Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Evaluation of Revenue and Expenditure Hypotheses: Evidence from Pakistan (1972-2015) Using EViews Software Muhammad Baqir Ali1 and Dr. Amjad Amin2 M.S. Scholar, Department of Economics, University of Peshawar, KPK, Pakistan 2 Assistant Professor, Department of Economics, University of Peshawar, KPK, Pakistan

1

ABSTRACT In the field of public finance economics, the potential links between government revenue and government expenditure is still an unsettled issue. Due to increased budget deficits and defaults in many developed and developing countries, it has intensely attracted the attention of policy analysts, theorists and politicians. On a theoretical front, four major hypotheses have resulted about the inter-temporal association between taxation (revenue) and spending. The first hypothesis proposed by Friedman (1978) and Buchanan and Wagner (1978) is the tax-spend hypothesis (where there is a unidirectional causality from tax to expenditure); the spend-tax hypothesis proposed by Peacock and Wiseman (1961), (where there is a unidirectional causality running from spend to tax); the fiscal synchronization hypothesis (where there is bidirectional causality and decisions are made concurrently)(Meltzer & Richard, 1981; Musgrave, 1966), and the institutional separation hypothesis or fiscal neutrality introduced by Beghestani and McNown (1994) relates that there is no connection between government revenue and expenditure such that no causal relation between revenue and spending is to be expected. This paper aims to investigate the causal relationship between government revenue and government expenditure of Pakistan, using time series data from 1972 to 2015. To forecast empirical and dynamic impact of random disturbances on the systems of variables, five steps VAR technique has been used. After testing empirical results shows no relationship between government revenue and expenditure, providing evidence of institutional separation in Pakistan. The study suggests that government should take expenditure and revenues decisions separately and target long run economic growth. Keywords: Government Revenue, Government Expenditure, Causality INTRODUCTION Fiscal policy, which requires an appropriate adjustment in government revenue and expenditure, is of vital importance in promoting balanced growth, mobilization of resources, capital formation, income and employment. By means of fiscal policy, an economy can adjust and monitor its spending and tax rates. Furthermore, various problems like continuous budget deficits and fiscal imbalances can be prevented if policy makers understand the relationship between government revenue and expenditure. Therefore, a study of the direction of causality between revenues and expenditures of the government is important to determine the suitable strategy for deficit reduction. Fiscal policy can be considered expansionary in the short run, when government expenditures exceeds government revenues and the emerging deficit can be depicted as a means to finance additional government expenditures. If these additional expenditures are growth encouraging, then a government deficit reveal an indirect effect on long-term economic growth. However, in a Ricardian world, deficit is viewed as taxes delayed, there must be no difference between deficit finance and taxes of government expenditures, as long as the tax structure is not changed in the future (Ludvigson 1996). In addition, if the economy is non-Ricardian, due to credit constraints, then public deficits can change private incentives to accumulation and thus directly influence economic growth.

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Developing countries like Pakistan, face dual challenges while planning and implementing fiscal balancing policies. On one side it faces public demand to boost public expenditure for infrastructure and social investment, and on the other side its ineptitude to raise revenues to finance increased expenditures. By boosting the aggregate demand government expenditures can be a good tool in stimulating a stagnant economy. Keynesians argue that by borrowing from the private sector, governments can reverse the economic downturns and then return it through various spending programs. Government investment in physical infrastructure, health, education may raise productivity of labor, increase in private investment, thus profiteering firms which may help in long-run economic growth (Barro, Martini 1990). According to Keynes, at first government deficits and increased government debt can only be good for stimulating economic growth if country’s GDP is lower than its potential output. But, when country is functioning near its potential level of output, fiscal deficits can result in high inflation. Putting it differently, expenditure is an external factor which can be handled as a policy instrument to encourage economic growth (Keynes 1936).Poor countries fail to prosper due to limitation of constrained resources, low per capita GDP, dependency on capital inflows, informal market transactions, small public sectors, poorly regulated financial markets and under-developed democratic institutions. FISCAL CONDITION OF PAKISTAN Government of Pakistan collects sizable portion of its revenue through taxes and surcharges which constitute 65 percent to 70 percent of overall revenue collection. It has been successful so far in bringing down in reducing fiscal deficit from 7 percent in the average in 1980 and 1990’s to 3.5 percent in 2007 and since than its fluctuating in 3 to 5 percent. Presently in 2015-16 the fiscal deficit is 4.3 percent lower than 5 percent in 2014-15. Policy makers vows to bring it more down to 3.5 percent of GDP in coming years. The increasing gap between revenues collected and total expenditure has been a problem for many economists and policy makers. The gap was Rs.150 Billion in 1991 when the ‘Resource Mobilization and Tax Reform Commission’ was founded. In 2003, the gap widened to Rs. 515 billion. Previously the government succeeded in reducing budget deficit recorded at 515.2 billion (1.7 percent of the GDP) during (July-December) of fiscal year 2016 which is below 2 percent of GDP (Ministry of Finance). Expenditures of the country remained at Rs 2520.1 billion against the revenues of Rs 2005 billion. The targets were achieved due to the decrease in developmental projects and revenue generation measures by Federal Board of Revenue (FBR). Pakistan comes close to the bottom of the countries on the basis of revenue collection with a narrow tax base of 1.5 percent of those who pay their taxes. Furthermore, two third of the revenue is being generated from indirect taxes. According to statistics provided by Federal Board of Revenue of Pakistan, it has been able to collect Rs. 2,590 billion PKR during 201415, yielding 15 percent growth rate over the collection of 2,254 billion during FY, 2013-14. The revenue target for FY 2014-15 of Rs. 2,605 billion has been achieved to the extent of 99.4 percent. Due to declining prices of various commodities especially oil prices the revenue collection was adversely affected. The data shows that 42 percent of the collection was contributed by sales tax followed by 40 percent direct taxes, customs 11.8 percent and federal excise duties 6.2 percent during 2014-15.In absolute terms the share of sales tax has declined from 44.2 percent to 42 percent in federal taxes 2014-15. On the other hand the share of direct taxes has been improved from 38.9 percent to 39.9 percent. The net revenues receipts are Rs. 2378 billion PKR showing an increase of 6.9 percent in 2015. Whereas, public expenditure (current and development) decreased by 1.6 percent (Pakistan Federal budgets). HYPOTHESIS OF TAX-SPENDING DEBATE AND PREVIOUS WORK The first hypothesis proposed by Friedman (1978, 79, 82) is “tax-spend hypothesis”, which states that when government revenues increases, the government expenditures will also get Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

