Fiscal decentralisation and economic growth: role of democratic institutions.
Iqbal, Nasir ; ud Din, Musleh ; Ghani, Ejaz 等
This study attempts to analyse the impact of fiscal
decentralisation on economic growth. It also examines the
complementarity between fiscal decentralisation and democratic
institutions in promoting growth. The modelling framework is the
endogenous growth model augmented with measures of fiscal
decentralisation through democratic institutions. To capture the
multidimensionality, three different measures of fiscal decentralisation
are used. The overall analysis shows that revenue decentralisation
promotes economic growth while expenditure decentralisation retards
economic growth. Composite decentralisation positively influences
economic growth implying that simultaneous decentralisation of revenue
and expenditure reinforce each other to promote economic growth.
Analysis also shows that democratic institutions play a significant role
in realising the benefits of fiscal decentralisation.
Various policy implications emerge from this study.
JEL Classification: C26, E02, H11, H72, O11
Keywords: Fiscal Decentralisation, Democracy, Economic Growth,
Pakistan
1. INTRODUCTION
Over the past three decades, there has been a growing tendency
towards fiscal decentralisation (FD) in emerging and developing
economies. FD occurs through devolution of fiscal responsibilities for
public spending and revenue generation or collection from the central
government to the provincial or local governments. FD is an effective
strategy to promote economic growth by increasing the efficiency of the
public sector. FD promotes sound macroeconomic management through: (i)
efforts that streamline public sector activities, (ii) reducing
operational and informational costs of service delivery, and (iii)
increasing competition among sub-national governments in providing
public services. This process strengthens government accountability
towards its citizens by involving them in monitoring its performance and
demanding corrective measures. This process also makes governments
responsive and accountable, leading to curbing corruption and improving
delivery of public services.
The implicit assumption behind the positive contribution of FD is
the existence of a well-defined institutional mechanism. This increases
the accountability and transparency in the political system and hence
lowering corruption. That ultimately leads to efficient allocation of
public resources and hence economic growth. The recent advancement in
the field of FD strengthens this assumption and gives a role to
institutions in the theorem of fiscal decentralisation.
The government of Pakistan has taken various steps towards
strengthening the process of FD. The process of revenue sharing started
right from the inception of Pakistan. Since independence, the Niemeyer
Award 1947, the Raisman Award 1952, the One Unit Formula 1961 and 1965
and seven NFC awards based on the 1973 Constitution for revenue sharing
have been announced. Recently, government of Pakistan has undertaken two
major developments by signing the 7th National Finance Commission (NFC)
award (through which a bulk of resources has been transferred to the
provinces) and by passing the 18th Constitutional Amendment (through
which a wide range of fiscal responsibilities have been shifted from the
centre to the provinces). These developments would result in a
fundamental shift in the division of powers between the centre and the
provinces. The latter would have more autonomy in performing various
functions like the provision of public goods and services, and
macroeconomic management.
Consequently, various questions arise: What would be the effect of
implementing FD in Pakistan? Can Pakistan, with a weak institutional
framework, attain its objective of bringing prosperity to Pakistani
people through FD? Can each province with its particular local receipts
generate and expand the economy? Malik, et al. (2007) and Faridi (2011)
investigate the growth effects of FD in Pakistan and find positive
contributions of FD. However, these studies suffer from various
shortcomings. Firstly the studies ignore the possibility of reverse
causality and endogeneity among fiscal variables as pointed out in the
literature [see e.g. Zhang and Zou (1998); Xie, et al. (1999); Thiessen
(2003); Jin, et al. (2005); limi (2005)]. Secondly, the studies ignore
the multidimensional perspectives of FD [see e.g. Martinez-Vazquez and
McNab (2003)]. Thirdly, the studies ignore the potential role of
democratic institutions in making FD process effective and growth
enhancing [see e.g. Iimi (2005); Neyapti (2010)].
This study offers an empirical assessment of the growth effects of
fiscal decentralisation using various measures of decentralisation.
Secondly, the role of democratic institutions in explaining the growth
effects of fiscal decentralisation is examined. To the best of our
knowledge, no study to date has investigated the role of democratic
institutions in explaining the growth process of fiscal
decentralisation. This study's modelling framework is the
endogenous growth model augmented with the measures of fiscal
decentralisation and democratic institutions. The possibility of reverse
causality and endogeneity among fiscal measures leads to the use of a
GMM approach to estimation.
The rest of this paper is structured as follows: Section 2
summarises the existing literature concerned with the growth effects of
FD; Section 3 provides an overview of the FD process in Pakistan; the
modelling framework and the data and econometric issues are explained in
Section 4 and Section 5 respectively; Section 6 presents the results of
this study and Section 7 the conclusion.
2. LITERATURE REVIEW: THEORETICAL AND EMPIRICAL
Before proceeding with this study, it is important to have a broad
idea of the current developments in the theoretical and empirical
literature on FD.
The impact of FD on economic growth is derived from the traditional
theory of fiscal federalism which presents a general normative framework
for the assignment of functions to different levels of governments.
Under the traditional theory, the process of FD may generate greater
economic efficiency in the allocation of resources in the public sector.
(1) There are various theoretical explanations available in the
literature that spell out how FD generates economic efficiency in public
sectors.
First, economic efficiency can be generated through resource
mobilisation which occurs through FD. FD grants greater autonomy and
funds to the sub-national governments. With the availability of more
funds and autonomy in decision making process, sub-national governments
are compelled into mobilising the available resources in their own
jurisdictions, rather than waiting for the provision of public goods and
services to come from the central government. This leads to greater
emphasis on economic efficiency across jurisdictions within a country
and also to tapping into what otherwise may have been untapped potential
[Rodriguez-Pose and Ezcurra (2010)].
Second, the "Theorem of Decentralisation" provides a
well-known mechanism through which FD may lead to greater economic
efficiency. According to this theorem, the preferences for public goods
and services differ across individuals and regions. The level of welfare
achieved by a national government through providing uniform public goods
and services is always inferior to that which can be achieved by
providing public goods and services in a decentralised setup which
allows for provision of goods and services across the different regions
[Oates (1972)]. It is because the sub-national governments are better
informed about the preferences of citizens than the national government.
Therefore, sub-national governments always perform better in providing
public goods and services according to the needs of local communities.
Similarly economic efficiency can be enhanced if the citizens are
mobile so that they can locate themselves to the jurisdictions that best
match their preferences [Tiebout (1956)]. Oates (1993) argues that
expenditures for social and infrastructure sectors are likely to be more
growth enhancing if carried out by sub-national governments than the
central government which may ignore the differences in preference. The
growth enhancing advantages linked with the FD process are more visible
in larger and more heterogeneous countries. In a small country with
homogenous characteristics the informational advantages of implementing
policies and providing different public goods and services at the
regional or local level may be limited. The benefits of FD increase
because internal heterogeneity causes the preferences of individuals to
be more diverse. Hence the benefits of FD can only be realised beyond a
certain threshold of country size [Rodriguez-Pose and Ezcurra (2010)].
Third, the competition among the jurisdictions is seen as an
important mechanism to encourage efficiency in taxation, regulation and
supply of goods and services [Tiebout (1956); Brennan and Buchanan
(1980)]. In the Public Choice Approach, FD may lead to competition among
the jurisdictions for mobile factors of productions. This forces
discipline upon public officials who tend to pursue their own interest
and seek to maximise their revenues. Similarly, fiscal competition among
different levels of government leads to a market-preserving federalism
which minimises the extent of government interventions, hence
maintaining market efficiency [Weingast (1995)].
