Analysis of revenue potential and revenue effort in developing Asian countries.
Javid, Attiya Y. ; Arif, Umaima
I. INTRODUCTION
Countries around the world are increasingly recognising that the
effective revenue system is the most important factor for economic
development. Factors effecting revenue potential measured as the revenue
to GDP has been one of the most important issues that concerns to
policy-makers from last three decades. Many developing countries face
difficulties in generating sufficient revenues for public expenditure.
In some countries budget deficits and the unproductive use of public
expenditures have narrow the vital investments in both human resources
and basic infrastructure that are necessary for providing base for
sustainable economic growth and development. Too much dependence on
foreign financing may cause problems of debt sustainability; therefore
developing countries will need Jo depend substantially on domestic
revenue generation.
There is a large body of literature on tax revenue potential in
developing countries [Bahl (1971); Tanzi (1987); Leuthold (1991); and
Stotsky and Mariam (1997); Gupta (2007)]. However, there is few studies
that examine institutional and governance quality as a factor
influencing tax collection and tax revenue potential. According to Tanzi
and Davoodi (1997) and Gupta (2007) these factors are responsible for
low tax collection in developing countries by allowing citizens
inappropriate tax exemptions and enabling tax evasion due to bad tax
administration. Therefore, a precondition for ensuring adequate revenue
collection is a legitimate and responsive institutions following the
rule of law with control on corruption and having high quality
bureaucracy to administer. Studies also confirm that a strong political
will to reform is required for successful reform process [Bird (2004)].
Aim, et al. (2003) suggest that tax records of countries are reflection
of their political or societal institutions.
The present study analyses the idea of taxable potential and tax
effort by extending to measure (fiscal) revenue potential and (fiscal)
revenue effort. Total fiscal revenue is sum of both tax and non-tax
revenue collection consisting of cash receipts from taxes, social
contributions, and non-tax sources such as fines, fees, rent, and income
from property or sales.
The main aim of the present study is to empirically investigate the
sources of resource mobilisation for developing Asian countries for the
period 1984 to 2010. The sample of the countries include: Pakistan,
Bangladesh, India and Sri Lanka, Indonesia, Malaysia, Thailand, China,
Philippines, Singapore, China, Singapore and Vietnam as these countries
have common characteristics of large and persistent as well as instable
budget deficit. More specifically, the study look at the main
determinants of revenues of the central government, and analyse the
extent to which factors such as the structure of the economy,
macroeconomic policy and institutions and the level of development
explain their variation. The study assesses the revenue effort of the
sample countries that is defined as index of the ratio of the actual
revenue collection to GDP and the predicted revenue capacity.
The resources available for fiscal policy are inadequate for South
Asian countries in particular and developing countries in general and
this will make difficult to meet all public expenditures and government
can focus on specific expenditures due to political pressure [Jha
(2009)]. India has shown an upward trend in revenue to expenditure ratio
overtime whereas Pakistan, Bangladesh and Sri Lanka have recorded a
decline in this ratio. So public deficit in South Asian countries
remains high for Pakistan and Sri Lanka and countries face considerable
resource constraints on financing of the deficit that result from their
expenditure in excess of revenues. India has shown good revenue
performance among South Asian countries but has shown no progress in its
performance between 2005 and 2008. Bangladesh's score enhanced
after 2006 but remained still thereafter. The most disappointing
performance has been by Pakistan among the South Asian countries,
however Sri Lanka's performance was comparable to India's in
2005 and 2006 but then has again worsen. China has registered an
increase in their revenue to GDP ratio from 5 percent in 1990 to 11
percent in 2011 whereas Singapore, Malaysia, Indonesia, Philippines and
Thailand show a decline in this ratio during this period. Although China
shows a rising trend compared to other countries but even then there is
no significant difference in revenue to GDP ratio of China and the rest
of the countries in ASEAN region. In General falling tax/GDP ratios in
these countries leads to structurally unshakable fiscal deficits and
necessitates investigating the main factors that may explain the
variation in resource mobilisation of developing Asian countries.
Furthermore, quality of institution that creates economic stability and
a move towards democratic regimes is also essential for the increasing
the revenue collection capacity developing Asian region.
This paper undertakes panel data analysis to estimate revenue
potential for a sample of developing Asian countries during 1984-2010
following the empirical methodology suggested by Bird, Vazquez, and
Torgler (2004) and Gupta (2007). The estimation results are used as
benchmarks to compare revenue potential and revenue effort in Asian
countries. Revenue potential is defined as the estimated revenue to GDP
with the regression, considering a country's specific
macroeconomic, demographic, and institutional features. Revenue effort
is an index of the actual revenue GDP and the predicted revenue
potential.
The study adds to existing empirical literature by comparing fiscal
capacity and fiscal effort among the developing countries of Asian
region over longer period of time from 1984 to 2010 and for almost three
decades separately: 1984 to 1990, 1991 to 2000 and 2001 to 2010. Second
besides the traditional supply side determinants like GDP per capita,
international trade, agricultural value added debt as a fraction of GDP
the impact of quality of institution and policy variable on a
country's revenue capacity are analysed. The corruption index, the
law and order and bureaucratic quality scores are used for this purpose.
The indexes are obtained from the International Country Risk Guide
(ICRG).
The study is organised as follows. Section 2 discusses the
theoretical and empirical literature in this area. Methodological
framework, data/sample and estimation technique are presented in Section
3. The empirical results of regression analysis to estimate fiscal
potential and index of fiscal effort analysis is presented in Section 4
and the last section concludes the study.
2. LITERATURE REVIEW
Regression Analysis focused on possible determinants of taxes are
used in the literature to estimate taxable capacity and the tax effort
of countries. Taxable capacity is defined as predicted tax-to-GDP ratio
calculated by the estimated coefficients of a regression specification
that takes into account the country specific characteristics [Gupta
(2007); Bird, et al. (2007), Le, Moreno- Dodson, and Rojchaichaninthorn
(2008)]. Tax effort is defined as index of the ratio of the share of the
actual collection to GDP and the predicted taxable capacity. A high tax
effort points to a situation when a tax effort index is above 1,
entailing that the country optimally uses its tax base to augment tax
revenues [Stotsky, et al. (1997)]. Likewise, a low tax effort means that
tax effort index is below 1, implying that the country may have
potential to increase tax revenues.
