Structural Gravity Model and Globalization: An Empirical

Analysis between Bangladesh, China, India and Pakistan

Mubashir Hussain 1
Mehak Ejaz 2
ABSTRACT
The globalization process during last two hundred years has influenced the world’s social,
cultural and economic positions. Consequently, world’s output, trade and population have
been increased significantly. As a result, wellbeing of nations has been improved. However, at
the same time in equality among the rich and poor nations of the world has also increased
which is evident from the “Gini coefficient” reported by the World Trade Organization
(WTO), 2013. Therefore this study aims to investigate export determining factors for four
neighboring countries namely China, India, Pakistan and Bangladesh. It is an empirical
analysis based on the structural gravity model. The study employees a relatively new
technique of panel data (PPML-Estimator) for the comparison between countries. The
findings reveal that factors such as income, expenditure and geographical distance in all four
countries are in line with the theoretical literature as well as consistent with gravity theory.
However, the dummy variables for language, contiguity and the Regional Trade Agreement
(RTAs) have mixed results. Notably, the RTAs in case of China, India and Pakistan have nega-
tive and significant impact. In another model, in order to capture the impact of globalization,
this study used overall globalization index (GI) and its proxies like Political Globalization
Index (PGI), Social Globalization Index (SGI), Economic Globalization Index (EGI) and
Information Flow Index (IFI). The results suggest that the proxies have mixed results whereas;
the overall globalization has positive and significant effect on the exports. On the basis of
these finding, it may be concluded that political conflicts among these neighboring nations are
the main hurdle for their mutual beneficial trade. Although, Pakistan’s socio-economic set up
is responsive towards globalization wave however, factors like poor management, energy
crises, corruption and terrorism have hampered its trade performance.

Keywords: Structural Gravity Model, Globalization, PPML-Estimator, International Trade

INTRODUCTION
Since mid-1800s the population of world has been increased by six folds, the world output rose
by sixty fold and international trade grew upto140 fold. “This virtuous circle of deepening
integration and expanding growth is what we now refer as “globalization” (Madisson, 2008).
Though, there is no consensus over the definition of globalization. However, in economics
perspective globalization is a process in which the commodity market, labor market and
capital market of an economy is integrated into the world economy. The history of globaliza-
tion may be traced back more prominently at least over the past 200 years. In early 1800s the
international monetary system “Gold Standard” backed by British pound sterling joined forces
with industrial revolution. This phase of globalization is referred as “first age of globalization”
in which migration, communication, technology and capital flow were moving across the
nations. But earlier to this; there was an age of “steam-power” in which advancement in steam
ship, steam railways engine industry and navigation led to explore new sea route to America,
1 - PhD Scholar at Shaheed Zulfiqar Ali Bhutto Institute of Science and Technology, Karachi. mubashir.hussain32@yahoo.com
2 - Associate Professor at Shaheed Zulfiqar Ali Bhutto Institute of Science and Technology, Karachi.

JISR-MSSE Volume 15 Number 2 July-Dec 2017 107

Asia and Africa for Europe, Madisson, (2008).

The first phase of globalization was interrupted by World War-1, Great Depression and World
War-II. The recovery from the World War-II brought extensive development in the field of
science and technical innovation that speeded up the process of globalization. The main driver
of globalization in the context of technical innovation; are rapid inventions in transportation
and communication. Another important driver is revolutionary innovations in the field of
information and communication technology. These innovations have reduced considerably the
cost of communication and transportation for international trade, WTO, (2013).

The recent phase of globalization has brought significant improvement in output, international
trade and wellbeing of societies around the world. According to World Trade Report, 2013
during the period of 1950-73, the world GDP grew 3 percent and merchandize trade increased
by 8 percent per annum in real terms. Hence, the period from 1950 to 1973 is known as
“golden age”. Moreover, during the period 1950 to 2007 world’s merchandize growth rose by
6.2 percent. Table-1 provides the percentage share of world export of world GDP over the
period of 1870 to 1998 which indicates that integrated world economy increased the world’s
trade massively.

Table 1 Share of world export in world GDP 1870-1998 (Percentage)

1870 4.6
1913 7.9
1950 5.5
1973 10.5
1998 17.2
Source: OECD (2001)
The globalization phenomenon contains the seeds of free trade which brings mutually benefi-
cial gains from trade for all societies of the world. According to WTO, (2013) during the last
three decades; the world merchandize trade has been increased from US$ 2.03 trillion in1980
to US$ 18.26 in 2011 which is four-fold increase in the world merchandize and is 7.2 percent
growth per year on average in current dollar terms. This upturn in world’s trade is mainly
attributed to the reduction in trade barriers (tariff and non-tariff) through preferences trade
agreements (PTA) under the umbrella of WTO.

According to the WTO’s world trade report, 2008, Globalization brought considerable chang-
es in economic growth, international trade and development and has reduced poverty around
the world. As it is clear from (Table-2 and Figure-2) that the current wave of globalization
emerging economies like BRICS i.e. Brazil, Russia, India China and South Africa have experi-
enced massive increase in their growth rate and trade than developed and less developed
nations. However, the world trade is distributed unevenly among the nations of the world.

