Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 164

Transcription

1 Pakistan s Exports Efficiency: An Application of the Stochastic Frontier Gravity Model Hira Liaquat Instructor (Economics), Virtual University of Pakistan Naila Gul Lecturer (Economics), Virtual University of Pakistan Asmara Irfan Faculty of Management, Universiti Teknologi Malaysia Abdul Sami Faculty of Management, Universiti Teknologi Malaysia Abstract International trade is the backbone of our modern and commercial world and is considered as an engine of economic growth. Pakistan is one of those countries whose exports are considered as best quality in international markets. According to Economic Survey of Pakistan ( ), about 60% of Pakistan s exports destinations are to ten countries namely, Afghanistan, Bangladesh, China, France, Germany, Italy, Spain, UAE, UK, and USA. The present study analyzes the efficiency of Pakistan s bilateral exports with these ten major export destination countries by using panel data for years by employing the new estimation technique of Stochastic Frontier Gravity Model. Result of the estimated coefficients show that bilateral export flows of Pakistan to its trading partners are significantly positively affected by income of the major importing countries. Population or market size of importing countries shows negative and insignificant impact on exports flows. Exports flows are found negatively correlated with the distance between the countries. The mean technical efficiency for all sample countries is found to be quite low which suggests the large deviations of actual observed exports flows from the potential exports flows estimated by the gravity equation. Technical efficiency of bilateral exports is found to range from 49% to 57% during the given periods. China is found the most efficient country in the sample which recorded 57% technical efficiency. The estimated results show that 99 percent variation between actual trade and potential trade in Pakistan is due to the inefficiency factors. Keywords: Bilateral Exports, Stochastic Frontier, Gravity Model, Efficiency, Inefficiency INTRODUCTION At the time of establishment of Pakistan, major export items were raw cotton, raw jute, raw wool, tea and hides, and major import items were consumer goods, cotton yarn, and textiles. Major trading partners of Pakistan at that time were USA, UK, Belgium, Germany, Japan, and Italy. Change began to occur after the process of industrialization from producing and exporting primary commodities to exporting manufactured products, and agricultural products, particularly cotton. Pakistan s trade policy and trade pattern have constantly been moving towards lower tariffs, more trade openness, more facilities, greater advantages, and tax relieves for foreign investors. Being a WTO member, Pakistan has many bilateral and multilateral trade agreements with many countries and international organizations such as ASEAN, ECO, SAFTA, etc. According to Pakistan Economic Survey ( ), Pakistan s economic growth was remained at 4.71 percent in fiscal year In current year, Pakistan also implemented the Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 164

2 mega project of China-Pakistan Economic Corridor (CPEC) worth of US$ 46 billion with Chinese Government. This project will boost the infrastructure development; improve communication ways, enhance energy resources, and develop Gawadar port. Table 1.1 shows the trend of major exports of Pakistan in different years. Table 1.1. Pakistan s Major Exports (Percentage Share) Commodity July-March P Cotton Manufactures Leather** Rice Sub-Total of three Items Other Items Total P: Provisional, **Leather & Leather Manufactured. Source: Pakistan Economic Survey ( ) Pakistan trades with many countries but its exports are highly focused in few countries. According to Pakistan Economic Survey ( ), about 60 percent of Pakistan s exports go to 10 countries namely; Afghanistan, Bangladesh, China, France, Germany, Italy, Spain, UAE, UK, and USA. Among these countries, USA has largest share of 17 percent in exports. Table 1.2 shows the major exports markets of Pakistan. Table 1.2. Pakistan s Major Exports Markets July-March Country P Rs. US$ % Share Rs. US$ % Share U.S.A China U.A.E Afghanistan U.K Germany France Bangladesh Italy Spain All Other Total Source: Pakistan Economic Survey ( ) Cotton contributes 55% of exports, while leather, rice, and synthetic items contribute 14% of exports. Pakistan s exports have shown decreasing trend during last years. Number of Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 165

