Financing the Budget Deficit of Pakistan: A study of the Revenue and Expenditure Nexus



Financing the Budget Deficit of Pakistan: A study of the Revenue and Expenditure Nexus

This paper examines the Revenue-Expenditure nexus for Pakistan over a period of 39 years,
from 1976 – 2014. It studies whether there exists causality between total revenue and total
expenditure, as well as their components: non tax revenue, tax revenue, development expenditure
and current expenditure.
It employs a five-step procedure which includes the stationarity testing using both ADF and PP
tests, there determination of optimal lags using AIC criteria, and the long run and short run
relationships using the Ordinary Least Square (OLS) method and Vector Auto Regressive (VAR)
framework respectively. At the end, the Granger causality is used to determine the existence of
the causality.
The paper comes to find that for Pakistan, there is no causality between total revenue and total
expenditure, total expenditure and tax revenue, development expenditure and tax revenue,
current expenditure and tax revenue, and, total revenue and development expenditure thereby
showing that the institutional separation hypothesis holds true in this context.
However, there does exist a positive uni-directional causality running from Total Revenue to
Current Expenditure, Non-Tax Revenue to Total Expenditure, Development Expenditure and
Current Expenditure, which suggests that the government should target revenues before
spending, i.e. Spend-and-tax hypothesis.
Keywords: Government Revenue, Government Expenditure, Tax Revenue, Non-Tax Revenue,
Current Expenditure, OLS, VAR, Granger-Causality
In the simplest terms, a budget deficit occurs when a Government’s expenditure exceeds its
revenue that is; government spends more than the revenues it collects from its potential
resources. Within economics, there are variations amongst different economists and school of
thoughts on the understanding and handling of problems. In this context the Neoclassical,
Keynesian and Ricardian paradigms hold the greatest significance.
The Neo-classical view is such that an individual plans his/her consumption over their life cycle,
therefore when consumption increases the saving rate is going down. Interest rate is then
required to rise in order for the money market to be at equilibrium. This view does not hold
importance in the context of this paper because given the current economic conditions, such a
persistent deficit would crowd out the savings, creating havoc in the economy.
According to the Keynesian school of thought, an individual has a high propensity to consume
out of the current income, therefore a tax reduction would have a spontaneous and substantial
impact on the aggregate demand. In this school of thought, the budget deficit was regarded as the
definitive propellant of economic growth; these economists believed that a high rate of
employment and stability could be achieved through appropriate maneuvering of the government
The Ricardian school of thought explains that individuals save over generations by transferring
resources and therefore consumption could be a function of these dynamic resources. Since these
deficits shift the payment of taxes to future generations by prolonging them, the dynamics
resources aren’t affected.
More recently these deficits are being blamed for various economic problems, including high
inflation rates, unemployment, stagnant economic growth, stock market crashes, high interest
rates and balance of payment issues as well as exchange rates and other economics concerns. A
budget deficit when viewed as a percentage of the GDP decreases as the economy progresses.
This is because increased tax revenues coupled with a high employment lead to economic growth
and therefore lessen the need for government support, especially for welfare purposes.
Governments counter these budget deficits by reducing the government taxes and increasing the
taxation, hence promoting economic growth.
The Ricardian equivalence model is built on the notion that increasing the amount a government
borrows would not have impact consumer expenditure since these rational consumers would
predict tax cuts and they realize that this higher expenditure would be penalized later. In short,
this means that a cut in taxation, which is financed by further borrowing, should have no affect
on the aggregate demand. This is because the rational individual realizes that extra consumption
today would cost them higher future taxation and the concept of consumption smoothing comes
into play. If this model is assumed to be an accurate depiction of the ground reality then it makes
fiscal policy redundant.
Contrary to the Barro-Ricardo equivalence theorem, budget deficits and their proceeding fiscal
policy measures do matter substantially. Hence, understanding the “revenue-expenditure nexus”
has important implications for the formation of fiscal policies. In order to finance and eventually
reduce the budget deficit, it is essential for the country’s government to know the relationship
between revenues and expenditures.
Whether and how the revenues should be increased to finance the budget or whether the
expenditures should be reduced is a major question faced by the decision makers. In order to
answer this question, it is imperative to know whether a causal relationship exists amongst the
two factors. Whether increasing resource mobilization to increase revenues also increases
expenditure or is it the increased expenditure on resource mobilization, which increases
revenues, is the key question in this context.
The objective of this paper is to study the relationship between expenditures and revenues, one
that is crucial to Pakistan’s current fiscal situation. The Government’s budget deficit is in a
deplorable condition and every year large sums have to be borrowed externally or domestically
in order to meet this deficit. This continuous borrowing and accumulation has had detrimental
effects on the economy, leading to internal and external debt. Moreover, factors such as
exchange rate, trade gap and the balance of payments have also been negatively impacted by this
increasing debt.
Our study objective is to answer the fundamental question of budget deficit financing by
investigating the variables causing the deficit. The paper observes, through empirical testing the
data, whether there exists a unidirectional or bidirectional relationship between the variables, of
budget deficit, both in the long run and short run and whether this is positive or negative
The study aims to identify the type of causality that exists amongst the variables, or whether one
exists at all. This is of core importance as Revenue and Expenditure are the foundations that set
course for the budget policy. The study also decomposes these variables into their constituents,
that is, non-tax and tax revenues as well as current and development expenditure.
Eventually, this paper makes some conclusion regarding what policy should the government of
Pakistan adopt while setting its Revenues and spending targets based on the results of the
econometric analysis that it employs. It is hoped that these will aid in policy making and will rid
the country of the nuisance that the budget deficit holds.

Questions this paper aims to answer:
• Does an increase resource mobilization also raise expenditures as a consequence,
reducing the effect of increased revenue?
• Does a change in non-tax revenue result in a change of development, current or total
• Does a change in tax revenue result in a change of development, current or total
• Does the reduction in expenditures reduce the revenues as a consequence, reducing the
impact of minimized spending?
• Should the government target both revenue and expenditure simultaneously since they
both affect and cause each other?
• Should the government work regardless of the relation between the revenues and
expenditures, that is, independently because they have no causal relation?
• Should the government set targets for revenue before spending or vice versa because
there is a uni-directional relation between them?