During the period of 1980 to 2011 the “Gini coefficient” has not been improved which implies
that inequality is increasing among the nations and societies WTO, (2013). In such a
perspective, there are countries especially less developed who have lagged behind in the race
of globalization and their trade performance is poor in the international arena. Pakistan is
among one of those countries whose exports percentage to GDP is declining persistently more
108 July-Dec 2017 Volume 15 Number 2 JISR-MSSE
prominently from 2003. Since 2003, the percentage to GDP is suffering from downturn
because of energy crises and war against terrorism in Pakistan. On the other hand, since from
same point of time the percentage of exports to GDP of its two neighboring countries, India
and Bangladesh who share fairly similar socioeconomic, cultural and historical linkage; India
and Bangladesh is improving extensively which is depicted in Figure-3.

World’s merchandize trade and real GD

GDP growth (left scale) Merchandise trade volume growth (left scale) Elasticity (right scale)

Figure 1. WTO Secretariat.

P, 1980-2011(annual percentage change)

TABLE-2 Trade to GDP Ratio

Economies 1995 2000 2005
World 43.3 50.3 56.4
Developed Economies 38.3 44.5 49.4
Emerging Economies 43.3 50.3 56.4
(LDCs) 46.7 53.1 63.8
Source: WTO,(2013)

Figure-2 Source: World Trade Report, 2013 (WTO)

JISR-MSSE Volume 15 Number 2 July-Dec 2017 109

According to World Trade Report WTO, (2013), the China was a largest merchandize export-
er in the world in 2011 with its export’s value US$ 18,255.2 billion and India’s export value
was US$ 304.6 billion. Furthermore, during last one and half decade Bangladesh’s trade
performance been improved and has emerged as one of the top exporter of textile goods.

Figure-3 Exports percentage to GDP (1988-2016) for Pakistan, India and Bangladesh.

Source: Generated by authors using WDI, 2016 data

Objective of the Study

This study aims to investigate the determining factors of exports of China, India, Pakistan and
Bangladesh (which are located in vicinity of one another). Separate econometric models have
been regressed for comparison among these countries. For this purpose structural gravity
model of trade has been employed using relatively an advanced panel data technique
“PPML-Estimator” developed by Santos and Silvana, (2011). Pakistan, China and India share
connecting border countries with one another and Bangladesh has common border with India.
So this research study focuses on comparative analysis among these four countries due to their
significant socio-economic role in the world. Furthermore, to arrest the impact of globalization
on their exports flow directed toward their 15 major trading partners, this study uses overall
globalization index (GI) and its proxies like Political Globalization Index (PGI), Social
Globalization Index (SGI), Economic Globalization Index (EGI) and Information Flow Index
(IFI) as it was first used by Husain, (2017) in gravity model analysis. The detail and computa-
tion of these indices is provided in Appendex-1

110 July-Dec 2017 Volume 15 Number 2 JISR-MSSE

LITRATURE REVIEW

Theoretical Framework of Gravity Model

The analogy and idea of gravity model of trade has been derived from the Newton’s “Univer-
sal law of Gravitational Force” which states that everybody in this universe attracts with a
force to another body in a way that this force of attraction is directly proportional to the
product of their masses and inversely related to the square distance between them. This
relationship in equation form is as under:

F = α ( m1 × m2 ) / r 2 (1)
Tinbergen, (1962) and Ravenstein, (1885), were the pioneer to use gravity equation in interna-
tional trade. The gravity model of trade in log linear form is written as:

log (Tradeijt ) = β0 + β1log ( GDPit . GDPjt ) + log ( Distanceijt ) + µt (2)

Equation (2) represents that economic size (GDPs) of trading partners which have positive and
geographical distance between them has negative impact on bilateral trade flows of trading
partners.

At this initial stage gravity model lacked appropriate theoretical foundations but afterwards
economist began to work on it and applied several theories from the literature of international
trade in gravity model of trade. The remarkable contribution of economist provided the strong
many theoretical foundations to the literature of gravity model. In this journey Anderson,
(1979) was first who tried to provide theoretical foundations to gravity model on the basis of
product differentiation by the place of origin so called “Armington Assumption” in the context
of constant elasticity substitution (CES). Armington, (1969), pointed to consider the goods for
their different kinds e.g. merchandise goods, chemical goods, petroleum and wooden products
etc. but also differentiated homogeneous goods on the basis of different place of production.
As the same kind of goods produced at two different places makes them imperfect substitute
due to different cost of production and cost of trade.