Several studies show that variables such as per capita GDP, the
sector wise composition of output, the degree of trade and financial
openness, the ratio of foreign aid to GDP, the ratio of overall debt to
GDP, a measure for the informal economy, and institutional factors such
as the degree of political stability and corruption plays an important
role in determining revenue performance of any economy [Gupta (2007);
Bird, et al. (2008) and Le, et al. (2008)]. Lotz and Morss (1967) find
that per capita income and trade share are important determinants of the
tax share. Chelliah (1971) relates the tax share to explanatory
variables such as mining share; non-mineral export ratio and agriculture
share and obtain similar results. In a related study covering developing
countries, Tanzi (1992) finds that half of the variation in the tax
ratio is explained by per capita income, import share, agriculture share
and foreign debt share.
The effect of trade liberalisation is considered as important
determinant that occurs primarily through reduction in tariffs, then one
expects losses in tariff revenue, however revenue may increase provided
trade liberalisation occurs through tariffication of quotas,
eliminations of exemptions, reduction in tariff peaks and improvement in
customs procedure [Keen and Simone (2004)]. Several studies find that
there is a positive relationship between trade openness and the size of
the government [Gupta (2007); Bird, et al. (2007) and Le, et al.
(2008)]. Rodrik (1998) also conclude that as societies seem to demand
(and receive) an expanded role for the government in providing social
insurance in more open economies are subject to external risks.
The degree of external indebtedness of a country is also examined
as factor that affects revenue performance of an economy [Gupta (2007)].
For generating necessary foreign exchange to service the debt, a country
may choose to reduce imports that lead to lower import tax otherwise the
country may choose to increase import tariffs or other taxes to generate
a primary budget surplus for debt servicing. The composition of aid has
an important effect on revenue performance, for example, concessional
loans are associated with higher domestic revenue mobilisation, while
grants have the opposite affect [Gupta, et al. (2004)].
Recently, some studies have explored the importance of
institutional factors in determining revenue performance. For example,
Bird, Martinez-Vasquez, and Torgler (2004) find that factors such as
corruption, rule of law, entry regulations play key roles. Several
regional studies have looked into quality of institution and governance
as determinants of resource mobilisation. Leuthold (1991) uses panel
data to find a positive impact from trade share and Stotsky and Mariam
(1997) find that both agriculture and mining share are negatively
related to the tax ratio, while export share and per capita income have
a positive effect.
Ghura (1998) concludes that the tax ratio rises with income and
degree of openness, and with the share of agriculture in GDP. He also
finds that other factors like corruption, structural reforms and human
capital development affect the tax ratio. Most studies find that per
capita GDP and degree of openness is positively related to revenue
performance, but a higher agriculture share lowers it. Studies such as
Tanzi (1992) and Eltony (2002) found that foreign debt is positively
related to resource mobilisation.
The present study provides comparison of fiscal capacity and fiscal
effort among the developing countries of Asian region. This study checks
the robustness of quality of institutions and macroeconomic policy
variables in determining the fiscal performance of countries in this
region over a long period of time 1984 to 2010 that is further divided
into three sub samples 1984 to 1990, 1991 to 2000 and 2001 to 2010.
3. METHODOLOGY AND DATA
The present study analyses revenue performance by estimating
revenue potential and calculating revenue effort index for developing
Asian countries over the period of 1984 to 2010. (1) The empirical
methodology applied by Bird, Vazquez, and Torgler (2004) and Gupta
(2007) is adopted to examine the potential revenue capacity of
developing Asian countries. Revenue (fiscal) potential is the predicted
revenue to GDP ratio estimated from the regression based on the country
specific characteristics and revenue (fiscal) effort is ratio between
the actual collection to GDP and predicted revenue capacity [Bird, et
al. (2004) and Le, et al. (2008)]. (2) The empirical specification of
the model that measures the revenue potential by estimating the
determinants of revenue is express as:
Revenue/GDP = F (Economic, Demographic, Institutional, Policy)
More specifically the basic specification of the model takes the
following form: (3)
Revnue / [GDP.sub.it] = [[alpha].sub.i] + [[alpha].sub.l]
[GDPC.sub.it] + [[alpha].sub.2] [Trade.sub.it] + [[alpha].sub.3]
[Debt.sub.it] + [beta][Popg.sub.it] + [gamma][Ins.sub.it] +
[delta][Inf.sub.it] + [[epsilon].sub.it] (1)
Revnue / [GDP.sub.it] = [alpha] + [[alpha].sub.l] [GDPC.sub.it] +
[[alpha].sub.2] [Trade.sub.it] + [[alpha].sub.3] [Debt.sub.it] +
[beta][Popg.sub.it] + [gamma][Ins.sub.it] + [delta][Inf.sub.it] +
[v.sub.i] + [[epsilon].sub.it] (2)
Where revenue to GDP ratio for the country i for the period t which
is function of economic variables, demographic, institutional/governance
quality and policy variables, The vector of economic variables measures
the structural characteristics of countries and it includes GDP per
capita, trade to GDP, external debt to GDP in the basic specification.
The share of agriculture to GDP and share of manufacture to GDP (4) are
also examined as the determinants of revenue potential of the Asian
countries. The population growth is taken as demographic variables. The
vector of institution includes the variables that capture institutions
and quality of governance such as control of corruption, high
bureaucracy quality and law and order scores. (5) The inflation rate is
used as macroeconomic policy variable which effects the investment and
income level of the country.