The positive impact of FD has been challenged in the previous
literature [see for example Prud'homme (1995); Tanzi (1996)]. The
critiques are based on the assumptions that underlie the
decentralisation models and the problems faced by local governments. The
proponents of decentralisation claim that local governments have an
informational advantage over the central government. However, this
assumption can be challenged on the grounds that central governments can
and do assign government officials to local offices. Apparently there is
no compelling reason to believe that the information obtained by these
representatives will be less accurate than the ones gathered by the
local governments [Tanzi (1996)]. Similarly, it is also argued that
local governments take into account the needs and preferences of the
local population and provide public goods and services accordingly.
Tanzi (1996) criticizes this assumption by saying that the local
populations may not have the power to actually influence the actions of
the local officials. This may result in local goods being produced
without taking into account the needs and preferences of the local
population. This is because local democracy is relatively weak and
ineffective especially in developing countries. Prud'homme (1995)
also argues that local preferences are complex and manifold. They cannot
be expressed in a single vote. The outcomes of local elections generally
depend on personal and/or political loyalties and rarely reflect the
preferences of the local population.
The opponents of decentralisation argue that there is a lack of
capacity to execute the responsibility for public services at
sub-national levels. The sub-national governments are usually less
efficient than the national government and this may undermine the
benefits of decentralisation [Tanzi (1996)]. There are problems like low
investment in technology and innovation because of the limited capacity,
both financially and technically, of the sub-national governments
[Prud'homme (1995)]. Due to the inefficiency of local
bureaucracies, local governments often lack good public expenditure
management systems to assist them in their tax and budget choice [Tanzi
(1996)].
Another potential problem usually associated with FD is the raiding
of the fiscal commons by the local governments due to the presence of a
soft-budget constraint. (2) In the case of a decentralised system,
sub-national governments may expect that their fiscal deficits are
covered by the central government. This in turn undermines the incentive
for sub-national governments to behave responsibly in handling finances.
The soft budget constraints have "a multiplicity of sources that
are associated with the prevailing fiscal institutions, with the
existing political structure, the weakness or even absence of various
important markets, and more importantly, the historical background of
intergovernmental fiscal affairs in the country" [Rodden, et al.
(2003)].
Most of the criticism against decentralisation does not dismiss the
idea of decentralisation per se, but is rather meant to highlight the
need for augmenting the decentralisation process with certain types of
institutions. According to the critics, only when these institutions are
present does decentralisation bear the fruits that are promised by its
proponents. The benefits of decentralisation largely depend on
institutional arrangements that govern the design and implementation of
decentralisation.
Given the lack of theoretical consensus on the impact of FD,
numerous studies have empirically examined the impact of FD on economic
growth.There are numerous studies that find a positive and significant
relationship between FD and economic growth [Oates (1995); Yilmaz
(1999); Thiessen (2003); limi (2005)]. However, various other studies,
have found a negative or even no relationship between FD and economic
growth [Os*es (1972, 1985); Davoodi and Zou (1998); Woller and Phillips
(1998); Martinez-Vazquez and McNab (2006); Thornton (2007); Baskaran and
Feld (2012); RodriguezPose and Ezcurra (2010)].
There are at least five possible reasons why the studies have
failed to come up with conclusive results on the role of FD. First, the
differences in the outcomes of these studies may be because different
studies have employed different measures of FD. The literature indicates
that it is difficult to measure the allocation of authority with
precision. If ambiguous or inappropriate measures of FD are employed,
wrong judgments about the growth effects of FD can be made [Ebel and
Yilmaz (2003)]. Akai and Sakata (2002) argue that studies which find a
negative association between FD and economic growth employ incorrect
measures of FD. Second, the differences in the outcome of empirical
studies that are based on a cross-country analysis may be due to the
differences in the economic, cultural, geographical and institutional
set-ups. In order to overcome these difficulties, single-country studies
have also been conducted. However, the outcome of these studies is still
inconclusive: some find a positive and significant association [see e.g.
Akai and Sakata (2002); Malik, et al. (2007); Carrion-i-Silvestre, et
al. (2008); Samimi, et al. (2010); Nguygen and Anwar (2011)] while
others find a negative or even no relationship between FD and economic
growth [see e.g. Xie, et al. (1999)]. Third, different countries have
different levels of FD, making it difficult to get consistent and robust
estimates based on a cross-country analysis. Fourth, the literature
identifies the possibility of reverse causality and endogeneity among FD
and economic growth [see e.g. Zhang and Zou (1998); Xie, et al. (1999);
Lin and Liu (2000); Thiessen (2003); Jin, et al. (2005)].
Martinez-Vazquez and McNab (2003) argue that reverse causality occurs
because efficiency gains from FD emerge as the economy's growth or
more decentralisation is demanded at relatively higher level of
development. However, existing literature does not control endogeneity
due to small sample sizes or the difficulty in finding valid instruments
with the only exception of limi (2005). Last, existing literature mainly
ignores the role of democratic institutions in making the FD process
effective with a few exceptions. For example, limi (2005) incorporates
the role of political institutions in analysing the role of FD. That
study finds that political institutions and FD complement each other in
promoting economic growth.
There is thus a clear need to re-examine the growth effects of FD,
especially at the country level using appropriate estimation methodology
and measures of FD.
3. FISCAL DECENTRALISATION IN PAKISTAN: AN OVERVIEW
The need for FD arose due to the mismatch between expenditure
requirements and the revenue generation capacity. This mismatch
necessitates the inter-governmental transfer among the federation and
provinces which is a vital part of the decentralisation process. The
horizontal as well as vertical mismatch between revenue and expenditure
requires legislative arrangement on financial transfers among different
levels of government. In both developed and developing countries, the
difference between revenue generation and actual expenditure across
national and sub-national governments is commonly observed.
Cross-country data on revenue and expenditure shows that there is a huge
mismatch between the revenue generation capacity of the national
government and the sub-national governments. A similar mismatch is
observed between national and sub-national government from the point of
view of expenditures. In the case of Pakistan, there is a serious
imbalance in the sub-national expenditures and revenue generation. The
statistics indicate that the revenue generation capacity of provincial
governments is nearly. 13 percent of the total revenue. On the other
hand, the expenditure needs of provincial governments are approximately
28 percent of the total expenditure (Table 1).
These imbalances between expenditure obligations and revenue among
federal and provincial governments leads to a large amount of transfers
of financial resources from the former to the latter level. Such
transfers and sharing of resources are embedded within the constitution
and supported by a series of legislative rules and regulations.
Inter-governmental transfers typically include revenue shares, grants,
straight transfers, loans and provincial revenues collected by federal
government and transferred to provinces after deducting collection
charges (e.g. royalties on gas and crude oil). There is a well-defined
mechanism for the distribution of resources from the federation to the
provinces in Pakistan. The resources are transferred from the federal to
the provincial level through the National Finance Commission (NFC). NFC
is an autonomous body established under the Constitution of Pakistan for
the re-distribution of resources from the federation to the provinces.
The resources are collected by the federal government and distributed
among the provinces according to their needs.
The amount of resources transferred from the central government to
the lower level government is determined on the basis of a certain
agreed formula. In Pakistan, the only criterion for resource
distribution has been the population since independence up to 2009. For
the first time a new criterion was designed for resource distribution
among the provinces in the 7th NFC award. In this award, four different
indicators are used to define the share of each province in the total
share to provinces, including (i) population, (ii) backwardness/poverty,
(iii) revenue generation/collection capacity, and (iv) inverse
population density (IPD) (Table 2).
In this formula, the population, once again, has the major share of
82 percent in total while poverty/backwardness has 10.3 percent share,
revenue generation/collection has 5 percent and inverse population
density (IPD) 2.7 percent.
The share of each province in the divisible pool has also changed
over time (Table 3). The share of Punjab was 57.87 in the 1990 NFC award
based on its population, whereas there was a minor decrease in 2006.