Another contribution by Helpman and Krugman, (1989), on the basis of monopolistic compe-
tition and differentiated products in accordance with increasing returns to scale. Deardorff,
(1998) explained gravity model on the grounds of “Heckchier-Ohlin Theorem” of internati
onal trade theory on the assuptions of “frictionless trade and trade with impediments”, Sarah,
(2012). Eaton and Kortum, (2002), proved gravity model envelopes the Recardian theory of
“Comperative Advantage”. This theory states that in two commodity two nations assumption
even if a country is less efficient (having absolute disadvantage) in production in both
commodities than other tarde partner, there are still basis for mutual beneficial trade. The
theory in opportunity cost terms states that “the cost of a commodity is the amount of a second
commodity that must be given up to release just enough resources to produce one additional
unit of first commodity”, Salvatore, (1998). Therfore, Eaton and Kortum, (2002) provided that
heterogenirty of production between trade partners are the basis of “Comperative Advantage”.
This theory states that in two commodity two nations assumption even if a country is less
efficient (having absolute disadvantage) in production of both commodities than other country
there are still basis for mutual beneficial trade.

JISR-MSSE Volume 15 Number 2 July-Dec 2017 111

Review of Structural Gravity Model
Exporting countries (i) produce differentiated goods which they trade with their partners. The
supply of each good is fixed at Qi, and factory gate price is represented for each commodity is
pi, hence the total nominal GDP is stated as Yi = pi Qi . On the other hand, Expenditures of
country i may be expressed in nominal income terms Ei. It is represents Ei = ϕiYi , where
ϕ i > 1 indicates deficit in balance of trade, whereas, while 1 > ϕ i > 0, implies a trade surplus in
an economy. Next we derive the structural gravity model like Dekle et al. (2007; 2008), from
demand side.

On the demand side, consumer preferences are assumed to be homothetic, identical across
countries, is and given by a CES-utility function for country j:
1−σ
1−σ 1−σ
⎧⎪ ⎫σ
⎨∑ ai
σ
cij
σ
⎬ (3)
⎪⎩ 0i ⎭

Teachers Where σ >1 is the elasticity of substitution among different varieties, i.e. goods from
different countries, αi > 0 is the CES preference parameter, which will remain treated as an
exogenous taste parameter and cij denotes consumption of varieties from country i in country
j. Consumers maximize equation (3) subject to the following standard budget constraint:

∑pc
i
ij ij = Ej (4)

Equation (4), show that expenditures Ej in country j are equal to total spending on varieties of
goods from which it imports the goods at delivered prices pij = pi . tij which are defined
conveniently as a function of factory-gate prices in the country of origin, pi, marked up by
bilateral trade costs, tij ≥ 1 , between trading partners i and j.

112 July-Dec 2017 Volume 15 Number 2 JISR-MSSE

Solving the consumer’s optimization problem yields the expenditures on goods shipped from
origin i to destination j as:
1−σ
⎛ aipitij ⎞
Xij = ⎜ ⎟ Ej (5)
⎝ pj ⎠
Where X ij denotes trade flows from exporter i to destination j and, for now, Pi can be
interpreted as a CES consumer price index:
1−σ
⎡ 1−σ ⎤
(6)
σ
pj = ⎢∑ ( aipitij ) ⎥
⎣ i ⎦
Given that the elasticity of substitution is greater than one, σ >1, equation (6) captures several
intuitive relationships. In particular, expenditure in country j on goods from source i, X ij , is
Proportional to total expenditure, E j, in destination j. The simple intuition is that, all else equal,
larger/richer markets consume more of all varieties, including goods from i.

It is Inversely related to the (delivered) prices of varieties from origin i to destination j, pij = pi tij.
This is a direct reflection of the law of demand, which depends not only on factory-gate price
pi but also on bilateral trade cost between partners i and j. The ideal combination that favors
bilateral trade is an efficient producer, characterized by low factory-gate price, and low
bilateral trade cost between countries “Directly related to the CES price aggregator ”. This
relationship reflects the substitution effects across varieties from different countries. All else
equal, the relatively more expensive the rest of the varieties in the world are, the more
consumers in country j will substitute away from them and toward the goods from country i.
contingent on the elasticity of substitution when factory-gate prices or the aggregate CES
prices (Or in the combination of those as a relative price) change. All else equal, a higher
elasticity of substitution will magnify the trade diversion effects from the more expensive
commodities to the cheaper ones.

The final step in the derivation of the structural gravity model is to impose market clearance
for goods from each origin:
1−σ
⎛ aipitij ⎞ (7)
Yi = ∑ ⎜ ⎟ Ej
j ⎝ Pj ⎠

Equation (7) states that, at delivered prices (because part of the shipments melt “en route”), the
value of output in country i, , should be equal to the total expenditure of this country’s
variety in all countries in the world, including i itself. To see this intuition more clearly, note
that the right-hand side expression in equation (7) can be replaced with the sum of all bilateral
shipments from i as defined in equation (5), so that Yi ≡ Σ j X ij ∀ j .
Defining Y ≡ ΣiYi and dividing equation (7) by Y, the terms can be rearranged to obtain:

Yi
( aipi )
1−σ Y
= 1−σ (8)
⎛ ⎞ Ej
t ij
∑ ⎜⎝ Pj ⎟⎠ Y
JISR-MSSE Volume 15 Number 2 July-Dec 2017 113
Following (Anderson and van Wincoop, 2003), the term in the denominator of equation (8)
1−σ
can be defined as: ∏ j1−σ = ∑( tij
Pj ) Ej Y , and be substituted into equation (8):
Yi
( aipi )
1−σ
= Y (9)
∏ i1−σ
Recent developments in gravity model are inspired by the work of (McCallum, 1995), who
provided empirical findings that “Canadian provinces trade more than 20 times as much
among each other than Canadian provinces and U.S. states do”. This outcome is referred as
“trade puzzle”. (Anderson and van Wincoop, 2003), based on his earlier work (Anderson,
1979) having assumptions that goods are differenciated from the place of origin, consumers
preferences are homothetic, identical across countries, and approximated by a CES utility
function provided the following systm of structural gravity model:
1−σ
Xij = Yi ,tEj ,t ⎜⎜ tij ,t ⎟⎟
⎛ ⎞
(10)
Yt ⎝ Pj ,t ∏ i ,t ⎠
1−σ
⎛ tij , t ⎞ Ej , t
(11)
∏1−σ = ∑ ⎜ ⎟
j ⎝ Pj , t ⎠ Yt
1−σ
1−σ ⎛ tij , t ⎞ Yi , t
(12)
P j, t = ∑ ⎜ ⎟
i ⎝ ∏ i, t ⎠ Yt

X ij ,t , denotes trade flows from exporter i to importer j

E j ,t , is the total expenditure in importer j
Yi ,t , is the value of total production in exporter i
Yt , is the value of world output;
tij ,t , denotes bilateral trade frictions between partners i and j;
σ>1 is the elasticity of substitution among goods from different countries;
Finally, the more important the terms coined by (Anderson and Van Wincoop, 2003) Pj,t and
Π i ,t as the inward and the outward multilateral resistances, respectively.
“The multilateral resistances are the vehicles that translate the initial, partial equilibrium
effects of trade policy at the bilateral level to country-specific effects on consumer and
producer prices”, (Piermartini and Yotov, 2016).

Log-linear Equation (1) and expanding it with an additive error term, obtains the follow-
ing estimating gravity equation:

ln Xij = lnEj, t+ lnYi, t lnYt+ (1 )lntijt (1 ) Pj, t (1 )ln i, t + μ ij, t (13)

Specification (1-11) is the most popular version of the empirical gravity equation, and it has
been used routinely in the trade literature to study the effects of various determinants of
bilateral trade. Hundreds of papers have used the gravity equation to study the effects of
geography, demographics, RTAs, tariffs, exports subsidies, embargoes, trade sanctions, the

114 July-Dec 2017 Volume 15 Number 2 JISR-MSSE

World Trade Organization membership, currency unions, foreign aid, immigration, foreign
direct investment, cultural ties, trust, reputation, mega sporting events (Olympic Games and
World Cup), melting ice caps, etc. on international trade.

Empirical Literature Review

Batra,(2004), using data for the year of 2000 and applying OLS technique of estimation.
Estimation results reveal that the gravity equation fits the data and prices, income and
geographical, cultural and historical variables are consistent with theory and are significant.

Panda, etal. ( 2016),With the obective of coparision of trade flow determinants between India
and China used the data from 2004 to 2013 for penal data anaylysis and found that both coun-
try’s trade flows more with neigboring countires. In the case of China common language and
high per capita income are determining factor of its trade and geographical distance is having
negative and significant impact as theory suggests. Whereas in the case of India higher GDP
level and low per capita income. Moreover when analyzing for pre and post financial crisis of
2007-08, common colony beacame a determining factor of India.

Caporale and Sova, (2015), analayze the trade flow of China with its major trading partners in
Asia. North America and Europe. The annual data from 1992 to 2012 was used for analysis
empolying a recent econometric technique “the fixed effect vector decomposition” (FEVD)
proposed by Plumper and Troeger, (2007). The econometric outcome suggest that economic
size and geographical distance are consistent with theory and FDI and WTO have positive
impact on its trade. Whereas, the dummy variable for finanacial crises for the years 2007-008
is impacting negatively on its bilateral trade.

Wang, (2016) using data ranging from 2000 to 2013 analyzed balance panel on PPML estima-
tion method through the gravity model covering 80 countries for the trade vegetable oil and
empirical findings reveal that incomes of importer countries have statistically positive impact
on trade of vegetable oil and geographical distance having statistically negative coefficient.

Rehman, (2003), investigated for determining factors trade for Bangladesh with 35 trading
partners through panel data analysis found that economic size, PCI differential and openness
have positive and significant impact on its trade and multilateral resistance factors also
influence Bangladesh’s trade positively.

Tripathi, S and Leitao, N. C. (2013), using data ranging from 1998 to 2012 applied Tobit and
GMM panel data technique for analysis of gravity model in the case of India and its main
trading partners. The econometric results suggest that higher GDPs have positive impact and
noteably geographical distance has also positive impact on bilateral trade of India. The
positive coefficient of distance is not consistent with basic gravity model. Moreover, cultural
proximity and political globalization have positive and significant impact on indian trade.