Income level, measured as GDP per capita, is used as a proxy for
the level of a country's development, and it is expected to be
positively related with the government's ability to collect
revenues and the citizens' ability to pay revenue. Thus, it is
expected that GDP per capita to have a positive and significant impact
on fiscal revenue [Bahl (1971); Fox, et al. (2005); Piancastelli (2001);
Gupta (2007); Bird, et al. (2004) and Le, et al. (2008)]. Trade tax
revenue being a major source of tax revenue in developing countries
[Rodrik (1998); Piancastelli (2001); Norregaard and Khan (2007); Gupta
(2007); Aizenman and JinJarak (2009)] lowers the overall tax-to-GDP
ratio in post trade liberalisation era under the Uruguay Round of World
Trade Organisation. The effect of trade liberalisation may be ambiguous
due to two opposite effects on taxes. On the one hand, it may have a
negative impact on taxes and fiscal revenue as higher trade openness is
expected to lower taxes collected on imports and export. On the other
hand, given that higher trade openness leads to higher economic growth
rates, open economies grow faster; and as a result, more taxes can be
collected with increasing this tax base. It is expected that the second
effect outweigh in case of Asian countries and trade openness has a
positive impact on taxes and total fiscal revenue. Further, Gupta (2007)
documents that if this liberalisation is undertaken through reduction in
tariffs then it is expected that tariff revenue will be reduced. On the
other hand, Keen and Simone (2004) argue revenue may increase if trade
liberalisation takes place through tariffication of quotas, eliminations
of exemptions, reduction in tariff peaks and improvement in customs
procedure. Rodrik (1998) also comes to conclusion that there is a
positive association between trade openness and the government
consumption, as people demand (and receive) increasing amount of public
goods in more open economies subject to external risks.
The revenue potential is effected by the debt of a country as to
generate the necessary foreign exchange to service the debt, a country
may choose to reduce imports and import taxes will be lower.
Alternatively, the policy may be to increase import tariffs or other
taxes in order to register budget surplus to service the debt [Gupta
(2007)]. Therefore, it is expected that level of indebtedness of the
country is positively associated with revenue potential of the country.
The recent empirical literature finds non-traditional variables
like institutional and governance quality as important determinants of
revenue potential for developing countries. The institutional and
governance factors impact revenue collection potential by influencing
tax evasion, inappropriate revenue exemptions, and weak revenue
collection administration [Tanzi and Davoodi (1997)]. Bird, et al.
(2004) argue that any successful tax reform should be rooted in a strong
political will to reform, and Aim and Martinez-Vazquez (2004) document
that a country's tax record is reflection of its political or
societal institutions. Bird, Martinez-Vazquez, and Torgler (2004)
conclude that rule of law and control of corruption is necessary
prerequisite for a more satisfactory revenue effort. For example poor
law and order conditions in the economy induce people to avoid the tax
and non-tax payments. If corruption is high in an economy, large part of
business community would prefer to work underground by paying bribes to
avoiding high revenue payments. If societies have feelings that their
interests are well represented at government level and they are
satisfied with quality and quantity of public goods like health,
education etc., there would be willingness to pay revenues. To evaluate
the impact of these institutional variables on revenue performance three
governance indicators computed by International Country Risk Data Guide
are included; corruption index, bureaucracy quality and law and order
scores. It is expected that control of corruption, high quality of
bureaucracy and strong law and order enforcement are positively,
associated with the revenue potential of developing Asian countries.
The inflation is policy variable that is included to measure the
quality of a country's macroeconomic policies. It allows capturing
direct effect of inflation on revenue collection through its impact on
consumption and investment, and subsequently on their related tax
categories. It is expected that inflation has negative effect on revenue
collection capacity.
3.1. Data and Sample
The study used annual data on economic, political and institutional
variables, from 1984 to 2010. The source of economic data is
International Financial Statistics and World Development Indicators.
Institutional variables are obtained from International Country Risk
Data Guide (ICRG). Economic variables revealing structural
distinctiveness of the countries include real GDP per capita,
agriculture value addition to GDP, trade to GDP, debt to GDP. The GDP
per capita is expected to have positive impact on revenue collection
capacity of a country with level of income and citizens also demand more
public goods and services. On the other hand large agriculture sector is
difficult to tax because of large share of subsistence and politically
infeasibility, and reduction in need of public goods and services which
are urban based. It is relative easier to tax foreign trade compared to
domestic activities as goods enter and leave the country at specific
places. Therefore, it is expected that trade openness has a positive
impact on revenue collection. Inflation is measured as percentage change
in consumer price index and it is expected that inflation has negative
impact on revenue collection capacity of the country.
The demographic variables include population growth and as the rate
of population growth increases, the revenue collection system finds
difficult to capture new revenue payers especially when revenue
collection administration capacity is weak. Therefore, the population
growth rate is expected to be negatively related to the revenue
potential of a country. Inflation measures the quality of a
country's macroeconomic policies. The quality of fiscal and
monetary policies in terms of revenue is measured by Inflation rate as
high level of inflation would reduce the revenue to GDP ratio due to
negatively effecting consumption and investing capacity and thus
decreasing tax revenue generated from these categories.
The quality of institutions captures various aspects of the
governance of the public sector, such as control of corruption, rule of
law; high bureaucracy quality and these factors are expected to be
positively associated with revenue collection capacity of a country. A
higher value of institutional indicates a higher quality of
institutions. The corruption index measures the extent of corruption by
assigning a numerical value to a country. The index ranges from 1 to 6,
where a higher number means lower corruption. Similarly the law and
order index also ranges from 1 to 6. The bureaucracy quality index is an
alternative institutional indicator of governance and it ranges from 1
to 4. Following Tanzi and Davoodi (1997) in this analysis institutional
variables are used after rescaling the original 1CRG corruption index,
law and order index and bureaucracy quality indicator to a range of -6
(least corrupt or best bureaucratic quality and best law and order
condition?) and -1 (most corrupt or worst bureaucratic quality and law
and order conditions).
3.2. Estimation Technique
The panel data estimation techniques fixed effect and random effect
models and dynamic panel data model are used. The econometric issues
related with these techniques are the presence of country specific fixed
effects and endogeneity. To deal with these issues, Arellono and Bond
(1991) introduced the Generalised Method of Moments after first
differencing the equation. Latter, Blundell and Bond (1998) suggest
efficiency can be increased by adding the original equation in the level
to the system, if the first difference of the explanatory variables is
uncorrelated with original effects. Lagged dependent and exogenous
variables can be used as instrument variables. Multicollinearity is
another problem which arises when two or more explanatory variables
appeared to be highly correlated with each other and to resolve this
problem the highly correlated explanatory variables are used in separate
specifications.
