Globalization and Foreign Trade of Pakistan with D-8 Group

Globalization has made the world smaller and flatter which is called “the death of distance”. This phenomenon has brought considerable increase in international trade in recent past. Many developing countries have been benefitted from the fruits of globalization and many other like Pakistan have lagged behind in the race. Pakistan is suffering persistently from trade deficit since 2003 and by the same point of time India, Sri Lanka and Bangladesh which share similar socio-economic conditions to Pakistan: are experiencing upward trend. This study investigates the popular theory “the gravity model of trade” in the context of Pakistan’s export flow to D-8 group. The gravity trade model has been innovated by introducing overall Globalization Index (GI) which led to improve explanatory power of gravity model. This research, in panel setting has used annual data ranging from 2003 to 2013, by employing advance estimation technique PPML, Estimator. The empirical results of the study infer that GDP, population and distance confirm the basic gravity model. While the globalization Index and contiguity variables are against the expected signs. Therfore, it is concluded that Pakistan need to explore new destinations specially should target the developing countries for its exports.

Developing nations, when looking for the destination of their exports find trade barriers and restrictions mostly from the high income nations due to protectionism of trade unions amongst them. Furthermore, the access of Pakistan's export to the markets of its major trading partners like U.S and European Union are somewhat conditional to social and political issues such as labor laws, Human rights and their self-interested political policies (Hussain, 2016).

Statement of the Problem
The globalization and international competitiveness in international trade has affected adversely the trade balance and economic growth of developing countries. In this state of affairs regional cooperation and integration can bring the developing economies at the stage of international trade (Zahra & Leili, 2011). Therefore, there arises a need for developing nations to come together in the trade unions and trade blocs. Pakistan's export flow directed toward to D-8 countries.
H 1 = Income, Population, connecting border, common official languages, Geographical distance among trade partners and globalization process of Pakistan and her trading partners of D-8 affect Pakistan's export flow directed toward to D-8 countries.

Theoretical Framework
The philosophy of gravity model of trade has been employed from the Newton's law of universal gravitational force. According to this law "the gravitational force between the two masses is directly proportional and inversely to square distance between them". F = α (m 1 . m 2 )/r 2 (1) Tinbergen (1962), Poyhonen (1963), and Pulliainen (1963, were pioneers who envisioned the idea of gravity model of international trade from Newton's law of gravitational force and transforming the formula into linear form for usual regression analyses as under: This regression states that bilateral trade between two nations i and j, is an increasing function of their GDP and a decreasing function of geographical distance (mostly using transportation cost) between them. The trade ij is bilateral trade flow between countries i and j, GDP i is supply capacity of home country measured in millions of real GDP, GDP j indicates demand potential of market and D ij is bilateral distance.
The trade gravity model has been used extensively by the researchers and remained successful in explaining the bilateral trade flows and volume in quantitative way, but theoretical justification in its earlier stage was poor but with the passage of time many economist provided theoretical justification by augmenting exogenous and dummy variables in the basic gravity model of trade (Mohmand & Wang, 2013).
There is a long history of gravity model of trade. Revenstein (1885) presented the idea of early cogent narrative of gravity model. Israd and Pack (1954) empirically estimated the negative impact of geographical distance. Tinbergen (1962) proposed a reduced form of gravity model but it had weak theoretical foundation lacking price specification. Learner (1974) improved the gravity model by introducing in it factor endowment to define the impact of income and population of trading partners.
After that, Anderson (1979) contributed by including utility function and product differentiation.
Then it was left for (Bergstrand, 1985) who afterward provided theoretical foundation on the grounds of constant elasticity and monopolistic competition. Kuregman and Helpman (1985) explained the gravity trade model on the criterion of increasing returns to scale, production criterion derived the model under the assumption of increasing returns to scale in production. Deardoff (1985) contributed by providing evidence that gravity model of trade is consistent with the Hickchier-Ohlin-theory of trade. Anderson and Wincoop (2003) provided some further strength to the theoretical foundation of gravity model which was based on their earlier work on Constant Elasticity of Substitution (CES) expenditure; they explained border puzzle on bilateral trade cost differences. Zurzoso and Lehmann (2003), estimated trade flows between two blocs Mercosur and European Union through Gravity Trade Model and findings suggest that infrastructure, income differences and exchange rate are major determinants of trade flows.

Empirical Literature Estimated in the Context of D-8 Group
Ismail and Kouhestani (2011)   suggest that supply capacity, potential demand and geographical distance along with relative price, common language are critical and consistent with augmented gravity model in the case of Pakistan whereas free trade agreements have negative insignificant impact that means FTA harm Pakistan's export.

Model Specification
Keeping in view the literature review discussed earlier, the trade pattern and essential social and political characteristics of Pakistan; following extended or augmented gravity model of trade is taken into consideration. The equation (3) then becomes:

PCI it × PCI jt ; is log product of per capita GDP of Pakistan and per capita GDP of trading partner.
Dist ij ; is geographical distance between the capitals of Pakistan to the capital of trading partner and j, and is expected it is having inverse relation with dependent variable.
BDR ij and Lang ij are dummy variables show contiguity and for sharing common official language respectively. The expected signs of parameters of variables are given as below in Table 1, Expected relationship of independent variables with the dependent variables.
GI i is overall globalization Index maintained by KOG Globalization, its method of computation and variable weight is provided in Appendex-1.

Data Set
This study determines the statistically significant trade variable by applying Augmented Gravity Trade dummy variables contiguity and common official language has been taken from the CEPII gravity dataset.