However, after the 7th NFC award in 2009, the share of Punjab has gone
down to 51.74 percent, mainly due to a change in the distribution
formula. The share of Sindh was 23.29 percent in 1990 on the basis of
its population; now it has increased to 24.55 percent in 2009. The share
of KPK was 13.54 in 1990 which has increased to 14.62 in 2009. Similarly
the share of Balochistan has increased from 5.3 percent in 1990 to 9.09
percent in 2009 on the basis of the revised formula.
4. MODELING FRAMEWORK
Fiscal decentralisation, the subject matter of this study, refers
to the devolution of policy responsibilities for public spending and
revenue collection from the central to the provincial governments.
Davoodi and Zou (1998) use the endogenous growth framework to analyse
the growth effects of FD. This study extends Barro's (1990)
endogenous growth model by assuming that public spending is carried out
at three levels of government: federal, state, and local. Later on,
various studies use this analytical framework to quantify the impact of
FD on economic growth [see e.g. Xie, et al. (1999); Iimi (2005)]. In
Pakistan, there are two levels of government: the federal and the
provincial which carry out public spending. Thus total government
spending is divided into two components: federal level and provincial
level government spending.
The benefits of FD can only be realised if the process is
complemented with good institutions which enhance the efficiency of the
public goods and services by meeting the preferred needs of the local
citizen; by increasing competition among provincial governments; by
reducing corruption and by enhancing accountability The role of
institutions is very crucial in making the theorem of decentralisation
applicable, limi (2005) further extends this framework by incorporating
the interactive term of FD and political institutions in the model.
Following limi (2005), the following model is defined to capture the
link among FD, democratic institutions and economic growth:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
Where GDPg is the per capita output growth rate, [tau] is the tax
rate, FD is the measure of fiscal decentralisation, INS represents
democratic institutions, X is the vector of control variables, [epsilon]
is the disturbance term that is assumed to be serially uncorrelated and
orthogonal to the explanatory variables and t (=1,2 ... ... .... N).
[[delta].sub.0], [[delta].sub.1], [[delta].sub.2], [[delta].sub.3] and
[[delta].sub.4] are the scalar parameters while [delta] is the vector of
parameters to be estimated. The vector X consists of control variables
that have frequently been used in growth literature as identified by
Mankiw, et al. (1992), Levine and Renelt (1992), Barro and Lee (1996)
and Sala-L-Martin (1997).
In this model, the interaction term, FD * INS should be of
particular interest since it allows us to test the hypothesis of FD and
democratic institutions being complementary. Based on this model, we aim
to empirically examine the following hypotheses:
(i) Fiscal decentralisation influences the evolution of per capita
output.
(ii) Fiscal decentralisation and democratic institutions are
complementary.
5. DATA AND ECONOMETRIC ISSUES
Our empirical analysis is based on time series data covering the
period 1972-2010. Data on fiscal decentralisation variables is collected
from the Fifty Year Economy of Pakistan and various annual reports
published by the State Bank of Pakistan. Data on other economic
variables is mainly taken from the Economic Survey of Pakistan (various
issues). Data on human capital is taken from the Barro and Lee Dataset
2011 and data on democratic institutions is taken from the Polity IV
Dataset.
5.1. Fiscal Decentralisation Measures
To empirically examine the role of FD, it is necessary to develop
measures of FD. There are two widely used measures of fiscal
decentralisation, namely the revenue decentralisation and the
expenditure decentralisation based on "Budget Data'. Revenue
decentralisation (RD) is measured as a ratio of the sub-national
government's revenue to the total government revenue (national plus
sub-national). Expenditure decentralisation (ED) is measured as a ratio
of sub-national government's expenditures to the total government
expenditures (national plus sub-national). Oates (1972) defines
expenditure centralisation as the share of the central government
spending in the total public spending and revenue centralisation as the
share of central government revenue in the total revenue. Woller and
Phillips (1998) re-define fiscal decentralisation measures after making
a few adjustments. First, in measuring revenue decentralisation, they
subtract the grant-in-aid given to sub-national government from the
total revenue and treat it as an expense to avoid double counting.
Second, in measuring expenditure decentralisation, they exclude social
security and defence spending from the total public spending as these
are considered to be the main parts of the non-decentralised government
spending.
These standard indicators have been used in a number of studies to
quantify the impact of FD. (3) However, the approaches to measure the
degree of FD and the reliability of the data have long been debated in
theoretical as well as in empirical literature. The data for FD measures
are obtained from the Government Finance Statistics (GFS) of the
International Monetary Fund (IMF). Ebel and Yilmaz (2003) identify three
major issues with GFS data. First, it is not possible to identify the
degree of local expenditure autonomy because the expenditures are
reported at the level of government that receives the amount. In this
way, the local spending that is directed by the central government is
added in the sub-national spending. Second, it is not possible to
identify the main source of revenues of the sub-national government,
whether collected through shared taxes, own taxes or piggybacked taxes.
Third, GFS does not distinguish between the different types of
intergovernmental transfers, whether these are conditional or
distributed according to some criteria. Therefore, the GFS data ignores
the degree of control of the central government over the revenues and
expenditure of the sub-national governments. These shortcomings
considerably overestimate the degree of FD [Stegarescu (2005)].
According to Martinez-Vazquez and McNab (2003), these measures are
defined on the basis of a single dimension of FD--expenditures going
through the sub-national budgets or revenue generated by the
sub-national governments. FD, however, is a multidimensional phenomenon
and it requires multidimensional measures to depict a true picture of
decentralisation. Martinez-Vazquez and Timofeev (2010) develop a
composite indicator of FD that captures the multidimensionality nature
of the FD process. The 'Composite Ratio', developed by
Martinez-Vazquez and Timofeev (2010), essentially combines the
information contained in expenditure and revenue ratios. Taking into
account the existing literature and availability of data, three
indicators are constructed to measure the level of FD for Pakistan. (4)
Revenue Decentralisation (RD)
The revenue decentralisation (RD) is measured as the ratio of the
provincial government's revenue to the total government revenue
(federal plus provincial)
RD = PR/[PR + FR]
Where RD, PR and FR are the 'Revenue Decentralisation',
'Provincial Revenue' and 'Federal Revenue'
respectively. Figure 1 shows the trend in revenue decentralisation in
Pakistan. The share of provincial government revenue in total government
revenue ranges from 10 to 25 percent. The share of provincial
governments' revenue was 15 percent in total government revenue in
1980, thereafter showing an increasing trend to reach 23 percent in
1987. After this period, there is a decreasing trend in revenue
decentralisation whereby provincial revenue share in total government
revenue reaches 10 percent in 2010.
[FIGURE 1 OMITTED]
Expenditure Decentralisation (ED)
The expenditure decentralisation (ED) is defined as the ratio of
provincial government expenditures to the total government expenditures
(federal plus provincial) less the defence expenditures and interest
payments on debt. These expenditures are mainly considered to be part of
the non-decentralised government expenditures.
ED = PE/PE + FE - (DE + IE)
Where ED, PE and FE are the 'Expenditure
Decentralisation', 'Provincial Expenditure' and
'Federal Expenditure' respectively. While DE and IE are
defence expenditure and interest payments respectively. Figure 2
represents the historical trend in expenditure decentralisation in
Pakistan. The share of provincial government expenditure in total
government expenditure ranges from 30 to 60 percent during the last
three decades. After reaching 50 percent in 1982, the share of
provincial government expenditure shows a declining trend reaching 39
percent in 1989. For the greater part of the 1990s, expenditure
decentralisation shows an increasing trend. However, after 1998 once
again, provincial shares in total expenditure show a decreasing trend,
declining from 55 percent in 1998 to 35 percent in 2010.