increased and increase in taxes will not lower the budget deficits. Thus, increased government spending leads to more deficit which if financed through external resources i.e., borrowings or printing money will higher the future interest payments which will lead to another increased deficit by paying more debts. The hypothesis entail a positive relation between changes in taxes and expenditure programs. On the other hand, Buchanan and Wagner (1977, 1978) gave alternative version of this hypothesis they argue that there is a negative causality between government expenditure and government revenues and increase in taxes will lead to reduction in spending. In contrast to Friedman they found increasing tax revenues as an appropriate policy for reducing budget deficit otherwise, the tax payers will suffer from fiscal illusion and will demand more public goods and services.Regarding fiscal illusion this hypothesis is true as evident from the confirmation of Blackley (1986, 144), Hoover and Sheffrin (1992, 230) and Eita and Mbazima (2008). The second hypothesis is “spend-revenue or spend-tax” hypothesis which argues that government spending raises government revenues. It means that that government will first decide the spending patterns and then it will take necessary steps to determine tax policy to finance further public expenditure programs. This proposition was put forwarded by Peacock and Wiseman (1961, 1979). Under this proposition the empirical results will show unidirectional causality functioning from expenditure to revenues. According to P-W the changes in temporary rise in government spending will lead to changes in government revenues due to ‘social disturbances’ like natural disasters or floods and other external shocks like great depression. After the shocks are over people will get used to the levied taxes and thus will pay them which will change government revenues on permanent basis. The study was latterly substantiated and supported by Robert J. Barro (1974, 1978, 1986,) in his tax smoothing hypothesis, increase in taxes are due to the effect of increased government expenditure and government expenditure is considered as an external dependent variable to which taxes adjust ( Aisha and Khatoon 2009). Following proposition of Ricardian equivalence theorem, he also explains that taxpayers are sane and rules out fiscal illusion. More precisely, economy is influenced by government purchases in numerous ways such as budget deficits and current account balances (see Barro 2005). Interest groups in some governments prefer to borrow at first and then finance increased government expenditures and then increase taxes to compensate accumulated debt and rising deficits. The third hypothesis which is known as “Fiscal Synchronization” which is mainly led by (Musgrave 1966) and (Meltzer, Richard 1981).According to them government expenditure and government revenues changes concurrently or simultaneously. In addition, the decisions of optimal level of spending and taxation depends on voter’s attitude towards welfare maximizing demand for public services.Manage and Marlow (1986), (Ram 1988) and (Li 2001)supports fiscal synchronization hypothesis and found bi-directional causation running between revenue and expenditure. Finally, the fourth school “ Institutional separation” also called fiscal neutrality or fiscal independence introduced by Beghestani and McNown (1994) proposed that, there exists no causal relationship between government revenues and expenditures and decisions on revenues and allocation of government expenditures are taken independently by the legislative and management authorities of the state due to their different functions. According to this hypothesis the level of government spending will be determined on the requirement of state population needs and on the ability of tax burden they can take over (see Darrat 1998). Furthermore, the absence of the causal is due to difference in the agendas of important actors [Hoover and Sheffrin (1992)] and because of the lack of agreement between the groups and parties for decision making process is the cause of rising public debt [Drazen (2001)]. Major promoter of this view is Wildavsky (1988) who argues that budgeting can be incremental and adjustments can be made on the margin if these separate institutions or groups reach on an agreement on the fundamentals.