In light of the budget deficit there are four extensively used hypothesis, as shown in the
preceding literature, as follows:
• The revenue-spend hypothesis: whereby there exists a unidirectional causality from
government revenue to government expenditure
• The spend-revenue hypothesis: whereby there exists a unidirectional causality from
government expenditure to government revenue
• The fiscal synchronization hypothesis: whereby there exists bidirectional causality
between government revenue and government expenditure
• The institutional separation hypothesis: Where there is no causality between government
revenue and government expenditure.
Theoretical Framework
Revenue and Expenditure are the two key factors of a Government Budget Equation, this
relationship has been a central issue for researchers in the realm of economics as it has important
implications on policy making. Within this context, there are four main hypotheses that help to
explain the relationship between government revenues and expenditure.
Tax-and-Spend hypothesis
This hypothesis is based on the argument that changes in taxation take lead on variation in
spending: that is the level of spending adjusts to the level of taxation. This theory was led by
Friedman (1978) who argued that as Taxes increase; government spending will also increase,
thus showing a positive causal relationship. Buchanan and Wagner (1977), also worked along the
same hypothesis and came to the conclusion that government spending will lead to higher budget
deficits and an increase in taxation is the cure to deficit.
Wagner’s version of the hypothesis was based on the concept of fiscal illusion, whereby the
consumers perceived prices of goods provided by the government to be lowered, hence
increasing their demand. The public, through this means actually faces higher costs because of
inflation associated with money creation and because the government has to pay a greater
interest on the cost it incurs. Through their interpretation, the remedies proposed by Wagner and
Buchanan were to increase the taxes in order to cure deficit in the budget. They were also in the
favor of curtailing the government’s ability to resort to deficit financing in its attempt to reduce
expenditure. Friedman on the other hand, argued that decreasing taxes would be an appropriate
measure and a long-term solution.
Spend-and-Tax hypothesis
This hypothesis is based on the notion that government decides expenditures first and then does
the changes in revenues; that taxes react to level of government spending. Robert (1978)
suggested a version of this hypothesis which was furthered by Peacock and Wiseman (1979),
they suggested that during a time of political or economic crisis, the government will increase
taxes in order to pay for the increased expenditures; this escalation in taxes might become
permanent. This is because the citizens’ tolerance towards a high tax rate develops and they are
thus willing to pay more, because of this the government expenditure also increases.
This interpretation is in line with Barro’s (1979) notion that expenditure financed by debt would
invariably mean an increased tax rate in the future in context to the Ricardian equivalence that
rules out the concept of fiscal illusion as discussed in the previous hypothesis. The Ricardian
Equivalence proposition states that the government cannot stimulate private consumption
through tax cuts and finance the deficit through bonds. That is, forward looking individuals will
understand that these tax cuts will be offset by greater interest rates plus the debt incurred at a
later point, hence they will have to pay higher taxes in the future. This hypothesis suggests that
expenditure decreases are the desired solution to reducing budget deficit.
The Ricardian modification to standard budget deficit observes that for a given path of
government spending, a cut in current taxes will lead to higher future taxes, which has the same
value as the present value of the initial cut. The Ricardian model therefore states that the national
saving does not change as the households’ demands for goods depend on the expected present
value of taxes.
Fiscal Synchronization hypothesis
This hypothesis suggests that the government decides its revenue and expenditures
simultaneously. It suggests that government spending and revenues are jointly determined by the
electorate in order to maximize the voter’s welfares. Musgrave (1966) and Meltzer and Richard
(1981) reflect the traditional theories of Public goods, they show that the level of taxation needed
is based on the comparisons of marginal costs and benefits that are associated with the
government spending programs. In essence, the hypothesis shows that the causal relationship
between government revenue and spending is bidirectional.
Institutional Separation hypothesis
This hypothesis states that government revenues and expenditures are independent of each other.
Wildaskvy (1988) says that separate institutions such as the executive and legislative branches of
the governments are part of the budgetary proceedings in order to determine the optimum level
of taxation and spending. This budgeting can be incremental and changes can be made in order
to reach a consensus amongst the stake holders.
Baghestani and McNown (1994), and further worked upon by Darrat (1998) also relates the
institutional separation of this expenditure and revenue decisions of the government. The
expenditure is defined on the basis of the requirements expressed by the citizens and revenue
would depend on the maximum tax burden as tolerated by the population.


Budget Deficit with context to Pakistan
Since independence in 1947, Pakistan has experienced a constant budget deficit, with the
exception of few years. The annual fiscal deficit has staggered around 6 percent of the GDP
since 1990, over a period of time; this deficit has translated into a debt, which is now a
permanent part of the economy. Although, growth was notable for a few years, the
accompanying fiscal deficit has increased the public debt, thus causing the vicious circle to
become further pronounced.
The extent of budget deficit and its consequences has been a hotly debated issue in recent years
due to the reason that Pakistan has experienced a continuous and growing political influence in
all the sectors of the economy. Political instability and the lack of an effective democracy is
routine in Pakistan and since the country has not been able to enjoy a stable socio-political
scenario the macro economic framework has suffered massively.
The government of Pakistan formulates a budget every year, which indicates its expected returns
and expenditures in the coming financial year. Receipts of the Government are expected from
different sectors, such from financial institutes, interest from loans given to other Governments,
tax revenues, local bodies’ etc. Expenses of Government consist of different projects,
developmental and non-developmental projects. If returns of the Government are equal to its
expenditures then the budget is balanced. The budget is in deficit if the expenses of the
Government are more than its receipts.
Amongst the developing countries, Pakistan is facing a large budget deficit, with the fiscal deficit
to GDP ratio hovering between 6 -8% over the decades. In 1990’s the situation was especially
dismal, with a chronic budget deficit. Accepting the gravity of the situation, a concentrated set of
reforms were initiated in early 2000, the continuous rise in revenues and fall in expenditures
from 2000 through to 2005 resulted in the deficit being reduced to 3% of the GDP.
When a Government finances its budget deficit through public debt (non-bank borrowing) and
bank borrowing, it is termed as deficit financing. There are many reasons why deficit financing
in Pakistan requires a sound fiscal policy in place, which is an essential pre requisite for a
functioning, stable macroeconomic environment.
The Government covers its budget deficit by borrowing from the State Bank in terms of printing
money or use State Bank to float various financial instruments to the Commercial banks and
borrow from them.. The effects of the both financing methods are an increase the money supply
in the country and this creates an inflationary pressure in the economy. Non-bank borrowing by
the Government is done through the sale of short-term federal bonds, treasury bills, defense
saving certificates etc.
This continuous process of borrowing form banks and non-bank sources increases interest rate,
inflation and discourages the private investment. There are also adverse effects of continuous
external borrowing these include imbalances in balance of trade, low GDP, capital flight from
the country.
The Government of Pakistan decided to face the challenge of deficit by employing a rule based
fiscal policy in 2002-2003. At that time the country had a fiscal deficit at 7% of GDP. This plan
was later on approved as a law in the parliament in June 2005 and was referred to as the “Fiscal
Responsibility and Debt Limitation Act”. The aim of the act was to ensure strict fiscal
discipline in future for the country.

Given below are some key postulates of the Act:
1) Within a period of 10 years beginning from July 1, 2003 public debt should not exceed
60% of GDP beyond June 2013.
2) Eliminating revenue deficit by June 2008.
3) Reducing public debt by at least 2.5 percentage point of GDP each year until June 2013
4) Not issuing guarantees to the borrowing of public service enterprises (PSEs) by more
than 2.0 percent of GDP in a given year.
5) Maintaining the level of spending on social sector and poverty-related programs above
4.5 percent of GDP in a given year and ensuring that expenditure on education and health
is doubled in terms of percentage of GDP by June 2013.
After the introduction of this act, all the rules were followed rigorously until June 2007 and
showed impressive results. The public debt as a percentage of GDP decreased from 75.1% (in
2002-2003) to 55.4% (in 2006-2007). The Revenue Balance decreased to 0.9% of GDP by 2006-
07 and was before on average at 3.1% of GDP during 1997-98 to 1999-2000.
Unfortunately after 2007 these rules were violated, and this resulted in a worsening situation and
public debt, for example, increased to 61.5% of GDP by 2011-2012 (which was 55.4% of GDP
in 2006-2007) and revenue deficit also worsened and increased to over 3% of GDP

Literature review:
We aim to study the strategy of budget deficit reduction for Pakistan by thoroughly studying its
components and their relation. For this purpose, we have composed our study in four portions
whereby we first take into account the scenario of budget deficits in Pakistan, followed by the
rest of the world including studies of budget deficit and relation between revenues and
expenditures under all four hypotheses mentioned above. Once, having studied the cases
throughout the world, we will empirically test our findings across Pakistan.
The literature that we have undertaken is varied and shows various contexts in which the budget
deficit has been studied. It shows us that Fiscal Synchronization is the most significant
hypothesis, with the majority of the countries under study are facing this sort of a relationship
amongst the two variables under scrutiny.
Using a Granger Test, Daniel K. Moulasi tested for an empirical relationship between revenue
and expenditure for Botswana. He found that a uni-directional relationship existed from revenues
to expenditures in the Short Run. Similarly, Ewing and Payne examined the inter-temporal
relationship amongst revenues in expenditures for five Latin American countries, for three of
these countries the Tax Spend hypothesis was shown to hold true.
Nwosu and Okafor analyzed data for the Government Revenue and Expenditure in Nigeria, they
used the VAR model, coming to the conclusion that there exists a Long Run relationship
between Government Revenues and Expenditures. For Ghana, a similar analysis yielded the
same result, whereby a Short Run causality exists, that is the Tax Spend Hypothesis is being