[FIGURE 2 OMITTED]
Composite Decentralisation (CD)
Composite decentralisation is measured using both revenue
decentralisation and expenditures decentralisation. It is more useful in
terms of analysing the impact of FD on economic growth.
CD = RD/[1 - ED]
Where CD, RD and ED are the 'Composite Decentralisation',
'Revenue Decentralisation' and 'Expenditure
Decentralisation' respectively. Figure 3 shows the composite of
revenue and expenditure decentralisation in Pakistan. This represents
the combined outcome of both processes. The trend shows that the
'Composite Decentralisation' measure ranges from 13 to 40
percent. 5.2. Other Control Variables
[FIGURE 3 OMITTED]
5.2 Other Control Variables
The dependent variable is GDP per capita growth rate. Descriptive
statistics show that the average GDP per capita is 451 US$ at constant
2000 prices. The average growth rate of GDP per capita is 2.234. Human
capital (HC) is measured using total secondary school enrolment without
considering age and gender composition. The average human capital is
20.02 and it moves from 7.1 in 1972 to 34.6 in 2010. Openness (OPN) is
defined as the ratio of total trade (imports plus exports) as percent of
GDP. Trade openness varies from 27 percent to 42 percent with the
average of 34 percent. Tax to GDP ratio is measured as the ratio of the
total consolidated tax receipts of government to GDP. The average tax to
GDP ratio is 12 percent with the range of 9 to 15 percent. The
contribution of taxes in economic growth crucially depends upon the
structure of the taxes. The impact of taxation on economic growth is
positive if private capital is less productive than public capital and
is negative if additional taxation is very expensive (limi, 2005).
Inflation is measured as the growth rate of CPI. The average inflation
rate is 9.6 varying from 3.1 percent to 30 percent. The overall budget
deficit (BD) fluctuates between 2.3 and 10.2. On average the overall
budget deficit is 6.5 in Pakistan. Democracy is used as a proxy for
measuring the quality of institutions in Pakistan. The data on democracy
is taken from the Polity IV dataset published by Marshall and Jaggers
(2011). The democracy index ranges from +10 (full democracy) to -10
(full autocracy). The descriptive statistics show that the average
quality of institution is 0.85 with the range of -7 to +8 in Pakistan.
There are several studies that have used the Ordinary Least Squares
(OLS) estimation technique to empirically investigate the impact of FD
on economic growth. A number of studies identify the possibility of
reverse causality and endogeneity among FD and economic growth [see e.g.
Zhang and Zou (1998); Xie, et al. (1999); Lin and Liu (2000); Thiessen
(2003); Jin, et al. (2005)]. Martinez-Vazquez and McNab (2003) argue
that reverse causality exists because efficiency gains from FD emerge as
economies grow or more decentralisation is demanded at relatively higher
levels of development. However, the existing literature does not control
endogeneity due to small sample sizes or the difficulty in finding valid
instruments with the only exception of Iimi (2005). Under this
situation, OLS estimates become biased and inconsistent. To tackle
endogeneity, the instrumental variables (IV) methods are used in the
empirical estimations. The IV methods are used to solve the problems of
simultaneity bias between explanatory variables, the dependent variable
and the error measurement.
The application of the generalised method of moments (GMM) can be
considered as an extension of the IV estimation method. The main
advantage of the GMM estimation method is that the model need not be
serially independent and homoscedastic. Another benefit of the GMM
estimation technique is that it generates parameters through maximising
the objective function which includes the moment restrictions in which
correlation between the lagged regressor and the error term is zero.
Keeping the advantages of the GMM estimation technique to overcome
endogeneity and omitted variable bias, the GMM estimation procedure
developed by Arellano and Bond (1991), Arellano and Bover (1995) has
been applied to estimate growth and stability equations using lagged
values of the variables as instruments. The STATA v11 has been used for
estimation.
The standard approach to determine the stationarity of the time
series data is checking the existence of unit roots in the given series.
The most commonly employed test for unit root analysis is called
Augmented Dickey Fuller (ADF) test [Dickey and Fuller (1981)]. The
results of the ADF test are reported in Table 5. The test statistics
indicate that inflation, budget deficit, GDP per capita growth rate,
openness and M2 to GDP ratio are stationary at level. While revenue
decentralisation, expenditure decentralisation, composite
decentralisation, macroeconomic instability index, human capital,
capital stock per worker, tax to GDP ratio and democratic institutions
are non-stationary at level and become stationary at first difference
which implies that these variables are difference stationary with one
order of integration.
6. EMPIRICAL RESULTS
This study has estimated the impact of various dimensions of FD on
economic growth. In Table 6, the impact of revenue decentralisation on
economic growth is shown. Various specifications to test the robustness
of results have been used.
Revenue decentralisation has a positive and significant impact on
economic growth in all specifications which are consistent with the
theory of decentralisation. This positive association indicates that the
higher the level of decentralisation on revenue side, the higher the GDP
per capita. The transfer of revenue enhancing responsibilities to
provincial governments is conducive for economic growth in Pakistan. As
shown in table 6, this result is robust, regardless of the inclusion of
other control variables; the estimated impact of revenue
decentralisation on economic growth remains positive and significant.
The impact of expenditure decentralisation on economic growth is
measured using five different specifications and results are reported in
Table 7. Expenditure decentralisation has a negative and significant
impact on economic growth in all specifications. (5) As shown in Table
7, these results are robust, regardless of the inclusion of other
control variables; the estimated impact of ED on economic growth remains
negative and significant. The negative association between ED and
economic growth implies that ED has growth retarding effects in
Pakistan. These results are in contrast to the theory of
decentralisation. Davoodi and Zou (1998) find similar results for
developing countries. There are several justifications that explain the
negative association of expenditure decentralisation with economic
growth in Pakistan.
First, the composition of public spending carried out by provincial
governments may explain the growth retarding effects of ED. The
expenditure decentralisation measure in this dissertation does not
indicate the composition of the public spending of the provincial
governments. Provincial governments generally allocate excessive amounts
to current expenditure instead of capital and infrastructure spending.
The literature suggests that the growth effects of capital and
infrastructure spending are positive and that of current spending are
negative.
Second, the institutional weaknesses at the provincial level may
lead to more corruption and hence lower economic growth. The third
reason may be the lack of autonomy in decision making by the provincial
governments that in turn can lead to inefficient outcome. The process of
FD may not materialise in its true sense because the decisions by
provincial governments may still be influenced by the federal
government. Fourth, the provincial governments may be unable to execute
proficient policies and organise efficient governance due to lack of
human as well physical resources. Fifth, the provincial government may
not be able to achieve economies of scale for the reason that they may
be too small to efficiently carry large scale infrastructure development
projects. Finally, the provincial governments often lack the
institutional framework that is required to gain the benefits of FD. The
lack of institutional framework can contribute to more corruption, less
accountability and inefficiency in the policy making processes, causing
a slowdown in the growth process. Similar arguments are put forward by
Martinez-Vazquez and McNab (2006) to explain the negative relationship
between expenditure decentralisation and economic growth for developing
countries.
Similar to RD and ED, the impact of composite decentralisation (CD)
on economic growth can be estimated. In CD, revenue decentralisation and
expenditure decentralisation reinforce each other. Table 8 presents the
results obtained from GMM estimation. The impact of composite
decentralisation on economic growth is positive and significant in all
models. The positive association reveals that composite decentralisation
(CD) is beneficial for Pakistan.