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Furstenburg et al. (1986) probed connection between government revenues and expenditures in the US for the period of 1955-81 and found that there exist unidirectional Granger causality running from expenditure to revenue. This study was also supported by Anderson, Wallace and Warner (1986). Chang and Ho (2002) examined tax-spend, spend-tax and fiscal synchronization for Taiwan by using data covering the period of 1967 to 1999. Their result showed that uni-directional causality between revenues and expenditure of the government validating tax-spend hypothesis for their country. For policy implication they stated that fiscal authorities must focus on spending cuts and not to raise revenues from taxes or by other sources. Carneiro et al. (2004) checked causality and relationship between government revenues and expenditure for low income country Guinea-Bissau. Their results concludes that GuineaBissau follows “spend-tax” behavior and adopts the approach of spending first than raise funds rather than raising first and then finance spending. The authors mentioned that the country should control government spending for fiscal sustainability and size of public deficit, in short term. Taha, et al. (2008) studied causal relationship between tax revenues and government spending in Malaysia by applying VAR model for the sample period of 1970-2006. The study provides evidence for the existence of bi-directional Granger causality running from direct tax revenues, indirect tax revenues to government spending. It is also observed that the direct and indirect taxes have major impact on the growth of government revenues than non-tax revenue and any reduction in the variables may lead to fall in government spending in the future. Haider (2008), checked relationship between expenditures of the government and tax revenue and used GDP deflator as the general price level from 1973 to 2003. The analysis confirms uni-directional causality running from expenditure to revenue. While, revenues quickly responding to changes in government expenditures proposes the preference of controlling the spending decisions to reduce the tax revenue-expenditure deficit. Other studies contain Aslan and Tasdemir (2009) for Turkey using EG and GH test for co-integration and found bidirectional causality between GR and GE. Afonso, et al. (2009) studied 25 European countries covering period of 1960-2006 by employing Bootstrap panel analysis. The results indicates uni-directional causality running from GR to GE in Austria, Belgium, Germany, UK and Finland and for other several European new members. However, in Spain, Italy, Greece, Portugal and France the causality runs from GE to GR. Aisha, et al. (2010) examined the causal relationship between government expenditures and revenues and concluded unilateral causality is running from expenditures to revenues in Pakistan. That is, government of Pakistan first spend and then raise tax revenues to finance expenditures of the country, rather than yielding revenues first and then spend to overcome the requests for grants and high budget deficit. Obioma, et al. (2010) empirically analyzed government revenue and government expenditure relationship in Nigeria, using time series data from 1970-2007, indicated that there is a longrun relationship between government expenditure and government revenue in Nigeria. Their findings support the revenue-spend hypothesis indicating that changes in government revenue influence changes in government expenditure. Ghartey (2010) used stepwise Granger causality technique to determine effective way of reducing deficits and debt burden in four Caribbean countries. The results confirm that only in Belize, taxes cause spending and independent for rest of the countries. Error correction model shows long-run bi-directional causation in all countries. Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

For forty Asian countries Mehrara et al. (2011) examined cointegration and causal relationship and found bi-directional short run and long run causal relationship between revenues and expenditures of the government which reveals validity of fiscal synchronization hypothesis in the region. They added that fiscal authorities in these countries should give attention to reduce their budget deficits by raising their revenues instead of increasing their expenditures to control budget deficit. Kurniawan (2012) following Hamilton and Flavin (1986) fiscal sustainability analysis highlighted major fiscal variables that is, expenditures, revenue, deficit and debt- for Indonesia. He used intertemporal budget constraint (IBC) framework to study the problem of fiscal sustainability in Indonesia with these variables. After satisfying the condition of transversality, his empirical findings confirms positive uni-directional causation from revenue to expenditure which satisfies Friedman (1978) tax-spend hypothesis. In conclusion, he suggests reforms in public expenditure to avoid exploding debt to GDP ratio by enhancing revenue collection efforts. Furthermore, Dogan (2013) examined the direction of linkage between government expenditure and revenue in Turkey for the long period of 1924-2012 in 89 pairs of observations. Results showed that there is one-way casual effect from GE to GR and supported spend-tax hypothesis as opposed to results of Payne et al. (2008) and Darrat (1998). Similarly, Takumah (2014) checked relationship between government revenue and expenditure with addition of GDP as a control variable for Ghana and found out cointegration and bidirectional causal relationship between the variables. Thus, confirming Fiscal synchronization hypothesis and policy implication to raise revues and decrease spending simultaneously in order to curb budget deficit problem in the country. Several studies have been carried out in developing and developed world by taking oil revenues as a dependent and independent variable due to its importance in country’s aggregate income or whose oil revenues compose the largest in budgetary revenues. For example study of (Nwosu and Okafor 2014)for Nigerian economy who followed method of Fasano and Wang (2002) examined expenditure and revenue nexus with disaggregated approach i.e., relationship of current and capital expenditures and oil and non- oil revenues from 1970-2011. The findings of their studies supports spend–tax hypothesis for Nigerian economy and suggests government to consider other sources of revenue especially the non-oil mineral sector and to reduce the size of large re-current expenditure. Furthermore, they believe that the government should bring about reforms in their expenditure (oil and non-oil) sectors to help set targets for better revenue mobilization and utilization. More recently, Abbas Ali Rezaei (2015) supported Friedman (1978) by analyzing long run and short run relationship for Iran using time series data from 1978-2012 and employed model of Toda-Yamamoto granger causality and ARDL techniques. His results predicted unidirectional causation from government revenue to government expenditure and positive relationship in both long run and in short run among the variables. In Pakistan, higher level of budget deficit are financed significantly by government borrowings these, raise the interest burden, future expenditure as well as expectation of higher taxes. To re-examine and test the validity of all four hypothesis, this paper examines four hypotheses tax-spend, spend-tax, fiscal synchronization and institutional separation in context of Pakistan. Secondly, five steps econometric methodology i.e. unit root test, lag length criterion, Johansen co-integration test, VAR analysis and Granger causality tests is used in this research. Thirdly, it investigates long run relationship between both of the variables with current data using multivariate model using real GDP growth as a third variable.