Rasoulinezhad, (2017) to investigate the specification of china’s external trade with its 13
major trading partners from OPEC member countries, used the annual data ranging from 1998
to 2014 and employed three panel data estimations with fixed effects, random effects and fully
modified ordinary least squares (FMOLS). The findings of the study reveal that gravity model

JISR-MSSE Volume 15 Number 2 July-Dec 2017 115

fits the data well and GDP, GDP difference, exchange rate, the trade openness, geographical
distance and WTO membership are trade determining factors of China with OPEC members.
Furthermore, he concluded that China’s trade patteren with OPEC is consistent with
Heckscer-Ohlin.

Hussain, (2017), used Pakistan and its 15 major trading partners’ data for the period 2003 to
2013 to augmented gravity model and estimated it with PPML-Estimator technique of panel
data. Empirical findings provide that GDP and PCI have positive significant impact on
Pakistan’s export flow and geographical distance is statistically negative impact. Furthermore,
the contiguity is statistically positive but the common official language is negative against the
expected sign. More significantly, the study first time introduces information flow index (IFI)
in the gravity model which is used as a proxy for overall globalization index to capture the
impact of globalization on exports flow and finding reveal that IFI has statistically positive
impact on exports and is reduces negative value of distance.

Hypothesis for Model (1)

H1: Economic size, geographical distance, contiguity, common official language and common
colony are determining factors of exports flow of China, India, Pakistan and Bangladesh.

Hypothesis for Model (2)

H1: Overall globalization index and its proxies (Cultural, Social, Economic, Social and Infor-
mation Flow index (IFI) don’t have impact on export of China, India, Pakistan and Bangla-
desh.

METHODOLOGY
Model Specification: Model (1) to determine the variables that explain export
Considering literature review of gravity model and taking into account the socio-economic
conditions of four countries which are understudy. The following models have been selected
for empirical estimation:

Model for PPML estimation with multilateral resistance term: as proposed by Piermartini and
Yotov, (2016).
lnX ijt = β0+ β1 ln ( DISTij ) + β2ln(Yi ) +β3ln ( E j ) +β4 (CNTG ) + β5 ( LANG ) +β6 (CLNY ) + β7 ( RTA) +uij + ηt + ε ijt (1)
Variables used in the models (1) are discussed as under:

Xijt, is export flow of countries i (exporters), China, India, Pakistan and Bangladesh directed
to their 15 major trade partners j (importers) for the time t. The data of exports of these coun-
tries has been taken from UNCOMTRADE database.

DISTij, is geographical distance from capitals the capital of an exporting nation i to the
capitals of its export destination j in kilometers. This data has been taken from the CEPII gravi-
ty dataset.

116 July-Dec 2017 Volume 15 Number 2 JISR-MSSE

Yi, is income, is Gross Domestic Product (GDP) in current US dollar of the exporting country
i which has been imported from World Development Indicator (WDI, 2016) of World Bank.
The GDP size represents the economic size and supply capacity of exporting nation.

Ej, is the expenditure (GDP in current US dollar) of importing countries j which represents the
economic size and demand capacity of an importing nation. The data is also taken from (WDI,
2016).

CNTG, is a dummy variable for contiguity implies border connections between trade partners
i and j. Its value is 1 if trade partners share common borders and 0 otherwise.

LANG is a dummy variable for common official language used in trading partners in i and j.
Its value is 1 if trade partners share common official language and 0 otherwise.

RTA, is a dummy variable if the trade partners i and j are co-signatory of regional trade
agreements and its value is 1 if trade partners share common borders and 0 otherwise.

CLNY, is a dummy variable for being a common colony. Its value is 1 if trade partners share
common characteristic of being common colony, and 0 otherwise.

The data for dummy variables in the model of this study i.e. contiguity, common official
language that is spoken in trade partners, common colony and RTAs has been taken from the
CEPII gravity dataset.

, is country-pair fixed effects.

, is exporter and importer time fixed effects.

Model Specification: Model (2) to capture the impact of globalization

To show the impact of globalization the following model has been estimated as used by (Hus-
sain, 2017).

lnX ijt = β0+ β1ln ( DISTij ) + β2ln(Yi ) +β3ln ( E j ) +β4 ( EGI ) + β5 ( PGI ) +β6 ( SGI ) + β7 ( IFI ) + β8 (GI ) +ε ijt (2)

Variables used in the models (1) are discussed as under:

Xijt, is export flow of countries i (exporters), China, India, Pakistan and Bangladesh directed
to their 15 major trade partners j (importers) for the time t. The data of exports of these
countries has been taken from UNCOMTRADE database.

DISTij, is geographical distance in kilometers between exporter nation’s capital to the capital
of their trading partners.

Ej, is the expenditure (GDP in current US dollar) of importing countries j which represents the
economic size and demand capacity of an importing nation. The data is also taken from (WDI,
2016).

JISR-MSSE Volume 15 Number 2 July-Dec 2017 117

The detail of computation and explanation of globalization indices and variables EGIi, SGIi,
PGIi, GIi, IFIi is provided in (Appendics-1) which has been used in the model (2). These
variables and indices are maintained by KOF Index of Globalization (2016).