Mehrara, Pahalvani and Elyasi look at the relationship in 40 Asian country and they showed that
in the long run and the short run the same causality, that is bi-directional causality exists.
Similarly, for Malaysia and Switzerland the same hypothesis is being followed, as shown by the
Granger Causality testing. For the USA, using a bivariate model and Building on the argument of
Darrat and Suliman they show that there exists no definitive causality amongst the two variables.
We see that amongst the sample taken, there exists a definitive relationship for every country that
is studied; these results however, vary over the long run and short run. The literature review is
arranged in a manner whereby the countries are divided by the hypothesis that is being followed
in the economy.
It is important to understand this crucial relationship between the variables, as it has very
significant implications for the policy-making regime. Pakistan, being one of Asia’s developing
countries has a significant budget deficit and this has not been extensively studied. Over the
recent past, the trends in the budget deficit equation have changed drastically, the motivation to
work on the topic stems from the face that the existing data is outdated.
Monetary expansion, especially in developing countries is usually associated with a large amount
government borrowing from within the banking system as well as international aid agencies in
order to finance the existing gap in the budget. This borrowing takes place because the
government is incapable of domestic resource mobilization, its narrow taxation base and
stringent tax structure. Moreover, the capital markets are not developed and the environment is
financially biased towards the expansion of supply of money.
Pakistan also has a fairly large budget deficit that is growing over time and although this is
usually blamed on the high inflation, low growth there are other factors such as a lack of private
investment and consumption which have contributed to this short fall. Over the past two decades,
the budget deficit has varied between 5.4% and 8.7% of the GDP, although many measures have
been taken in order to reduce this deficit but all efforts have proved fruitless.
In the late 1990’s the deficit went down to 6.4% because the government reduced its
expenditures in development, but this was cosmetic as this change was not achieved by
increasing the tax-to-GDP ratio but by reducing the amount spent on infrastructure work. This
was, therefore unsustainable and this was shown when the deficit grew in the proceeding year.
Until Pakistan broadens its taxation base and implements stringent policies to monitor tax
evasion, the menace will keep on growing.
In the past fifty years, the country has had a constant budget deficit resulting in a current account
deficit and this has largely been filled in using loans from abroad, creating a large pool of
international debt-beyond what can be repaid. And this debt servicing has cost us in excess to 5%
of Pakistan’s GDP growth.

The budget deficit that is so big requires the swift expansion of domestic credit, as this change
would cause the foreign reserves to attain equilibrium at a new level. In countries such as
Pakistan, the free capital market intervention is very minimal and not very well developed.
The authors Abdul Waheed, Bashir Ahmad Khilji & Mubashir Ahmad in their article titled
“Revenue and Expenditure: What causes what? Empirical Evidence from Pakistan” tried to
analyze the causal relationship between expenditure and revenue of Pakistan in the period 1981
to 2010.They analyzed the spending and income of the Pakistani government. Here they used
three variables Government Revenues, Government Expenditure and Gross Domestic Product of
Pakistan (which is treated as a control variable).The hypothesis of the study was that revenue
changes the expenditure. They carried out the unit root tests with Augmented Dickey Fuller and
Phillips and Perron tests to test for causality and co-integration between expenditure and revenue
the Johansen co-integration test and VECM were used. The results showed that no causal
relationship existed between the variables, which were a different result from other studies in the
Tahir Sadiq examines the causal relationship between federal and provincial taxes as well as
expenditure. This analysis was done in order to find a viable strategy to control budget deficit.
Methodology used included the Granger causality test which was applied on the data spanning
from 1980-81 to 2009-10. Granger method says that future revenues and expenditure can be
predicted based on past revenues and expenditures. If past revenue values can be used to explain
present expenditure values than we can say that causality exits and its direction is from revenue
to expenditure (unidirectional causality).If the opposite case occurs then we say that the direction
of causality is from expenditure to revenue.

Mr. Sadiq had two equations in there empirical testing, one for revenue and the other for
The hypothesis for the expenditure equation was
Ho: Revenue does not Granger Cause Expenditure
H1: Revenue does Granger Cause expenditure
The hypothesis for the second revenue equation was
Ho: Expenditure does not Granger cause Revenue.
H1: Expenditure does Granger Cause Revenue
The data used in this particular research was collected from the Pakistan economic survey and
the state Bank of Pakistan (1980-81 to 2009-10). Data collected for empirical testing included
the federal and provincial tax revenues as well as current and development expenditure. Non tax
revenues (like interest income, profits and dividends) were not included in the estimation
because they are mostly exogenous in nature. To avoid the problem of non-stationary the data
was transformed into real per capita values.
In the first part, regression to test for causality was run between” total tax revenue and total
expenditure”, where total expenditure includes federal and provincial current and development
expenditure. Total revenue variable consisted of federal and provincial total tax revenues.
The results indicated that there is no causal relationship between total government revenue and
expenditure. So for both equations the null hypothesis was accepted (at 5% significance level)
which say that “Ho: total Revenue does not Granger Cause total Expenditure” and “Ho: total
Expenditure does not Granger cause total Revenue.”
Another researcher, Hussain (2005), in his article published in 2005 for Pakistani data showed
that in the data up to 2002-03 a causal relationship did indeed exist between revenue and
expenditure, but the results of this research by Tahir Sadiq indicate that the past relationship does
no longer exist due to the severe changes in our expenses and lack of taxation revenue.
In the second part of the research regression to test for causality, was executed between “total tax
revenue and total current expenditure”. Here too no causal relationship existed between the two
variables. The null hypothesis that “Ho: total Revenue does not Granger Cause total current
Expenditure” and “Ho: total current expenditure does not Granger cause total revenue” are both
Thirdly a causality test was conducted between “Tax Revenue and Development Expenditure”.
Here the null hypothesis that “Ho: total revenue does not granger cause total development
expenditure” and “Ho: total development expenditure does not Granger cause total revenue” are
both accepted at 5% significance level.
The results showed that there is no strong causal relationship (from any direction) between
federal and provincial tax revenue and expenditure which implied that governments in the past
have failed to control the budget deficit. Either the expenditures increased out of control or they
were unable to mobilize tax revenues.
As there is no causality between revenue and expenditure according to this research then an
increase in the tax to GDP ratio will not result into increase in the expenditure, so this way the
fiscal deficit will decline. Conversely, if expenditure decreases then this will not affect the efforts
made on the fiscal front and therefore the fiscal deficit will tend to decrease. This research
recommends that we should concentrate on reducing the fiscal deficit by focusing on both
revenue and expenditure which eventually will lead to reduction in inflation in Pakistani
Nadeem Iqbal, in his paper on studying the impact and relation of revenues and expenditures for
Pakistan concludes that budget deficit and debt have close relationship. Budget deficit is
financed through borrowing; it has effect neither on government expenditures nor on taxes. So a
deficit does not generate long run stabilizing effect on total revenues and government
By taking data from 1961 to 2008, he finds that Government expenditures have insignificant
effect on future taxes and similarly lag value of taxes has no effect on future taxes. So neither
‘spend-and-tax hypothesis’ and nor ‘tax-and-spend hypothesis’ is satisfied.
He finds that taxes and spending decision are taken independently and there is no long run cointegration
between taxes and expenditures. He gets another interesting result that in case of
Pakistan budget imbalances are reduced either through borrowing or through monetization of
debt. Results show that budget deficit has no impact on the behavior of government expenditures
and taxes.
Moreover change in taxes is not followed by change in government expenditures and vice versa.
It means there is no co-integration between taxes and spending. The historical behavior of
Pakistan’s inter-temporal budget constraint shows that taxes and spending decisions are
independent of each other.