3 exports products has not much increased and high technology products are not yet explored for exports. Need of the time is to explore new export markets in Africa, ASEAN Region, South America, Eastern Europe, and Russia etc. Decline in exports is due to the reduction in international prices of commodities, energy crises, increase in the cost of production, lack of research and development etc. Due to the importance of international trade and decline in exports during recent years, the objective of present study is to examine the Pakistan s bilateral exports efficiency with its major trading partners from years 2006 to 2015 as well as to find out the factors which are responsible to create inefficiency in Pakistan s bilateral exports flows. REVIEW OF EMPIRICAL STUDIES Various studies have used the Stochastic Frontier Gravity Model to find out the imports, exports and trade efficiency both in developed and developing countries. But this technique has received very little attention in Pakistan in the field of international trade. Present study will fill this gap and would be a good contribution in literature. Following are the some empirical studies based on the Stochastic Frontier Gravity Model. Miankhel (2015) empirically investigated the effects of some behind the border constraints to Pakistan exports by using Stochastic Frontier Gravity Model for the years and The results revealed that behind the border constraints in Pakistan significantly affected the total exports during The findings suggested that for improving trade efficiency and promoting competitive exports, Pakistan needs to grow its institutional capacity given beyond the border trade barriers. It was also recommended that Pakistan should work to reduce political obstacles to regional trade. Nasir and Kalirajan (2014) selected modern services like computer and information, business and telecommunications to examine the export performance of developed Asian economies by using Stochastic Frontier Gravity Model. The finding showed that services export potential of South Asia and ASEAN economies were weaker as compared to developed economies in Europe and North America. It was observed that the important factors which increases the services export potential in emerging countries were number of graduates and the quality of ICT infrastructure. It was suggested that these countries need to remove the behind the border constraints as well as to adopt latest technologies in order to compete the developed countries. Ravi Shankar and Stack (2014) conducted a study to examine trade potential efficiency of Eastern European countries by using Stochastic Frontier Gravity Model for the years Results showed that GDP, per capita GDP differential, colony and member of EU were positively correlated with exports while distance, exchange rate and landlocked were negatively correlated with exports. Findings also showed high efficiency scores and high degree of trade integration for the Czech Republic, Hungary, Poland and Slovakia against Western partners. Hungary s performance was found to closer to four fifths of frontier trade levels. Mixed efficiency scores were also observed for Hungary and Slovakia against Greece, Iceland and Norway. Deluna & Cruz (2013) used the Stochastic Frontier Gravity Model to investigate the Philippine merchandise export efficiency and potential with trading countries for the years The outcomes of the study showed that income and market size of importing countries positively affected the merchandise export flows of Philippines. The technical Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 166

4 efficiency of all sample countries was ranged from 38% to 42% and potential was changed within the period. Result of technical inefficiency model predicted that potential is increased by membership of the Philippines to ASEAN, APEC and WTO. The important factors which increased the export potential of Philippine merchandise exports were free labor market, reduction in corruption in importing countries and common language. Koh (2013) used the Stochastic Frontier Gravity Model to examine the trade determinants and trade potential of Brunei Darussalam for the years Finding revealed that GDP, population, colonial relationship, and trade agreements were positively correlated with the level of trade, while distance was negatively correlated with the level of trade. Brunei s trade potential was found relatively low due to the presence of behind the borders constraints but these constraints were found decreasing over time for ASEAN. Kang and Fratianni (2006) employed the Stochastic Frontier Gravity Model of international trade and regressed bilateral trade flows on trading partner s income and distance to assess the potential bilateral trade flows. Trade efficiencies were computed and ranked for individual countries, 10 geographical regions, and 11 regional trade agreements. Low efficiency was observed in international trade as the gap between actual trade and potential trade was larger. Conceptual and Theoretical Review Gravity Trade Model: Gravity trade model has been used widely in empirical literature of international trade by researchers in thousands of research papers and published articles covering all areas of trade. This model describes bilateral trade flows among different countries or regions in terms of their economic sizes, distance among them and other factors affecting the trade. Gravity trade model can be used in different dimensions like to determine the FDI flows among different countries, to find migration flows, to examine the effects of regional trade agreements on trade etc. Gravity trade model rooted in Physics was initiated from the law of gravity of Newton. This law states that force of gravity is positively correlated with the product of the two bodies and negatively correlated with the squared distance between these two bodies. This law can be written as: F ij = C.M i M j /D ij ² (1) F ij = Gravity force between two bodies i and j C = Gravity constant D ij = Distance M i, M j = Masses of the two objects This approach shows that goods flow from country i to country j is equal to the product of potential trade volumes of the two countries divided by any resistance to trade i-e distance. Tinbergen (1962) applied the gravity model to describe the trade flows between different countries which can be written as: T ij = C.G i G j /D ij (2) Tij = Bilateral trade flows Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 167