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

JUSTIFICATION OF THE PROBLEM The key to development of a country is high government revenues and less non developmental expenditure. Now-a-days Pakistan economy is facing a critical situation besides significant economic growth. This is due to the unsatisfactory indicators of fiscal balances. In such situation, nearly all economic indicators are declining. During past decades, government preferred non-developmental expenditure over government revenues, despite of low tax to GDP ratio, which resulted in larger budget deficit. Government expenditures requires comprehensive analysis of the tax revenues in the economy because tax revenues provide the larger pool than any other source of revenue for the government under the main heads of direct and indirect taxes. Budgetary policies are based on the calculation, estimation and forecasting of tax revenues in order to formulate the government expenses in mode which can be put on to achieve the stated targets. However, the roots of government expenditure plan are primarily linked to the expenditure-revenues nexus for the economy. This study is all over the fiscal expenditure and fiscal revenues relations in Pakistan since 1972 to 2015. OBJECTIVES OF THE STUDY • To check and explain the causal relationship between revenues and expenditure variables in the context of Pakistan for the period 1972 to 2015; • To seek the impact of relationship between revenues and expenditure in multivariate models using real GDP as the third variable; • To check long run equilibrium relationship between government expenditure and government revenues. HYPOTHESES The study will have to test the following hypotheses; H01: Freidman and Wagner’s Tax-spend Hypothesis does not apply to Pakistan; H02: Peacock and Wiseman’s Spend-Tax Hypothesis does not apply to Pakistan; H03: Musgrave’s fiscal synchronization Hypothesis does not apply to Pakistan; H04: Beghestani and McNown’s Institutional Separation Hypothesis does not apply to Pakistan. DATA In this study, secondary time series data is employed and econometrics approach has been used to investigate the causality and cointegration between government revenues and expenditure of Pakistan during the period 1972 to 2015. Data is obtained from various sources such as, Federal Board of Revenue annual tax directories, various issues of economic survey of Pakistan and Handbook of statistics Pakistan. The variables include government revenues, government revenues and real GDP growth. METHODOLOGY AND DATA ANALYTICAL TECHNIQUES To study empirically, the relationship between government revenues and government expenditure a single model is estimated. First of all Augmented Dickey Fuller (1981) is used to check the stationarity of the variables. The optimal lag length has been determined using Vector Auto Regressive (VAR) method. To find co-integration, Johansen cointegration is used to check long run relationship among the variables. To forecast the empirical relevance to the theory and the dynamic impact of random disturbances on the systems of variables VAR technique has been used. Impulse response and variance decomposition will be checked, which helps in tracking the impact of any variable on other and causal analysis in the system. In last, pair wise Granger Causality (1969) test has been applied to check any causal linkage among the variables. Variables have been transformed in natural logarithms (Ln).

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

EMPIRICAL MODELS The research estimated a single equation. The equation will use determinants of government revenue as a function of government expenditure and real domestic product as: TX = α+ β1PX+ β2RGDP+ µ …….. (I) Initial model for the granger relationship between Expenditure-Revenues nexus can be specified as. ln TX = α+ ln β1PXt + ln β2RGDPt+ µ t ……..

(II)

For testing stationarity Unit root test will be used. So, the model for this test become as: ∆PXt = ἅ1+ ἅ2 + ß PXt-1 + ∑ Þ∆ PXt-1 + µ t ∆TXt = ἅ1+ ἅ2 + ß TXt-1 + ∑ Þ∆ TXt-1 + µ t ∆RGDPt = ἅ1+ ἅ2 + ßRGDPt-1 + ∑ Þ∆ RGDPt-1 + µ t For long run relationship Co-integration test will be used. Here we will regress TX on PX and RGDP as follows: LnTXt = ἅ1+ ἅ2LnPXt+ ἅ3LnRGDPt + µ t Where, Ln is the natural log of variables. Tax revenues, Public expenditures, Real GDP TX are the fiscal tax revenues (direct and indirect combined) The PX is the public expenditure for fiscal year in Pakistan RGDP is the real GDP annual growth rate in Pakistan µ t is the uncorrelated error term. AUGMENTED DICKEY FULLER TEST To tackle autocorrelation problem, Augmented Dickey Fuller Test has been used to solve the problem of non-stationarity. After conducting ADF test at level and first difference the results are given in Table 1.The results given in Table 1 shows that LTX is already at level, because its probability value is less than 5 %. LRGDP and LPX are non-stationary at level. However, at first difference, both variables becomes stationary. Table 1 Stationarity results of ADF test statistics 1972-2015 Variable Level (Trend & Intercept) First Difference Result LTX

0.0488

_

I(0)

LPX

0.6277

0.0000**

I(1)

LRGDP

0.9114

0.0000**

I(1)

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

OPTIMAL LAG SELECTION CRITERIA The optimal lag length is determined using Vector Auto Regressive (VAR) method and the “Akaike information criteria (AIC)” shows that the optimal lag length as 1. For optimal lag selection one of the ways, is to choose the lowest Akaike Information Criterion (AIC) value. The lower the AIC value, the better the model.

