military spending would be a significant policy option and may be considered a means to improve income distribution. Wang et al (2012) integrate Malmquist productivity index (MPI) with bootstrapping to establish statistical inferences that provide a complete and effective analysis of the impact of defense expenditure on economic productivity between 1993 and 2009 for OECD member countries. Their findings indicate that the average MPI with defense expenditure is higher than that without defense expenditure. Also, their results prove that the effective defense expenditure strategies undertaken by government are important to improve economic productivity of countries. Apparently, a limited number of studies about defense expenditures in Iran have been carried out too. For instance, Beyzaei (2001) dealt with the effects of defense expenditures on economic growth during the period and showed that there is a significant and negative relationship between the military burden and economic growth. He also concluded that there is a relationship between the military burden and war conditions. Moreover, Hassani Sadrabadi and Kashmari (2007) investigated the influence of defense expenditures on economic growth and private consumption in Iran during For this purpose, the Feder Model (which consists of private consumption, exports, non-defense government, and defense expenditure) was used. The results of this study showed that the direct effect of defense expenditures on economic growth is positive, and indirect effect of defense expenditure on the private consumption is negative. Based on this brief review of literature, no study investigated the relationship between defense expenditures, budget deficits and transfer payments focusing on Iran. So, this study strives to fill this gap. 24

5 In the above mentioned studies except Özsoy (2008) and Elveren (2012) the role of defense spending and budget deficit in reducing the income inequality has been neglected. Moreover, it is worth mentioning that, in the studies related to Iran, only the impact of defense spending on economic growth is considered. 3. Methodology The mixed impacts (positive and negative effects) of defense expenditure on macroeconomic variables are accepted, whether in the third-world countries or in developed ones. It is possible to use the defense expenditures as expansionary fiscal policy by the government to stimulate national economy and redistribution of national income among different social groups. Defense expenditures may influence economy through many ways. From Keynesian point of view, the increase of defense expenditures - as a part of government s expenditures - leads to the total demand increase and by stimulating demand, employment increases and jobless peoples are employed, and all of above mentioned factors cause an increase in production, employment, and economic growth. Nonetheless, an increase in government activities, in addition to its consequences, causes government to encounter problems in financing its expenditures, as it has encountered budget deficit in many years. Since, defense expenditures include a great part of government s expenditures, we can claim that it is one of the main reasons for budget deficit. With regard to endogenity of the variables, the ECM and Johansen-Juselius cointegration tests have been used. These methods are interrelated. IRF is also a suitable method to measure the reaction of variables to changes in another. Accordingly, ECM and IRF methods have been used in this study Model Error Correction Model (ECM) is a dynamic model that relates the short-term behavior of variables to their equilibrium amounts. The existence of cointegration relation between economic variables is the statistical basis for the use of ECMs. These models have substantial empirical appeal and the main reason is their ability to relate short-term fluctuations of variables to their long-term equilibrium status. Hence, in this paper the Vector Error Correction Model (VECM) is employed to determine the relationship between variables. The general form of VECM is: ( ) Where Y is appointed variables of model and is defined as: DEFGNP: the ratio of government s budget deficit to GNP DSGNP: the ratio of government s defense expenditures to GNP EDGNP: the ratio of government s educational expenditures to GNP HEGNP: the ratio of government s health expenditures to GNP 25

6 TPGNP: the ratio of government s transfer payments expenditures to GNP B i (i = 1, 2 p-1) are coefficients and U t is the vector of disturbances terms. The matrix includes information about the long-run balanced relationships. Also = in which is the adjustment coefficient of imbalance and shows the speed of adjustment to the long-run balance and matrix is the coefficient of long-run equilibrium relationships. The sentence Y t-p is the error correction term (ECT). In this paper, annual time-series data, obtained from Central Bank of Iran cover the period of We also use the ratio of private saving variable to GNP (PSAVING) is used to test RET. The trend of variables are showed in Figure (1). Statistical characteristics of variables are also depicted in Table (1). It is concluded from the table (1) that budget deficit with the mean of 4.5%, has the standard deviation of 3.89% and the maximum amount of 14.33%. Defense expenditures of government lies between 1.87% and 9.05% and have the total mean of 4.9% and standard deviation of 2.21%. Figure (1): trends of variables ( ) 1. Data is available only during this period. 26