5 C = Constant G i G j = GDP or GNP of country i and country j D ij = Distance between country i and country j GDP/GNP and distance show economic size and trade costs respectively. Equation (2) is non-linear form so, log form was used to convert this non-linear equation into a linear equation; thus, the gravity equation now can be written as follows: logx = c b loggdp b loggdp b log Dis tan ce e (3) ij 1 i 2 j 3 ij ij X ij = Exports from country i to country j GDP i = Gross Domestic Product of country i GDP j = Gross Domestic Product of country j Distance = Trade costs between country i and country j e ij = Residual or random error term c = Regression constant b s = Estimated coefficients/elasticities Thus, gravity model states that trade volume are directly proportional to the GDP of importing and exporting countries and indirectly proportional to the distance between them. Anderson and Wincoop (2003) developed the structural model and analyzed the border effect on the welfare and trade by using the general equilibrium gravity model to solve the famous border puzzle of McCallum (1995). This puzzle stated the trade between Canadian provinces after controlling for size and distance is a factor 22 (2,200%) times trade between states and provinces. The gravity model thus developed was a demand function and can be written as: П logx logy logy logy 1 logt log logp ij i j ij i j (4) logt b log distance b contig b comlang _ off b colony b comcol k ij 1 ij (5) This model has important implications because along with economic size and distance, it has included many dummy variables of border, common official language, colony, and common language that were not included in the intuitive model. STOCHASTIC FRONTIER ANALYSIS Stochastic frontier Analysis was developed independently by Aigner et al. (1977) and Meeusen and van den Broeck (1977). SFA has been used extensively for observing firm performance. Traditionally SFA specifies a production frontier which represents the maximum output that can be obtained from a given level of inputs. The firms operate on frontier are called technically efficient firms and the firms which operate within the frontier Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 168

6 are called inefficient firms. Proposed model of Stochastic Frontier Production is given below. This model has two random components so it is also known as composed error model. Y i = X i β + (v i u i ) (6) Y i = output for the i th firm X i = explanatory variables associated with the i th firm. β = Vector of unknown parameters to be estimated v i = Random errors which are assumed to be iid~n(0, σ v 2 ) and independent of u i u i = A non-negative random variables introduce to measure magnitude of technical inefficiency in production assumed to be σ u 2 (u i ~iid N(u i, σ u 2 ) reflects the fact that each firm s output must lie on or below its stochastic frontierf(x i ; β + v i ). However u i can have other distributions such as gamma and exponential. Mean of this distribution assumed to be function of firm specific characteristics or explanatory variables. Technical inefficiency model is defined as: μ i = δ i Z i + w i (7) z i = p 1 vector of firm specific variables which influence the efficiency of firm δ = 1 p vector of unknown parameter to be estimated w i = Random variable defined by the truncation of the normal distribution with zero mean and variance σ i 2 Technical efficiency can be measured by taking the ratio of observed output to maximum attainable output given set of inputs. (Kumbhaker and Lovell (2000), Taymaz and Saatci (1997) as given below: TE = f(x i;β) exp(v i u i ) f(x i ;β) exp(v i ) (8) 0 TE 1 = exp ( μ i ) Technical efficiency score lies between 0 and 1.TE = 1 implies efficient firm while TE = 0 implies inefficient firm. The maximum likelihood method can be used to estimate the coefficients of stochastic frontier model and inefficiency model simultaneously. Aigner et al. (1977) proposed a loglikelihood function for the model by assuming half normal distribution for technical inefficiency effects. Aigner et al. expressed the likelihood function in terms of two variance Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 169