Lag

Table 2 Optimal lags selection through Akaike Information Criterion Log L LR FPE AIC SC

0

-105.9543

1

31.49372

2

NA

HQ

0.085386

6.053016

6.184976

6.099074

244.3520*

6.81e-05*

-1.082984*

-0.555145*

-0.898754*

35.13357

5.864202

9.29e-05

-0.785198

0.138521

-0.462795

3

40.61860

7.922817

0.000116

-0.589922

0.729677

-0.129347

4

43.49335

3.673295

0.000173

-0.249630

1.465848

0.349118

JOHANSEN COINTEGRATION TEST In order to find out the presence of long run equilibrium relationship Johansen cointegration test is used to find the number of cointegration vectors. To run this particular test, all the variables must be integrated of the same order. Our variables are integrated of the same order, so, we can easily apply this test. Table 3 Results of Cointegration Test No. of CE(s) Eigenvalue Trace Statistics 0.05 Critical Value Prob. ** None

0.286587

22.42172

29.79707

0.2756

At most 1

0.163275

8.913924

15.49471

0.3734

At most 2

0.43608

1.783508

3.841466

0.1817

The null hypothesis is that the number of co-integrated equation is zero which means that there is no co-integrated equation in the system. The table 3 above and trace statistics shows that there is no cointegration among the variables, government revenues, government expenditures and real gross domestic product. So, we will accept null hypothesis which says that there is zero co-integrated vector in the whole system. IMPULSE RESPONSE FUNCTION Impulse response function is a shock to a VAR system and identify the responsiveness of the dependent variable (endogenous variable) in the VAR. The middle blue line indicates the reaction of a shock to the variables. The red line while shows maximum possible deviation in positive or negative direction. Furthermore, an impulse cites reaction of any dynamic system Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

in response to some external change. Below graph shows confirmation of VAR results and the relationship existing between all fiscal variables. •

•

•

Response of government expenditure to government expenditure: The first graph shows response of government expenditure to government expenditure when one standard deviation shock is given to government expenditure. As we can see the blue line is above the line which means that the response is positive, this shock positively affected LPX. Response of government expenditure to government revenues: The effect of shock of government expenditures is initially zero on government revenues and deviates to positive direction till the end up to period 10. Response of government expenditure to real gross domestic product: The effect of the shock here is also positive, initially the line is at zero but after some time it deviates to upward positive direction. Hence, when one standard deviation shock is given, real gross domestic product shows a positive change as shown in in the above given graph.

Response to Cholesky One S.D. Innovations ± 2 S.E. Response of LPX to LPX

Response of LPX to LTX

Response of LPX to LRGDP

.4

.4

.4

.3

.3

.3

.2

.2

.2

.1

.1

.1

.0

.0

.0

-.1

-.1

-.1

-.2

-.2 1

2

3

4

5

6

7

8

9

10

-.2 1

2

Response of LTX to LPX

3

4

5

6

7

8

9

10

1

Response of LTX to LTX .15

.15

.10

.10

.10

.05

.05

.05

.00

.00

.00

-.05

-.05

-.05

-.10

-.10 2

3

4

5

6

7

8

9

10

2

Response of LRGDP to LPX

3

4

5

6

7

8

9

10

1

.4

.4

.4

.2

.2

.2

.0

.0

.0

-.2 3

4

5

6

7

8

9

10

5

6

7

8

9

10

3

4

5

6

7

8

9

10

9

10

Response of LRGDP to LRGDP .6

2

2

Response of LRGDP to LTX .6

1

4

-.10 1

.6

-.2

3

Response of LTX to LRGDP

.15

1

2

-.2 1

2

3

4

5

6

7

8

9

10

1

2

3

4

5

6

Graph of the impulse response function GRANGER CAUSALITY TEST The empirical results appear in Table 4 were obtained with one lag of each variable. Using pairwise Granger causality test and annual data for government expenditures, government revenues and real gross domestic product over the period 1972 to 2015. We accept null hypothesis. We have found no causality among the variables i.e.,uni-directional or bidirectional types of causalities in the system.

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

7

8

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

It can be seen in the below given table that probability is 0.39, which is greater than 5 %, which means that RGDP does not granger cause PX. Thus it can be concluded that none of the variable granger causes one another showing no causal relationship.

Null Hypothesis

Table 4 Results for Granger Causality Test Obs F-Statistics

LRGDP does not Granger Cause LPX LPX does not Granger Cause LRGDP

42

LTX does not Granger Cause LPX LPX does not Granger Cause LTX LTX does not Granger Cause LRGDP LRGDP does not Granger Cause LTX

0.95881

0.3927

0.75382

0.4777

2.40774 42

42

Prob.

0.1040

0.19738

0.8217

2.10876

0.1357

1.44190

0.2495

CONCLUSION & RECOMMENDATION The study attempted to find if relationship and direction of causality exists between government revenues and expenditure of Pakistan from 1972 to 2015.The result shows that there is no causal evidence and no co-integration among the variables. It indicates that there is a lack of agreement between executive and legislative branches of Pakistan government in taking spending and revenue decisions. The results are consistent with the fourth hypothesis of ‘Institutional separation’ given by Beghestani and McNown (1994) reflecting interdependence of revenue and expenditures in Pakistan. Based on VAR model estimates, neither revenues nor expenditures respond to budgetary disequilibria, leading to rejection of previous hypotheses of tax-and spend, spend-and-tax and fiscal synchronization models of the budgetary process. The government needs to regulate its macroeconomic policies particularly, increasing tax to GDP ratio, avoid debts (foreign or domestic grants and loans) and to spend on the basis of revenue yields to reduce large fiscal deficits that are unsustainable for long-run economic growth. The results of this study indicates that government expenditure does not explain government revenue in Pakistan, which supports “Institutional separation hypothesis”. The implication is that the budget deficit for Pakistan can be cured by independently addressing government expenditure and government revenues. Hence, the government should make expenditure and revenue decisions separately and policies meant to improve the government budget position should hence target other factors like economic growth which affect government expenditure and revenues. REFERENCES Aisha, Z., & Khatoon, S. (2009). Government expenditure and tax revenue, causality and cointegration: The experience of Pakistan (1972-2007). The Pakistan Development Review, 951-959.