4. EMPIRICAL RESULTS
The analysis begins with basic specification of the revenue model 1
and determinants include the log of per capita GDP, trade to GDP, debt
in GDP, population growth and control of corruption and inflation.
Generalised method of Moments of Blundell and Bond (1998) is used as
estimation technique that allows to deal with country specific effects
and any edogeneity that may be due to the correlation of the country
specific effects and dependent variable. The result of Hausman test
indicates that fixed effects specifications best describes the data in
almost all specifications. Latter in model 2 and 3 bureaucracy quality
and law and order score are included one by one. Then GDP per capita is
replaced by agriculture value added to GDP in model 4, 5 and 6. The
results of fixed effect models 1 to 6 are presented in Table 1.
The per capita GDP has significantly positive impact in basic
specification of revenue potential model 1 suggesting that the capacity
to collect and pay revenue increases with the level of development of
sample countries. This result is consistent with earlier studies
[Chelliah (1971); Bahl (1971); Fox, et al. (2005); Gupta (2007)]. The
trade openness is positive and significant determinant of revenue to GDP
because trade-related taxes are easier to impose [Gupta (2007)]. The
result indicates that debt has a positive effect on revenue potential;
Gupta (2007) finds that debt is negatively related with revenue
performance. The population growth rate is negatively related to the
revenue potential. Bird, et al. (2004) also finds the inverse
relationship between population growth and resource mobilisation
suggesting that as the rate of population growth increases, the tax
system may lag behind in its ability to capture new taxpayers. Inflation
has negative and significant impact on the revenue capacity of the
sample countries. The negative relationship of inflation confirms
inflation detrimental impact on revenue collection potential of
countries and is consistent with the results reported by Agbeyegbe,
Stotsky, and Mariam (2004). The high inflation rate reduce the
purchasing power and investing ability of consumers and therefore
negatively impact the revenue collection from these heads.
The impact of institutional quality on revenue collection is
positive and significant as expected as they are added one by one in
model 1, 2 and 3 (Table 1). This is consistent with the findings of
Tanzi and Davoodi (2000) and Bird, Martinez-Vazquez, and Torgler (2004)
and Gupta (2007). These results support that quality of institution and
governance increase the revenue capacity and this is a direct channel
for the impact of institutions on revenue collection. There is indirect
impact that institutions have through shadow economic activity.
In model 4, 5 and 6 (Table 2) agriculture value added to GDP is
replaced by GDP per capita. The agriculture to GDP has negative and
significant relationship with revenue potential of sample countries. The
presence of large agriculture sector is considered administratively and
politically difficult taxing agriculture and government rather wants to
either provide tax exemptions or subsidies. This also reduces the demand
for government services, since many public sector activities are urban
based [Tanzi (1992); Gupta (2007)]. Most of the variables have expected
relationship with revenue potential of Asian countries; however the
effect of inflation and trade openness turns out to be insignificant.
The trade openness has less role in revenue generation in countries
which are more agriculture may be due to fact that have a negative
impact on taxes and fiscal revenue as higher trade openness is expected
to lower taxes collected on imports and export that offsets the positive
effect because of fact that higher trade openness leads to higher
economic growth rates [Combes and Saadi-Sadeq (2006)]. (6)
Table 2 reports the results from the dynamic panel models. The
results of models 7, 8 and 9 confirm that lagged revenue to GDP is a
strong and significant predictor of current revenue potential. Gupta
(2007) also finds random walk type of result in cross country analysis
in this regard. The result indicates that per capita GDP, debt to GDP
are significant predictor of revenue potential. However, the impact of
per capita GDP is substantially smaller in the dynamic specification.
The impact of both agriculture value added to GDP and population growth
are negative as expected in model 10, 11 and 12 but impacts are
marginally smaller in the dynamic specification. The indebtness of
country has positive significant effect on revenue potential in all six
models. The trade openness and inflation are no more significant
determinants of revenue potential of sample countries.
This above panel regression provide a simple empirical analysis of
the predicted values of the revenue to GDP obtained through Equation 1
that measure the revenue potential of Asian countries. The ratio of the
actual to predicted revenue is calculated to measure the level of
revenue effort of sample Asian countries [Bird, et al. (2004) and Gupta
(2007)].
Tax Effort Analysis
The above analysis has focused on finding the main factors that
affect revenue potential in a sample of developing Asian countries.
However, this does not tell whether a country could not, if it wanted,
attain higher revenue potential [Chelliah (1971); Chelliah, et al.
(1975) and Gupta (2007)].
Different countries have different potential to raise revenues that
must be taken into consideration while making cross-country revenue
comparisons [Gupta (2007), Bird, et al. (2004) and Le, et al. (2008)].
The selection of regression results to estimate the predicted values of
revenue ratios are made on the base of their significance and economic
rationale in this analysis [Teera and Hudson (2004)]. Several studies
have followed the same approach to measure revenue effort across
countries [Gupta (2007); Bird, et al. (2004)]. The predicted values of
the revenue ratio is obtained through model 1 and 4, thus measure the
country's revenue potential, while the ratio of the actual to
predicted revenue is calculated for the level of revenue effort. Thus, a
country that lies on the regression line have a revenue effort index
equal to 1, and countries that have actual revenue effort above
predicted revenue performance have a revenue effort index higher than
one, in reverse case revenue effort index is less than 1. The results of
revenue effort are presented in Table 3 for sample countries Malaysia,
Indonesia, Thailand, Philippines and Singapore have exhibited
significant revenue performance compared to other countries, having
revenue effort index greater than 1. These countries have probably
largely used their revenue potential. On the other hand, countries like
Pakistan, Bangladesh and Sri Lanka have revenue effort indices well
below 1 which suggests that they have yet to achieve their full revenue
potential, as they are constrained by low per capita GDP, a dominant
agriculture sector.