7 parameters as σ s 2 = σ u 2 + σ v 2 and λ = σ u /σ v (0 λ ) however the parameterization of Battese and Corra (1977) will be used replacing λ with γ. γ has value between zero and one where as it could be any non-negative value. (0 γ 1). σ s 2 = σ v 2 + σ u 2 (9) γ = σ u 2 σ v 2 +σ u 2 = σ u 2 /σ s 2 (10) σ v 2 = Variance of statistical noise σ u 2 = Variance of inefficiency effect If the value of σ 2 is equal to zero then u i will be zero it means all firms are fully efficient. If σ 2 is larger than zero, then all firms are not fully efficient. The value of gamma γ lies between zero and one. If the value of γ is one, the deviations from the frontier are due to random error. SFA can be used to define a trade frontier whereby inefficient trade performance refers to the degree to which actual trade falls short of the maximal, frontier level of trade. This is achieved by modifying the conventional gravity model. Stochastic Frontier Gravity Model: The Gravity Stochastic Frontier Model is the Integration of Gravity Model and Stochastic Frontier Production Function Model which was formally introduced by Kalirajan (2000) to address the inherent bias of the conventional gravity model of trade and to estimate potential trade flows. With a stochastic frontier approach, the gravity equation can be written as: lnx ijt = lnf(y ijt ; β)exp (v ijt u ijt ) (11) Xijt = Actual exports from country i to country j f (Yijt; β) = Function of the determinants of potential trade (Yijt) and is a vector of unknown parameters uijt = Single sided error term, the economic distance bias referred by Anderson (1979), which is due to the influence of the behind the border measures of the importing country. This bias creates the difference between actual and potential trade between two countries. uijt takes value between 0 and 1 and it is usually assumed to follow a truncated (at 0) normal distribution, (μ, σu2). When uijt takes the value 0, this indicates that the bias or countryspecific beyond the border constraints are not important and the actual exports and potential exports are the same, assuming there are no statistical errors. When uijt take the value other than 0 (but less than or equal to 1), this indicates that the bias or country-specific beyond the Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 170

8 border constraints are important and they constrain the actual exports from reaching potential exports. vijt = Double sided error term, which is usually assumed to be N(0,σv2), captures the influence on trade flows of other left out variables, including measurement error that are randomly distributed across observations in the sample. Model, Data and Methodology Traditionally, bilateral trade flows have been estimated through cross sectional data but cross sectional and time series data generates biased results due to heterogeneity. Panel data estimation has many advantages over cross sectional and time series because it controls individual heterogeneity, reduces collinearity among explanatory variables through large degree of freedom and increases the efficiency of econometric estimates. Following model specification is used to fulfil the objective of the present study. log( EXP ) log(gdp ) log( POP ) log( DST ) uji vji (12) jit 1 it 2 it 3 jit i = Importing country i.e. Afghanistan, Bangladesh, China, France, Germany, Italy, Spain, UAE, USA, UK j = Exporting country Pakistan EXP jit= Bilateral exports from Pakistan (j) to importing country (i) at time t GDP i = Gross Domestic Product of country i at time t as proxy for income POP i = Population of country i as proxy for market size DST ji = is the geographical distance between the capital cities of country j and i measured in kilometers. The inefficiency effect model is specified in the following equation which captures the natural factors of importing countries that contribute to Pakistan s exports inefficiency. uij ( LNG ) ( LNL ) ( AREA ) w (13) 0 i 1 i 2 i ij LNG i = Common official language LNL i = Landlocked Area i = Country area Equations (12) and (13) are estimated simultaneously using Frontier 4.1 software of Tim Coelli (2004). Brief description of each variable is given as follows: Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 171

9 Bilateral Exports Flows: Bilateral exports flows of Pakistan with its ten major export destination countries are used as dependent variable. Data of bilateral exports in US$ are obtained from UN COMTRADE database of United Nations. Gross Domestic Product: Gross Domestic Product (GDP) of importing countries is used as independent variable as a proxy for income. GDP measures the total value of all goods and services produced in an economy during a given time period. It represents total volume of production and income of a country. Exports flows and GDP are positively correlated; higher the income of importing country, more will be the demand for importing goods, and higher will be the exports from Pakistan; so estimated GDP coefficient is expected to have positive sign. Data of GDP in current US$ are obtained from World Development Indicators (WDI). Population: Total population of importing countries is used as independent variable as a proxy for market size. Exports flows and population are positively correlated; higher the population of importing country more will be the demand for importing goods, and higher will be the exports from Pakistan; so estimated population coefficient is expected to have positive sign. Data of population in total numbers are obtained from World Development Indicators (WDI). Distance: Distance is used to capture the effect of transportation costs and transaction costs on exports flows. Exports flows and distance are negatively correlated; higher the distance, higher will be the trade cost, and lower will be the exports volume; so estimated distance coefficient is expected to have negative sign. Data of distance in kilometers are taken from CEPII, Research and Expertise on the World Economy. Data of distance in kilometers is measured from one country s capital to other country s capital by using great circle distance 1. Great circle formula is often used to measure the distance between two countries as it considers Economic Centre of each country or longitude and latitude of the capital. Language: Language is used as dummy variable which takes the value of 1 if country shares the same common official language (English) and the value of zero otherwise. Exports flows and language are positively correlated; countries sharing same language tend to trade more than countries having different languages; so estimated language coefficient is expected to have positive sign. Information about languages of countries is obtained from CEPII, Research and Expertise on the World Economy. Landlocked: Landlocked is used as dummy variable which takes the value of 1 if country is landlocked and the value of zero otherwise. Landlocked country is almost all surrounded by land and lack of oceans and sea. Exports flows and landlocked are negatively correlated; landlocked countries tend to trade less due to lack of oceans and sea ports; so estimated landlocked coefficient is expected to have negative sign. Information about landlocked countries is obtained from CEPII, Research and Expertise on the World Economy. Area: Total land area is used as dummy variable to capture the country size which takes the value of 1 if country area is greater than 1.5 million kilometer square and the value of zero otherwise. According to CIA Fact book, countries are ranked geographically as greater or less than 1.5 million kilometer square. Land area does not include in-land water bodies like lakes, reservoirs, and rivers. Exports flows and land area are positively correlated; larger the 1 See Notes on CEPII s Distances Measures: The GeoDist database by Mayer and Zignago (2011) Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 172