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Aslan, M., & Taşdemir, M. (2009). Is fiscal synchronization hypothesis relevant for Turkey? Evidence from cointegration and causality tests with endogenous structural breaks. Journal of Money, Investment and Banking, 12, 14-25. Anderson, W., Wallace, M. S., & Warner, J. T. (1986). Government spending and taxation: What causes what? Southern Economic Journal, 630-639. Barro, R. J. (1974). Are government bonds net wealth?. Journal of political economy, 82(6), 10951117. Barro, R. J. (1979). On the determination of the public debt. Journal of political Economy, 87(5, Part 1), 940-971. Barro, R. J., & Sala-i-Martin, X. (1990). Economic growth and convergence across the United States (No. w3419). National Bureau of Economic Research. Baghestani, H., & McNown, R. (1994). Do revenues or expenditures respond to budgetary disequilibria?. Southern Economic Journal, 311-322. Buchanan, J. M., & Wagner, R. E. (1977). Democracy in deficit. Acad. Press. Buchanan, J. M., & Wagner, R. E. (1978). Dialogues concerning fiscal religion. Journal of Monetary economics, 4(3), 627-636 Chang, T., & Ho, Y. H. (2002). Tax or spend, what causes what: Taiwan's experience. International Journal of Business and economics, 1(2), 157. Carneiro, F., Faria, J. R., & Barry, B. S. (2004). Government revenues and expenditures in GuineaBissau: causality and cointegration. Darrat, A. F. (1998). Tax and spend, or spend and tax? An inquiry into the Turkish budgetary process. Southern Economic Journal, 940-956. Dogan, E. (2013). Does “Revenue-led Spending” or “Spending-led Revenue” Hypothesis Exist in Turkey. British Journal of Economics, Finance and Management Sciences, 8(2), 62-75. Dwivedi, D. N. (2002). Microeconomics: Theory and Applications. Pearson Education India. Federal Board of Revenue Tax Directory. Granger, C. W. (1969). Investigating causal relations by econometric models and cross-spectral methods. Econometrica: Journal of the Econometric Society, 424-438. Ghartey, E. E. (2010). Government expenditures and revenues causation: some Caribbean empirical evidence. Applied Econometrics and International Development, 10(2), 149-165. Hamilton, J. D., & Flavin, M. (1985). On the limitations of government borrowing: A framework for empirical testing. Hoover, K. D., & Sheffrin, S. M. (1992). Causation, spending, and taxes: Sand in the sandbox or tax collector for the welfare state?. The American Economic Review, 225-248. Kurniawan, R. (2012). Sustainability of Fiscal Policy and Government Revenue-Expenditure Nexus: Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

The Experience of Indonesia. Li, X. (2001). Government revenue, government expenditure, and temporal causality: evidence from China. Applied Economics, 33(4), 485-497. Ludvigson, S. (1996). The macroeconomic effects of government debt in a stochastic growth model. Journal of Monetary Economics, 38(1), 25-45. Ministry of Finance News Bulletin 2014-15. Marlow, M. L., & Manage, N. (1987). Expenditures and receipts: Testing for causality in state and local government finances. Public Choice, 53(3), 243-255. Meltzer, A. H., & Richard, S. F. (1981). A rational theory of the size of government. Journal of political Economy, 89(5), 914-927. Mehrara, M., Pahlavani, M., &Elyasi, Y. (2011). Government revenue and government expenditure nexus in Asian countries: panel cointegration and causality. International Journal of Business and Social Science, 2(7), 199-207. Musgrave, R. (1966). Principles of budget determination. Public finance: Selected readings, 15-27. Nwosu, D. C., & Okafor, H. O. (2014). Government revenue and expenditure in Nigeria: A disaggregated analysis. Asian Economic and Financial Review, 4(7), 877. Obioma, E. C., & Ozughalu, U. M. (2010). An examination of the relationship between government revenue and government expenditure in Nigeria: Cointegration and causality approach. CENTRAL BANK OF NIGERIA, 48(2), 35. Pakistan Budgets 1998-1915. Pakistan Economic Surveys 1975 to 2015. Wiseman, J., & Veverka, J. (1961). The growth of public expenditure in the United Kingdom. Princeton University Press. Ram, R. (1988). Additional evidence on causality between government revenue and government expenditure. Southern Economic Journal, 763-769. Ricardo, D. (1891). Principles of political economy and taxation. G. Bell. Rault, C., & Afonso, A. (2009). Bootstrap panel Granger-causality between government spending and revenue in the EU. Rezaei, A. A. (2015). Tax-Spend, Spend-Tax or Fiscal synchronization hypothesis: Evidence from Iran. International Journal of Innovation and Applied Studies, 10(3), 844. Samuelson, P. A. (1954). The pure theory of public expenditure. The review of economics and statistics, 36(4), 387-389. Statistical Year Book of Pakistan’s Economy. Takumah, W. (2014). The dynamic causal relationship between government revenue and government expenditure nexus in Ghana. Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Von Furstenberg, G. M., Green, R. J., &Jeong, J. H. (1986). Tax and spend, or spend and tax?. The Review of Economics and Statistics, 179-188. Vergne, C. (2009). Democracy, elections and allocation of public expenditures in developing countries. European Journal of Political Economy, 25(1), 63-77. Wildavsky, A. B. (1997). The new politics of the budgetary process. Addison Wesley Publishing Company