Numerous control variables have been used to estimate the impact of
FD on economic growth. Tax to GDP ratio (T/GDP) has a positive and
significant relationship with economic growth. This implies that the
higher the tax to GDP ratio, the higher the GDP per capita growth. Trade
openness (OPN) has a positive and significant impact on economic growth,
implying that trade is beneficial for economic growth in Pakistan. The
positive association of trade openness and economic growth is due to the
benefits emerging from specialisation, competition and economies of
scale. It is also due to productivity improvements made possible through
the access to advanced technologies [Din, et al. (2003)]. Various
empirical studies also provide evidence that trade promotes economic
growth in Pakistan [Khan, et al. (1995); Iqbal and Zahid (1998); Din, et
al. (2003)]. Human Capital (HC) has a positive and significant impact on
per capita GDP growth, implying that Pakistan could increase its per
capita growth rate by investing more in human capital. This finding
confirms the traditional view that the countries that invest more in
their human capital do better in terms of economic growth. These results
are broadly in line with the other studies that have found a positive
association between human capital and economic growth in Pakistan [Abbas
(2001); Abbas and Foreman-Peek (2008); Qadri and Waheed (2011)].
Inflation has a negative and significant impact on economic growth,
implying that inflation hurts the growth process. A negative and
significant relationship between budget deficit and economic growth has
been found.
6.1. Role of Democratic Institutions
The literature suggests that FD may positively affect economic
growth in the presence of strong democratic institutions. In order to
check the role of institutions in FD process, the interactive term of
democratic institutions is added. Neyapti (2004, 2010) similarly
suggests the use of expenditure decentralisation with other
institutions, such as central bank independence, local accountability,
and governance quality, to test for the effectiveness of expenditure
decentralisation. In Table 9, democratic institutions and interactive
term of democratic institutions is added with FD.
The interactive term of revenue decentralisation and expenditure
decentralisation with democratic institutions has a positive and
significant impact on economic growth implying that FD and democratic
institutions are complemented by each other. However, Brambor, et al.
(2006) shows that it is incorrect to decide on the inclusion of the
interactive term simply by looking at the significance of the
coefficient of the interactive variable. The marginal effect of FD on
economic growth should be observed by constructing confidence intervals
for the estimates of coefficient of ED and interactive term of ED and
institutions over the possible values of the institutions. Similarly for
RD, if the interval lies above the zero line, then the effect is
significantly positive and vice versa. Through this, the range of
institutional values for which the effect of RD and ED can be said to be
significant, can be found.
[FIGURE 4 OMITTED]
Figure 5 shows that with the low quality of institutions, the
growth effect of expenditure decentralisation is negative. However, as
the quality of institutions improves, the expenditure decentralisation
exerts a positive impact on economic growth. The institutional school of
thought argues that the quality of institutions increases the efficiency
of the economic factors of production. It reduces the level of
corruption and enhances the accountability of the governments. (6)
[FIGURE 5 OMITTED]
7. CONCLUDING REMARKS AND POLICY IMPLICATIONS
In this study, the growth effects of fiscal decentralisation in
Pakistan over the period 1972-2010 using the GMM estimation procedure
have been analysed. The empirical analysis shows that revenue
decentralisation is growth enhancing in Pakistan. Decentralisation of
revenue generation responsibilities generates positive externalities
which increase the per capita income of the country. On the other hand,
it is found that expenditure decentralisation has a negative association
with the growth rate of per capita income. This is mainly due to the low
institutional quality which may increase the corruption level and make
public officials less accountable. Lack of human and physical
infrastructure may also lead to inefficient outcome of expenditure
decentralisation in Pakistan. Composite decentralisation also has a
positive association with growth mainly due to the positive effect of
revenue decentralisation. This implies that if Pakistan focuses
simultaneously on both types of decentralisation then it will be helpful
in enhancing the per capita income. Only expenditure decentralisation is
not helpful in achieving high and sustainable economic growth. The
empirical analysis also reveals that the tax to GDP ratio has a positive
association with economic growth. Trade openness has positive linkages
with the growth rate of per capita income in Pakistan. Human capital
also positively influences economic growth. Analysis reveals that FD
becomes effective in the growth process if it is complemented with good
quality institutions. It is observed that the interaction of expenditure
decentralisation and revenue decentralisation with democratic
institutions has a positive impact on economic growth.
Few policy implications emerge from the empirical analysis:
(i) The tax to GDP ratio has a positive association with economic
growth. This finding has important implications for Pakistan. In
Pakistan the tax to GDP ratio is very low as compared to other developed
and developing countries. Due to a low tax base, Pakistan is
consistently facing the problem of a high budget deficit. Increasing the
tax to GDP ratio has two advantages: firstly, it directly contributes to
economic growth and, secondly, it mitigates the negative impact of
budget deficit on economic growth through reducing budget deficit. In
Pakistan the main source of tax is the general sales tax on goods and
services (GST) which is non-distortionary in nature. Taking into account
the growth and stability effect of taxation, there is a need to further
broaden the tax base and tax rates. To widen the tax base, all sources
of income--including services, real estate and agriculture--must be
brought under the tax net. The implementation of the Reformed General
Sales Tax (RGST) can be an option for increasing the tax base and tax
revenue. Implementation of RGST is essential to fully tap the revenue
generation capacity as well as to help the documentation process in the
economy.
(ii) The process of fiscal decentralisation, especially revenue
decentralisation, is beneficial for the economy of Pakistan. To achieve
long term economic growth, revenue decentralisation should be better
streamlined through making the provinces more reliant on their own
resources. The positive association of revenue decentralisation with
economic growth has an important implication for the design of fiscal
decentralisation in Pakistan because the process of restructuring
government (which began with the passage of 7th NFC ward and 18th
Constitutional Amendment) is in the early stage. This requires a serious
effort both in terms of strengthening the institutions and promoting
fiscal decentralisation to achieve the objective of better economic
growth. The benefits of fiscal decentralisation can only accrue when
provincial governments have a real fiscal autonomy, adequate
accountability and sufficient capacity to respond to the local
requirements.
(iii) Expenditure decentralisation can only be effective when the
provinces have sufficient administrative capacity and have been made
accountable and transparent through good institutions. The expenditure
decentralisation can make positive contribution to economic growth if
steps are taken to improve the administrative capacity of the provincial
governments. This requires initiating programmes that provide technical
and administrative skills to the public officials at the provincial
level. These programmes are more likely to enhance the spending
management skills of the provincial governments.
(iv) The present initiatives taken by the government in
strengthening the provinces through providing more autonomy and
resources have a clear implication for Pakistan's long term
economic prosperity and macroeconomic stability. However, the outcome of
these reforms crucially depends upon the institutional framework of the
country. Strengthening of democracy is a prerequisite for achieving the
fruits of fiscal decentralisation.
REFERENCES
Abbas, Q. (2001) Endogenous Growth and Human Capital: A Comparative
Study of Pakistan and Sri Lanka. The Pakistan Development Review 40:4,
987-1007.
Abbas, Q. and J. Foreman-Peck (2008) The Mincer Human Capital Model
in Pakistan: Implications for Education Policy. South Asia Economic
Journal 9:2, 435-462.
Akai, N. and M. Sakata (2002) Fiscal Decentralisation Contributes
to Economic Growth: Evidence from State-Level Cross-Section Data for the
United States. Journal of Urban Economics 52, 93-108.
Arellano, M. and O. Bover (1995) Another Look at the Instrumental
Variables Estimation of Error-components Models. The Journal of
Econometrics 68:1, 29-51.
Arellano, M. and S. Bond (1991) Some Tests of Specification for
Panel Data: Monte Carlo Evidence and an Application to Employment
Equations. Review of Economic Studies 58:2, 277-297.
Barro, R. J. (1990) Government Spending in a Simple Model of
Endogenous Growth. Journal of Political Economy 98:5, 108-125.