10 size/area of country more will be the trade; so estimated area coefficient is expected to have positive sign. Information about area of countries is obtained from CEPII, Research and Expertise on the World Economy. Results and Discussions Equations (12) and (13) are estimated by using the Stochastic Frontier Gravity Model to measure the Pakistan s exports efficiency for the years by using Software FRONTIER 4.1 by Tim Coelli (2004). Both equations are estimated simultaneously. The idea to estimate all the parameters simultaneously in a single step procedure were given by Kumbhakhar and McGuckan (1991) and Reifschneider and Stevenson (1991). This idea was consistent with assumption that inefficiencies are identically and independently distributed. The estimation provides the export efficiency scores of all trading partners which is the objective of present study. Table 3 shows the estimation results of Gravity model. Table 1.3: Maximum Likelihood Estimates of the Coefficients of Stochastic Frontier Gravity Model for Pakistan s Exports with Major Trading Partners ( ) Variables Est. Coefficients t-ratio β0 Constant 28.89* 25.3 β1 GDP i 0.314* 4.95 β2 Population i * β3 Distance ij * * Represents 5% level of significance Source: Authors Calculations through SFA GDP coefficient is positive and significant at 5% level of significance with the income elasticity of bilateral exports of 3.14%. This shows that Pakistan s bilateral exports flows will increase by 3.14% on average with 1% increase in GDP of importing countries. Economic growth is measured by the level of GDP; higher the GDP of a country, higher will be the economic growth of that country and higher will be the trade volume including imports and exports. People demand more exotic foreign varieties which may lead to the innovation or invention of new products. Thus, large economies tend to import more because of their higher incomes and also tend to export more because of their large variety of output or production; so, larger the economy, larger will be the trade (Krugman, 2012). Population is used as a proxy for the market size of importing countries. Population coefficient is significant but negative. This shows that Pakistan s bilateral exports flows will decrease by 3.74% on average with 1% increase in population of importing countries. The possible reason of negative coefficient may be that increasing population with increasing demand and increasing consumption of domestic goods lead to greater part of the production absorbed by the domestic market which shrinks the surplus available for exports resulting in lower trade. During previous few years, Pakistan s major trading partners like USA, China, and EU observed a sluggish economic growth. Major portion of Pakistan s exports go to US, China and EU. Therefore, due to slowdown in China and European economies and due to less demand of imports in these economies, Pakistan s exports are declined significantly. Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 173