APPENDIX List of Variables Symbol

Variable

TX

Total Government Revenues

PX

Total Government Expenditures

RGDP

Real Gross Domestic Product

Government Revenue (LTX) at level Null Hypothesis: LTX has a unit root Exogenous: Constant Lag Length: 0 (Automatic - based on SIC, maxlag=2) t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-2.942247

0.0488

Test critical values:

1% level

-3.592462

5% level

-2.931404

10% level

-2.603944

Source: Eviews 9 , Software for Econometric techniques GOVERNMENT EXPENDITURE LPX (LEVEL WITH INTERCEPT) Null Hypothesis: LPX has a unit root t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-0.124134

0.9401

Test critical values:

1% level

-3.592462

5% level

-2.931404

10% level

-2.603944

LPX (LEVEL WITH TREND AND INTERCEPT) Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-1.918430

0.6277

Test critical values:

1% level

-4.186481

5% level

-3.518090

10% level

-3.189732

RGDP (LEVEL WITH INTERCEPT) t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

1.226751

0.9979

Test critical values:

1% level

-3.592462

5% level

-2.931404

10% level

-2.603944

RGDP ( LEVEL WITH TREND AND INTERCEPT ) Null Hypothesis: RGDP has a unit root t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-1.132213

0.9114

Test critical values:

1% level

-4.186481

5% level

-3.518090

10% level

-3.189732

LRGDP AT FIRST DIFFERENCE t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-6.702127

0.0000

Test critical values:

1% level

-3.596616

5% level

-2.933158

10% level

-2.604867

LPX AT FIRST DIFFERENCE Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-6.022477

0.0000

Test critical values:

1% level

-3.596616

5% level

-2.933158

10% level

-2.604867

LAG LEGNTH CRITERIA USING AKAIKE INFORMATION CRITERION VAR Lag Order Selection Criteria Lag

LogL

LR

FPE

AIC

SC

HQ

0

-105.9543

NA

0.085386

6.053016

6.184976

6.099074

1

31.49372

244.3520*

6.81e-05*

-1.082984*

-0.555145*

-0.898754*

2

35.13357

5.864202

9.29e-05

-0.785198

0.138521

-0.462795

3

40.61860

7.922817

0.000116

-0.589922

0.729677

-0.129347

4

43.49335

3.673295

0.000173

-0.249630

1.465848

0.349118

5

53.99890

11.67283

0.000176

-0.333272

1.778087

0.403649

6

69.37380

14.52074

0.000144

-0.687433

1.819805

0.187660

7

73.83908

3.473002

0.000236

-0.435505

2.467613

0.577761

8

93.39818

11.95278

0.000191

-1.022121

2.276877

0.129317

* indicates lag order selected by the criterion Table 5: Results of Augmented Dickey-Fuller Test 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 JOHANSEN LONG RUN COINTEGRATION TEST Unrestricted Cointegration Rank Test (Trace) Hypothesized

Trace

0.05

No. of CE(s)

Eigenvalue

Statistic

Critical Value

Prob.**

None

0.286587

22.42172

29.79707

0.2756

At most 1

0.163275

8.913924

15.49471

0.3734

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

At most 2

0.043608

1.783508

3.841466

0.1817

Trace test indicates no cointegration at the 0.05 level

Unrestricted Cointegration Rank Test (Maximum Eigenvalue) Hypothesized

Max-Eigen

0.05

No. of CE(s)

Eigenvalue

Statistic

Critical Value

Prob.**

None

0.286587

13.50779

21.13162

0.4067

At most 1

0.163275

7.130416

14.26460

0.4738

At most 2

0.043608

1.783508

3.841466

0.1817

Max-eigenvalue test indicates no cointegration at the 0.05 level * denotes rejection of the hypothesis at the 0.05 level Unrestricted Cointegration Coefficients (normalized by b'*S11*b=I): LPX

LTX

LRGDP

0.278570

-3.162295

2.213692

2.152035

0.258833

-1.753337

2.128266

-1.688903

0.675837

Unrestricted Adjustment Coefficients (alpha): D(LPX)

-0.030575

-0.045559

-0.032448

D(LTX)

0.035399

-0.003354

-0.013135

D(LRGDP)

-0.048405

0.106770

-0.041238

Log likelihood

42.79115

1 Cointegrating Equation(s):

Normalized cointegrating coefficients (standard error in parentheses) LPX

LTX

LRGDP

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

1.000000

-11.35189

7.946628

(3.52884)

(2.99549)

Adjustment coefficients (standard error in parentheses) D(LPX)

-0.008517 (0.01024)

D(LTX)

0.009861 (0.00437)

D(LRGDP)

-0.013484 (0.01752)