Barro, R. J. and J. W. Lee (1996) International Measures of
Schooling Years and Schooling Quality. American Economic Review 86:2,
218-223.
Baskaran, T. and L. P. Feld (2012) Fiscal Decentralisation and
Economic Growth in OECD Countries; Is there a Relationship? Public
Finance Review doi: 10.1177/1091142112463726.
Brambor, T., W. R. Clark, and M. Golder (2006) Understanding
Interaction Models: Improving Empirical Analysis. Political Analysis
14:1, 63-82.
Brennan, G. and J. M. Buchanan (1980) The Power to Tax--Analytical
Foundation of a Fiscal Constitutions. Cambridge: Cambridge University
Press.
Cantarero, D. and P. P. Gonzalez (2009) Fiscal Decentralisation and
Economic Growth: Evidence from Spanish Regions. Public Budgeting and
Finance 29:4, 24-44.
Carrion-iSilvestre, J. L., M. Espasa, and T. Mora (2008) Fiscal
Decentralisation and Economic Growth in Spain. Public Finance Review
36:2, 194-218.
Davoodi, H. and H. Zou (1998) Fiscal Decentralisation and Economic
Growth: A Cross-country Study. Journal of Urban Economics 43, 224-257.
Dickey, D. A. and W. A. Fuller (1981) Likelihood Ratio Statistics
for Autoregressive Time Series with a Unit Root. Econometrica 49:4,
1057-1072.
Din, M., E. Ghani, and O. Siddique (2003) Openness and Economic
Growth in Pakistan. The Pakistan Development Review 42:4, 795-807.
Ebel. R. D. and S. Yilmaz (2003) On the Measurement and Impact of
Fiscal Decentralisation. In Jorge Martinez-Vazquez and James Aim (eds.)
Public Finance in Developing and Transitional Countries. Cheltenham:
Edward Elgar.
Eller, M. (2004) The Determinants of Fiscal Decentralisation and
Its Impact on Economic Growth: Empirical Evidence from a Penal of OECD
Countries, Diploma in Economics, Vienna University of Economics and
Business Administration.
Faridi, M. Z. (2011) Contribution of Fiscal Decentralisation to
Economic Growth: Evidence from Pakistan. Pakistan Journal of Social
Science 31:1, 1-13.
Feltenstein, A. and S. Iwata (2005) Decentralisation and
Macroeconomic Performance in China: Regional Autonomy has Its Costs.
Journal of Development Economics 76:2, 481-501.
Giugale, M. and S. B. Webb (2000) Achievements and Challenges of
Fiscal Decentralisation: Lessons from Mexico. World Bank Publications.
Iimi, A. (2005) Decentralisation and Economic Growth Revisited: An
Empirical Note. Journal of Urban Economics 57:3, 449-461.
Iqbal, Z. and G. M. Zahid (1998) Macroeconomic Determinants of
Economic Growth in Pakistan. The Pakistan Development Review 37:2,
125-148.
Jin, H., Y. Quian, and B. Weingast (2005) Regional Decentralisation
and Fiscal Incentives: Federalism, Chinese Style. Journal of Public
Economics 89:9-10, 1719-1742.
Khan, A. H., A. Malik, and L. Hasan (1995) Exports, Growth and
Causality: An Application of Co-integration and Error-correction
Modelling. The Pakistan Development Review 34:4, 1001-1012.
Kornai, J. (1979) Resource-constrained versus Demand-constrained
System. Econometrica 47:4, 801-819.
Levine, R. and D. Renelt (1992) A Sensitivity Analysis of Cross
Country Growth Regressions. American Economic Review 82:4, 942-963.
Lin, J. Y. and Z. Liu (2000) Fiscal Decentralisation and Economic
Growth in China. Economic Development and Cultural Change 49:1, 1-21.
Malik, S., M. Hassan, and S. Hussain (2006) Fiscal Decentralisation
and Economic Growth in Pakistan. The Pakistan Development Review 45:4,
845-854.
Mankiw, G. N., D. Romer, and D. N. Weil (1992) A Contribution to
the Empirics of Economic Growth. Quarterly Journal of Economics 107:2,
407-437.
Marshall, M., T. Gurr, and K. Jaggers (2013) Polity IV: Political
Regime Characteristics and Transactions, 1800-2012. Centre for
Systematic Peace, Polity IV.
Martinez-Vazquez, J. and A. Timofeev (2010) Decentralisation
Measures Revisited. Public Finance and Management 10:1.
Martinez-Vazquez, J. and R. M. McNab (2003) Fiscal Decentralisation
and Economic Growth. World Development 39, 1597-1661.
Martinez-Vazquez, J. and R. M. McNab (2006) Fiscal
Decentralisation, Macrostability and Growth. Hacienda Publica
Espanola/Revista de EconomiaPublica 179:4, 25-49.
Neyapti, B. (2004) Fiscal Decentralisation, Central Bank
Independence and Inflation: A Panel Investigation. Economics Letters
82:2, 227-230.
Neyapti, B. (2006) Revenue Decentralisation and Income Distribution
.Economics Letters 92:2, 409-416.
Neyapti, B. (2010) Fiscal Decentralisation and Deficit:
International Evidence. European Journal of Political Economy 26:2,
155-166.
Nguyen, L. P. and S. Anwar (2011) Fiscal Decentralisation and
Economic Growth in Vietnam Journal of the Asia Pacific Economy 16:1,
3-14.
North, D. C. (1981) Structure and Change in Economic History. New
York: Norton and Co.
Oates, W. E. (1972) Fiscal Federalism. New York: Harcourt Brace
Jovanovich.
Oates, W. E. (1985) Searching for Leviathan: An Empirical Study.
American Economic Review 75:4, 748-757.
Oates, W. E. (1993) Fiscal Decentralisation and Economic
Development. National Tax Journal 46:2, 237-243.
Oates, W. E. (1995) Comment on 'Conflict and Dilemmas of
Decentralisation' by Rudolf Hommes.ln M. Bruni and B. Pleskovic
(eds.) Annual World Bank Conference on Development Economics. 351-353.
Oates, W. E. (2005) Toward a Second-Generation Theory of Fiscal
Federalism. Journal of International Tax and Public Finance 12:4,
349-373.
Prud'homme, R. (1995) On Danger of Decentralisation. World
Bank Research Observer 10:2, 201-220.
Qadri, F. S. and A. Waheed (2011) Human Capital and Economic
Growth: Time Series Evidence from Pakistan. Pakistan Business Review
12:4, 815-833.
Rodden, J., G. S. Eskeland, and J. Litvack (2003) Fiscal
Decentralisation and the Challenges of Hard Budget Constraints.
Cambridge and London: MIT Press.
Rodriguez-Pose, A. and A. Kroijer (2009) Fiscal Decentralisation
and Economic Growth in Central and Eastern Europe. Growth and Change
40:3, 387-^417.
Rodriguez-Pose, A. and R. Ezcurra (2010) Is Fiscal Decentralisation
Harmful for Economic Growth? Evidence from the OECD Countries. Journal
of Economic Geography 11:4, 619-644.
Sala-i-Martin, X. (1997) I Just Ran Two Million Regressions.
American Economic Review 87:2, 178-183.
Samimi, A. J., S. K. P. Lar, G. K. Haddad, and M. Alizadeh (2010)
Fiscal Decentralisation and Economic Growth in Iran, Australian. Journal
of Basic and Applied Sciences 4:11, 5490-5495.
Stegarescu, D. (2005) Public Sector Decentralisation: Measurement
Concepts and Recent International Trends. Fiscal Studies 26:3, 301-333.
Tanzi, V. (1996) Fiscal Federalism and Decentralisation: A Review
of Some Efficiency and Macroeconomic Aspects. In M. Bruno and B.