5. CONCLUSION
The development of revenue effort index that relates the actual
revenues of a country to its estimated revenue capacity provide an
appealing measure that considers country specific fiscal, demographic,
and institutional characteristics. This study analyses revenue
performance across developing Asian countries over the period 1984 to
2010 and also for the sub periods 1984 to 1990, 1991 to 2000 and 2001 to
2010. The results indicate that per capita GDP, share of agriculture in
GDP and foreign debt are statistically significant and strong
determinants of revenue performance in almost all specifications of the
model. The trade openness and inflation are also having impact on
revenue performance in some specifications. Among the institutional
factors, control of corruption and high bureaucracy quality and improved
law and order conditions have a significantly positive effect on revenue
performance in all model specifications. The results confirm that
countries that depend on agriculture value addition tend to have poorer
revenue performance. The analysis highlights that revenue performance
depends on level of development of country, its institutional and
governance quality and to macroeconomic policy and political will for
reforms. This analysis can be considered complimentary providing a
broader picture of revenue performance but detailed analysis of a
country's revenue system that takes account of the country's
overall fiscal policy, public expenditures needs and the overall level
of development in Asian region is needed for future research. The
results imply that architect of revenue reforms must be country specific
that requires broad investigation of the country's revenue
capacity, revenue performance, and institutional structure.
APPENDIX
Table A2
Determinants of Revenue Potential in Developing Asian Countries
1984-1990
Mod 1 Mod 2 Mod 3
Constant 0.21 * 0.22 * 0.23 *
(6.39) (7.2) (6.2)
GDP per Capita 0.02 * 0.02 * 0.02 *
(3.8) (5.02) (4.95)
Trade/GDP 0.02 * -0.02 * 0.03 *
(2.88) (2.86) (2.02)
Debt/GDP 0.18 0.19 0.188
(4.44) (4.8) (4.23)
Population Growth -0.07 -0.01 * 0.005 *
(-2.01) (-2.00) (3.15)
Inflation -0.002 -0.001 -0.001
(-1.46) (-1.11) (-0.87)
Control of Corruption 0.012 *
(3.25)
High Quality Bureaucracy 0.013 *
(3.87)
Law and Order 0.008
Sargan Test (p value) (0.25) (0.18) (0.21)
Hausman Test (p value) (0.43) (0.39) (0.28)
[R.sup.2] 0.73 0.72 0.66
1991-2000
Mod 1 Mod 2 Mod 3
Constant 0 11 * 0.10 * 0.10 *
(4.11) (4.11) (3.9)
GDP per Capita 0.005 * 0.001 * 0.001 *
(3.12) (2.37) (3.31)
Trade/GDP 0.06 * 0.05 * 0.05
(7.54) (7.7) (7.7)
Debt/GDP 0.10 0.10 0.10
(4.2) (3.84) (3.89)
Population Growth -0.005 * -0.06 * -0.006 *
(-2.12) (-2.4) (-2.5)
Inflation -0.002 -0.002 -0.002
(-2.36) (-2.62) (-2.4)
Control of Corruption 0.005
(2.11)
High Quality Bureaucracy 0.002 *
(2.64)
Law and Order 0.01 *
(2.28)
Sargan Test (p value) (0.32) (0.23) (0.21)
Hausman Test (p value) (0.13) (0.12) (0.14)
[R.sup.2] 0.70 0.65 0.65
2001-2010
Mod 1 Mod 2 Mod 3
Constant 0.13 * 0.16 * 0.06 *
(2.34) (2.42) (0.83)
GDP per Capita 0.04 * 0.001 * 0.01 *
(2.04) (2.17) (2.81)
Trade/GDP 0.01 * 0.02 * 0.01 *
(2.53) (3.01) (3.12)
Debt/GDP 0.186 0.17 0.17
(4.35) (3.99) (4.26)
Population Growth -0.01 * -0.01 * -0.01 *
(-3.2) (-2.60) (-3.17)
Inflation 0.001 0.001 0.002
(0.15) (0.74) (1.02)
Control of Corruption 0.01 **
(1.96)
High Quality Bureaucracy 0.01 *
(2.51)
Law and Order 0.014 *
(2.78)
Sargan Test (p value) (0.20) (0.31) (0.28)
Hausman Test (p value) (0.09) (0.76) (0.67)
[R.sup.2] 0.72 0.65 0.67
Note: * Indicates significance at 1 percent, ** at 5 percent and
*** at 10 percent level. The Hausman Test supports fixed effect
model. The GMM is estimation technique and lag exogenous are used
as instruments.
Table A3
Determinants of Revenue Potential in Developing Asian Countries
1984-1990
Mod 1 Mod 2 Mod 3
Constant 0.01 -0.01 -0.03
(0.67) (-0.5) (-0.85)
Agriculture Value -0.02 * -0.03 * -0.03 *
Addition to GDP (-3.07) (-3.7) (3.10)
Trade/GDP 0.04 * 0.08 * 0.06 *
(2.4) (5.6) (3.8)
Debt/GDP 0.16 * 0.17 * 0.15 *
(3.9) (4.1) (3.17)
Population Growth -0.01 * -0.01 * -0.01 *
(-2.13) (-2.22) (-2.83)
Inflation -0.02 * -0.02 * -0.01 *
(-1.89) (-1.89) (-2.77)
Control of Corruption 0.01 *
(5.31)
High Quality Bureaucracy 0.02 *
(5.27)
Law and Order 0.01 *
(3.02)
Sargan Test(p value) 0.32) (0.29) (0.31)
Hausman Test (0.18) (0.11) (0.21)
(p value)
[R.sup.2] 0.69 0.68 0.61
1991-2000
Mod 1 Mod 2 Mod 3
Constant 0.12 * 0.12 * 0.12 *
(4.8) (4.3) (3.33)
Agriculture Value -0.02 * -0.04 * -0.03
Addition to GDP (-2.2) (-2.43) (-0.35)
Trade/GDP 0.06 * 0.05 * 0.05 *
(5.3) (5.07) (4.91)
Debt/GDP 0.10 * 0.10 * 0.10 *
(4.7) (4.39) (4.39)
Population Growth -0.01 ** -0.01 * 0.01 *
(-1.86) (-1.96) (-2.15)
Inflation -0.02 * -0.03 * -0.02 *
(-2.42) (-2.77) (-2.6)
Control of Corruption 0.05 *
(2.14)
High Quality Bureaucracy 0.03 * 0.01 *
(2.70) (2.37)
Law and Order
Sargan Test(p value) (0.19) (0.22) (0.21)
Hausman Test (0.16) (0.13) (0.14)
(p value)
[R.sup.2] 0.66 0.65 0.64
2001-2010
Mod 1 Mod 2 Mod 3
Constant 0.24 * 0.29 * 0.23 *
(5.10) (5.35) (3.93)
Agriculture Value -0.05 -0.05 -0.05
Addition to GDP (-2.5) (-2.95) (-2.44)
Trade/GDP 0.03 ** 0.02 ** 0.02 *
(1.9) (1.8) (1.85)
Debt/GDP 0.16 * 0.15 * 0.16 *
(4.2) (3.8) (4.0)
Population Growth -0.01 * -0.01 * -0.01 *
(-2.7) (-2.08) (-2.6)
Inflation -0.01 * -0.02 * -0.02 *
(-2.97) (-2.15) (-2.35)
Control of Corruption 0.01 *
(2.05)
High Quality Bureaucracy 0.01 *
(2.87)
Law and Order 0.05 *
(2.64)
Sargan Test(p value) (0.23) (0.25) (0.35)
Hausman Test (0.09) (0.24) (0.08)
(p value)
[R.sup.2] 0.67 0.66 0.67
Note: * Indicates significance at 1 percent, * at 5 percent and
** at 10 percent level. The Hausman Test supports fixed effect
model. The GMM is estimation technique and lag exogenous are used
as instruments.