11 Moreover, large population (large market size of importing country) may indicate a large resource endowment, self-sufficiency, and less reliance on international trade. Distance coefficient is negative and significant. This shows that bilateral exports flows will decrease by 9.54% on average with 1% increase in distance of both countries. Higher the distance among trading countries, lower will be the trade volume as transportation cost increases. There is very little that economists fully understand about global trade but there is one thing that we do know commerce declines dramatically with the distance (Leamer 2007). The further the distance between two trading countries, the lesser the bilateral trade between them (Tinbergen, 1962; Poyhonen, 1963; Bergstrand, 1985). All empirical studies conducted on the gravity trade model confirmed the negative relationship between trade volume and distance. Table 1.4 shows the estimation results of technical inefficiency model. Table 1.4: Maximum Likelihood Estimates of Coefficients of the Inefficiency Effect Model for Pakistan s Exports with Major Trading Partners ( ) Variables Est. Coefficients t-ratio λ0 Constant 0.270* 31.6 λ1 Language ij * λ2 Landlocked ij λ3 Area i -1.51* Sigma Squared σ * 9.96 Gamma γ 0.99* 2.30E+06 Log Likelihood Function * Represents 5% level of significance Source: Authors Calculations through SFA Present study has incorporated some dummy variables affecting the level of bilateral exports flow of Pakistan i.e. common language, landlockedness, and the area of country. Language coefficient is negative and significant at 5% level of significance. Common language between the two trading countries significantly increases technical efficiency of exports flows. This increases technical efficiency by 94%. (Negative sign of the coefficient indicates reduction of technical inefficiency). It suggests that technical/trade efficiency increases with common language of countries. Landlocked coefficient is positive and insignificant. If the country is landlocked i.e. lack of oceans and sea ports (as is the case of Afghanistan), increases transportation costs which may lower the exports flows. Area coefficient is negative and significant at 5% level of significance. Area is used as a proxy for country size. This implies that area is significantly increases technical/trade efficiency. This increases technical efficiency by 151%. Thus, trade with larger importing countries increases technical efficiency. Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 174

12 The estimated σ 2 is highly significant. This is a measure of the mean total variation over the ten years time period. This implies that the exports flows of Pakistan during this period have been changing. The estimated Gamma (γ) value is also significant which employs that 99 percent variation between actual trade and potential trade is due to the inefficiency factors. The gamma coefficient (γ) which is the ratio of the variation in exports due to the natural specific factors to total variation in exports is close to one and significant. A significant and larger gamma coefficient shows that use of the Stochastic Frontier method is appropriate for the sample data. This also shows that there are country specific natural constraints, which are not captured by the other explanatory variables. For example, different language in trading countries and area of importing countries restrain Pakistan from reaching their export potential. It is also confirmed by conducting hypothesis testing. The null hypothesis of no technical inefficiency effects also rejected at 5 percent level of significance. This hypothesis result indicates that countries are supposed to be technically inefficient. Table 1.5 shows the estimation results of technical efficiency scores as measure of exports performance. Table 1.5: Technical Efficiency Scores for Sample Countries ( ) Importing Countries Mean Efficiency Afghanistan Bangladesh China France Germany Italy Spain UAE USA UK Average TE Mean Efficiency = 0.18 Source: Authors Calculations through SFA Technical Efficiency scores in Table 5 predict that technical efficiency is changing during the 10 years period. The given results show that Pakistan is not showing 100% export efficiency with any of its trading partner. It also shows that given the determinants of our model, Pakistan s export performance with its major trading partners is not good. Pakistan is lack behind to showing maximum exports to its trading partners and there is lot of potential exits in Pakistan to increase its exports to its trading destinations. The mean technical efficiency for all sample countries range from 49% to 57% during these periods. Mean technical efficiency among the country groups in year 2014 are relatively high, which is above the mean technical efficiency. Export flows are more efficient in China with mean technical efficiency of 57%, Germany with mean technical efficiency of 54%. China is the most efficient country in the sample which recorded 57% technical efficiency. This is Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 175