Estimation Technique
There are constraints when dealing with cross section and time series data. The cross sectional data involves one or more than one variable at the same point of time. On the other hand, in time series data we observe the values of one or more variables over a period time. On the other hand, "Panel Data has space as well as time dimension" (Gujrati, 2003). Panel data has the "observations on the same units in several different time periods, Panel data may have individual (group) effect, time effect, or both, which are analyzed by fixed effect and/or random effect models". Penal data estimation has econometric superiorities over cross-sectional and time series data. According to Baltagi and Badi (2001), "Panel data deals with more informative data, more variability, less collinearity among the variables, more degrees of freedom and more efficiency". The panel data has two more important features namely "Fixesd effect Model" and Random "Effect Model". Historically gravity model has been estimated on OLS estimation technique which suffers from inefficient paramaters estimates or even asymptotically inefficient. The OLS model has been extensively used by researchers. The problem with OLS is that zero trade flows which are mostly contained in the gravity data; in logarithm forms are dropped from estimation.
Hence, hetroskedasity is a major issue of gravity model which plagues the data by making the estimates biased and inconsistent. To overcome the problem, Silva (2006)  Another problem with estimation of gravity model is adjustment in trade policy changes. "Fixed-effects estimation is sometimes criticized when applied to data pooled over consecutive years on the grounds that dependent and independent variables cannot fully adjust in a single year's time" (Cheng & Wall, 2005). Therefore, researchers suggest the data with interval to capture the trade policy changes in the model. This study takes the data for the alternative years (Model "B") and for comparison also estimated the model with consecutive years (Model "A") in Table 3. This comparison show that R 2 value in Panel (B) has been increased than Panel (A), which implies that expalnatory power of model has been increased. The estimation with panel samples pooled over consecutive years in model (A) is not having reliable estimates of distance or trade cost parameters (Piermartini & Yotov, 2016).   This study has taken the logarithm product of geographical distance from Pakistan to its destination (distance×distance) in the true spirit of Newton's law of gravitation as reported in model (C). Its economics explanation may be given as the distance covered by means of transportation is "up" and "down", i.e., from home country to destination of export "up" and then back to exporting country "down". By comparing model (B) with (C) it may be noted that the overstated negative impact of distance elasticity parameter has been reduced to exactly half without affecting the overall performance of model (B). Which is more useful in interpretation of distance or cost elasticity. The inclusion of GI in model (C) and is reported in model (D) in Table 3 shows that R-square value has been improved considerably, from .82178148 to .8692646 this implies that the model has become more good fit model.
As the theory suggests that globalization has made the world "smaller and flatter" and researchers having in view the impact of globalization call it "the death of distance". In this context it is interesting to note that with the inclusion of globalization Index the negative impact of distance has declined and connecting border has become more irrelevant (compare model "C" with "D").

The Panel Unit Root Test
The panel unit root test for the data show that all variables are integrated. Exports and overall globalization index variables are integrated at I (0) and GDP is integrated at I (1) as represented in Table 2. The empirically estimated result in Table 4 using PPML method reveal that as expected the product of GDPs of   Zahra and Leili (2011). The product of population of Pakistan and trading partner has also positive as well as is highly significant. In the context of such empirical finding; one billion populations of D-8 countries that accounts 17 percent of world population is a contributing factor in export from Pakistan. This finding is in accordance with Sherifand (2013), and Zahra and Leili (2011). More interestingly, the coefficient of globalization Index is negative against the expected sign and is statistically significant. The popular reason may be given for this outcome in the case of Pakistan that in short runs; the cost associated with the globalization to outweigh the benefits from it. Free trade brings the economic gain for everyone (positive sum-game). But the small industries cannot compete globally with the industries performing economies of scale. Furthermore, the economic globalization may increase the cost of doing business due to higher wages of skilled labor (www.investopedia.com). The international transportation cost of Pakistan maintained by WDI is available for few recent has increasing trend with time that do also support the above arguments in favor of negative impact of globalization Index on Pakistan's exports. Furthermore, it may also be concluded that globalization in case of developing countries' trade incur increasing cost in the initial stages. However, the results are for globalization proxy variable are not consistent with the findings of Mubashir (2017a, 2017b). The coefficient value of geographical distance between Pakistan and its trading partner is negative and is significant at 5 percent level which is conformity with basic gravity model of trade. However, the significant and negative impact of dummy variable for connecting border on dependent variable is not consistent with theory. The Iran is only border connecting country with Pakistan from D-8 group. Although, there is not any direct political conflict between Pakistan and Iran but this negative and significant sign indicate that because of differences on some world political issues and UN's economic sanctions; Pakistan could not trade with Iran as per the trade potential that exist between the two neighboring countries. And finally the dummy variable for common official language with D-8 is also negative and statistically significant at 5 percent level which represents that common official language is not determining factor in explaining the export flow of Pakistan when dealing with D-8.

Discussion and Policy Recommendations
On the basis of concerned literature review and estimated model of this study; the conclusion is drawn here that GDP, population size and geographical distance are determining factor for Pakistan's export flow to D-8. Globalization has negative and significant impact on Pakistan's export; particularly in the case of D-8 group. It might be concluded that with the use tool of globalization Pakistan moves away from developing to developed nation for their exports destinations. Keeping in view the above emprical finding of large population size of D-8 group with Pakistan trade policy maker for international trade should also take interest in trading with developing countries rather relying on high income nations.
Against the expected positive sign of contiguity and common language, this study observes negative impact of these dummy variables on trade which suggests that political forces should develop the harmony and take appropriate confidence building measures with neighboring countries to take the advantages of trade with connecting border countries.