The Data Set

This study uses data for the period of 1992-2013 for four countries namely China, India,
Pakistan and Bangladesh. The data for geographical distance, common official language,
sharing common colony, RTAs of each country has been taken from the (CEPII) gravity
dataset.The bilateral export data of four countries and their 15 major partners has been taken
from UNCOMTRADE database. The data for overall globalization index and its
sub-categories i.e. economic, social and political globalization is taken from KOF Index

Globalization, 2016 and finally the data for gross domestic products (GDP) of four exporting
countries understudy and their 15 major trading partners has taken from the World
Development Index (WDI), of World Bank database.

Estimation Technique
This study employs relatively advance econometric estimation technique PPML-Estimator for
panel data analysis of the model which has developed by Santos and Silvana, (2006, 2011b.).
This technique is having estimation superirity over traditional approaches of panel data
techniques like OLS etc. First it mangaes usefully large number of zeros which is a problem in
trade data and addresses the issue of heteroskedasity appropriatly.

Baldwin and Taglioni, (2006) explained the importance of multilateral resistence terms (MRs)
which are theoritical construct and if not contolling the MRs in the estimation; they referred as
to commit as “Gold Medal Mistake”. Therefore, to control multilateral resistance terms
( Pj ,t andΠ i ,t ) properly, this study uses exporter time and importer time fixed effects Feenstra,
(2004). Furthermore, the trade policy variables like RTA and tariff are endogenous which may
cause the estimates unreliable, in this regard as proposed by Agnosteva, etal. (2014) a variable
of pair-fixed effects has been used in the estimation of gravity model.

Cheng and Wall (2005), noted that “fixed-effects estimation applied to data pooled over
consecutive years, is sometimes criticized on the grounds that dependent and independent
variables cannot fully adjust in a single year’s time”. In this connection researchers have
proposed to use the data with 3-years to 5-years interval rather taking the data with consecutive
years. Therefore, this study takes the data for three year interval in order to capture the
adjustment in policy changes in trade regimes.

RESULTS AND DISCUSSION

The table-3 provides the estimated results of four models for four countries. In all four cases,
the coefficients of variables of basic gravity model; like income, expenditures and
geographical distance are consistent with the theory. These results are in accordance with
empirical results of the study of Batra, (2006) Hussain, (2017) Panda atel, (2016). The
coefficients of income and expenditure represent demand and supply capacity of trade
partners. These are positive and significant at 1% level in all four models of China, India,
Pakistan and Bangladesh. This result implies that larger is the size of economies of trade

118 July-Dec 2017 Volume 15 Number 2 JISR-MSSE

partners more they trade. The coefficients of bilateral geographical distance is having negative
and is significant at 5% level in the case of China, India and Pakistan and is significant at 1%
level in the case of Bangladesh. This empirical outcome reveal that with distant destinations
when they trade incur more cost and consequently attract less export. The coefficients for
expenditure are all positive and significant which provides the insight that more the per capita
income of the people of importer country they spend more on foreign goods. Common official
language is significant and has positive impact on exports in the case of China and Pakistan
and Bangladesh which is consistent with the finding of Shujaat, (2015) and Mohmand and
Wang, (2013). Furthermore, the dummy variables border are all insignificant and in the case
of Pakistan and India the RTAs have negative and significant impact which suggest that the
disputes of these countries with their neighboring nations are hurdles to get benefit from
Regional Trade Agreements. Being the common colony in the case of India the dummy has
positive and significant impact on its trade, because the small economies of SAARC who
share common colony characteristic with India rely more on India. Whereas it is having
negative and significant in the case of Pakistan insignificant in the case of Bangladesh and
China does not share common colony with their trading partners.

TABLE-3 Four Gravity models estimated PPML-Estimator

Names of Variables Names of countries above their respective models
Dependent variable India China Bangladesh Pakistan
(Export)
Distance -1.618 -0.359 -1.612 -0.571
(0.012)** (0.040)** (0.002)** (0.048)**
Income 1.038 0.774 1.502 .723
(0.0023)*** (0.0042)*** (0.007)*** (0.001)***
Expenditure 0.845 .784 0.439 -0.244
(0.001)*** (0.001)*** (0.005)*** (0.003)***
Border -0.606 0.918 -0.522 (0.157)
(0.205) (0.021)** (0.346) (0.319)
Language 0.691 1.817 ---------- 1.577
(0.071)* (0.007)*** (0.008)***
RTA -0.432 0.333 0.940 -1.424
(0.045)** (0.050) (0.108) (0.015)**
Colony 0.966 ----------- -0.276 -0.593
(0.024)** (0.535) (0.002)***
R2 0.861 0.876 .816 0.594

The values in parenthesis represent with three steric*** that it is significant at 1% level, with two steric**
significant at 5% level of significant and one steric* indicate that it is significant at 10% level of signifi-
cant.