REFERENCES
Agbeyegbe, T. D., J. Stotsky, and A. Wolde Mariam (2004) Trade
Liberalisation, Exchange Rate Changes and Tax Revenue in Sub-Saharan
Africa. International Monetary Fund, Washington, DC. (IMF Working Paper
No. 178).
Aizenman, Joshua and Yothin Jin Jarak (2009) Globalisation and
Developing Countries--A Shrinking Tax Base? Journal of Development
Studies 45:5, 653-671.
Aim, James, Jorge Martinez-Vazquez and Friedrich Schneider (2004)
Sizing' the Problem of the Hard-to-Tax. In Martinez-Vazquez Aim and
S. Wallace (eds.) Taxing the Hard to Tax.
Arellano, Manuel and Steve Bond (1991) Some Tests of Specification
for Panel Data: Monte Carlo Evidence and an Application to Employment
Equations. The Review of Economic Studies 58, 277-97.
Bahl, Roy (1971) A Regression Approach to Tax Effort and Tax Ratio
Analysis. International Monetary Fund, Staff Paper 18, 570-612.
Bird, Richard M. and Pierre-Pascal Gendron (2007) The VAT in
Developing and Transitional Countries. Cambridge University Press.
Bird, Richard M., Jorge Martinez-Vazquez, and Benno Torgler (2004)
Societal Institutions and Tax Efforts in Developing Countries. Andrew
Young School of Policy Studies, Georgia State University: Georgia,
Atlanta. (Working Paper 04-06).
Blundell, Richard and Steve Bond (1998) Initial Conditions and
Moment Restrictions in Dynamic Panel Data Models. Journal of
Econometrics 87, 115-43.
Chelliah, Raja J. (1971) Trends in Taxation in Developing
Countries. International Monetary Fund, Staff Papers 18, 254-0331.
Chelliah, Raja J., Hessel J. Baas, and Margaret R. Kelly (1975) Tax
Ratios and Tax Effort in Developing Countries, 1969-71. International
Monetary Fund, Staff Papers 22, 187-205.
Davoodi, H. and D. Grigorian (2007) Tax Potential vs. Tax Effort: A
Cross-Country Analysis of Armenia's Stubbornly Low Tax Collection.
International Monetary Fund, Washington, DC. (IMF Working Paper No.
106).
Eltony, M. Nagy (2002) Determinants of Tax Efforts in Arab
Countries. Arab Planning (Institute Working Paper 207).
Fox, William F. and Tami Gurley (2005) An Exploration of Tax
Patterns around the World. Tax Notes International: Special Reports
37:9, 793-807.
Ghura, Dhaneshwar (1998) Tax Revenue in Sub Saharan Africa: Effects
of Economic Policies and Corruption. International Monetary Fund.
Washington, DC. (IMF Working Paper 98/135).
Gupta, Abhijit Sen (2007) Determinants of Tax Revenue Efforts in
Developing Countries. (IMF Working Paper, No: WP/07/184).
Gupta, Sanjeev, Benedict Clements, Alexender Pivovarsky, and Erwin
R. Tiongson (2004) Foreign Aid and Revenue Response: Does the
Composition of Aid Matter? In Gupta, Clements, and Gabriela Inchauste
(eds.) Helping Countries Develop: The Role of Fiscal Policy. Washington,
DC: International Monetary Fund.
ICRG (2011) International Country Risk Guide Index. New York: The
PRS Group.
International Monetary Fund (n.d.) IMF Financial Statistics (IFS).
International Monetary Fund, Washington, DC. CD-ROM database.
Jha, R. (2009b) Fiscal Policy. In A. Dutt and J. Ros (eds.)
International Handbook of Development Economics. London: Edward Elgar
Publishers. 305-318.
Keen, Michael and Alejandro Simone (2004) Tax Policy in Developing
Countries: Some Lessons from the 1990s and Some Challenges Ahead. In
Sanjeev Gupta, Benedict Clements, and Gabriela Inchauste (ed.) Helping
Countries Develop: The Role of Fiscal Policy. Washington, DC:
International Monetary Fund.
Le, Tuan Minh, Blanca Moreno-Dodson, and Jeep Rojchaichaninthom
(2008) Expanding Taxable Capacity and Reaching Revenue Potential:
Cross-Country Analysis. World Bank.
Leuthold, Jane H. (1991) Tax Shares in Developing Countries: A
Panel Study. Journal of Development Economics 35, 173-185.
Lotz, Jorgen R. and Elliott R. Morss (1967) Measuring "Tax
Effort" in Developing Countries (Evaluation de effort fiscal
dandles pays en voie de development) (Medicion del
"esfuerzotributario" de los paises en desarrollo).
International Monetary Fund, Staff Papers 14:3, 478-99.