The Tax-Spend hypothesis
This study aims to find the relationship between revenue and expenditure in the ASEAN
countries using data for 22 years fro 1980-2012 in its ten member states in order to asses the
relationship in these countries.
The study finds that a sound fiscal policy is imperative to promote internal equilibrium such as
stable price and sustainable growth in employment and output.
The study, after checking for stationary and cross dependence shows that panel co-integration
and cross-dependence is present in the long run relationship between expenditure and revenue.
The tax and spend hypothesis holds for five out of the ten countries whereby the changes in
government revenues lead to changes in expenditure.
In their case study over Botswana, the author Daniel K. Moalusi tested causality between
government expenditures and revenue using Granger test and finds a unidirectional link running
from revenues to spending in the short run and a bi-directional link in the long run. They
conclude that Botswana must raise taxes to reduce the deficit in budget.
Latin America
In their paper, Government Revenue-Expenditure Nexus: Evidence from Latin America, Ewing
and Payne examine the inter-temporal relationship amongst revenues and expenditures across
five Latin American countries. They emphasize that understanding the link between spending
and taxation is primary in understanding where and how the resources are allocated by the
government. The paper uses the Engle-Granger Bivariate co-integration approach to test several
hypotheses. For two of the countries, Chile and Paraguay a bi-directional causality is found to
support the fiscal synchronization hypothesis whilst the other three are based on the tax-spend
hypothesis, meaning that these countries should focus on fixing revenues to control the the
Cebula (1989) measures another impact of Budget deficit, which is that if the government sells
more bonds to finance the deficit, this will most likely cause interest rates to increase. This is
because the government needs to increase interest rates in order to attract investors and compete
with private institutions for the available funds. If government interest rates increase, this will
push up other interest rates as well; therefore this study focuses on the fiscal framework of the
government policies. The paper shows that the budget deficit increases the long term interest
rate, even with the high visibility of Ricardian Equivalence.
In this paper the impact of seasonal behavior is examined with respect to the government
revenue and expenditure. The paper provides a detailed study of the effect of taxation and
expenditure reform on the fiscal deficit. In the paper, quarterly data is used to empirically test for
interdependence between these two parts of the deficit that Barbados faces from 1973-1989. The
Granger testing assumes that the time series are weakly stationary, and non-deterministic,
meaning that the comparison of forecasts can be done through a bivariate causality between
government expenditure and revenue.
The statistical steps are a three staged process, whereby the orders of integrability of the data
series are determined, then transformed in to their stationary counterparts. And then the test for
co-integration for the appropriate frequencies was done. The results show that there is zero
frequency co-integration between country’s expenditure and revenue, and a unidirectional
causation from revenue to expenditure. The results show that the government revenue
information obtained during the time period could have been used to improve the forecasts of the
government expenditure in Barbados.
This is consistent with the Buchanan-Wagner view that revenue leads to expenditure.


The Spend-Tax hypothesis
Authors Damian C. Nwosu and Harrison O. Okafor in their article “Government Revenue and
Expenditure in Nigeria: A Disaggregated Analysis” examined the relationship between total
expenditure and disaggregated government expenditure with the total revenue and disaggregated
revenue of Nigeria. The data used was time series from the time period 1970 to 2011. The
empirical method focused on co-integration techniques and VAR models like the Error
Correction Mechanism (ECM) for analyses purposes. The results show that a long run
relationship is at work between government expenditure and revenue.
The results from VAR analyses show a unidirectional relationship between revenue and
expenditure variables. These result advocate spend-tax Hypothesis in Nigeria. This shows that
changes in expenditure initiate changes in revenue.
In this study the authors have tried to analyze the relationship between expenditure and revenue
in Ghana. For their empirical analyses they have used Engle-Granger bivariate co-integration and
error correction model.
The data used was taken from “Bank of Ghana, Ministry of Finance and Economic Planning, and
Ghana Statistical Service”. The data spanned from 1983 to 2007 and was converted from annual
data into quarterly data.
Engle-Granger bivariate test showed that revenue and expenditure were co-integrated. In real
terms if the expenditure increased in Ghana by 1% (as a percentage of GDP) that would increase
the revenues by 0.86 percent (as a % of GDP).On the other hand a 1% increase in government
revenue would increase government expenditures by 1.03%.
If there is a 1% nominal increase in government revenue that will elicit an increase of 0.88% in
government expenditure. On the other hand a nominal rise of 1% in government expenditure
would lead to 0.99% increase in government revenues. The results showed that both government
revenue and expenditures were positively related.
Long run analyses showed that a unidirectional causality existed from expenditure to revenue.
Thus it supported the spend-tax hypothesis in the long run for Ghana.
The author suggests that the medium budgeting framework in Ghana has to be improved, which
can be done if more attention is given to controlling expenditure rather than increasing revenues.
So while making the budget they must focus on ways to control spending.
Portugal, Ireland, Italy, Greece and Spain (PIIGS)
The paper looks at the government revenue-expenditure nexus in the case or Portugal, Ireland,
Italy, Greece and Spain (PIIGS) and the analysis covers a period from 1988-2014 using a
modified Granger causality, as proposed by Konya in 2006. Cross Sectional Dependence is
checked using three different tests, the given test assumes cross sectional dependence and cross country
The countries given in this context have strong influence through globalization and have a
comparable economic climate, since they all are part of the European union. The data is used on
level and both expenditure and revenue are expressed as a percentage of the GDP, the results
show that there is cross sectional dependence in the countries, any shock in one country will
affect the other country. The second result is that there is no cross country homogeneity and
every important significant relationship is independent to the country in focus.
The results show that there is a uni-directional causality that runs from revenues to expenditure
in the case of Greece and Italy, the same causality is also seen for Portugal, however it runs in
the opposite direction. Spain and Ireland show no Granger Causality.

The Fiscal Synchronization hypothesis
The study analyses the relationship between expenditures and revenues of the Swiss government
from 1872 to 2002. The data was checked for structural breaks and it showed that during the two
world wars there was a significant surge in expenditures couples with a drop in revenues, thereby
showing a clear budget deficit. The paper firstly inspects the cause behind a budget deficit,
checking whether it is the increase in revenues or the drop in expenditures. It uses a multivariate
error correction model in order to estimate short run as well as long run causal relations, it also
checks whether there was a structural break(or breaks) present which formed this causality.
The paper’s outcome is that it shows revenues and expenditures to be closely co-integrated. It
shows that there were major changes to the budgetary policy framework following the Second
World War. The error correction model shows a bi-directional causality that runs between
revenues and expenditures, thus relation them both in the long and short-term relationships. The
results conclude that Swiss policy makers consider the allocation policy when making decisions
regarding the taxation policy. Since the causality is bi-directional, the converse also holds true.
Asian Countries
In their article “Government Revenue and Government Expenditure Nexus in Asian Countries:
Panel Co-integration and Causality” authors Mohsen Mehrara , Mosayeb Pahlavani and Yousef
Elyasi focus on the relationship between government revenue and expenditure which plays a
major role in the policy making. The authors investigated the relationship between government
revenue and expenditure of 40 Asian countries .The time period under focus is from 1995-
2008.GDP was treated as a control variable in the model. The research found a co-integration
relation between government expenditure and revenue. The method the used was the Kao panel
co-integration test. They showed that in long run as well as short run a bidirectional causal
relationship exist government revenue and expenditure. According to this research
interdependence existed between the expenditure and revenues. The government must make its
decisions of expenditure and revenues at the same countries in this region must decrease
their expenses and increase their revenues at the same time in order to overcome the budget
The authors used government expenditure, revenue and GDP as the three variables in its model.
They used the panel unit root test proposed by Levin et al. (2002), panel co-integration using the
approach suggested by Kao (1999) and finally they used the Granger causality
Using yearly direct tax revenues, indirect tax revenues, non-tax revenues and government
spending of Malaysia, from period of 1970 to 2006, with 36 observations on each of the
variables, the results show that Granger-causality is running in both directions between
government spending and tax revenues. This means that for the case of Malaysia direct and
indirect tax revenues have a causal impact on government spending. That is, when the
government raises salaries, the purchasing power will also increase, indirectly affecting the sales
tax or service tax collection. Therefore, there is a long run bidirectional causality relationship
between government spending, direct tax revenue and indirect tax revenues.