JISR-MSSE Volume 15 Number 2 July-Dec 2017 119

Table-4 provides the results regarding the impact of overall globalization (GI) and its proxies
i.e. economic (EGI), social (SGI), cultural (CGI), political (PGI) and information flow (IFI)
indices on exports of four countries. These models have been estimated with basic gravity
variables i.e. income, expenditure and geographical distance, however, have been excluded
from the table-4 because of brevity purpose. The results suggest that economic globalization
has positive impact and significant impact on the exports of China, Bangladesh and Pakistan
and cultural globalization is an explaining variable in the case of India which has positive and
significant impact. Social globalization has negative and significant impact on Pakistan’s
export and has positive and significant impact in the case of Bangladesh. Political
globalization is having negative and highly significant impact on China’s trade. However, it
has positive and highly significant impact in the case of Pakistan and Bangladesh.

The variable information flow index (IFI) has negative and high significant impact in the case
of India and Bangladesh. Whereas, in the case of China and Pakistan information flow
(through internet, print and electronic media) is highly positive and significant similar results
were found by Hussain, (2017).

TABLE-4 Four models to capture the impact of globalization using PPML

Names of Variables Names of countries above their respective models
Dependent variable India China Bangladesh Pakistan
(Export)
Economics ______ .0322 0.026 1.616
Globalization (0.002)*** (0.035)** (0.025)**
Cultural 0.019 0.0001 ______ 0.004
Globalization (0.008)*** (0.717) (0.378)
Social Globalization ______ ______ 0.261 -0.026
(0.000) (0.031)
Political -0.040 -0.001 0.015 0.025
Globalization (0.572) (0.004)*** (0.036)** (0.0025)***
Information -0.079 0.005 -0.088 0.214
flow index (0.001)*** (0.001)*** (0.001)*** (0.001)***
Globalization 0.317 0.013 0.184 0.040
index (0.001) (0.001)*** (0.001)*** (0.001)***

The values in parenthesis represent with three steric*** that it is significant at 1% level, with two steric**
significant at 5% level of significant and one steric* indicate that it is significant at 10% level of signifi-
cant.

Overall globalization index is positive and significant in the case of India, China, Bangladesh
and Pakistan. Which implies that globalization process overall is an export determining factor
for all these four countries.

120 July-Dec 2017 Volume 15 Number 2 JISR-MSSE

CONCLUSION AND POLICY RECOMMENDATIONS
This study using data from 1992 to 2011 employed PPML-Estimator technique and analyzed
the structural gravity model. On the basis of empirical outcomes it is concludes the variables
of basic gravity model of trade like income, expenditures and distance are consistent with the
theory and the hypothesis of the study. The political disputes among these four neighboring
countries are not only main hurdle in boosting their mutual beneficial trade but also hampering
the capability to trade with rest of the world. Resolving the political dispute they can gain from
mutual trade because of natural close geographical location and due to low trade cost. In this
regard, consequently they can also reap the fruits of RTAs signed among them like Global
System of trade Preferences among Developing Countries (GSTP) came into force from 19th
April, 1989, South Asian Preferential Trade Agreement (SAPTA) effective from 07th
December, 1995 and South Asian Free Trade Agreement (SAFTA) which was put into
practice 1st January, 2006, because the RTAs have negative and significant impact on exports
of Pakistan, India and Bangladesh in this study. However, process of globalization, overall is
very much supporting phenomenon. Pakistan’s socio-economic set up is responsive with
globalization, however, the energy crises, high cost of production, weak trade policy
implementation and low value addition in its exports is undermining its trade performance
than its neighboring countries India and Bangladesh.

REFERENCES
Agnosteva, D.E., Anderson, J.E., & Yotov, Y.V. (2014). Intra-national trade costs:
(Measurement and aggregation Working Paper No. 19872). NBER
Anderson, J., & Wincoop, E. (2003). Gravity with Gravitas, A Solution to the Border Puzzle.
The American Economic Review, 93(1), 170-192.
Anderson, J. (1979). The Theoretical Foundation of the Gravity Equation. The American
Economic Review, 69(1), 106-16.
Armington, P. S. (1969). A Theory of Demand for Products Distinguished by Place of
Production, (IMF Staff papers, No. 16). Washington DC.
Baldwin, R., & Taglioni, D. (2006). Gravity for dummies and dummies for gravity equations.
(Working paper 12516). National Bureau of Economic Research.
Batra, A. (2006). India`s Global Trade Potential: the gravity Model Approach.
Global Economic Review, 35(3), 327-361.
Caporale, M., & Sova, A. (2015). Trade flows and trade specialisation: The case of China.
China Economic Review, 34(2015), 261-273.
Cheng, H., & Wall, J. (2005). Controlling for Heterogeneity in Gravity Model of Trade and
Integration. Federal Reserve Bank of St. Louis Review, 87(1), 49-63.
Deardorff, A.V. (1982). The General Validity of the Heckscher-Ohlin Theorem.
American Economic Review. 72, ( 4), 683-94.
Dekle, Robert, Jonathan E., & Samuel K. (2007). Unbalanced Trade, The American
Economic Review. 97 (2), 351–355.
Eaton, J., & Kortum, S. (2002). Technology, geography, and trade. Econometrica,
70(5), 1741-1779.
Feenstra, R. C. (2004). Advanced International Trade: Theory and Evidence.
Princeton University Press.