Norregaard, John and Tehmina S. Khan (2007) Tax Policy: Recent
Trends and Coming Challenges. (IMF Working Paper, No. WP/07/274).
Piancastelli, Marcelo (2001) Measuring the Tax Effort of Developed
and Developing Countries: Cross Country Panel data Analysis--1985-95.
(Texto Para Discussion No. 818).
Rodrik, Dani (1998) Why Do More Open Economies Have Bigger
Governments? Journal of Political Economy 106:5, 997-1032.
Stotsky, Janet G. and Asegedech Wolde Mariam (1997) Tax Effort in
Sub-Saharan Africa. International Monetary Fund, Washington, DC. (IMF
Working Paper).
Tait, Alan A., Barry J. Eichengreen, and Wilfrid L. M. Gratz (1979)
International Comparisons of Taxation for Selected Developing Countries,
1972-76. Staff Papers 26, 123-156. International Monetary Fund.
Tanzi, V. (1987) Quantitative Characteristics of the Tax System of
Developing Countries. In Newbery and Stern (eds.) The Theory of Taxation
for Developing Countries. New York: Oxford University Press.
Tanzi, Vito (1992) Structural Factors and Tax Revenue in Developing
Countries: A Decade of Evidence. In Vito Tanzi and H. R. Davoodi (eds.)
Open Economies: Structural Adjustment and Agriculture Corruption, Growth
and Public Finances. Washington, DC: International Monetary Fund. (IMF
Working Paper No. 182).
Tanzi, Vito and Hamid Davoodi (1997) Corruption, Public Investment,
and Growth. Washington, DC, International Monetary Fund. (Policy
Research Working Paper No. 4559, March 2008).
Teera, J. M. and J. Hudson (2004) Tax Performance: A Comparative
Study. Journal of International Development 16, 785-802.
World Bank (2010) World Development Indicators (WDI). Washington,
DC: The World Bank.
Comments
The paper titled "Analysis of Revenue Potential and Revenue
Effort in Developing Asian Countries" by Attiya Y. Javid and Umaima
Arif provides a comprehensive study on the revenue performance
determinants across developing Asian countries and calculated the
revenue effort indices using the predicted and actual revenue ratios..
Among the other factors contributing towards revenue realisations
authors have rightly pointed out the possibility of including the
institutional quality and overall governance measure too.
Typo mistakes are there.
GMM; J-test to see over/exact identification, Panel data time
series issues, low R2
Why Debt to GDP ratio is positively related with tax efforts and
why inflation negatively?
Total revenues account for both the tax and non-tax revenues, in
developing countries such as Pakistan they also present a large share
which may not be described better by variables which are used best to
describe the tax revenues.
Further in some context of Fiscal policy instrumentation the
causality between Revenues and Expenditures is also very strong, e.g. in
case of Pakistan Husain, et al. (2010) found that Spend-Tax hypothesis
is very strong. So including the size of the government as measured by
Government Expenditures to GDP ratio may be added as an explanatory
variable in the regression.
Country specific structural changes in the revenue streams needs to
be accounted for, such as the enactment of WTO in 1995 as also
identified by the authors and others such as in case of Pakistan
FRDL-2005 etc. As also pointed out by the authors that the Hausman test
prefers the fixed effects model as compared to the random effect or
both.
Over all the paper is a good contribution to the existing knowledge
on the subject.
Mahmood Khalid
Pakistan Institute of Development Economics, Islamabad.
Attiya Y. Javid <
[email protected]> is Professor of
Economics at Pakistan Institute of Development Economics, Islamabad.
Umaima Arif <
[email protected]> is Lecturer, Economics
Department, Quaid-i-Azam University, Islamabad.
Authors' Note: Any remaining errors and omissions are the
authors' sole responsibility.
(1) Revenue potential refers to the predicted revenue to GDP ratio
that can be estimated with the regression, taking into account a
country's specific macroeconomic, demographic, and institutional
features. While lacking solid theoretical foundations, actual tax to GDP
(likewise revenue to GDP) is one of the most commonly used measures for
cross country comparison of tax (fiscal) effort. The advantages of this
measure are that it is easy to obtain and gives quick overview of
revenue performance across countries. The problem is that, the
measurement of the potential revenue capacity is based on, a priori, set
of explanatory variables that determine the potential capacity of a
country to collect revenue, but it does not reflect either the demand
for higher public expenditures or the political willingness to collect
revenue as pointed out by Bird (1978) and Toye (1978).
(2) A high fiscal effort is the case when effort index is above 1,
indicating that the country well utilises its revenue base to increase
revenues and a low fiscal effort is the case when effort index is below
1 implying that the country may have relatively substantial potential to
raise revenues [Stotsky, et al. (1997); Bird, et al. (2004) and Le, et
at. (2008)]. This index allows to compare country's revenue effort
vis-a-vis that of its peers [Tanzi and Zee (2000)].
(3) The fixed effect panel regression specification is given in
Equation (I) and random effect specification on Equation (2)
respectively.
(4) The sector-wise composition of GDP also affects revenue
collection capacity because in some sectors of the economy it is easy to
impose tax than others. For example, the agriculture sector is
considered as difficult to tax, especially if there are a large number
of subsistence farmers. On the other hand, manufacturing sector
consisting of a few large firms can generate large tax. These components
of GDP are added one by one to avoid multicolinearity.
(5) Due to high correlation between institutional variables one
variable is included in the specification at a time.
(6) Appendix Tables A2 and A3 reports die results of models 1 to 12
for sub-periods: 1984 to 1990, 1991 to 2000 and 2001 to 2010. The
results of regression analysis are almost the same for the sub-periods
as well.