Yashobanta and Smruti analyze the causal relationship between central government revenue and
expenditure for India using annual data over the period 1970-2008. The
Johansen co-integration test suggested that there is a long-run relationship between central
government revenue and expenditure. In the short run, unidirectional causality from expenditure
to revenue supporting “Spend-and-Tax” hypothesis exists but the result from Granger causality
test based on Vector Error Correction Models (VECM) suggests bidirectional causality between
central government revenues and expenditures in the long-run supporting Fiscal Synchronization
To solve the problem of continuously increasing budget deficits, Chang and Ho (2002) state that
the government of China needs to know if raising revenue, cutting expenditures, or simply
changing both sides without taking into account of the interdependence between the two, should
be the way to continue on fiscal situation in China. In this study, the hypothesis of tax-andspend,
spend-and-tax, or fiscal synchronization was tested using annual time series data for
China over the period 1977 to 1999. The results from Granger causality test show feedback
between government revenues and government expenditures, supporting the fiscal
synchronization hypothesis for China and recommending it to raise revenues and reduce
expenditures in order to tackle deficits.

In this particular article the researchers Yousef Elyasi and Mohammad Rahimi have tried to find
the causal relationship, for Iran, between the two variables government expenditure and
government revenue. They used time series yearly data which spanned from 1963 to 2007.
The results of the empirical testing showed that a bidirectional causal relationship exists between
the two variables and this relationship exists both in the long run as well as the short run. So this
research supports the fiscal synchronization hypothesis. Hence for the case of Iran, expenditure
should be decreased and at the same time government revenue should be increased, this action
plan will control the budget deficit. In Iran’s case, fiscal synchronization might be the reason
why government’s spending plans for the year are dependent its oil revenues.
These revenues then directly affect the expenditure of the government and also the rate at which
the economy grows. If we look at the expenditure side, if the government expenditure increases
it boosts up the economy which then leads to increase in the government’s revenue from other
sources besides oil revenues. The author concludes by saying that such bi-directional causality
might even make it harder for the government to control budget deficit.
Gulf Cooperation Council
In the paper to examine the relationship between government revenues, government spending
and economic growth for Gulf Cooperation Council (GCC) countries for the period from 1990 to
2010, the authors find that government expenditures Granger cause government revenues for
Qatar and the United Arab Emirates only. Hence, there is no evidence for spend-and-tax
hypothesis for 4 countries of sample. Moreover, government revenues Granger cause
government expenditures for Saudi Arabia only. Therefore, there is no evidence for tax-and38
spend hypothesis for 5 countries of sample. They also found a unidirectional causality running
from government expenditures to GDP in Bahrain only. For Kuwait, Qatar and Saudi Arabia,
GDP Granger causes government revenues while GDP Granger causes government expenditures
for Oman and Qatar.
Researchers in the article tried to find the relationship between national public budget revenue
and expenditure in Romania. They used the granger causality test as an empirical tool.
Controlling the budget deficit has been a very difficult task for the Romanian governments.
Different methods have been used to control this problem. They either focus on cutting down
expenditure or on increasing taxes. In order to examine the relationship between expenditure and
revenue this article used the Granger causality test.
The monthly data set for this analysis was taken from National Bank of Romania; during 1991-
2005.The two variables under analyses were “national public budget revenues” and “national
expenditure”. Co-integration and VAR was used to examine the relationship between these two
variables. They also applied the unit root test on the data and it showed that the variables were
not stationary (integrated of order one). Absence of stationarity means that co-integration tests
could be applied on the given Romanian data.
They used the following regressions to estimate for co-integration and to test for the stationarity
for residuals.
Rt = β0 + β1Gt + πt, and Gt = α0 + α1Rt + λt,
The results showed that the residuals and deficit in this case would be stationary. Any shocks in
revenue and expenditure will temporarily impact the fiscal deficit.
The granger test results showed a bidirectional causality which meant that fiscal synchronization
exists in Romania. VAR was also used to find out the impact of shocks in revenue and
expenditure. First they saw that a fluctuation in state budget revenue affected state budget
expenditure a lot and then the same strong impact was of shocks in state budget expenditure on
the state budget revenue.
So the results showed that government revenue causes the expenditure and government
expenditure causes revenue in other words the synchronization hypothesis was accepted in the
case of Romania. Only decreasing the expenditure as a fiscal policy will not help in this case
because that will lead to a fall in revenue as well.

The Institutional Separation hypothesis


United States of America
This paper examines the causality between revenues and expenditures through co-integration and
Hsiao’s modified Granger Causality test for the data post World War 2,that is 1946 up till 1996.
A bivariate model is used in the study, building on the argument of Darrat and Suliman (1994)
who argue that if variable x is not causing variable y, over a time period t then we cannot infer
that this causality holds true in the larger economic context.
The data testing for stationarity using three different tests was done and it was found that
expenditures do not granger cause revenues and hence taxes, this is contrary to Barro’s views
which state that higher government expenditure would result in a surge in taxes.
The results show that neither of our two variables depends on the other, instead taxes and
expenditure adjust to the disequilibrium independently of each other. Changes in expenditure
results due to a change in the economic growth, it shows a growth when there is a rise in the
demand for Government provided services such as higher incomes, growth in the population and
due to some local or external event. An increase in tax revenues, on the other hand is usually
caused due to an increase the GDP. In simpler words: in the United States spending is not
proceeded by taxation.
The paper concludes that there is no evidence of any causality between revenue and expenditure
and it is shown that GDP causes both these variables. This shows that using a bivariate model
excludes possible relevant variables and these may generate spurious results. This is because of
the institutional separation of the functions allocation and collection of tax of the Government.

European Union
This paper analyses panel data from 1976-2006 the European Union in order to test the four
hypothesis of budget deficit. Panel data is used because most countries are very similar in the
composition of their institutional structures; moreover the EU countries have to meet a predetermined
target so they can comply with the primary economic policy. This bootstrap method
is used on the general government spending and revenue collection. Through this procedure, the
panel is homogenous and thus we can test each individual panel using the Granger Causality test.
It requires a lag and it is not contradictory in results and allows the authors to detect how many
and what causality exists.
All four hypotheses have been analyzed using data from developed and developing countries and
the Granger causality has been examined. Using data modeling it was seen that government
expenditure Granger causes government revenue for France and Portugal, consistent with the
“spend-tax hypothesis”, “tax-spend hypothesis” for Germany, Italy, and Netherlands and finally,
and “institutional separation hypothesis” or “fiscal independence hypothesis” for Austria,
Belgium, Denmark, Finland, and UK.
The study aimed at finding whether there was a relationship between government revenues and
expenditures in the long run over the period 1961-2010. They used Augmented Dickey Fuller
and Phillip-Perron testing followed by an Engle-Granger Causality test. The results show that
there is no relationship amongst the variables. They took up to three lags in their data and were
unable to find some causality amongst the variables.
A model is specified, which shows the functional relationship between Government Expenditure
and Revenues as well as the real GDP, showing that changes in expenditure could be due to
changes in real GDP as well, therefore the model incorporates the changes in the output, and it’s
implications on the budget deficit. The result showed that there was no bidirectional or
unidirectional causality amongst the two variables and it shows that the government’s spending
decision and the government’s sources of revenue function completely independently of each
other. Hence the institutional independence hypothesis is being followed in Nigeria.