7 Table (1): the statistical characteristics of variables DEFGNP DSGNP EDGNP HEGNP TPGNP PSAVING Mean Maximum Minimum Std. Dev Jarque-Bera Major Findings and Discussion To avoid spurious results, after determining the stationarity of variables and cointegration relation between variables, we estimate the model. IRF is also used and RET is tested, as well Unit Root Tests To determine the stationarity of variables, the Augmented Dickey-Fuller (ADF) test is used. The results of this test have been depicted in Table (2). The results show that all variables are non-stationary, but the difference of them is stationary. So, all the variables are I (1). Table (2): the results of ADF test Variable Level First difference Interpretation DEFGNP ** I(1) DSGNP ** I(1) EDGNP * I(1) HEGNP ** I(1) TPGNP ** I(1) *: significant at the 5% level **: significant at the 1% level 4.2. Cointegration Test For reviewing the presence or lack of cointergration relation between variables and generally to determine the long-term relationships between variables, the Johansen-Juselius method is used. For this purpose we use two tests, namely, the trace test and the maximum eigenvalue test. Results have been shown in Table (3). According to the results in Table 3, the number of cointergration vectors according to both tests is 1. So the variables of model have cointergration relationship and there is a long-term relationship between them. The cointegration vector is shown in table 1 in appendix 2. 27

8 Table 3: results of the trace test and the maximum Eigenvalue test for determining the number of cointergration vectors Null hypothesis for both tests Alternative Hypothesis for trace test Alternative Hypothesis for maximum Eigenvalue test Statistics of trace test Statistics of maximum Eigenvalue test Critical values (95%) maximum trace Eigenvalue test test r=0 r 1 r= r=1 r 2 r= r=2 r 3 r= r=3 r 4 r= r=4 r 5 r= Ricardian Equivalence Hypothesis Testing The RET hypothesis shows that the increase of government s debt causes a decrease in private consumption or its equivalence, increase in private saving. But from Keynesian point of view an increase in government s debt leads to the increase in income of private sector and stimulates consumption and decreases saving. Therefore, to understand whether budget deficit can be used as a policy for income transfer among different groups and generations, we review private saving in comparison to government s budget deficit. The continuum of changes of DEFGNP and PSAVING variables in Figure 1 illustrate that changes in private saving and government s budget deficit, in this example, are against each other. This suggests the rejection of RET hypothesis in Iran s economy during the period in question. For the purpose of gaining more reliable results, we estimate the effect of the budget deficit on private saving. For such a purpose we specify and estimate a model like the one specified below. In this model we consider private sector saving as endogenous variable and government s budget deficit as exogenous variable. The results of estimation are showed below: PSAVING = * DEFGNP (1.6167) (0.2713) [ ] [ ] 28

9 The results reveal that, the effect of government s budget deficit on private saving is significant and negative 2. In other words, an increase of the budget deficit causes a decrease in saving and consequently an increase in consumption and this shows a rejection of the RET hypothesis in Iran during this period. The results are similar to those of Abbasian and Nuri (2007) and FarzinVash and Farahbakhsh (2009). The reason could be that consumers in Iran have a shorter spending time horizon in comparison to the government s spending time horizon and also the presence of an imperfect capital market that causes consumers to become sensitive to current income. Furthermore, since most of government s income is from the oil industry, it is inferred that government has no obligation to supplying its own budget deficit which then causes unreliability in the future incomes and taxes VECM Estimation Results As all the variables are I (1) and regarding the cointergration test results, VECM is a suitable method for evaluating the relationship between variables. As shown in Table 4, all the lag-selection criteria except AIC pick lag 1 are the optimal lag length. AIC criterion shows the number of lags as 3. While adding variables lags causes no improvement in dispelling problems such as serial correlation, Heteroskedasticity and specification bias, and more lags of variables leads to loss of degrees of freedom, so we choose lag 1 as the optimal lag length. Table 4: VECM lag length selection Lag LogL LR FPE AIC SC HQ NA * * * * * LR: sequential modified LR test statistic (each test at 5% level) FPE: Final prediction error AIC: Akaike information criterion SC: Schwarz information criterion HQ: Hannan-Quinn information criterion In order to show the structural break associated with the effect of 1978 revolution on the trend of variables, a dummy variable "DUM" is used. It takes a value of 0 during and 1 during The results indicate a negative but not significant relation between government s budget deficit and its first lag. Government s defense expenditures in the previous period have a significant and negative 2 standard deviation are in parenthesis and t-statistic are in brackets 29