13 followed by Germany (54%), Bangladesh (52%), and Italy (52%). Spain, UAE, and USA recorded a low technical efficiency (49%). As a whole, the technical efficiency measure of exports flows is quite low (Mean Efficiency 18%) which suggests large deviations of actual observed exports flows from the potential exports flows estimated by the gravity equation. 6. CONCLUSION This paper has investigated the bilateral exports efficiency of Pakistan to its 10 major trading partners by adopting new approach of Stochastic Frontier Gravity Model for the years It is observed that Pakistan is not showing maximum possible exports to its major trading partner and there exist lot of potential to increase its exports. The results of the study predicts that Pakistan s bilateral exports flows increases with increase in GDP of importing countries and will decrease by increase in population or market size and distance of importing countries. Study has also incorporated some dummy variables affecting the level of bilateral exports. Common language between the two trading countries and larger country size in terms of area significantly increases the technical efficiency of exports flows. It is also anticipated that if the country is landlocked then it increases transportation costs which may lower the exports flows. The estimated results show that 99 percent variation between actual trade and potential trade in Pakistan is due to the inefficiency factors. The mean technical efficiency for all sample countries is found to range from 49% to 57% during the study period. China is found the most efficient country in the sample which recorded 57% technical efficiency. On average, the technical efficiency measure of exports flows is quite low that is 18% which suggests the large deviations of actual observed exports flows from the potential exports flows estimated by the gravity equation. References Aigner, D. C. (1977). Formulation and Estimation of Stochastic Frontier Production Function Models. Journal of Econometrics, Anderson, J. E. (1979). A theoretical foundation for the gravity equation. The American Economic Review, 69(1), Anderson, J. E., & Van Wincoop, E. (2003). Gravity with gravitas: a solution to the border puzzle. The American Economic Review, 93(1), Battese, G. E., & CORM, G. S. (1977). Estimation Of A Production Frontier Model: With Application To The Pastoral Zone Of Eastern Australia. Australian Journal of Agricultural Economics, Bergstrand, J. H. (1985). The gravity equation in international trade: Some microeconomic foundations and empirical evidence. The Review of Economics and Statistics, C.Kumbakhar, S., & Lovell, C. (2000). Stochastic Frontier Analysis. University of Cambridge. Coelli, T. (1994). A Guide to Frontier Version 4.1: A Computer Program for Stochastic Frontier Production and Cost Function Estimation. Center for Efficiency and productivity Analysis. Working paper No. 7/96, University of New England Australia. Frankel, J. A., & Romer, D. (1999). Does trade cause growth?. American Economic Review, 89(3), Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 176

14 Jr, R. S., & Cruz, E. D. (2014). Philippine Export Efficiency and Potential: An Application of Stochastic Frontier Gravity Model. World Journal of Economic and Finance, Kalirajan, K. (2000). Indian Ocean Rim Association for Regional Cooperation (IORARC):Impact on Australia s trade. Journal of Economic Integration. Kang, H., & Fratianni, M. (2006). International Trade Efficiency, the Gravity Equation, and the Stochastic Frontier. Kelley School of Business. Koh, W. C. (2013). Brunei Darussalam s Trade Potential and ASEAN. Southeast Asian Journal of Economics, Krugman, P. R. (2012). International economics: Theory and policy, 9/E. Pearson Education India. Kumbhakar SC, S. G. (1991). A Generalized Production Frontier Approach for Estimating Determinants of Inefficiency in United-States Dairy Farms. Business and Economic Statistics, Leamer, E. E. (2007). A flat world, a level playing field, a small world after all, or none of the above? A review of Thomas L. Friedman's the World is flat. Journal of Economic Literature, 45(1), Mayer, T., & Zignago, S. (2011). Notes on CEPII s distances measures: The GeoDist database. McCallum, J. (1995). National borders matter: Canada-US regional trade patterns. The American Economic Review, 85(3), Miankhel, A. K. (2015). Pakistan s Potential Trade and Behind the Border Constraints. The Pakistan Strategy Support Programe. Nasir, S., & Kalirajan, K. (2014). Modern services export performances among emerging and developed asian economies. Asian development bank. Pakistan, Government of ( ). Pakistan Economic Survey. Islamabad: Finance Division. Pöyhönen, P. (1963). A tentative model for the volume of trade between countries. Prediction from a Gravity model. Weltwirtschaftliches Archiv, Qureshi, M. I., Khan, N. U., Rasli, A. M., & Zaman, K. (2015). The battle of health with environmental evils of Asian countries: promises to keep. Environmental Science and Pollution Research, 22(15), Ravishankar, G., & Stack, M. M. (2014). The Gravity model and trade efficiency: A Stochastic Frontier Analysis of Eastern European countries' potential trade. The World Economy, 37(5), Reifschneider, D., & Stevenson, R. (1991). Systematic Departures from the Frontier: A Framework for the Analysis of Firm Inefficiency. International Economic Review, Tinbergen, J. (1962). An analysis of world trade flows in shaping the world economy. Edited by Jan Tinbergen New York NY: Twentieth Century Fund. Taymaz, E., & Saatci, G. (1997). Technical Change and Efficiency in Turkish Manufacturing Industries. Journal of productivity Analysis, Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 177