This paper uses the data covering a period of 1976-2014, which includes a total of 38
observations. The data sources are from the Pakistan Economic Survey and State Bank of
Pakistan. For the previous year, that is 2014, we are using the data as targeted by the ministry of
finance. The revenue and expenditure variables are a taken as a sum of both federal and
provincial level, showing the holistic figures. Nominal variable series have been used to study
the correlation and causality.
Our variables include
Ø Tax Revenues
Ø Non-Tax Revenues
Ø Total Government Revenues
Ø Current Expenditure
Ø Development Expenditure
Ø Total Government Expenditure.

This paper uses a five-step methodology, whereby we first take the unit root test, followed by an
OLS regression analysis, optimal lag determination through AIC, VAR analysis, and finishing
with a Granger Causality test. We used the E-views software to carry out the testing.
We use a bi-variate regression equation to observe the relationships between the following sets
of variables Total Revenue and Total Expenditure, Total Revenue and Current Expenditure,
Total Revenue and Development Expenditure, Non tax Revenue And Total Expenditure, Non-
Tax Revenue and Current Expenditure, Non-Tax Revenue and Development Expenditure, Tax
Revenue and Current Expenditure, Tax Revenue and Development Expenditure and Tax
Revenue and Total Expenditure. This way we have explored all sets of possible combinations of
interplay between the collections and spending of government of Pakistan.
A bi-variate regression equation is used to investigate the relationship between Government
expenditure and government revenue. The relationship is determined both ways. Therefore, the
nine sets of equations are given as below, please note that ‘t’ in the OLS equation denotes the
first difference lag of all variables it accompanies.

1. Total Revenue and Total Expenditure
!”!! = !! + !!!”!! + !!(1a)
!”!! = ! + !!!”!! + !!(1b)
‘R’ and ‘E’ represent the total revenue and the total expenditure. !! & !! are the respective
coefficients of Revenue and Expenditure. Taking the natural log of the variables allows us to
interpret the relationship in term of elasticities. !! & !! are the respective error terms associated
with our equations.
2. Total Revenue and Current Expenditure
!”!”! = !! + !!!”!! + !!(2a)
!”!! = ! + !!!”!”! + !!(2b)
‘R’ and ‘CE’ represent the total revenue and the current expenditure. !! & !! are the respective
coefficients of Revenue and Expenditure. Taking the natural log of the variables allows us to
interpret the relationship in term of elasticities. !! & !! are the respective error terms associated
with our equations.
3. Non tax Revenue And Total Expenditure
!”!! = !! + !!!”#!! + !!(3a)
!”#!! = ! + !!!!!! + !!(3b)
‘NR’ and ‘E’ represent the non tax t revenue and the total expenditure. !! & !! are the respective
coefficients of Revenue and Expenditure. Taking the natural log of the variables allows us to
interpret the relationship in term of elasticities. !! & !! are the respective error terms associated
with our equations.


4. Tax Revenue and Total Expenditure
!”!! = !! + !!!”#!! + !!(4a)
!”#!! = ! + !!!”!! + !!(4b)
‘TR’ and ‘E’ represent the tax revenue and the total expenditure. !! & !! are the respective
coefficients of revenue and expenditure. Taking the natural log of the variables allows us to
interpret the relationship in term of elasticities. !! & !! are the respective error terms associated
with our equations.
5. Tax Revenue and Current Expenditure
!”!”! = !! + !!!”#!! + !!(5a)
!”#!! = ! + !!!”!”! + !!(5b)
‘TR’ and ‘CE’ represent the tax revenue and the current expenditure. !! & !! are the respective
coefficients of Revenue and Expenditure. Taking the natural log of the variables allows us to
interpret the relationship in term of elasticities. !! & !! are the respective error terms associated
with our equations.

6. Non-­‐Tax Revenue and Development Expenditure
!”!”! = !! + !!!””#!! + !!(6a)
!””#!! = ! + !!!”!”! + !!(6b)
‘nTR’ and ‘dE’ represent the non-tax revenue and the development expenditure. !! & !! are the
respective coefficients of Revenue and Expenditure. Taking the natural log of the variables
allows us to interpret the relationship in term of elasticities. !! & !! are the respective error terms
associated with our equations.
7. Non-­‐TaxRevenue and CurrentExpenditure
!”!”! = !! + !!!””#!! + !!(7a)
!””#!! = ! + !!!”!”! + !!(7b)
‘nTR’ and ‘CE’ represent the non-tax revenue and the current expenditure. !! & !! are the
respective coefficients of Revenue and Expenditure. Taking the natural log of the variables
allows us to interpret the relationship in term of elasticities. !! & !! are the respective error terms
associated with our equations.

8. Tax Revenue and Development Expenditure
!”!”! = !! + !!!”#!! + !!(8a)
!”#!! = ! + !!!”!”! + !!(8b)
‘TR’ and ‘DE’ represent the tax revenue and the development expenditure. !1 & !1 are the respective
coefficients of Revenue and Expenditure. Taking the natural log of the variables allows us to interpret the
relationship in term of elasticities. !! & !! are the respective error terms associated with our equation.
9. Total Revenue and Development Expenditure
!”!”! = !! + !!!”!! + !!(9a)
!”!! = ! + !!!”!”! + !!(9b)
‘R’ and ‘DE’ represent the total revenue and the development expenditure. !1 & !1 are the respective
coefficients of Revenue and Expenditure. Taking the natural log of the variables allows us to interpret the
relationship in term of elasticities. !! & !! are the respective error terms associated with our equations

Unit Root Test
To test for stationarity in the series, we used two tests, the Augmented Dickey Fuller test and the
Phillip Peron test. The ADF test is used to test for unit roots in time series data, the more
negative the number, the stronger is the rejection of the null hypothesis, that is, there is a unit
root at some level of confidence. The PP test builds on the ADF with the same null hypothesis, it
differs because it uses a lag in order to make the test endogenous, thereby nullifying the ADF
test. The test makes a non-parametric correction to the t-statistic and hence the test result is
The null hypothesis for a unit root (stationarity) is tested against the hypothesis of no unit root
(non-stationarity). If the mean, variance and covariance of a given series are not varying over
time then the series is used at level. If they become stationary after taking the first difference, it
is used as such.
For both the ADF and PP, we conduct tests at level and at first difference of both the series. We
also use both the constant and the constant and trend for both our variables. Finally, the Akaike
Information Criterion (AIC), which measures the relative quality of the statistical model for the
given data is used to determine the lag length. The maximum lag length that we take is 2.

Ordinary Least Square Regression (OLS)
The Ordinary Least square regression estimates unknown parameters in a liner model of
regression, its goal is to minimize the differences between the data sets. The result is an estimator
that can be expressed using a simple linear formula.
We are using OLS Regression to test the existence and validity of the long run relationship using
between the two variables in each combination. Not only do we see how much one variable, say
tax, is explained by the other, say development expenditure; but also we see the whether this
relationship is positive or negative which will strengthen the results by appointing direction to
the causality indicated by Granger test. We use Ordinary Least Square method to run the
regression and Granger Causality Test with a level of significance at 5% (**).
Vector Auto-regression (VAR)
In order to capture the linear interdependence amongst the variables in the time series, we use the
VAR model. This model is useful for determining the dynamic behavior of economic and
financial time series and for forecasting. It allows researchers to determine relationships between
variables and their lags in the short run.
If the coefficients found using the VAR technique are jointly significant then this could mean the
possible existence of causality amongst the two variables.
Granger Causality Test
We are using Granger Causality Testing to test the relationship between revenues and
expenditure for Pakistan. The data being used is from the period 1976-2014. The variables we
have used are total government expenditure, current expenditure, development expenditure and
total government revenues, non-tax revenues, tax revenues. The source of data is the State Bank
of Pakistan and Pakistan Economic survey.
The Granger model says that if the past values of expenditure explain the revenue then there is a
causal relation from expenditure to revenues. And if the past values of expenditure explain the
revenues then there is a causal relationship from revenues to expenditure. The Granger test
employs the following equations where Y is given as expenditure and X is the revenue.
1: Xt = ΣaXt-j + ΣbjYt-j + μ
2: Yt = ΣcXt-j + ΣdYt-j + μ
The first equation explains revenue as function of the lag of revenue and lag of expenditure. The
second equation shows expenditure as a function of the lags of revenue and expenditure.
The hypotheses from both equations are as follows:
Equation 1:
• H0: REV does not Granger cause EXP
• H1: REV does Granger cause EXP
Equation 2:
• H0: EXP does not Granger cause REV
• H1: EXP does Granger cause REV