Pleskovic (eds.) (1995) Annual World Bank Conference on Development
Economics. Washington, DC: World Bank. 295-316.
Thiessen, U. (2003) Fiscal Decentralisation and Economic Growth in
High Income OECD Countries. Fiscal Studies 24, 237-274.
Thornton, J. (2007) Fiscal Decentralisation and Economic Growth
Reconsidered. Journal of Urban Economics 61:1, 64-70.
Tiebout, C. M. (1956) A Pure Theory of Local Expenditures. Journal
of Political Economy 64:5, 416-424.
Treisman, D. (2000) Decentralisation and Inflation: Commitment,
Collective Action, or Continuity. American Political Science Review
94:4, 837-857.
Weingast, B. R. (1995) The Economic Role of Political Institutions:
Market-preserving Federalism and Economic Development. Journal of Law,
Economics and Organisations 15:1, 1-31.
Woller, G. M. and K. Phillips (1998) Fiscal Decentralisation and
LDC Economic Growth: An Empirical Investigation. Journal of Development
Studies 43, 139-148.
Xie, D., H. Zou, and H. Davoodi (1999) Fiscal Decentralisation and
Economic Growth in the United States. Journal of Urban Economics 45,
228-239.
Yilmaz (1999) The Impact of Fiscal Decentralisation on
Macroeconomic Performance, National Tax Association. Proceedings of the
92nd Annual Conference on Taxation, US. 251-260.
Zhang, T. and H. Zou (1998) Fiscal Decentralisation, Public
Spending and Economic Growth in China. Journal of Public Economics 67:2,
221-240.
(1) According to Giugale and Webb (2000) efficiency means
satisfying the needs and preferences of taxpayers at the lowest possible
cost.
(2) The idea of soft budget constraint was introduced by Kornai
(1979) to analyse the behaviour of state owned firms. The SBC is used in
a decentralisation system to refer to lower level governments that look
to a higher level government to recover or bailout their excessive
deficits. The term bailout refers to the additional funding that the
higher level government provides the lower level governments when it
would otherwise be unable to service its obligations. On the other hand,
hard budget constraint (HBC) implies that lower level governments have
to face the full costs of their expenditure decisions.
(3) Sce for example [Oates (1995): Zhang and Zou (1998); Xie, el
al. (1998); Yilmaz (1999); Lin and Liu (2000); Thiessen (2003); Akai and
Sakata (2002); Eller (2004); Iirni (2005); Feltensleina and Iwata
(2005); Cantarero and Gonzalez (2009); Neyapli (2010)].
(4) Due to unavailability of fiscal data at local level, this
analysis only focuses at aggregate level using time series data. This
analysis also ignores the other dimension of decentralisation namely
administrative and political dimensions of the decentralisation because
of the same reason.
(5) In terms of the negative association of expenditure
decentralisation with economic growth, our findings are in line with the
findings of other empirical studies such as Davoodi and Zou (1998),
Zhang and Zou (2001), Rodriguez-Pose and Kroijer (2009) and Nguygen and
Anwar (2011).
(6) See North (1981) for further elaboration on the role of
institutions in economic growth.
Nasir Iqbal<nasir84@pide.org.pk> is Staff Economist, Pakistan
Institute of Development Economics (PIDE), Islamabad. Musleh ud Din
<muslehuddin@pide.org.pk> is Joint Director, Pakistan Institute of
Development Economics (PIDE), Islamabad. Ejaz Ghani
<ejazg@yahoo.com> is Dean Faculty of Economics, Pakistan Institute
of Development Economics (PIDE), Islamabad.
Authors' Note: This paper is heavily drawn from PhD Thesis
entitled "Fiscal Decentralisation, Macroeconomic Stability and
Economic Growth". The authors are grateful for the comments
received from the seminar participants at PIDE.
Table 1
National vs. Sub-National Revenue and Expenditure Shares:
International Comparison
Country Revenue Share Expenditure Share
National Sub-National National Sub-National
Australia 69 31 54 46
Brazil 69 31 54 46
Canada 44 56 37 63
India 66 34 45 55
South Korea 95 05 50 50
Pakistan 92 08 72 28
Source: Watt (2005).
Table 2
Sharing Criterion in Various NFC Awards
Award Sharing Criteria (Weight)
NFC 1990 Population (100%)
NFC 1996 Population (100%)
NFC 2006 Population (100%)
NFC 2009 Population (82%), Poverty (10.3%), Revenue (5%),
IPD (2.7%)
Table 3
The Share of Each Province in the Divisible Pool (Percent)
Province NFC-1990 NFC-1996 NFC-2006 NFC-2009
Punjab 57.87 57.37 57.37 51.74
(57.87) (57.87) (57.36) (57.36)
Sindh 23.29 23.29 23.71 24.55
(23.29) (23.29) (23.71) (23.71)
KPK 13.54 13.54 13.82 14.62
(13.54) (13.54) (13.82) (13.82)
Balochistan 5.30 5.30 5.11 9.09
(5.30) (5.30) (5.11) (5.11)
TOTAL 100.0 100.0 100.0 100.0
Note: Population shares are reported in parenthesis based on Census
conducted before the NFC Award.
Table 4
Descriptive Statistics
Variables Obs. Mean Std. Dev
Revenue Decentralisation (RD) 39 0.130 0.041
Expenditure Decentralisation (ED) 39 0.465 0.067
Composite Decentralisation (CD) 39 0.247 0.089
Inflation (INF) 39 9.587 5.748
Budget Deficit (BD) 39 6.464 1.805
GDP per Capita (Constant 2000 US$) 39 451.7 113.3
GDP per Capita Growth Rale 39 2.234 2.002
Human Capital (HC) 39 20.02 7.111
Openness (OPN) 39 0.338 0.037
Tax to GDP Ratio (T/GDP) 39 0.123 0.015
Democratic Institution (INS) 39 0.846 6.745
Variables Min Max
Revenue Decentralisation (RD) 0.071 0.221
Expenditure Decentralisation (ED) 0.336 0.686
Composite Decentralisation (CD) 0.129 0.494
Inflation (INF) 03.10 30.00
Budget Deficit (BD) 02.30 10.20
GDP per Capita (Constant 2000 US$) 279.1 668.6
GDP per Capita Growth Rale -1.950 6.570
Human Capital (HC) 10.54 34.60
Openness (OPN) 0.273 0.432
Tax to GDP Ratio (T/GDP) 0.095 0.145
Democratic Institution (INS) -7.000 8.000
Table 5
Unit Root Test (ADF Test)
Level
No With Result
Variables Trend Trend
Revenue Decentralisation (RD) -2.13 -3.24 NS
Expenditure Decentralisation (ED) -1.72 -2.48 NS
Composite Decentralisation (CD) -1.69 -3.41 NS
Inflation (INF) -4.02 -3.62 S
Budget Deficit (BD) -2.95 -3.77 S
GDP per Capita Growth Rate -5.72 -5.63 S
Human Capital (HC) 1.29 -2.26 NS
Openness (OPN) -2.93 -3.56 S
Tax to GDP Ratio (T/GDP) -1.32 -2.02 NS
Democratic Institution (INS) -1.97 -1.91 NS
First Difference
No With Result
Variables Trend Trend
Revenue Decentralisation (RD) -4.63 -4.56 S
Expenditure Decentralisation (ED) -7.19 -7.02 S
Composite Decentralisation (CD) -5.49 -5.43 S
Inflation (INF)
Budget Deficit (BD)
GDP per Capita Growth Rate
Human Capital (HC) -4.19 -5.23 S
Openness (OPN)
Tax to GDP Ratio (T/GDP) -5.12 -5.71 S
Democratic Institution (INS) -5.71 -5.76 S
Note: 5 percent critical value is -2.87 for the case of no-trend, and
-3.42 when a trend is included. AIC is used for lag selection. S
stands for stationary series and NS stands for non-stationary series.