10 effect on government s budget deficit. Defense expenditures also have a negative and significant relation with its previous period s amount. The effect of budget deficit on health expenditures is negative and not significant. The effect of government s transfer payments also on its own previous period is significant and positive. The effect of educational expenditures also on its own previous period is significant and positive. The results of VECM indicate positive and significant effect of defense expenditures, educational expenditures, and government s transfer payments on the government s budget deficit in long-term. The results also indicate negative effect of health expenditures on budget deficit in the long-term. Furthermore, results indicate negative and not significant effect of revolution on the government s defense expenditures and positive and significant effect on government s transfer payment. The coefficient of error correction term (ECT) shows that 18 percentage of the difference in defense spending from its long run level is corrected each period significantly. Also, 20 percentage of the difference in budget deficits from its long run level is corrected to each period, but this coefficient is not significant. The VECM estimation results are presented in Appendix Impulse Response Function(IRF) IRF results show the dynamic behavior of each variable at the time of imposing impulse. These impulses usually are chosen as 1 unit of standard deviation. IRF shows the effect of shocks to budget deficit, defense expenditures, educational expenditures, health expenditures, and government s transfer payment on each other. The obtained results of IRF are shown in the Figure 2. The initial shock to budget deficit has had a positive and significant effect on itself during a period and then has had negative effect on itself for the next period. This shock also has negative and significant effect on defense expenditures for two periods, but has negative and not significant effect on educational expenditures. The effect of above mentioned shock on health expenditures and government s transfer payment is trivial. Shocking defense expenditures has positive and significant effect on government s defense expenditures for two periods. This shock has positive and significant effect on itself for a period and then has negative effect for the next period. This shock also has positive and not significant effect on the health expenditures and has negative and not significant effect on transfer payment. 30

11 Figure 2: Impulse Response Function (IRF) for all models variables Response of D(DEFGNP) to Cholesky One S.D. Innovations Response of D(DSGNP) to Cholesky One S.D. Innovations EFGNP D D SGNP EDGNP H EGNP TPGNP EFGNP D D SGNP EDGNP H EGNP TPGNP Response of D(EDGNP) to Cholesky One S.D. Innovations Response of D(HEGNP) to Cholesky One S.D. Innovations EFGNP D D SGNP EDGNP H EGNP TPGNP EFGNP D D SGNP EDGNP H EGNP TPGNP Response of D(TPGNP) to Cholesky One S.D. Innovations EFGNP D D SGNP EDGNP H EGNP TPGNP Shocking educational expenditures for a period has positive effect and for the next period has negative effect on budget deficit. This shock has negative effect for two periods on defense expenditures 31

12 and a period after which has positive and significant effect on defense expenditures and also has positive effect on health expenditures for two periods and negative effect on government s transfer payment. The above mentioned shock has positive and significant effect on itself for a period and then has negative and not significant effect on itself. Shocking health expenditures for a period has positive effect and for the next period has negative and not significant effect on budget deficit. The effect of this shock on defense expenditures is negative and significant for two periods and for the next period is positive, and also has positive and significant effect on educational expenditures for a period and has negative and not significant effect for the next period. Shocking health expenditures has positive and significant effect on itself for a period. Shocking transfer payment for a period has negative and not significant effect on budget deficit and defense expenditures and for the next period has positive and significant effect on which. This shock has positive and not significant effect on educational expenditures for two periods and has positive and significant effect on health expenditures for three periods. Also above mentioned shock has positive and significant effect on itself for four periods. 5. Summary and Conclusion In this paper, the relationships between government s defense expenditures and government s transfer payment during the period of have been studied. The results obtained from Impulse Response Function show that shocking budget deficit effects significantly only on itself and defense expenditures. IRF also shows that the positive and significant effect of the shock to defense expenditures on budget deficit is for two consecutive periods. Estimation of VECM indicates positive and significant effect of defense expenditures, educational expenditures, and transfer payments, and also negative and significant effect of health expenditures on budget deficit in the long run. This is done while defense expenditures in the previous period have a significant and negative effect on budget deficit. Defense expenditures also have a negative and significant relation with its previous term. The results indicates the RET hypothesis rejection in Iran during investigation period. The reason for such issue in Iran can be consumer s shorter time horizon in comparison to the government s time horizon and also the presence of imperfect capital market. Furthermore, as most of the government s income is from the oil industry, it is inferred that the government has no obligation to finance its own budget deficit which this issue causes unreliability in the future incomes and taxes. So, government can t use budget deficit and defense expenditures as a mean for income transfer among different social-income and inter-generation groups. 32