The Hypothesis can be checked for the existence of our aforementioned hypothesis using the
following as a guideline. (*)
Fiscal Synchronization
Ø If b≠0 and c≠0 and b≠0 and d≠0: Then there is bi-directional causality between revenues
and expenditures.
Fiscal Illusion Hypothesis
Ø If b=0 and c=0, and b=0 and d=0: Then there is no casualty between revenues and
Tax and Spend Hypothesis
Ø If b=0 in the equation 1 and d≠0 in Equation 2: Then there exists, a uni-directional
causality from revenues to expenditure.
Spend and Tax Hypothesis
Ø If b≠0 in Equation 1 and d=0 in Equation 2: Then there is a uni-directional casualty from
expenditure to revenue.
*Where a, b, c and d are elasticities.

Instead of the typical causality tests involving only the use of Granger Causality to conclude on
the relationship, we are using the aforementioned 5-step procedure which employs the following
1. Making the nominal data stationary.
2. Finding the optimal lag lengths using Akaike Information Criteria
3. Using OLS regression to find the elasticities i.e. how much one variable explains the
other and whether this relation is positive or negative. This gives the Long Run relation
between two variables.
4. Using VAR to find the Short Run correlation between the two variables and their lags,
taking both variables as endogenous and effecting each other. (This helps to predict if
causality can exist or not.)
5. Using the Granger causality to find the direction of causality between both variables
whether it is bi-directional or uni-directional.

Unit-root Test Results
We conducted the unit roots using both Augment Dickey Fuller and Phillips Perron Unit root
tests for the following 6 variables that were used in different combinations for our OLS
regression and VAR, and Granger causality tests.
Total Revenue
Using ADF-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Using PP-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Total Expenditure
Using ADF-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Using PP-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Non-Tax Revenue
Using ADF-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Using PP-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Tax Revenue
Using ADF-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Using PP-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Current Expenditure
Using ADF-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Using PP-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and

Development Expenditure
Using ADF-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and
Using PP-unit root test, the series were non-stationary at level, with constant and constant and
intercept. The series became stationary at first difference, with constant and constant and

We have organized our results in a way, whereby, every set of variables (as explained in the
methodology) is being assessed one by one, first by conducting the unit root tests, then running
OLS Regression, then testing for VAR and finally by applying the Granger Causality.
Total Revenue and Total Expenditure



Conclusion and Policy implications
The paper found the data to be stationary using the ADF and PP tests at first difference and the
optimal lag determined by the Akaike Information Criteria was found to be 2. The OLS
Regression explained the relationship between both variables in each of the nine permutations.
VAR testing was done amongst all endogenous variables in order to determine the short run
relation and indicate whether causality existed. Finally, Granger Causality testing was achieved
and the null hypothesis was accepted or rejected.
We can safely conclude that the results of our empirical analysis support the fourth hypothesis of
Institutional independence in Pakistan for the following five sets of combinations taken in the
study, these include. :
1. Total Expenditure and Total Revenue
2. Total Expenditure and Tax Revenue
3. Development Expenditure and Tax Revenue
4. Current Expenditure and Tax Revenue
5. Total Revenue and Development Expenditure
These results hold for both Short Run and Long Run in Pakistan, thus suggesting that the
Government of Pakistan can work to reduce and also finance fiscal deficit. This can be done
without regard to the relation amongst the aforementioned components because it has been
empirically proven that they are independent of each other.
Pakistan can work to increase resource mobilization by broadening its tax base as well as making
the taxation process simple, efficient, equitable, transparent and compliant. Since there is no
causality relation between the five pairs, these steps can be taken without any fear of increasing
the expenditures simultaneously.
It has also been found that a change in Tax Revenue collection does not cause any subsequent
changes in any of the components of expenditure. The tax revenues are independent of
expenditures and do not cause them. Hence, a budget deficit reduction strategy should aim at
increasing the resource mobilization through tax collection whilst simultaneously reducing the
unnecessary expenditures that the Government is incurring.
The results from our study are a true depiction of the reality that Pakistan’s budget formation has
faced over the years. Successive governments have formulated the budget policies without taking
into account either Friedman or Baro’s mechanisms of tax and spend or spend and tax. The
presence of Fiscal Independence show that the government has been spending and collecting
without taking into account the impact of these closely associated variables on each other.
Tax revenue has been a major component of total revenues as over the past 40 years, tax revenue
has yielded between 65-85% of the total revenue collection averaging at around 78% of total
revenue. Current Expenditure, on the other hand, comprises almost 75% of the total spending,
and is therefore a major component of total expenditure. There are thee main factors of current
expenditure, these are defense servicing, debt servicing and general administration expenditure.
Pakistan has been a fragile state ever since its inception, after the accession of East Pakistan, the
martial rule of the 1980’s and the constant transition from democracy to dictatorship has had a
negative impact on the macro economy of the country. Adding fuel to fire is the on-going war on
terror that has made defense spending a constant priority for every government. Defense
expenditure is a major component of current expenditure and each year it is set without regard to
the available resources. This shows that Pakistan doesn’t consider its revenue collection targets
whilst spending on defense.
Deficit, over a period of time translates into debt, and therefore, the same argument is applicable
to the current debt situation. Due to a constant balance of payment crisis, the permanent fiscal
deficit and the war on terror, the country is heavily indebted not only domestically but also
Debt servicing is a significant part of government spending and it is independent of the tax
collection, since debts have to be repaid as soon as the bond matures or the country would
default. Henceforth, this also suggests that there no attention is given to the empirical evidence
whilst deciding where to spend the revenues collected.
The study also shows the counter argument to the aforementioned results, for four of our
permutations; we deduce that there exists a positive uni-directional causality amongst the
following four pairs.
1. Total Revenue and Current Expenditure
2. Non-Tax Revenue and Total Expenditure and
3. Non-Tax Revenue and Development Expenditure
4. Non-Tax Revenue and Current Expenditure
It can be seen that a positive uni-directional causality is running from non-tax revenues to total
expenditures, from non-tax to development expenditure, from non-tax revenues to current
expenditures and from total revenues to current expenditure.