Table 6
The GMM Estimates: Dependent Variable (GDP per Capita Growth)
Variables (1) (2) (3)
RD 0.0206 * 0.0455 *** 0.0461 ***
(0.0120) (0.0167) (0.0176)
OPN 0.0414 ** 0.0705 **
(0.0204) (0.0327)
T/GDP 0.0475 * 0.0592 *
(0.0274) (0.0312)
HC 0.0505 *** 0.0515 ***
(0.0159) (0.0190)
INF -0.00966 *
(0.00529)
BD
Constant 0.0658 ** 0.113 * 0.112 *
(0.0263) (0.0642) (0.0640)
Observations 37 37 37
R-squared 0.247 0.409 0.408
Wald Chi2 Test 3.92 10.31 11.67
Normality Test 0.97(0.61) 0.70(0.71) 0.71(0.70)
Endogeneity Test 0.0685 0.0885 0.0711
Over Identification test 0.7070 0.9423 0.9638
D. W. Test 1.89 2.42 2.43
Variables (4) (5)
RD 0.0487 *** 0.0530 ***
(0.0160) (0.0173)
OPN 0.0625 * 0.0245
(0.0337) (0.0317)
T/GDP 0.0675 ** 0.0808 **
(0.0276) (0.0348)
HC 0.0381 ** 0.0426 **
(0.0157) (0.0185)
INF -0.00687 *
(0.00399)
BD -0.0292 *** -0.0337 ***
(0.00852) (0.00939)
Constant 0.243 *** 0.251 ***
(0.0690) (0.0698)
Observations 37 37
R-squared 0.532 0.546
Wald Chi2 Test 31.41 36.38
Normality Test 0.77(0.68) 0.88(0.64)
Endogeneity Test 0.0625 0.0305
Over Identification test 0.5625 0.6446
D. W. Test 2.59 2.71
Robust standard errors in parentheses.
*** p<0.01, ** p<0.05, * p<0.1.
Table 7
The GMM Estimates: Dependent Variable (GDP per Capita Growth)
Variables (1) (2) (3)
ED -0.0922 ** -0.116 *** -0.129 ***
(0.0400) (0.0392) (0.0317)
OPN 0.0385 * 0.0274 *
(0.0215) (0.0162)
T/GDP 0.0371 * 0.0387 *
(0.0196) (0.0201)
HC 0.0241 * 0.0183 *
(0.0128) (0.0103)
INF -0.00980 *
(0.00577)
BD
Constant -0.0509 * 0.0289 * 0.0547
(0.0300) (0.0171) (0.0581)
Observations 37 37 37
R-squared 0.207 0.421 0.493
Wald Chi2 Test 5.32 11.54 27.28
Normality Test 0.31(0.85) 0.67(0.72) 0.37(0.70)
Endogeneity Test 0.0395 0.0154 0.0265
Over Identification test 0.6341 0.6149 0.5225
D.W Test 2.29 2.52 2.54
Variables (4) (5)
ED -0.115 *** -0.122 ***
(0.0341) (0.0338)
OPN 0.0251 * 0.0238 *
(0.0135) (0.0127)
T/GDP 0.0497 * 0.0498 *
(0.0285) (0.0291)
HC 0.0266 * 0.0279 *
(0.0144) (0.0149)
INF -0.00598 *
(0.00332)
BD -0.0368 *** -0.0346 ***
(0.0118) (0.0130)
Constant 0.190 *** 0.194 ***
(0.0552) (0.0555)
Observations 37 37
R-squared 0.451 0.537
Wald Chi2 Test 19.73 25.22
Normality Test 0.24(0.88) 0.16(0.92)
Endogeneity Test 0.0495 0.0028
Over Identification test 0.7243 0.7903
D.W Test 2.68 2.65
Robust standard errors in parentheses.
*** p<0.01, ** p<0.05, * p<0.1.
Table 8
The GMM Estimates: Dependent Variable (GDP per Capita Growth)
Variables (1) (2) (3)
CD 0.0190 * 0.0444 *** 0.0452 **
(0.0113) (0.0166) (0.0176)
OPN 0.0392 * 0.0382 *
(0.0218) (0.0222)
T/GDP 0.0494 * 0.0514
(0.0273) (0.0316)
HC 0.0519 *** 0.0532 ***
(0.0162) (0.0193)
INF -0.0108 **
(0.00517)
BD
Constant 0.0570 *** 0.0953 0.0940
(0.0215) (0.0610) (0.0611)
Observations 37 37 37
R-squared 0.248 0.420 0.419
Wald Chi2 Test 2.85 10.48 11.69
J.B. Normality Test 0.91(0.63) 0.69(0.71) 0.69(0.71)
Endogeneity Test 0.0462 0.0733 0.0613
Over Identification test 0.7536 0.8955 0.9176
Durban Watson Test 1.88 2.39 2.40
Variables (4) (5)
CD 0.0478 *** 0.0528 ***
(0.0159) (0.0171)
OPN 0.0285 * 0.0207
(0.0158) (0.0317)
T/GDP 0.0692 ** 0.0837 **
(0.0276) (0.0344)
HC 0.0403 ** 0.0455 **
(0.0157) (0.0185)
INF -0.00713 **
(0.00379)
BD -0.0283 *** -0.0330 ***
(0.00821) (0.00890)
Constant 0.218 *** 0.227 ***
(0.0665) (0.0661)
Observations 37 37
R-squared 0.538 0.553
Wald Chi2 Test 33.15 39.10
J.B. Normality Test 0.72(0.69) 0.81(0.66)
Endogeneity Test 0.0548 0.0767
Over Identification test 0.5239 0.5983
Durban Watson Test 2.55 2.68
Robust standard errors in parentheses.
*** p<0.01, ** p<0.05, * p<0.1.
Table 9
The GMM Estimates: Dependent Variable (GDP per Capita Growth)
(1) (2)
RD 0.00426
(0.0117)
ED -0.117 ***
(0.0305)
INS 0.00117 ** 0.00150 *
(0.000491) (0.000836)
RD*INS 0.0132 ***
(0.00330)
ED*INS 0.0449 ***
(0.0129)
OPN
T/GDP
HC
Constant 0.0450 * -0.0546 **
(0.0231) (0.0245)
Observations 37 37
R-squared 0.250 0.240
Wald Chi2 Test 29.18 33.54
J.B. Normality Test 1.02(0.60) 0.17(0.91)
Endogeneity Test 0.0376 0.0064
Over Identification test 0.6695 0.8442
Durban Watson Test 1.93 2.24
(3) (4)
RD 0.0271
(0.0194)
ED -0.151 ***
(0.0387)
INS 0.000813 * 0.00162 *
(0.000492) (0.000894)
RD*INS 0.00914 **
(0.00412)
ED*INS 0.0446 ***
(0.0156)
OPN 0.0463 0.00304
(0.0369) (0.0506)
T/GDP 0.0409 * 0.0205
(0.0246) (0.0268)
HC 0.0597 ** -0.0150
(0.0164) (0.0138)
Constant 0.108 * 0.00899
(0.0614) (0.0606)
Observations 37 37
R-squared 0.318 0.224
Wald Chi2 Test 51.22 29.00
J.B. Normality Test 0.45(0.80) 0.23(0.89)
Endogeneity Test 0.0144 0.0012
Over Identification test 0.6302 0.5745
Durban Watson Test 2.36 2.29
Robust standard errors in parentheses.
** p<0.01, ** p<0.05, * p<0.1.