13 6. References Abbasian, E. and Nuri, R. (2007) Ricardian Equivalence Testing in Iran, Tahghighat e eghtesadi, No.79, pp (in Persian). Abel, J.D. (1994) Military spending and income inequality, Journal of Peace Research, vol. 31, 1, pp Barro, R.J. (1974) Are government bonds net wealth?, Journal of Political Economy, vol. 82, 6, pp Beyzaei, E. (2001) Relationship between military expenditure and some economic variables in Iran in , Humanism journal of Alzahra university, No.37, 38 (in Persian). Brzoska, M. (1983) The Military Related External Debt of Third World Countries, Journal of Peace Research, vol. 20, 3, pp Central Bank of the Islamic Republic of Iran, National Accounts of Iran, different years. Cukierman, A. and Meltzer, A.H. (1989) A political theory of government debt and deficits in a neo- Ricardian framework, American Economic Review, vol. 79, 4, pp Elveren. A.Y. (2012) Military Spending and Income Inequality: Evidence on Cointegration and Causality for Turkey, , Defense and Peace Economics, Vol. 23, 3, pp FarzinVash, A. and Farahbakhsh, N. (2009) Hypothesis tests of two deficits in Iran, Journal of Economic science (JES), No.11, pp (in Persian). Frederickson, P.C. and Looney, R. (1994) Budgetary consequences of defense expenditures in Pakistan: short-run impacts and long-run adjustments, Journal of Peace Research, vol. 31, 1, pp Hasani SadrAbadi, M.H. and Kashmari, A. (2007) Impact of government expenditure on economic growth and its indirect effect on private consumption in Iran (investigation of supply side economic model), The Economic Research, No.2, pp (in Persian). Hess, P. (1989) The military burden, economic growth and human suffering index: evidence from LDCs, Cambridge Journal of Economics, vol. 13, pp Looney, R. (1990) Militarization, military regimes and general quality of life in the Third World, Armed Forces and Society, vol. 17, 1, pp Melman, S. (1974) The Permanent War Economy: American Capitalism in Decline, New York: Simon & Schuster. Meltzer, A.H. and Richard, S.F. (1981) A rational theory of the size of government, Journal of Political Economy, vol. 89, 5, pp Noferesti, M. (1999) Unit Roots and Cointegration in Econometrics, Rasa Cultural Services institute (in Persian). Özsoy, O. (2008) Government Budget Deficits, Defense Expenditure and Income Distribution: The Case of Turkey, Defense and Peace Economics, Vol. 19, 1, February, pp Özsoy, O. (2001) Budgetary trade-offs between defense, education and health expenditures: the case of Turkey, Defense and Peace Economics, vol. 12, 6, pp Samadi, A.H. and Pahlavani, M. (2009) Cointegration and structural break in economic, University of sistan and baluchestan (in Persian). Seiglie, C. (1997) Deficits, defense, and income redistribution, CATO Journal, vol. 17, 1, pp Tashkini, A. (2006) Applied Econometrics with Microfit, 1th edition, Tehran s Dibagaran publisher (in Persian). Verner, J.G. (1983) Budgetary trade-offs between education and defense in Latin America: a research note, Journal of Developing Areas, vol. 18, 1, pp Wang. T.P., Shyu. S.H.P. and Chou. H.C. (2012) The impact of defense expenditure on economic productivity in OECD countries, Economic Modeling, vol 29, pp Appendix 1: source of data 33

14 Appendix 2: VECM Estimation Results 34