These relationships have important policy implications for Pakistan. A positive uni-directional
from Total Revenues to Current Expenditure shows that Friedman’s tax and spend hypothesis
exists in Pakistan. This implies that the Government must set revenues before incurring any
expenditure, especially those associated with general administration. A rise in total revenue
collection would result in an increase in the current expenditure of the Government and vice
Therefore, when reducing it’s current expenditures the government has to reduce revenues as
well. Conversely, raising revenues means that the Government would end up spending more on
current expenditure. Hence, to finance the Revenue Deficit where by current expenditure is
greater than total revenue, the government can only rely on tax collection or broaden its tax base.
A rise in taxes has been empirically shown to be independent of current expenditure.
The results also suggest that the strategy to finance the budget deficit through increasing non-tax
receipts, causes the expenditures to rise as a consequence. A change in non tax revenues causes
the same change in the total expenditure, current expenditure and development expenditure. Thus
the Government of Pakistan has to rely heavily on tax revenue as a source for expenditure
The taxation sector is the major source of revenue. Streamlining, strengthening and reforming
the taxation process as well as simplifying it will result in increased tax payments, as tax evaders
and there would be a greater population paying their due tax burden. Another major hurdle in the
tax collection process is the vice of corruption, eliminating corruption and making processes
more transparent would increase taxes manifold. Another important implication would be to
reform the entitlement programs, alter the defense budget and improve welfare programs –
minimizing these and other non-necessary payments such as protocol expenditure can reduce the
amount the government spends.
The aforementioned measures, if implemented would reduce the budget deficit by reforming not
only the revenue collection structures but also the expenditure outlets, hence leading to long term
macro economic stability


1. Amoah.B., Loloh.F.R.( August 2008): Causal Linkages Between Government
Revenue And Spending: Evidence From Ghana. Monetary Policy Analysis and
Financial Stability Department .Bank of Ghana.
2. Antonio Afonso and Cristophe Rault (2008): Bootstrap Panel Granger-
Causality Between Government Spending and Revenue. <;
3. B. Dougalas Bernheim: A Neoclassical Persepctive on Budget Deficits. <
4. Baghestani, H., McNown, R., (1994): Do Revenues or Expenditures Respond To
Budgetary Disequilibria? Southern Economic Journal, 61, (2), 311-322.
5. Barro R.J. (1979): On the Determination of the Public Debt; Journal of Political
Economy; 81, 940-971
6. Barro, R.J., (1978): On the Determination of the Public Debt; Journal of
Political Economy; 81, 940-971.
7. Benjamin S. Chen and Ashagre Yigletu: Causality between taxes and
expenditures in the USA, Multivariate approach
8. Blackley, P.R., (1986): Causality between revenues and expenditures and the size
of the federal budget. Public Finance Quarterly, 139–56.
9. Bradley T. Ewing and James E. Payne: Government Revenue-Expenditure
Nexus: Evidence from Latin America <
10. Buchanan J.M. and Wagner R.W. (1978): Dialogues concerning Fiscal
Religion; Journal of Monetary Economics, 4,627-636
11. Buchanan J.M., Wagner, R.W.( 1978): Dialogues concerning Fiscal Religion;
Journal of Monetary Economics, 4,627-636
12. Chang.T, Ho.Y, (2002): A Note on Testing “Tax-and-Spend, Spend-and-Tax or
Fiscal Synchronization”: The Case of China
13. Clark, J., Mirza, S.A., (2006): Comparison of some common methods of
detecting Granger noncausality. Journal of Statistical Computation and
Simulation 76, 207–231.
14. Cosimo Maggazino(2014): The relationship between revenue and expenditure in
the ASEAN countries. An international Quarterly, ISSN 1096-6838, Volume 31,
Number 3
15. Darrat A. F. (1998): Tax and Spend, or Spend and Tax? An Inquiry into the
Turkish Budgetary Process; Southern Economic Journal; 1998, 64(4), 940-956
16. Dr Ashfaque H Khan. (Tuesday, November 27, 2012): Fiscal responsibility.
The News. <;
17. Fasano, U., Wang, Q., (2002): Testing the Relationship Between Government
Spending and Revenue: Evidence from GCC Countries. IMF Working Paper
18. Friedman M. (1982): Supply side policies: where do we go from here in supplyside
economics in the 80’s; conference proceedings; Westport, CT
19. Friedman, M., (1978): The Limitation of Tax Limitation. Policy Review, 7-14.
20. Furstenberg, V., Green, R.J., Jeong, J., (1986): Tax and spend, or spend and
tax?. Review of Economics and Statistics, 179–88.
21. Gebhard Kirchgassner: The causality between Expenditures and Revenues:
Empirical Evidence from Switzerland, Grbhard Kirchgassner.
22. Hoover K.D and Sheffrin S.M. (1992): Causation Spending and taxes: Sand in
the sandbox or tax collector for the welfare state? American Economic Review,
82: 225-248
23. Jones J.D. and Joulfaian D. (1991): Federal Government Expenditure and
Revenue in the early years of the Republic: Evidence from 1792-1860; Journal of
Macroeconomics, 13(1), 133-155
24. Jones, J.D., Joulfain, D., (1991): Federal Government Expenditures and
Revenues in the Early Years of the American Republic: Evidence from 1792-1860.
Journal of Macroeconomics, 13, (1), 133-155.
25. Kivilicim Metin: The relationship between inflation and the budget deficit in
26. Mathew Abidofun Dada (2013): Impirical investigation of Government
Expenditure and Revenue Nexus: Implication for Fiscal sustainability in Nigeria.
27. Mehrara .M; Pahlavani .M and Elyasi.Y. (April 2011): Government Revenue
and Government Expenditure Nexus in Asian Countries: Panel Cointegration and
Causality. International Journal Of Business And Social Science, Vol. 2 No. 7;
Special Issue-April 2011.
28. Meltzer A.H. and Richard S. (1991): A rational Size of Government; Journal of
Political Economy, 89(5), 914-927
29. Mihai Mutascu: A bootstrap panel Granger causality analysis of government
revenues and expenditures in the PIIGS countries. 2015
30. Mioara.C.E.(2005): Causality Between Government Revenues And Expenditures
In Romania. Finance Chair, Academy of Economic Studies, Piata Romană
31. Mohammad Aslam Chaudhary and Ghulam Shabbir: Macroeconomic
impacts of budget deficit on Pakistan’s foreign sector
32. Mohsen Mehrara, Mosayeb Pahlavani and Yousef Elyas: Government
Revenue and Government Expenditure Nexus in Asian Countries: Panel
Cointegration and Causality..
33. Mohsen Mehrara: Government Revenue and Government Expenditure Nexus in
Asian Countries: Panel Cointegration and Causality
34. Musgrave R. (1966): Principles of Budget Determination (in Public Finance:
Selected Readings by Cameron and Henderson, 15-27
35. Nwosu.DC and Okafor.HO.(2014): Government Revenue And Expenditure In
Nigeria: A Disaggregated Analysis. Asian Economic and Financial Review, 2014,
4(7): 877-892. <;
36. Obeng, Samuel (2015): A causality of the Revenue Expenditure Nexus in Ghana
37. Peacock A. and Wiseman J. (1961): The Growth of Public expenditure in the
United Kingdom; Princeton, N.J (Princeton University Press)
38. Richard J. Cebula: A contemporary investigation of causality between the
primary government budget deficit and the ex ante real long term interest rate in
the US.
39. Robert J Barro: The Ricardian Approach to budget deficits
40. Ronald C. Craigwell, Hyginus Leon and Clyde Mascoll: Government revenue
and expenditure causality in the presence of seasonality in Barbados
41. Roshaiza Taha (2008): Causality between Tax Revenue and Government
Spending in Malaysia. <The International Journal of Business and Finance
42. Sadiq.T.(2010): The Causality between Revenues and Expenditure of the Federal
and Provincial Governments of Pakistan. The Pakistan Development Review 49:4
Part II (Winter 2010) pp. 651–662.<;
43. Vamvoukas G. (1997): Budget expenditures and Revenues: An Application of the
Error Correction Modeling; Public Finance; 1997, 52(1), 139-143
44. Waheed.A; Khilji .BA; Ahmad.M : Revenue and Expenditure: What causes
what? Empirical Evidence from Pakistan. Swiss Journal of Research in Business
and Social Sciences, Vol. 1, No.4, pp 48-59. <
45. Yashobanta.Y, Smruti.S, (2012): Causal Link between Central Government
Revenue and Expenditure: Evidence for India < http://mpra.ub.unimuenchen.
46. Yousef Elyasi, Mohammad Rahimi.(2007): The Causality between Government
Revenue and Government Expenditure in Iran. International Journal of Economic
Sciences and Applied Research 5 (1): 129-145.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s