JISR-MSSE Volume 15 Number 2 July-Dec 2017 121

Helpman, E., & Krugman, P. (1985), Market structure and international trade.
Increasing returns, imperfect competition and the international economy (MIT Press).
Cambridge).
Hussain, M. (2017). Globalization and Gravity Model of Trade of Pakistan:
A PPML-Estimator Analysis. Management and Administrative Science Review,
6 (1), 15-27.
Madisson, A. (2008). Shares of the Rich and the Rest in the World Economy: Income
Divergence Between Nations, 1820–2030. Asian Economic Policy Review,
3(1), 67-82.
McCallum, J. (1995). National Borders Matter: Canada-U.S. Regional Trade Patterns.
The American Economic Review, 85(3), 615-623.
Mohmand, Y. T., & Wang, A. (2013). The gravity of Pakistan’s export performance in
stratified sampling. Pak. J. Stat., 29, (2), 203–216
Panda, R., & Sethi, M. (2016). A Study of Bilateral Trade Flows of China and India.
Indian Journal of Science and Technology, 9 (15), 01-07.
Piermartini, R., & Yoto V. (2016). Estimating Trade Policy Effects with Structural Gravity:
(CESEFO Working Paper No: 6009).p. 56.
Plumper, T., & Troeger, V.E. (2007). Efficient Estimation of Time-Invariant and Rarely
Changing Variables in Finite Sample Panel Analyses with Unit Fixed Effects.
Political Analysis, 15(2), 124-139.
Rasoulinezhad, E. (2017). China’s foreign trade policy with OPEC member countries.
Journal of Chinese Economic and Foreign Trade Studies, 10(1), 61-81.
Ravenstein, E. G. ( (1885). The Laws of Migration. The Journal of the Royal Statistical
Society, 47 (2),(241-305).
Rehman, M. M. (2003). A Panel Data Analysis of Bangladesh’s Trade: The Gravity Model
Approach. BISS Journal 19(1), p. 54.
Salvatore, D. (1998). International Economics. (6th ed.) Princeton Hall International.
Santos Silva, J.M.C., & Silvana T. (2006). The log of Gravity. Review of Economics and
Statistics, 88 (4), 641-305.
Shujaat, A., & Waheed, A. (2015). Potential Export Markets for Bahrain: A Panel data
Analysis. International Journal of Trade, Economics and finance, 6(3), 165-169.
Starck, S. C. (2012). The theoretical foundation of the Gravity Modeling: What are the
developments that have brought gravity modeling into mainstream economics?,
A Master Thesis, Department of Economics, Copenhagen Business School.
Tinbergen, J. (1962). Suggestions for an International Economic. New York: Shaping the
World Economy.
Trends in International Trade-B. (2013). Annual Report World Trade Organization (WTO,
2013). Retrived from https://www.wto.org/english/res_e/booksp_e/wtr13-2b_e.pdf
Tripathi, S., Leitao, N.C. (2013). India’s Trade and Gravity Model: A Static and Dynamic
Panel data, (MPRA Paper, 45502).
Wang, J. (2016). Analysis and Comparison of the Factors Influencing Worldwide Four
Kinds of Vegetable Oil Trade: Based on Gravity Model. Modern Economy, 7, 173-182.

122 July-Dec 2017 Volume 15 Number 2 JISR-MSSE

Appendex-1

2017 KOF Index of Globalization

Indices and Variables Weights

A. Economic Globalization [36%]

i) Actual Flows (50%) Trade (percent of GDP) (21%) Foreign Direct Investment, stocks
(percent of GDP) (28%) Portfolio Investment (percent of GDP) (24%) Income Payments
to Foreign Nationals (percent of GDP) (27%)

ii) Restrictions (50%) Hidden Import Barriers (22%) Mean Tariff Rate (28%) Taxes on
International Trade (percent of current revenue) (26%) Capital Account Restrictions
(24%)

B. Social Globalization [37%]

i) Data on Personal Contact (33%) Telephone Traffic (25%) Transfers (percent of GDP)
(2%) International Tourism (26%) Foreign Population (percent of total population)
(21%) International letters (per capita) (25%)

ii) Data on Information Flows (36%) Internet Users (per 1000 people) (37%) Television
(per 1000 people) (39%) Trade in Newspapers (percent of GDP) (25%)

iii) Data on Cultural Proximity (32%) Number of McDonald's Restaurants (per capita) (47%)
Number of Ikea (per capita) (47%) Trade in books (percent of GDP) (6%)

C. Political Globalization [27%] Embassies in Country (25%) Membership in

International Organizations (27%) Participation in U.N. Security Council Missions
(22%) International Treaties (26%).

JISR-MSSE Volume 15 Number 2 July-Dec 2017 123

124 July-Dec 2017 Volume 15 Number 2 JISR-MSSE