2 Cointegrating Equation(s):

Log likelihood

46.35636

Normalized cointegrating coefficients (standard error in parentheses) LPX

LTX

LRGDP

1.000000

0.000000

-0.722881 (0.12070)

0.000000

1.000000

-0.763706 (0.05814)

Adjustment coefficients (standard error in parentheses) D(LPX) D(LTX) D(LRGDP)

-0.106562

0.084894

(0.07762)

(0.11350)

0.002643

-0.112810

(0.03403)

(0.04976)

0.216288

0.180708

(0.12948)

(0.18932)

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

VECTOR AUTOREGRESSION ESTIMATES

LTX(-1)

LTX(-2)

LTX(-3)

LPX(-1)

LPX(-2)

LPX(-3)

LRGDP(-1)

LRGDP(-2)

LRGDP(-3)

LTX

LPX

LRGDP

0.956498

1.068670

0.141026

(0.25317)

(0.59713)

(0.99695)

[ 3.77809]

[ 1.78969]

[ 0.14146]

-0.155285

-0.887887

0.714403

(0.31510)

(0.74321)

(1.24084)

[-0.49280]

[-1.19467]

[ 0.57574]

0.117598

-0.073840

-0.567815

(0.22801)

(0.53778)

(0.89787)

[ 0.51576]

[-0.13730]

[-0.63240]

0.039188

0.978652

0.050779

(0.08107)

(0.19120)

(0.31923)

[ 0.48341]

[ 5.11841]

[ 0.15907]

0.008103

-0.227165

-0.196230

(0.10473)

(0.24703)

(0.41243)

[ 0.07737]

[-0.91959]

[-0.47579]

-0.077017

0.105075

0.227078

(0.07348)

(0.17330)

(0.28934)

[-1.04819]

[ 0.60632]

[ 0.78481]

0.001094

-0.001408

0.803226

(0.04539)

(0.10705)

(0.17874)

[ 0.02411]

[-0.01315]

[ 4.49392]

0.040256

0.118184

-0.038381

(0.05840)

(0.13775)

(0.22998)

[ 0.68930]

[ 0.85799]

[-0.16689]

0.028277

-0.123694

-0.079417

(0.04707)

(0.11101)

(0.18534)

[ 0.60081]

[-1.11427]

[-0.42850]

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

C

0.603870

0.581881

-0.309436

(0.25715)

(0.60651)

(1.01262)

[ 2.34832]

[ 0.95939]

[-0.30558]

R-squared

0.997410

0.969286

0.968363

Adj. R-squared

0.996658

0.960369

0.959178

Sum sq. resids

0.280188

1.558685

4.344801

S.E. equation

0.095070

0.224232

0.374373

F-statistic

1326.540

108.7019

105.4288

Log likelihood

44.03380

8.852978

-12.16234

Akaike AIC

-1.660185

0.055952

1.081090

Schwarz SC

-1.242241

0.473897

1.499034

Mean dependent

12.54295

13.32066

13.71118

S.D. dependent

1.644593

1.126373

1.852917

VARI ANCE DEC OMP OSITI ON OF LPX, LRG DP AND LTX

Determinant resid covariance (dof adj.) 4.85E-05 Determinant resid covariance

2.10E-05

Log likelihood

46.30386

Akaike information criterion

-0.795310

Schwarz criterion

0.458523

Variace Decomposi tion of LPX: Period

S.E.

LPX

LRGDP

LTX

1

0.216080

100.0000

0.000000

0.000000

2

0.335081

94.90452

0.180841

4.914638

3

0.409791

92.53047

0.582763

6.886765

4

0.457283

90.33952

1.853586

7.806895

5

0.489012

88.04299

3.787704

8.169308

6

0.511560

85.77366

5.919681

8.306663

7

0.528283

83.73902

7.880599

8.380382

8

0.541029

82.02265

9.517754

8.459594

9

0.550955

80.60835

10.82498

8.566672

10

0.558846

79.43937

11.85725

8.703380

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018

Variance Decomposi tion of LRGDP: Period

S.E.

LPX

LRGDP

LTX

1

0.358622

0.709717

99.29028

0.000000

2

0.461781

1.150734

98.84924

2.47E-05

3

0.510275

2.095662

97.67741

0.226929

4

0.538072

3.437339

95.67868

0.883978

5

0.558135

4.862990

93.25391

1.883099

6

0.574632

6.116261

90.86338

3.020356

7

0.588886

7.093753

88.76259

4.143657

8

0.601420

7.803125

87.01560

5.181275

9

0.612547

8.297288

85.58680

6.115913

10

0.622500

8.634241

84.41048

6.955275

Period

S.E.

LPX

LRGDP

LTX

1

0.091762

15.34917

1.557459

83.09337

2

0.132186

18.75358

1.442146

79.80428

3

0.162962

18.43334

4.405865

77.16080

4

0.187602

17.02881

9.204449

73.76674

5

0.208819

15.43128

14.44166

70.12706

6

0.227566

14.02835

19.12790

66.84375

7

0.244298

12.90507

22.94100

64.15393

8

0.259341

12.03418

25.92210

62.04372

9

0.272976

11.36131

28.23435

60.40434

10

0.285436

10.83562

30.04617

59.11821

Variance Decomposi tion of LTX:

Cholesky Ordering: LPX LRGDP LTX

Journal of computing and management studies ISSN 2516-2047. Issue 1. Volume 2. January 2018