Table 1
Determinants of Revenue Potential in
Developing Asian Countries: 1984 2010
Mod 1 Mod 2 Mod 3
Constant 0.12 * 0.19 * 0.14 *
(5.40) (7.22) (6.35)
GDP per Capita 0.08 * 0.07 * 0.07 *
(3.92) (3.5) (1.95)
Agriculture Value Addition to GDP
Trade/GDP 0.05 * 0.06 * 0.04 *
(2.26) (3.10) (3.85)
Debt/GDP 0.48 * 0.35 * 0.20 *
(4.14) (9.90) (9.90)
Population Growth -0.04 * -0.03 * -0.05 *
(-2.32) (-5.44) (-2.98)
Inflation -0.02 ** -0.01 ** -0.01 **
(-1.94) (-1.89) (-1.85)
Control of Corruption 0.01 *
(3.98)
High Quality Bureaucracy 0.01 *
(2.79)
Best Law and Order Conditions 0.008 *
(2.95)
Sargan Test (p-value) (0.18) (0.11) (0.13)
Hausman Test (p-value) (0.27) (0.21) (0.11)
[R.sup.2] 0.71 0.65 0.65
Mod 4 Mod 5 Mod 6
Constant 0.14 * 0.16 * 0.18 *
(3.41) (4.6) (5.8)
GDP per Capita
Agriculture Value Addition to GDP -0.05 * -0.04 * -0.02 *
(-3.3) (-2.04) (-1.6)
Trade/GDP 0.04 0.03 0.08
(0.7) (0.09) (0.56)
Debt/GDP 0.39 * 0.41 * 0.42 *
(7.3) (7.8) (7.7)
Population Growth -0.03 * -0.02 * -0.03 *
(-4.4) (-4.2) (-3.89)
Inflation -0.01 -0.01 0.01
(-0.44) (-0.57) (-0.66)
Control of Corruption 0.02 *
(3.9)
High Quality Bureaucracy 0.01 *
(2.1)
Best Law and Order Conditions 0.009 *
(1.8)
Sargan Test (p-value) (0.21) (0.17) (0.10)
Hausman Test (p-value) (0.58) (0.31) (0.44)
[R.sup.2] 0.69 0.64 0.65
Note: * Indicates significance at 1 percent, ** at 5 percent and
*** at 10 percent level. The Hausman Test supports fixed effect
model. The GMM is estimation technique and lag exogenous are used
as instruments and Sargan J-Test for overidentying restrictions
confirms that the error term is uncorrelated with the
instruments. The Hausman test supports that error terms are
uncorrelated with explanatory variables so fixed effect model is
better choice.
Table 2
Determinants of Revenue Potential/Capacity in Developing Asian
Countries: Dynamic Panel Model
Mod 7 Mod 8 Mod 9
Constant 0.05 0.06 0.05
(3.6) (3.76) (3.49)
Lag Revenue/GDP 0.72 0.74 0.72
(17.7) (17.01) (17.4)
GDP per Capita 0.03 0.04 0.03
(1.84) (1.89) (1.85)
Agriculture Value Addition to GDP
Trade to GDP 0.005 0.006 0.004
(1.13) (1.38) (1.03)
Debt/GDP 0.07 * 0.07 * 0.07 *
(5.37) (5.36) (5.29)
Population Growth -0.02 * -0.02 ** -0.02 *
(-2.33) (-1.93) (-2.34)
Inflation -0.05 -0.06 -0.03
(-0.76) (-0.83) (-0.55)
Control of Corruption 0.01 *
(2.31)
High quality Bureaucracy 0.02 **
(195)
Best Law and Order Conditions 0.02 *
(2.33)
Sargan Test (p-value) (0.15) (0.27) (0.13)
Hausman Test (p-value) (0.42) (0.35) (0.44)
[R.sup.2] 0.75 0.72 0.71
Mod 10 Mod 11 Mod 12
Constant 0.05 0.05 0.04
(3.5) (3.1) (2.05)
Lag Revenue/GDP 0.68 0.69 0.68
(11.6) (12.9) (12.8)
GDP per Capita
Agriculture Value Addition to GDP -0.03 ** -0.02 ** -0.01 **
(-1.88) (-1.90) (-1.92)
Trade to GDP 0.01 -0.005 0.008
(0.20) (-0.08) (0.12)
Debt/GDP 0.07 * 0.07 * 0.08 *
(3.47) (3.3) (3.6)
Population Growth -0.03 ** -0.03 * -0.03 *
(-1.85) (-1.97) (-1.94)
Inflation -0.01 -0.01 0.01
(-0.32) (-0.84) (-0.68)
Control of Corruption 0.01 *
(2.44)
High quality Bureaucracy 0.007 *
(2.23)
Best Law and Order Conditions 0.02 *
(2.05)
Sargan Test (p-value) (0.21) (0.20) (0.22)
Hausman Test (p-value) (0.15) (0.14) (0.19)
[R.sup.2] 0.73 0.70 0.70
Note: * Indicates significance at 1 percent, ** at 5 percent and
*** at 10 percent level. The Hausman Test supports fixed effect
model. The GMM is estimation technique and lag exogenous are used
as instruments and Sargan J-Test for overidentying restrictions
confirms that the error term is uncorrelated with the
instruments. The Hausman test supports that error terms are
uncorrelated with explanatory variables so fixed effect model is
better choice.
Table 3
Revenue Effort Index for Developing Asian Countries
Model 1: GDP per Capita
1984-90 1991-00 2001-10 1998-10
India 0.97 0.94 1.03 0.98
Pakistan 0.96 0.90 0.85 0.92
Bangladesh 0.84 0.80 0.82 0.85
Sri Lanka 1.05 0.89 0.84 0.87
Malaysia 1.00 0.92 1.02 0.91
Indonesia 0.85 0.93 1.11 1.07
Thailand 1.11 1.17 1.01 1.26
Philippines 0.95 1.03 0.94 1.29
China 1.12 0.83 1.11 1.01
Singapore 1.06 1.37 1.31 1.40
Vietnam 0.84 1.00 0.90 0.87
Model 4: Agriculture Value Added to GDP
1984-90 1991-00 2001-10 1984-10
India 0.96 0.87 1.04 0.98
Pakistan 1.00 0.93 0.85 0.91
Bangladesh 0.84 0.80 0.82 0.83
Sri Lanka 0.88 0.84 0.90 0.89
Malaysia 1.00 0.85 1.02 0.88
Indonesia 0.85 0.93 1.04 1.01
Thailand 1.00 0.92 1.05 1.05
Philippines 0.89 0.88 0.94 0.99
China 1.12 0.89 1.11 1.03
Singapore 1.05 1.14 1.11 1.15
Vietnam 0.82 0.84 0.99 0.90
